San Luis & Delta-Mendota Water v. Natural Resources Defense Coun

isolation, must, was, in meets the stan- we the inference that BP scrutinized at the Tellabs, least, very deliberately at 127 S.Ct. dard.” U.S. reckless as to the (emphasis original); in N.M. State false or misleading public nature of their Council, Inv. 641 F.3d at 1095. statements is “at compelling least as as Matrixx, any opposing inference.” The district court also concluded that Tellabs, at (citing S.Ct. 551 U.S. at periods, all the statements and con- “[a]t 2499). 324, 127 S.Ct. plaintiffs allege closely duct that more re- simple corporate mismanagement sembles CONCLUSION

than actionable securities fraud.” While agree exemplify that BP’s actions cor- we end, In the we conclude that after six porate mismanagement, the pleadings also years preliminary litigation, the allega- charge company that BP ais that has should tions now be tested on the merits. publicly responsibility shirked for its role We return the matter to the district court causing Bay spills every the Prudhoe for that purpose. foregoing For the rea- step way. may of the while And there sons, judgment of the district court is enough complaint be in the to establish IN PART REVERSED and AFFIRMED with respect scienter to CEO John IN PART. party Each shall bear their own 25, 2006, April Browne’s statement costs. gives pause only us weeks after a devastating oil spill, willing BP’s CEO was a public spill

to make statement “in spite

occurred of the fact that [BP has] monitoring

world class corrosion and leak systems, being

detection applied both with-

in regulations set the Alaskan authori- According allegations

ties.” SAN LUIS & DELTA-MENDOTA WA complaint, as well as BP’s later admis- AUTHORITY; TER Westlands Water sions, this statement had no in reali- basis District; Orchards; Jasper Stewart & ty- Arroyo Farms, LLC; King Pistachio extent, corporate mismanage- To some Grove; Contractors; State Water Met plausible explanation ment would be a ropolitan Water District of Southern corpo- how misinformation travels California; for a Coalition Sustaina case, rate suite. But in alleged this facts Delta; County Agen ble Kern Water complaint support the conclusion cy; Family Alliance, Farm Plaintiffs- that BP had been aware of corrosive condi- Appellees, Department California decade, yet tions for over a chose not Resources, Water Intervenor-Plain complaint to address them. The also al- tiff-Appellee, intentionally that BP leges adopted corro- monitoring practices sion that deviated Sally JEWELL, Secretary industry standards an effort to cut as of the De Interior; partment Depart suggests every costs. And it that BP had U.S. know, least, Interior; very they reason to Fish ment of U.S. & Service; Ashe, did not have accurate information Daniel M. as regard- Wildlife Fish and ing Bay pipe- the condition of the Prudhoe Director the U.S. Wildlife Service; Lohoefener, Regional allegations lines. When we consider the Ren as true, holistically accept them to be Director of the Fish and Wildlife U.S. California; for a Sus- Region, Coalition ern Service, Pacific Southwest Delta; County Water Kern tainable Interior; Department of the Unit- U.S. Alliance, Family Plain- Agency; Farm Reclamation; Mi- of Bureau ed States tiffs, Connor, of as Commissioner chael L. Reclamation, U.S. of Bureau U.S. Department of Water California Interior; David Department of the Resources, Intervenor- Bu- Murillo, of the as Director U.S. Plaintiff, Reclamation, Re- Mid-Pacific reau of Interior; Department gion, of U.S. Contractors, Water State Director, Cowin, De- California Mark Plaintiff-Appellant, Resources; United partment of Water Justice; Department U.S. of States Agency; Protection Environmental Secretary Depart Sally Jewell, of the as McCarthy, capaci- in her official Gina Interior; Fish & U.S. ment

ty Administrator of Environ- as Ashe, Service; Daniel M. as Wildlife Agency; De- mental Protection U.S. Fish and Wildlife Director of U.S. Anthony Transportation; partment of Lohoefener, Regional Service; Ren as capacity Foxx, in as Secre- his official Fish and Wildlife Director of the U.S. Transportation; Ad- tary Maritime of Region, Service, Pacific Southwest Sr., Jaenichen, ministration; Paul N. Interior; Department of the Unit U.S. Acting capacity Mar- Reclamation; as in his official Of ed Bureau States Administrator; Department Connor, U.S. itime L. as Commissioner Michael Johnson, Reclamation, Security; Jeh Bureau of of Homeland of the U.S. Interior; Secretary Department of capacity as of in his official U.S. Murillo, as Director of the U.S. Security; David Federal Emer- Homeland Reclamation, Mid-Pacific Bureau of Management Agency; gency William Region, Department of the Interi U.S. Fugate, capacity Craig in his official or; Department of Jus United States of the Federal as Administrator tice; Environmental Protection U.S. Management Agency; Emergency McCarthy, Agency; in her offi Gina Army Corps Engi- of States United capacity as Administrator of the cial Bostick, neers; P. Command- Thomas Agency; Protection Environmental Engineers, ing of General Chief Transportation; Department of U.S. Army Engi- Corps of United States Foxx, Anthony capacity in his official neers, Defendants, Transportation; Secretary Mari as of Administration; Paul N. Jaeni time chen, capacity Sr., as in his official Council; Natural Resources Defense Acting Deputy Maritime Administra Bay Institute, Intervenor- tor; Department of Homeland Se U.S. Defendants-Appellants. curity; Johnson, in his official Jeh Secretary

capacity of Homeland Se as Manage Luis Delta-Mendota Au- curity; Emergency San & Water Federal District; thority; Craig Fugate, Agency; Water Westlands William ment Orchards; Arroyo Jasper capacity Administra & in his official as Stewart Grove; Emergency Farms, LLC; King Man Pistachio of the Federal tor Army agement Agency; Metropolitan United States Water District of South- capacity Engineers; as Administrator of the Envi Corps Thomas P. Bos Agency; ronmental Protection Commanding U.S. tick, and Chief General Department Transportation; An Army Engineers, States United thony Foxx, capacity in his official Department Corps Engineers; U.S. Secretary Transportation; of Mari Interior, Defendants-Appellees, of the Administration; N. time Paul Jaeni Council; Defense Natural Resources chen, Sr., capacity in his official as Bay Institute, Intervenor- Acting Deputy Maritime Administra Defendants-Appellees. tor; Department U.S. of Homeland Se curity; Johnson, Jeh in his official Delta-Mendota Water Au- San Luis & capacity Secretary as of Homeland Se District; thority; Westlands Water curity; Manage Emergency Federal Orchards; Arroyo Jasper Stewart & Agency; Craig Fugate, ment William Grove; Farms, LLC; King Pistachio capacity in his official as Administra Delta; for a Sustainable Coalition Emergency tor of the Federal Man County Agency; Family Kern Water Army agement Agency; United States Alliance; Farm State Water Contrac- Engineers; Corps of Thomas P. Bos Plaintiffs, tors, tick, Commanding General and Chief Engineers, Army of United States Department of Water California Corps Engineers, Defendants-Ap of Resources, Intervenor- pellees, Plaintiff, Council; Natural Resources Defense Bay Institute, Intervenor- Metropolitan Water District Defendants-Appellees. California,

of Southern San Luis & Delta-Mendota Water Au Plaintiff-Appellant, thority; District; Westlands Water v. Jasper Orchards; Arroyo Stewart & Grove; Farms, LLC; King Pistachio Sally Jewell, Secretary Depart- of the as Contractors; Metropoli State Water Interior; Department ment of the U.S. of tan Water District Southern Cali Interior; of the Fish & Wildlife U.S. fornia; Coalition for a Sustainable Service; Ashe, M. Daniel as Director Delta; County Agency; Kern Water Service; of the U.S. Fish Wildlife Family Alliance, Plaintiffs-Ap Farm Lohoefener, Regional Ren as Director Department pellees, of California Wa Service, of the Fish and Wildlife U.S. Resources, ter Intervenor-Plaintiff- Region, Pacific U.S. De- Southwest Appellee, Interior; partment of the United v. Reclamation; Mi- States Bureau of Connor, chael L. as Commissioner of Sally Jewell, Secretary Depart as of the Reclamation, Bureau of U.S. U.S. Interior; Department ment of the U.S. Interior; Department David of Interior; Fish & of the U.S. Wildlife Murillo, Director of the Bu- as U.S. Service; Ashe, as Director Daniel M. Reclamation, reau of Mid-Pacific Re- Service; of the U.S. Fish and Wildlife Interior; gion, Department Lohoefener, of Regional U.S. Ren as Director U.S., Justice; Department Service, United States Fish and Wildlife Agen- Region, Environmental Protection De U.S. Pacific Southwest U.S. Interior; cy; McCarthy, partment in her of the United Gina official Reclamation; Department of Water Mi California Bureau of States Connor, Resources, L. as Commissioner chael Intervenor- Reclamation, Bureau of U.S. the U.S. Plaintiff, Interior; David Department of the Murillo, Bu Director of the U.S. as Reclamation, Re Mid-Pacific

reau of Jewell, Secretary Depart Sally Interior; gion, Department of the U.S. Interior; Department ment of the U.S. Justice; Department of United States Interior; & Wildlife U.S. Fish Agen Protection Environmental U.S. Ashe, Service; Daniel M. as Director McCarthy, cy; In her official Gina Service; Fish and of the U.S. Wildlife Envi capacity Administrator of the as Lohoefener, Regional Ren as Director Agency; Protection U.S. ronmental Service, Fish of the U.S. and Wildlife Transportation; Department An of Region, De Pacific U.S. Southwest thony Foxx, capacity in his official as Interior; Secretary Transportation; Mari partment of of United Administration; Paul N. Jaeni time Reclamation; of Mi States Bureau chen, Sr., capacity in his official as Connor, L. as of chael Commissioner Acting Deputy Maritime Administra Reclamation, Bureau of U.S. U.S. tor; Department of Se U.S. Homeland Interior; Department of the David Johnson, curity; in his official Jeh Murillo, Bu as Director of U.S. Secretary capacity of Homeland as Se Reclamation, Mid-Pacific Re reau of curity; Emergency Manage Federal Interior; Department gion, U.S. of Craig Fugate, Agency; ment William Justice; Department States of United capacity as Administra his official Agen Environmental Protection U.S. Emergency of the Federal Man tor cy; McCarthy, in her official Gina Army agement Agency; United States capacity of the Envi as Administrator Engineers; Corps P. Bos of Thomas Agency; ronmental Protection U.S. tick, Commanding General Chief Department Transportation; An of Army Engineers, of United States thony Foxx, capacity official as his Engineers, Defendants-Ap Corps of Secretary Transportation; Mari pellants, of Administration; Paul N. time Jaeni chen, Sr., capacity in his official as Council; Natural Resources Defense Acting Deputy Maritime Administra Bay Institute, Intervenor- tor; Department of Homeland U.S. Se Defendants. curity; Johnson, official Jeh his Au- San Luis & Delta-Mendota Water Secretary capacity as of Homeland Se District, thority; Westlands Water curity; Emergency Manage Federal Plaintiffs-Appellants, Fugate, Agency; Craig ment William capacity in his official as Administra Emergency tor of the Federal Man Jasper Orchards; Arroyo & Stewart agement Agency; Army United States Farms, LLC; King Grove; Pistachio Corps Engineers; Thomas P. Bos Delta; for a Coalition Sustainable tick, Commanding General Chief County Family Agency; Kern Water Army Engineers, United States Alliance; Farm State Water Contrac- Corps Engineers, Defendants-Ap tors; Metropolitan Water District of California, Plaintiffs, pellees, Southern Council; Resources Defense Natural Bay Institute, Intervenor-

Defendants-Appellees. Contractors,

State Water

Plaintiff-Appellant, Jewell, Secretary

Sally Depart as of the Interior;

ment of United States Justice; Daniel M.

Department

Ashe, Acting Director U.S. Service; Fish

Fish and Wildlife U.S. & Service; Cowin, Di Mark

Wildlife

rector, Department California of Wa Resources; Department

ter California Resources, Defendants-Ap

of Water

pellees.

Metropolitan District Water California,

of Southern

Plaintiff-Appellant, Service; Sally

U.S. Fish & Wildlife Jew

ell, Secretary Department Interior; Ashe, Acting Daniel M.

Director of the Fish and Wildlife U.S.

Service; Bureau United States

Reclamation; Mcdonald; J. William Department Re

California of Water

sources; Cowin, Director, Mark Cali Department

fornia of Water Re

sources, Defendants-Appellees. 11-15871, 11-16623, 11-16662,

Nos.

11-16617, 11-16624, 11-16621,

11-16660. Appeals,

United States Court

Ninth Circuit.

Argued Sept. and Submitted 2012.

Filed March Ethan Oakley (argued),

Robert H. Car- Scott, R. United Eddy, Charles son Justice, .Environ- Department States Division, Resources & Natural ment Moreno, D.C.; S. Ignaeia Washington, Justice, As- Department States United General, Washington, Attorney sistant Monroe, In- D.C.; Department Jim Sacramento, Solicitor, terior, Office Defendants-Appellants. CA, for Federal (argued) Doug (argued), Deputy Katherine Poole and Clifford T. Lee Attor- General; Harris, Obegi, ney Attorney Natural Resources Defense Coun- Kamala D. CA; cil, Francisco, California; San Trent W. Orr and General Kathleen A. Kenea- Torgun, Earthjustice, General; George ly, Attorney M. San Senior Assistant Francisco, CA, Byrne, for Defendant-Interve- Supervising Deputy Robert W. At- General; Dennis, nors-Appellants. torney Cecilia L. Alexander, Gary Deputy Attorneys Gener- (argued), Hanspeter Daniel J. O’Hanlon al, Francisco, CA, San for Plaintiff-Inter- Walter, Akroyd, Kroniek, and Rebecca R. venor-Appellee, Department California Moskovitz, Girard, Tiedemann & Sacra- Water Resources. CA; mento, Manson, Craig General Coun- Amelia Minaberrigarai, County Kern sel, District, Fresno, CA; Westlands Water Bakersfield, CA, Agency, Water for Appel- Sims, Kales, Steven O. Michelle C. County Agency. lee Kern Water Williamson, Geoffrey Hyatt M. Brownstein Thornton, Weiland, Robert D. Paul S. LLP, Denver, CO, Farber Schreck for Remillard, Ashley LLP, Nossaman Ir- Plaintiffs-Appellees-Cross-Appellants San vine, CA, Appellees for County Kern Wa- Authority Luis & Delta-Mendota Water Agnecy ter and the Coalition for a Sustain- and Westlands Water District. Delta. able Gregory K. (argued), Wilkinson Steven ARNOLD,* Before: MORRIS S. Anderson, Cushman, M. Melissa R. RAWLINSON, JOHNNIE B. and JAY S. Martin, Steven Krieger G. Best Best & BYBEE, Judges. Circuit LLP, Riverside, CA, for Plaintiffs-Appel- lees State Contractors. Water BYBEE; Opinion by Judge Partial Concurrence and Partial Dissent Judge Scully,

Marcia L. Interim General Coun- ARNOLD; Partial sel; Concurrence and Masouredis, Linus Deputy Chief Gen- Partial Dissent Counsel, Judge RAWLINSON. eral Metropolitan Water Dis- California, Sacramento, trict of Southern OPINION

CA; Carr, Christopher J. William N. Sloan, Brandon, BYBEE, and Travis Morrison & Judge, Circuit with whom LLP, Francisco, CA, Foerster ARNOLD, San for Judge, joins Circuit as to Parts I, II, IV.C, Plaintiff-Appellee-Cross-Appellant IV.D, V.A, V.C, and with Metropolitan Water District RAWLINSON, of Southern joins whom Judge, Circuit California. except as to Part V.C.2: * Arnold, Circuit, The Honorable Morris Eighth S. sitting by designation, Senior Cir- Judge Appeals cuit for the U.S. Court of biological opinion it, invalidated these court put court aptly As district that conclud- Service Fish and Wildlife over “continuing war from the cases arise Valley and State Wa- the Central Luis San ed of the delta smelt.” protection exis- Salazar, the continued Projects jeopardize ter Auth. Water & Delta-Mendota fish and its habitat. (E.D.Cal.2010). a three-inch tence in part. affirm part reverse fray. The district We joined to the We are Valley Project brought and the al Central suit under consumers2 — Project, operated respectively State Water against Administrative Procedure Act vari- (Reclama- by the Bureau of Reclamation defendants, ous federal including Reclama- tion) California, are per- and State tion, FWS, Secretary and the haps largest important the two and most Interior, prevent the federal defendants projects water in the United States. implementing the biological opinion projects These supply combined water and its proposed alternatives. The district originating northern California to more court, lengthy in a comprehensive and 20,000,000 than agricultural and domestic opinion, deeply biologi- critical of the consumers central and southern Califor- opinion cal and concluded that it was arbi- water, estuary nia. The source of this trary capricious. and The court accused at the Bay confluence the San Francisco repeatedly “ignoring the FWS of best [the] and Joaquin (Bay- Sacramento-San Delta science available” to reach a “results-driv- Delta), is also the lone habitat for the delta en choice.” 760 smelt, species a threatened En- under the “show[ing] fully no inclination' to and hon- § dangered Species Act. 16 U.S.C. 1531 et estly supply beyond address water needs seq. species,” even as it “interdicts] In requested biolog- Reclamation supply water for domestic human con- ical opinion (BiOp) from the U.S. Fish and sumption agricultural use for over (FWS), Wildlife Service in accord with the twenty people million depend who on the (ESA), Endangered Species Act on wheth- Projects for supply,” their water id. at er operations its continued would jeopard- FWS). (quoting 956-57 ize the smelt. a more than 400-page opinion by the FWS as the —described acutely We are aware the conse- most complex biological opinion pre- ever quences court, of this proceeding. As a pared FWS concluded that the Cen- —the however, we are limited in our review of Valley operations tral would threaten the matters within expertise of an agency. and, required by delta smelt the Endan- mayWe review the biological opin- FWS’s Act, gered Species proposed “reasonable ion and implementation Reclamation’s and prudent alternatives” that Reclama- arbitrariness, caprice, or actions otherwise tion should take to ameliorate the effect on in accordance with law. 5 U.S.C. the smelt. The alternatives recommended 706(2)(A). § Although the FWS must em- by the FWS would reduce the water ex- ploy “the best ported scientific and commercial from northern California to south- available,” ern through 1536(a)(2), California data Valley the Central 16 U.S.C. it is Projects. State Water Reclamation “not required support finding that a has notified the FWS that it intends to significant risk exists with anything ap- operate Projects in compliance with proaching certainty,” scientific Indus. Un- the biological opinion. Inst., Dep’t ion v. Am. Petroleum 448 U.S. 607, 656, plaintiffs-appellees water S.Ct. 65 L.Ed.2d 1010 —various districts, (1980) contractors, water agricultur- And, (plurality opinion). “[w]hen District, Glossary 1. See the of Terms at the end of this Jasper lands Water Stewart & Or- opinion complete chards, for a list of abbreviations Department the California of Water acronyms. Resources, Metropolitan and the Water Dis- trict of Southern California. plaintiffs-appellees 2. The include the San Luis Authority, & Delta-Mendota Water the West- AND PROCEEDINGS I. FACTS determina- scientific this kind of examining BELOW generally court must reviewing ... a tion Baltimore Gas most deferential.” be at its Background A. NRDC, 87,103,103 462 U.S. v. Elec. Co. & Joaquin Delta The Sacramento-San 1. (1983). 2246, For L.Ed.2d 437 76 S.Ct. water devel history of California “The below, we conclude explained reasons story sup of distribution is opment and these to observe court failed the district critical wa ply and demand. California’s judgment. its reverse and we standards a lack of water but problem ter is not im- practical enormous recognize the We resources.” distribution of water uneven But the conse- this decision. plications v. Res. Control States State Water United Congress Bd., 82, 98, when prescribed Cal.Rptr. were 227 quences Cal.App.3d 182 fish, (1986). Valley species of California’s Central that “these 161 determined productive esthetic, comprises some of most wildlife, ecologi- are of plants Extending 450 in the world. recreational, farmland historical, educational, cal, south, averaging north to miles from Nation and and scientific value west, the Central 100 miles wide east 1531(a)(3). theAs 16 U.S.C. people.” rivers: The principal two Valley includes in Tennessee observed Supreme Court in the northern begins River Sacramento may “It seem Valley Authority v. Hill: valley, past Sacra runs south part of the of a the survival to some curious mento, Feather and is fed fish of three-inch relatively small number Joaquin River Rivers. The San American halting require permanent ... would Nevadas, northeast of in the Sierra begins dam,” “the but virtually completed aof Fresno, into the and northwest runs west Spe- Endangered explicit provisions by smaller riv Valley, and is fed Central precisely that result.” require Act cies Calaveras, Chowchilla, ers, including the 2279,' 153, 172-73, 57 98 S.Ct. 437 U.S. Mokelumne, Merced, Fresno, Stan- Rings, (1978). have species 117 Such L.Ed.2d islaus, The two Tuolumne Rivers. highest priorities,” “afforded the been in the converge Sacramento-San rivers sacri- if it means “the Congress, even estuary that and form an Joaquin Delta pro- benefits of the anticipated fice of the Bay, and Bay, San Francisco joins Suisun millions of dollars ject many and of per Although over Ocean. the Pacific 174, at 98 S.Ct. public funds.” Id. originates north water of California’s cent omitted). (footnote us prohibits law Sacramento, percent of more than 70 utilitarian making “such fine calcula- is in the south. demand the state’s the smelt’s interests tions” balance irrigation region supplies this water from of the citizens of Cali- the interests against agriculture million acres of for seven 187, 2279. 98 S.Ct. Conse- fornia. Id. nearly twenty million people, more than “[rjesolution these any quently, other See Where residents. half of California’s questions” about policy From?, fundamental Water Come Does California’s resources California allocation of water Water News Aquafornia, The California agencies and the Congress a.m., ... with 13, 2008, http://www. “lies 9:29 Blog, Aug. authority, delegated Congress has to which aquafornia.com/index.php/where-does- and, ulti- general See legislatures as well as with state californias-water-come-from/. 609, 612-13, Rank, Balti- 372 U.S. ly Dugan as whole.” v. mately, populace (1963); United 999, Elec., 97,103 10 L.Ed.2d S.Ct. S.Ct. 462 U.S. more Gas & Co., 339 U.S. Live Stock v. Gerlach States 2246. 725, 728-29, 955, 70 S.Ct. 94 L.Ed. 1231 approved California what (1950); Project (SWP), known as the San Luis & Delta-Mendota State Water Water States, largest state-built project water Auth. United 681— United States. (9th San Luis & Cir.2012); Delta-Mendo Westlands Water Dist. Auth., ta Water 672 F.3d at Managed States, v. United 337 F.3d 1095-96 by the Department of Water California (9th Cir.2003); Bay-Delta In re Program (DWR), Resources “[t]he SWP serves the Impact Report Envtl. matic Coordinated domestic water approximately needs of Proceedings, 1143, 43 Cal.4th 77 Cal. Californians,” two-thirds all of principally 578, 709, Rptr.3d (2008); 184 P.3d 715-17 in Southern California. In re Bay-Delta, Bd., State Water Res. Control 182 Cal. 184 P.3d at 716. SWP “21 consists of 97-100, App.3d at 227 Cal.Rptr. 161. reservoirs, dams and power ... five In an effort to manage plants, the pumping plants, 16 increasing and 662 miles of aqueduct.” conflicting and Id. 716 n. placed 1994, eight at 2. demands the wa- agencies state and 10 federal flowing agencies ter through the Sacramento-San formed Bay-Delta the CALFED Program Joaquin region, Delta California and the (CALFED) to address comprehensively United States have embarked on two mas- the challenges managing Bay-Delta of the First, projects. 1933, sive California estuary. Id. at 717. proposed the Valley Project Central (CVP), plan The and operate transfer water from CVP SWP each the a ma- jor station for pumping Sacramento water from River to the water-deficient areas Bay-Delta to aqueducts canals and the Joaquin Valley San and from the carry will to the south. water Both Joaquin San River to regions the southern plants California, are near Tracy, located of Valley. the Central State Water Res. together they reverse the natural flow Bd., 98-100, Control 182 Cal.App.3d at 227 of the part southern Bay-Delta Cal.Rptr. 161. Reclamation took over the through two distributaries of the San Joa- project 1935, and it is largest now “the Old, quin, Rivers, and Middle referred to management federal water project in the Deltob-Mendota, as “OMR.” San Luis & United States.” Central Delta Water 760 863. operates at The CVP Agency States, 938, United Plant, the Jones Pumping capable of di- (9th Cir.2002). The CVP consists of a 4,600 verting (cfs). per feet cubic second dams, series Shasta, of including Folsom, Nearby, the operates SWP Harvey O. Dams; and Nimbus reservoirs; hy- Plant, Banks Pumping with a capacity of dropower plants; and 500 of miles canals 10,300 cfs, although generally operates aqueducts. Bay-Delta, In re 184 P.3d 6,680 at or 82, below BiOp 108, cfs. at at 716 n. 1. In Congress adopted the 159-60. plants have been constructed Valley Project Central Improvement Act with louvers that allow pass water (CVPIA), 102-575, Pub.L. No. 106 Stat. through into pumping plant, will but 4706, which Congress described as de- prevent most from entering fish the plants. signed “to achieve a reasonable balance The process of entering the fish the plants, among competing demands for use Cen- entrainment, known as traps some dif- Valley Project water, tral including species ferent of fish. BiOp at 67. requirements wildlife, of fish and agricul- salvaged fish are hauled in injected trucks tural, municipal and industrial power with oxygen and released at sites on the CVPIA, § 3402(f), contractors.” 106 Stat. Sacramento Joaquin San Rivers. at 4706. BiOp at 145. Over a 15-year recent Bd., 182 Control Resources Water were State million fish than 110 more period, Cal.Rptr. Cal.App.3d at facili- Banks the Jones salvaged from 1970s, and a raft however, Reclamation number, Since This BiOp at 160. ty. mitigate have worked agencies of fish state number underestimates greatly salinity on Sui- facilities, water because increased effects actually entrained upstream di inches) resulting from (1.2 Bay are not sun mm30 fish less than otherwise nat that would water louvers. version efficiently collected at 112- Suisun. urally through flow fish, especially those Smaller at 160-61. Bay high are Salinity in Suisun levels killed in 13. are stage, larval juvenile or the Delta. to diversion ly sensitive Those pumps. the sal- not survive frequently do salvaged de are used to standards Two related at 338. process. vage The Bay-Delta. salinity of the the scribe Salinity Zone or LSZ. Low is the first have the River and SWP The Colorado is the transition BiOp at 147. The LSZ supply of water major the historically been the inland the freshwater between point .of As the result California. for southern east flowing and brackish water 2003, rivers in signed agreement an interstate. the Bay Francisco from San ward from the less water will receive California Ocean, ranging water and includes Pacific Quantification River. Settlement Colorado per thousand salinity parts from 0.5 Authority, Water Diego San Agreement, BiOp at The thousand. 191. parts per six http://www.sdcwa.org/quantification- represents X2. X2 to as is referred (last second July visited settlement-agreement the at Bay-Delta the which point in the 2013). Cali- consequence, southern As a parts per thou than salinity is two less SWP. sought more water from fornia has Dep’t Dist. v. Water See Westlands sand. for wa- demand at 89-90. Increased (9th Interior, 876 Cir. effect predictable a has had ter from SWP 2004). LSZ, a encompasses which The Bay-Delta. the flowing into the water general Bay-Delta, is region of the larger from the rivers more is diverted As water these Together, X2. around ly centered the southern Cen- the Delta into that feed Bay- by determined largely are regions its salinity of the Delta the Valley, tral be outflow, the difference which is Delta threat to with the along increases estuaries exported. water and the the inflow tween there. that thrive species the marker for the X2 as a use agencies large describes generally The Delta fish indicator for a habitat as well as LSZ natural labyrinth lowland area BiOp at standard. regulatory as a the confluence in and around channels Delta-Mendota, 149-50, 236; Luis & San Riv- Joaquin San the Sacramento They nn. 4-5. & ers.... of X2 as distance the location express qual- affecting water major factor Bridge, Gate east of Golden kilometers Delta saltwater intrusion. ity in Dist., 376 F.3d at Water Westlands Lands, or sea below situated Delta subject ocean level, constantly smelt 2. The delta entering from Salt water tidal action. transpaci- (Hypomesus smelt The delta into the well Bay extends Francisco

San small, species inch two-to-three is a ficus) tidal Delta, of the saline and intrusion San Francisco of fish endemic natural only is checked waters Estu- Joaquin Delta Bay/Sacramento-San flowing fresh water barrier formed Once abundant 140-41. ary. BiOp at the Delta. out from species Bay-Delta ecosystem, formally or informally consult with the delta smelt is now in imminent danger of Secretary Interior, or his or her species extinction. In March delegee. § 1536(a)(4); 16 U.S.C. ESA, listed as threatened under the 402.14(a); Rivers, § C.F.R. see Am. designated BayDelta and the FWS found, F.3d at 1122. If no effect is consul- system a critical habitat for delta smelt tation is required. § 50 C.F.R. 402.14. 17.11; § 1994.3 50 C.F.R. at 140. Formal required consultation is when the Yet, decade, over past delta smelt acting agency consulting or agency deter- has been population decimated rela- even mines proposed likely action is levels, to these depleted tive with a meas- adversely affect species a listed or critical up ured decline since 2000 of to three 402.13, habitat. §§ 50 C.F.R. 402.14. of magnitude orders below historic lows.4 Formal requires consultation the consult- Delta-Mendota, San Luis & ing FWS, agency, here the to issue a bio- F.Supp.2d at consequence, 866. As a logical opinion stating pro- whether announced in 2010 reclassifying posed likely action is jeopardize such the delta smelt from a threatened to an species 1536(b); § or habitat. 16 U.S.C. endangered species was warranted but 50 C.F.R. 402.14. Should the jeop- action precluded by higher priority listings.5 Id. species habitat, ardize the or the consult- provides The ESA “both substantive and ing agency suggest any must “reasonable procedural provisions designed protect to prudent (RPA) and alternatives” endangered species and their habitat.” would projects allow the to opera- continue Am. Rivers v. Nat’l Marine Fisheries tion without causing jeopardy to spe- Serv., (9th 1118, Cir.1997).6 126 F.3d 1121 cies'or adverse modification to its critical 7(a)(2) protection, § One such ESA, 1536(b)(3)(A). § habitat. 16 U.S.C. Once requires agencies federal to “insure that it the "BiOp, acting receives agency authorized, any funded, action or carried “shall determine whether what out agency such ... likely is not proceed manner to with the light action in jeopardize the continued any existence of its section obligations 7 and the [FWS’s] endangered species or species threatened biological opinion.” 402.15(a). § 50 or C.F.R. result in the destruction or adverse If, consultation, after agency modification of thé habitat of deter- species.” such 1536(a)(2). § 16 U.S.C. mines it comply 7(a)(2), § cannot Should the with proposed find that its may action it may affect a apply exemption, can which species listed or critical habitat, only must be authorized the Endangered 1990, 3. the National Marine Fisheries Ser- 5. The assigned delta smelt prior- a listing vice listed the Sacramento River winter-run ity 12, of 2 on a scale being from 1 to with 1 chinook salmon species as a threatened under highest priority. San Luis & Delta-Men- Endangered Species 50 Act. C.F.R. dota, F.Supp.2d 760 at 866-67 n. 6. 227.4; States, § 596, v. United 545 U.S. Orff 599, 2606, (2005); 125 S.Ct. 162 L.Ed.2d 544 Among 6. things, other the CVPIA instructs States, (9th O'Neill v. United 681 Secretary "operate of the Interior to Cir.1995). This case does not involve the Valley Project Central to meet obligations all salmon. law, under State and Federal including but Endangered Spe- limited Federal 4. The 2008 population delta smelt was esti- Act, CVPIA, cies § 16 level, seq.” U.S.C. 1531 et mated at' Fed.Reg. 1.5% the 1980 75 3406(b), § (April 2010), 106 Stat. at 17667 4714.. and 2009 levels were estimated to be the lowest record. San Delta-Mendota, Luis & F.Supp.2d at 866. findings of fact im remedial order and Committee, com- panel an ad hoc Species covered, law, among and at which members conclusions of branch of executive posed in which the state appointee from on delta smelt of things, least one other the effects n 16 U.S.C. is to occur. project in OMR. See San Luis & negative flows 402.15(c), 1536(e); §§ 451. § 50 C.F.R. Delta-Mendota, F.Supp.2d district court ordered Reclamation Leading To Present Proceedings B. implement “pulse a winter and DWR Controversy - negative of no more than flow” OMR 2,000 cfs, Biological Opinion “operate the CVP and 1. The FWS’s daily average up- a net SWP to achieve opinion biological sought Reclamation (reverse) flow in the OMR not to stream of its continued part FWS as from the -5,000 running seven-day cfs on a exceed and its of the CVP long-term' operation during period a defined average agencies operations state coordinated Id.; Int. Rem. Order at 5-7. spring.”8 Following 8.7 BiOp at SWP. also ordered the The district court 7(a)(2) subsequent formal and a review months, consultation, biological just nine complete issued a the FWS new (2005 BiOp). ultimately The 2005 in 2005 that it would extend opinion a deadline Delta-Mendota, proposed coordinated BiOp found year. San Luis & to one and SWP would not of the CVP operations 865; Int. Rem. Order at on the continued adverse effect have an recovery of the delta smelt existence a new delta smelt The FWS issued bio- Del- San'Luis & critical habitat. deadline, logical opinion on the December ia-Mendota, at 863. The F.Süpp.2d (2008 1- BiOp BiOp).9 BiOp or at Resources Defense Council—de- Natural BiOp, to the 2005 396. In stark contrast in fendants-intervenors-appellants BiOp concluded that the “coordi- the 2008 con- challenged the FWS’s present case— SWP, as operations of the CVP and nated the East- in District Court for clusion U.S. likely jeopardize the con- proposed, are California, and the court of ern District of the delta smelt” tinued existence and arbitrary capri- and BiOp the 2005 found modify smelt critical “adversely delta habi- F.Supp.2d at 387. Kempthome, cious. respect to BiOp 276-78. With evidentiary tat.” conducting extensive After smelt, findings five the FWS entered an inter- delta court hearing, the district issued subjected BiOp, preparing this the FWS 9.In precipitating event for Reclamation Team, Oper- seeking biological opinion was a 2004 Review consist- a it to an Internal Peer (OCAP), reflecting ating development complex Criteria and Plan experts of ing in the of facilities, delivery require- changes water ESA biological opinions under from ments, regulatory and in- and restrictions FWS, throughout from a team comments creasing between coordination federal FWS experts from within the delta smelt Kempthome, agencies. NRDC v. state agencies, the review of PBS and other related (E.D.Cal.2007). firm, J, consulting as & an environmental Reclamation and the as from well feedback Bay-Delta water is diverted from the 8. When response BiOp In to the at i-vi. DWR. See course, contrary natural amount to its received, the FWS significant feedback expressed nega- exported usually as a water BiOp. changes See made substantial plants pumping re- because the tive number ii. example, Banks For OMR flow. if verse 2,500 per sec- facility pumping cubic feet Aqueduct, noted it is ond to the California -2,500 as cfs. (1) of fact: “Diversions of water from the smelt stage by life controlling OMR flows during Delta have increased since 1967 when the the vulnerable December to March conjunction began operation period. SWP It proposed has two actions. Ac- “designed the CVP.” at 276. The tion 1 is protect upmigrating CVP/SWP smelt, operations including have entrained delta smelt” and describes periods the two adults, larvae, juveniles, at the Banks when delta smelt are most vulnerable to facilities; habitat; and Jones reduced smelt entrainment: in December and when the outflows, altering and reduced the Delta appears. BiOp first flush at 280-81. Ac- (2) LSZ; location of “The delta tion 1 proposes limiting neg- therefore currently at smelt is its lowest level of ative “daily flows at OMR based on a monitoring abundance began salvage effect, since index.” Id. this means (3) 1967”; the proposed “Under that when “daily salvage index” reach- CVP/SWP (“the operations, inflows to the Delta likely threshold”), es a critical point risk reduced, to be further Projects water demands have to reduce their diversion increase, upstream of the Delta days. most nota- for 14 During period, OMR bly on the American River.” -2,000 flows can be negative “no more than likely This is to “cause increased relative a 14-day running average cfs” for and “no -2,500 entrainment of adult delta smelt in negative the more 5-day than cfs” for a spring, juve- winter and and of larval and running average. BiOp at 329. Ac- (4) spring”; nile delta smelt in the “Other tion follows from Action 1 but covers the baseline stressors will continue to period March, adverse- through December smelt, ly affect the delta such as contami- when- adult pre-spawning delta smelt are nants, microcystis, aquatic macrophytes, vulnerable to entrainment. BiOp at 352. (5) recover, species”; invasive To delta During period, OMR flows can be no smelt will need a more abundant -5,000 cfs, adult more negative than although population, an increase in the quality and the expected that flows gener- would quantity spawning, rearing, migra- ally -2,000 -3,500 be in range cfs to habitat, tory reduction contaminants cfs. pollutants, exposure reduction in (Action 3). RPA Component 2 Compo- blooms, algal disease and toxic and a re- 2 protects nent juvenile larval and delta duction in entrainment at water-diversion by limiting smelt OMR flows following the in Bay-Delta. BiOp facilities at 276- completion Component of 1 Bay- when the respect 77. With to delta smelt critical Delta temperatures 12°C, water reach or habitat, “past FWS found that and spent when a female smelt is detected in present operations of the have CVP/SWP trawls at Jones or Banks or is found in the degraded these habitat elements” such salvage requires facilities. Action 3 they that support are “insufficient to suc- projects to maintain their aver- CVP/SWP cessful delta smelt recruitment at levels age -1,250 -5,000 OMR flows between and provide

that will species’ for the conserva- 30, cfs until June or until the mean water BiOp tion.” at 278. level, temperature a target reaches which- The FWS compo- recommended five ever occurs earlier. BiOp at 357-58. nents and listed six separate actions as (Action 4). RPA Component 3 Compo- “reasonable and prudent alternatives” nent improves smelt habitat increas- (RPA): ing Bay-Delta during outflow the fall. (Actions 2). Component

RPA requires and Action 4 that in September and Component protects October, the adult delta in years when precipitation minimized when would above be “wet or entrainment is defined as runoff and according to regulated OMR flows were pro- DWR must normal,” and Reclamation BiOp at 285-86. proposed RPA. the FWS’s to maintain Delta outflow vide sufficient the FWS concluded that consequence, aAs km from the than 74 more eastward X2 no likely anticipated of take is not “this level 81 km years and wet Golden Gate species to the or jeopardy result 282-83, years.10 above-normal of crit- or modification destruction adverse 369. implement- when the RPA is ical habitat 11). (Action 6 Com Component RPA k ed.” by establishing a habitat 4 restores ponent and intertidal to create restore program or present 2. The case Bay- habitat subtidal associated complaints challenging first of six 6 re Marsh. Action and Suisun Delta filed in March FWS’s 2008 restore at least to create quires DWR or Delta-Mendota, 760 San Luis & 2009. 8,000 Marsh. in the Delta and Suisun acres “Plaintiffs moved for a F.Supp.2d at 865. BiOp at 283. injunction prevent ... Rec- preliminary Component 5 moni- Component 5. RPA implementing Component lamation implementation, reports on the tors and RPA, violated the alleging FWS of success, possible improvements Policy Act National Environmental 1-4.12 Components (“NEPA”), the ESA.” Id. district finding granted, part, court motion “incidental Finally, the FWS issued likely to suc- plaintiffs-appellees were (ITS) with 50 in accord statement” take claim, NEPA on the merits their ceed ITS, purposes For 402.02. C.F.R. specific to make requiring FWS that its reasonable presumed justify weekly decisions findings to written implement- would be alternatives prudent Id. OMR flow restrictions. regarding the FWS premise, Based on ed. preliminary that, opera- Plaintiffs-appellees sought as a result of found CVP/SWP implementation injunction against delta tions, a take of the there would be Following an evi- 3. Component the RPA Id. smelt, the- extent of although and that district court issued estimate, dentiary hearing,13 smelt would be difficult take permits which eral Rule of Evidence that “there the FWS noted In Action parties biological "appoint any expert ways achieve the may be court to other action,” choosing.” that it would evalu- any Fed. goals agree of this own on and 4 "consis- modify Action 706(a). pro- alternatives and appointed ate two The court R.Evid. action.” The this tent with the intention of Washington, University of from the fessors "adaptive process as an to this Quinn, referred to Thomas Punt and Dr. Dr. Andre BiOp management process.” at 283. complex and technical the court on advise the appointed court also matters. The scientific see, infra, 12. 5 Note For Action 11. experts: a Reclamation em- two additional CVP, and a knowledgeable on the ployee separate ac- Component did have a not 12. knowledgeable SWP. employee in the item, formally DWR and Action 5 was not tion Components. any RPA associated court-appointed ex- addition to these In of a specific the installation 5 was Action permitted substantial perts, the district court affected on the Old River that physical barrier (but experts from selected declarations juvenile delta smelt. of larval and entrainment to) deci- parties. In its written agreed by the BiOp at 377-78. sion, extensively on relied the district court by its submitted own opinions and evidence hearing, the district evidentiary At the 13. experts. parties’ experts by the and pursuant Fed- appointed experts four court injunction a preliminary confirming capricious and not the result of the best plaintiffs-appellees had succeeded on their available science. Id. at 968. These calcu- NEPA finding plaintiffs- claims and significantly lations influenced the upper appellees likely were to succeed on the 2, and lower OMR 1, flow limits in Actions merits of their claim. ESA Id. and 3. In December the district court Second, the district court found that the entered final judgment on the primary BiOp’s models, use of two different CAL- 115-page opinion. claims a Id. at 967- DAYFLOW, SIM II and predict Although the FWS’s BiOp X2, location of introduced bias requiring antipodal reached conclusions to the 2005 or, least, corrective very calibration at the BiOp a no jeopardy reached con- —which explanation. Id. The district court also clusion arbitrary and was found capri- found that the produced by bias the com- cious—the again district once court found parison of II CALSIM to DAYFLOW to be arbitrary capri- tainted the BiOp’s justification for Action cious under the ESA and the APA and 4, which management involves the of X2. RPA, remanded the BiOp, its and Recla- addition, the district court found that provisional mation’s acceptance of the RPA did not sufficiently explain itwhy agency. San Luis & Delta-Mendo- (Action 4) is essential in Component 3 ta, F.Supp.2d 970. The court’s maintain X2 specific at the locations of 74 required remand completion yet upstream km Bridge the Golden Gate third analyzing impact of CVP following years” “wet km and 81 following and SWP operations on the delta smelt. “above-normal years.” Id. at 969. Delta-Mendota, San Luis & Third, the district court found that the In March BiOp did not sufficiently explain why dif- district court entered final judgment on all ferent data sets were used to calculate the remaining claims. incidental juvenile take limit for and for Although the accepted district court smelt, adult why these limits were BiOp’s central “Project op- conclusion that using calculated average previous erations likely are to jeopardize the contin- (which years’ smelt salvage would be ex-

ued existence adversely modify the and/or pected to be exceeded 50% of all future critical smelt,” habitat of delta id. at years). 969, the district court determined that Fourth, the district court found the there were a number of specific flaws with BiOp did not adequately support its con- BiOp, id. at 967-70. briefly We will Project clusions that operations are set rea- forth the district court’s principal ob- sonably certain indirectly affect jections here —which highly technical delta smelt limiting delta smelt food and somewhat out of obtuse context—and supply, by increasing harmful pollution explain them in detail more in the discus- contaminants, increasing the sion section. detrimental impact of the “other stres- First, the district court found BiOp’s predation, sors” of macrophytes, and mi- reliance on analyses using raw salvage fig- crocystis on delta smelt. Id. e.g., those incorpo- calculations that ures — rated the absolute Fifth, “raw” number of smelt the district court held that stations, entrained in pumping opposed as BiOp failed to analyze feasibility, economic to the smelt entrained as a percentage of consistency with purpose action, of the the total population arbitrary be consistency with the action agency’s —to

601 sup the ESA nor NEPA Neither § 402.02. See 5 required by authority, as review, our so separate standard for ply § 402.02. 50 C.F.R. seq; § et 551 U.S.C. these Acts under review claims under we at 969-70. Id. of the APA. Bennett v. the- standards the NRDC Defendants Federal 1154, 154, 174, 117 S.Ct. Spear, 520 U.S. reversal of urging timely appealed, have (1997); Natural Oregon L.Ed.2d 281 137 address remand. We court’s the district Mgmt., Ass’n v. Bureau Land Desert in Part IV. their claims (9th Cir.2010); 1092, Pyra F.3d 1109 625 Water & Delta-Mendota The San Luis Indians v. U.S. mid Lake Paiute Tribe of timely appellees other Authority and (9th 1410, 1414 F.2d Dept. Navy, 898 the district arguing cross-appealed, 706(2) Cir.1990). pro APA Section They enough. raise go not far court did upheld agency action must be vides that an First, violated that the FWS three claims: “arbitrary, capri unless it is on review nondiscre- by separating not out the ESA discretion, cious, or otherwise an abuse of discretionary actions tionary actions 5 in accordance with law.” U.S.C. not baseline; sec- environmental setting in 706(2)(A). court, reviewing we As a arbitrarily ond, acted Reclamation whether the decision was “must consider the flawed adopting capriciously the relevant on a consideration of based third, and, FWS failed to BiOp; there has been a clear factors and whether by NEPA.14 required review conduct the judgment.” Citizens to Preserve error of V. their claims Part address We 402, Park, Volpe, 401 Inc. v. U.S. Overton history has (1971), ex- procedural 416, 814, This tortured 28 L.Ed.2d 136 91 S.Ct. years, and has led grounds over seven abrogated part tended on other Sanders, par- suits and one fully consolidated v. recognized five in Califano brought by 99, 105, various suit 51 L.Ed.2d tially consolidated U.S. S.Ct. (1977). by the must be supplied Although inquiry use water our groups who SWP, highly comple- standard of review is thorough, as well as CVP and deferential; BiOps— agency’s decision is “enti extensively researched tion of two regularity,” currently progress. presumption All the tled to a awith third judgment for may substitute our while, on the we delta smelt has teetered 415-16, 91 S.Ct. agency. that of the Id. brink of extinction. has relied on “rel agency 814. Where Defendants and agree with Federal We a reasonable [such that] evant evidence court’s and reverse the district the NRDC adequate support might accept as mind BiOp. affirm the district remand of the We conclusion,” supported its decision is failed that Reclamation judgment court’s Watch, Lake evidence.” Bear “substantial implement- NEPA before comply with (9th FERC, F.3d v. Inc. BiOp. ing the FWS’s Cir.2003). is sus the evidence “[i]f Even inter than one rational ceptible of more REVIEW OF II. STANDARDS uphold [the must pretation, court] [the rul the district court’s We review findings.” Id. agency’s] summary judgment de novo. ing on ESA, must Under Kempthome, McFarland Cir.2008). supported (9th actions on evidence base its ap- this direct quired by NEPA. We address also and the NRDC 14. Federal Defendants appeal. appellees’ cross appeal peal together court’s determination the district re- to conduct review Reclamation failed 602 Pitts, reviewing

“the best scientific and commercial data court.” v. Camp 411 U.S. 138, 142, 1241, § 402.14(g)(8); available.” 50 C.F.R. 16 93 S.Ct. 36 L.Ed.2d 106 (1973). 1536(a)(2). The U.S.C. determination of

what constitutes the “best scientific data If agency the record before the does belongs agency’s “special available” action, support agency if the expertise.... examining When this kind of agency has not considered all relevant determination, scientific opposed to sim- factors, reviewing or if the court simply fact, ple findings reviewing court must challenged cannot evaluate the agency generally be at its most deferential.” Bal- action on the basis of the record before Co., 103, timore 462 Gas & Elec. U.S. at it, course, proper except in rare data[,] superior 103 S.Ct. 2246. “Absent circumstances, is to remand to the agen- imperfections occasional do not violate” the cy for investigation additional or expla- ESA best available standard. Kern Cnty. nation. reviewing gen- court is not Allen, 1072, Farm Bureau v. 450 F.3d erally empowered to conduct a de novo (9th Cir.2006). 1080-81 inquiry being into the matter reviewed and to reach own its conclusions based “The best available data re on such inquiry.... quirement ‘merely prohibits agency] [an factfinding capacity of the district disregarding available scientific evi court is typically unnecessary thus way dence that is in some better than the ” judicial review of decisionmak- Cnty. evidence relies on.’ Kern [it] Farm ing. Bureau, 450 (emphasis F.3d 1080 add Lorion, Light Florida Power & ed) Co. v. 470 (quoting Biological Sw. Ctr. Diver for 729, 744, 1598, U.S. 105 S.Ct. 84 L.Ed.2d Babbitt, (D.C.Cir. sity 58, v. 215 F.3d 60 (1985). See also Citizens to 2000)). Preserve “Essentially, FWS ‘cannot ignore ” Park, 419-20, Overton 401 U.S. at 91 S.Ct. available biological information.’ Id. 1080-81 (quoting Burford, Conner v. (9th Cir.1988)). 1441, Thus,

F.2d “in- danger There is a when a re sufficient ... incomplete [or] information viewing goes court beyond the record be ... does not agency’s] excuse failure [an fore the agency. “When a reviewing comply statutory requirement with the of a court considers evidence that was not be comprehensive biological opinion using the fore agency, inevitably leads the best information available” where there reviewing court to substitute judgment superior some additional information Asarco, for agency.” that of the Inc. v. Conner, available. 848 F.2d at 1454-55. (9th Cir.1980). 616 F.2d EPA hand, On the other where the information See Biological Sw. Ctr. Diversity v. available, readily is not we cannot insist on Serv., (9th U.S. Forest 100 F.3d perfection: “[T]he ‘best scientific ... data Cir.1996) (“Judicial review of an agency ” available,’ does not mean “the best scien- typically decision focuses on the adminis tific data possible.” Building Indus. Ass’n trative record in existence at the time of Norton, (D.C.Cir. the decision and encompass any does not 2001). part of the initially record is made court.”). reviewing Accordingly, we III. THE SCOPE OF THE RECORD do not review “the evidence to determine Our review is limited to “the ad the correctness or wisdom of the agency’s already existence, ministrative record ... decision even if the court has also ex not some initially new record made in the amined the administrative record.” Asar-

603 material in the highly the technical reviewing explain co, 1160. If the F.2d at 616 BiOp. evidence find substantial court cannot for record, compensate “not it should

the however, court, did The district by undertaking its agency’s dereliction the court-appointed not limit itself to the ex id., merits,” but into the inquiry own objection of the vigorous Over the perts. further for agency remand to the should multiple the admitted appellants, court Ventura, 537 U.S. INS v. proceedings, see multiple experts hired declarations from 353, 12, 16, 154 L.Ed.2d 123 S.Ct. though party- the appellees, the even (2002). many of the appointed experts addressed however, have, ex- crafted narrow “We by the court- being issues same addressed rule.” Lands general to this ceptions By government’s the appointed experts. (9th 1019,1030 Powell, v. 395 F.3d verify, Council sought we not count—which have Cir.2005). explained, As we have have the declara although we examined record —the district court ad tions the administra- expansion of the allow [w]e forty expert declarations mitted more than narrowly construed in four tive record once the court denied appellees; from the (1) supplementation circumstances: to exclude the decla appellant’s the motion agency has necessary if the to determine rations, their appellants the submitted own explained all factors considered Yet, see we cannot experts’ declarations. (2) decision; agency relied on docu- parties’ experts added what (3) record; supplemen- not in the' ments court-appointed could have experts explain technical is needed tation court. reasonably provided to the district (4) subjects; plain- or complex or terms Hodel, Animal See Council Def. part on the have shown bad faith tiffs Cir.1988) (9th (plaintiff F.2d agency. why district court to show need failed Forest v. U.S. Fence Creek Cattle Co. record). the administrative go ed to outside Cir.2010). (9th Serv., was, to cre predictably, of this The effect con- Supreme in mind the Court’s Keeping Moreover, it experts. a battle of the ate reviewing factfinding, we with court cerns in the district court gave proceedings with exceptions these approached have the administrative appearance caution, ... undermine exception lest “the proceedings and that the open record Council, 395 rule.” Lands general merits of debating for were a forum F.3d Growers’ Ass’n BiOp. See Ariz. Cattle F.3d Wildlife, 273 v. U.S. Fish & concerns that the have serious We Cir.2001) (9th (“Considering evidence these rules. failed to observe district court would render the record ex First, four ex outside appointed court the district process traordinarily complex consultation understanding the techni to aid it in perts we will not meaningless.”). Just as ... BiOp. aspects scientific cal and post-hoc ration agency supply allow consistent experts appointed were These actions, “post-decision for its alizations so 706. See Federal Rule Evidence ... not be advanced as may information BiOp, we Having 13. read supra Note or sustaining for new rationalization either court’s need sympathetic district Sw. Ctr. agency’s decision.” attacking an party No has interpreter. for scientific Diversity, 100 F.3d Biological court’s objected appeal to the district cautioned, specialists “[w]hen reason As the Court we see no appointments, and can views, must conflicting express experts objection to the use able rely experts on the marginal pur- have discretion to reasonable and “served some opinions qualified experts pose”: of its own even

if, matter, might a court original as an find however, of expert testimony, Most the contrary persuasive.” views more Marsh should not have been admitted or at Council, 360, v. Ore. Natural Res. 490 U.S. least not have been considered for the 378, 1851, 104 109 S.Ct. L.Ed.2d 377 purpose judging the wisdom of the (1989). stack-testing requirement. EPA’s This testimony plainly technical was elicited quick

Even a review of the district purpose determining for the the sci- opinion appearance court’s shows that the entific merit of the EPA’s decision. Al- open reality. of an record was the though testimony may the have served extensively district court relied on the dec- marginal purpose allowing some in parties’ experts-as-advo- larations of the district court to evaluate the EPA’s rejecting BiOp. cates as the basis for court of inquiry, only we can conclude See, Delta-Mendota, e.g., San Luis & that the extent inquiry of the scientific 877-79, 881-84, 889-90, F.Supp.2d at 894- undertaken at trial necessarily led the 97, 903-07, 912, places, 922. In the district district court to judgement substitute its pits experts against court each other agency. for that of the contrary their positions resolves as a (9th Cir.1980). E.g., 1153,1160-61 matter of scientific fact. San Luis & 616 F.2d Delta-Mendota, 884, similarly The district court overstepped effect, 904-07. In the district court its bounds here. we Because review the opened post-hoc to a notice-and- novo, however, judgment court’s de we can proceeding comment involving parties’ scope confine our own of review to the experts, judged and then the BiOp against record, plus administrative that evidence received. The comments ESA consul- that satisfies the standards we have set process tation is not a rulemaking proceed- AgroSciences forth here. See Dow v. Nat’l ing, a request.from but one for the Serv., Marine Fisheries expertise agency. of a second Although (4th Cir.2013) (confining review to the ad- may we each of review those proceedings ministrative disregarding record and APA, under the agency’s obligations court). affidavit submitted to the district different, under each slightly and we We will consider the BiOp any other must account for that difference our evidence the record submitted review. parties that the FWS considered in mak- ing its decision. We will also consider the . Our opinion Asarco v. EPA addressed testimony of the four experts' the district question using experts a technical court appointed pursuant to Rule point case, case and is on In here. to whom objection. there is no the district court a had technical matter 1,000 involving before a foot stack at a it— IV. MERITS OF THE permitted parties smelter —and to call experts. their days The court held four Before we consider the challenges to the hearings in which at BiOp, least ten witnesses we preliminary have some observa- First, testified. We held that.some of the testi- tions. is a bit of a mess. mony, mainly experts from provided just mess, but, who And not a little bit of a “background material” on the op- smelter’s more than pages, big bit of a mess. rest, - erations, appropriate; of it And the FWS knew it. November was not appropriate, even if it came shortly before the FWS submitted [BiOp] Draft are extensive but are correct- draft, Department the California final por- able before -the December 15 court ordered commented on Resources of Water for its finalization. We are will- the title [deadline] draft. Under tion of the FWS’s closely Disorga- ing to work with the FWS to ad- Confusing and Document is “The quickly....” these concerns And the nized,” that “the docu- dress advised the DWR that its own peer panel complained text is review unrefined. The especially ment ..., four-day period “review was conducted in a many actions logical order often out Ass’n tight is un- under a schedule.” See Pac. and the document fairly vague, (9th EPA, wrote, .615 F.2d 811 DWR Fisheries portions, Some polished.” Cir.1980) (“The Agency recognized itself The FWS “largely unintelligible.” were thorough review that its -data collection was not peer its draft to a also submitted engineer- as it otherwise would have been: ‘The time by an outside coordinated panel, imposed by statutory dead- questions the FWS constraints firm. One of the ing Agency from conduct- BiOp precluded was lines panel was asked whether clear, ing sampling program.’”). con- an exhaustive in a manner that is “organized (i.e., challenging This deadline was not the fault complete is it understanda- cise and ble)?” by punches agency, and of the but was set the same panel pulled no later, “The Panel’s re- district court would hold as follows: responded n arbitrary and ca- BiOp is ‘no.’ The ver- FWS’s rushed was sponse question to this Delta-Mendota, Luis Analysis] provided pricious. to See San & sion of the [Effects (ordering at 865 the FWS a and had not been 760 the Panel was draft BiOp just in nine organization, revised adequately general edited for produce analyses in months —a deadline that would later be consistency across sections how by give three months to reported, and for re- extended described and were just year produce was that one a new The bottom line dundancies.” BiOp).15 have a difficult time.” “most readers would assessments. The peer and the We concur these

Both the California DWR jumble disjointed facts and commented on the FWS’s panel review appears to be the result of issuing BiOp. analyses. It imposing schedule imagine happens exactly, was under what we observed that the “FWS would DWR produce is ordered to an get difficult deadline to when an exceptionally extremely compli- on an important opinion in.” In' a letter this document comment subject matter cover- just cated and technical sent from the DWR to Reclamation deadline, agencies federal and state ing multiple the DWR two weeks before' the and tens affecting la- millions of acres land pressures on the the FWS reflected expect that people. of millions of We “Our concerns with the bored under: arising deciding problems from cases district court's 15. One of the ironies of the pre- time to deadlines is that the FWS had less Court was “late in the Term” when the produce opinion either the district than oppor- deadlines from "an cluded its own it: we will have had to review court or thorough tunity for more consideration in December 2008. was issued principles risk.” Robbins v. basic Califor- judgment summary deci- The district court's nia, 420, 435-36, 101 S.Ct. 453 U.S. years day, sion issued almost two J., (1981) (Powell, concurring L.Ed.2d nearly years opinion three This issues judgment). overruled the Robbins was later. Ross, following States v. Term in United adhering to deadlines is The concerns 798, 824, 2157, 72 L.Ed.2d 572 102 S.Ct. U.S. occasionally who must rush familiar to courts (1982). judgment. Powell once bemoaned Justice prior likely patched together was SWP and this is not the last document documents, quickly by respect individu- that the FWS will issue with assembled smelt, other,- legal to the delta nor is this the last working independent of each als redundancies, readability, challenge analy- that we will hear. Future not edited for given ses should be the time and ponderous, and flow. It is a chaotic docu- attention ment, size, overwhelming in that these serious issues deserve. and without roadmaps that signposts the kinds of A. The BiOp’s Reliance on Raw trained, intelligent even readers need in Salvage Figures Upper Set agency’s reasoning. order to follow the Lower Flow OMR Limits Was Not ultimately anyone We wonder whether Arbitrary and Capricious imposition tight -of well-served pumping opera deadlines in a matter of such Under normal conse- tions, enough pumped a water is from the quence.16 Deadlines become substantive OMR that the river’s flow reverses. BiOp constraint on what an can reason- case, quantities at 159-60. As vast ably only In water are do. this the FWS not river, pumped from the smelt and other compile report had to write and —a pumps, fish are entrained where substantial task in and of itself—but was they captured and counted in to, fish sal pressure under among things, other vage facilities. Id. Not all smelt are sal produce population a reliable estimate of however, vaged, juvenile as smelt smaller Kempthome, delta smelt. (0.79 inches), than 20 mm smelt still (faulting BiOp at 373 the 2005 stage, larval and some percentage adult estimate, failing produce for to such an smelt, by delta are killed the pumps. noting viability that “[t]he of Delta BiOp at prospects salvaged 338. The for scrutiny smelt has been under for over ten grim, rarely smelt are also as smelt sur years. party No has producing shown that id.; vive the salvage process, up sixty to a population reliable estimate is scienti- percent population of the smelt is lost each unfeasible”). fically Such scientific tasks year at the pumping plants, BiOp at 210. may not be as well suited to deadlines as Unsurprisingly, BiOp the 2008 found that producing copy; product written the final as the OMR increasingly flows became necessarily will reflect the time allotted to negative, entrainment risk and accom agency. currently is pre- FWS panying population loss increased accord paring BiOp, again its third under the ingly, thereby threatening popu the smelt orders of the district court and on the continuing viability. lation’s BiOp at 163. district Although court’s deadlines. ul-we timately conclude that we can discern the effect, mitigate To this BiOp RPA agency’s reasoning and that the FWS’s imposed pumping expressed limits BiOp adequately supported by flows,17 negative terms of ranging OMR - record arbitrary -1,250 (cfs) and not and capricious, per from cubic feet second to 5,000 we also recognize that Reclamation has cfs. at 280. Which limit is continuing responsibilities applicable under CVP and is determined the location of recognize 16. imposes We protect ESA itself 17. 1 and 2 Actions of the RPA adult consulting agency produce deadlines a migrating upstream delta smelt in winter BiOp, may but those deadlines be extended spawn, protects months to and Action 3 larval requesting agency consent of the or the juvenile spring delta smelt in the after applicant. concerned See 16 U.S.C. they hatch. 1536(b)(1), (2). Here, the district court fixed the deadline. relation renders its calculated flow turbidity,18 water and this population, the smelt way, other factors. at 280. limits unreliable. Put another ac- range limits, determining the OMR flow cording appellees, when there are more relied, in the number of part, on BayDelta, smelt more smelt are the fish screen- salvaged smelt salvaged. Any apparent relationship delta be- Although acknowledged ing facilities. salvage, tween OMR flows and smelt salvaged only of smelt therefore, the number may actually relationship be a percentage “a small of the actu- represents population between smelt size and smelt entrained,” good and “is not a al number Failing salvage. to account for this will mortality estimate of actual delta smelt skew, entirely, analy- if not invalidate entrainment,” BiOp at through sis. on this information because FWS relied agreed, finding The district court smelt is salvage on the of delta “[d]ata data, use of salvage opposed raw “[t]he an index of en-

typically provide used size, salvage population data scaled to pumps,” BiOp trainment into the diversion problematic salvage figures because raw (or size) do not account for the size relative great prac- limit has a The OMR flow population.” of the smelt San Luis & significance, merely to the delta tical Delta-Mendota, at 889. Californians, represents but to as it smelt The district court further found that “the the ultimate limit on the amount of water than salvage use of normalized data rather available to sustain California’s millions gross salvage accept- data is the standard agricultural urban and users. methodology.” ed scientific Id. It conclud- -5,000 challenge upper the cfs Appellees salvage raw data ed the FWS’s use of flows, claiming that the limit on OMR Figures in B-13 and B-14 to determine based its calculation of the flow limit FWS restrictions on OMR flows was “scienti- in solely fig- on information contained two at fically unacceptable.” Id. 891. Because in significant ures that show a increase the the FWS failed to use the best available approximate- salvaged number of smelt data, -5,000 cfs flow limit in scientific its -5,000 B-14. See ly Figures cfs: B-13 and arbitrary capri- Actions and 3 was figures, appellees BiOp at 350. These Id.; cious. see id. 891-94. -5,000 assert, limit justify failed to the cfs however, objections, These suffer of delta smelt sal- because number (raw First, appro- adjusted problems. the FWS vaged salvage) was not for several B-14 Figures B-13 and priately for that relied population the smelt’s estimated (normalized). -5,000 flow limits. objection justify cfs year Appellees’ B14, data, acknowl- Figures B-13 and the FWS BiOp’s salvage use of raw rather uncertainty inherent to model- data, calculating appro- edged in than normalized essentially ing an the relation between OMR flows priate flow limits is OMR model, a and chose a conservative argument; appellees bias smelt omitted variable discretion choice that is within the FWS’s salvaged that the number of smelt assert Alts, Coal, to Pesti- to make. See Nw. year highly in a influenced the total for (9th 1043, 1050 year, of smelt that and that cides population EPA Cir.2008). -5,000 Second, lim- cfs flow failure to account for BiOp’s therefore the pended in the water. turbidity the measurement of how 18. Water foreign particles or are sus- much sediment -5,000 solely half of the curve. At about in the RPA was not lower prescribed B-14, data,20 B-13 and Figures determined from there is a “break” in the and for cfs (mean- figures in those do not -5,000 and therefore flaws than cfs negative flows more BiOp’s conclusions. necessarily doom Bay- ing exported more water from the Third, flow limits work BiOp’s OMR Delta), upward slope increases at the incidental take statement tandem with increasing sought rate. The FWS to veri- (ITS), population-level which accounts fy data was whether the break impacts. analyses actual. It conducted additional verify in the data to that there was not a

That could have done more the FWS limits determining any point OMR flow is uncontro- natural break at other and that This, however, say is not to verted. or any error the OMR flow rates sal- arbitrarily capricious- acted FWS vage could not have caused the break. hold, court, ly; contrary to the district we BiOp at concluded that 349-51. FWS sup- flow limits are that the FWS’s OMR -1683, negative with “flows more than sal- evidence. ported substantial BiOp at vage increased.” major 1. The choice of a more conser- The FWS stated that as- “[a] FWS’s sumption analysis vative model to calculate flow limits of this as the is that declined, B-14 Figures sup- B-13 and population of Delta smelt ported by substantial evidence number of at risk of entrainment re- fish mained constant.” B-What Figure BiOp19 graph B-13 in the is a compare salvage 13 did was actual num- showing relationship salvaged between bers with flow. did OMR What flow, and the adult delta smelt OMR meas- appellees not do—and what the and the graph positive ured in cfs. The shows a district court claim the FWS should have salvaged correlation between smelt and prepare figure done—was an additional in is, the reverse flow of the river. That the compared salvage which it “normalized” greater pumped through the water the stations, numbers with OMR flows. Normalized pumping Jones and Banks salvage would be the measure of the sal- greater salvaged the count of smelt at vaged by population those stations. The FWS noted that the smelt divided the total smelt,21 graph upward sloping in effectively yielding figures linear Figures reproduced producing popula- 19. B-13 and B-14 are in has shown that a reliable published scientifically opinion. Kempt- tion estimate is unfeasible.... the district court’s home, estimates, population arbitrary Without F.Supp.2d it is at 386-87. project op- for the to conclude that appears 20. The in the data in break B-13 to jeopardy simply erations will not result in -3,000 -6,000 span just about cfs to under cfs. projects relatively will because the take few- BiOp attempt at 348. FWS The did not to they past, er smelt did in the in the than break; quantify simply it observed that undisputed fact of the fact that the smelt "appears there to be a ’break’ in the dataset population steadily declining has been in -5,000 approximately BiOp at OMR." at 347. years. Failing incorporate any recent population information about smelt abun- may 21. Part of the district court’s frustration setting dance into the of the take limits is a that, date, be no one has been able to rendering BiOp fundamental failure ar- produce population a reliable estimate for the bitrary capricious. delta smelt. In 2007 when the district court Kempthome, F.Supp.2d Despite FWS, BiOp sent the 2005 back to the it com- warning, the court's the FWS did not conduct mented that population study, explain why it nor did one Yet, FWS, viability of Delta smelt has been un- could not be conducted. reviewers, scrutiny years. party peer der for over ten No and the district court’s ex- rates, flow as is the distribution of the percentage of smelt each showing what salvaged pumps. population at the delta smelt relation to the year were 165, 331, pumps. BiOp at 353-55. A lack fact, itself had stated that the FWS of real-time information and variations in- “by normalizing verify its conclusion could systems to environmental herent make by popula- salvage data the estimated virtually precision impossible. BiOp at on the Midwater [Fall tion size based 165, 331, Yet, 353-55. as even the district peer at 349. The data.” Trawl] recognized, population court numbers of normaliz- similarly suggested had review low, perilously the delta smelt are San ing the data: Deltar-Mendota, Luis & at suggests pre- Panel that the use of by pumping and entrainment of adult smelt should be salvage dicted plants “sporadically significant has a in- population for size.... One normalized population dynamics,” fluence on id. at 877 population way salvage to normalize added). (emphasis previous to divide fall size is (MWT) A Trawl index. simi- Midwater circumstances, In such the FWS’s deci- regression model to the one fitted to lar salvage sion to use raw data rather than sal- salvage would relate the normalized population-adjusted salvage data reason- (OMR) River vage to Old and Middle ably protects the delta smelt population salvage as a nor- Expressing flows.... regard year-by-year without fluctuation may some of help malized index remove population size. The in determining appropriate during factors OMR delta smelt December to March” perts suggested proxies popula- in the [W]hat that there were for are the uncertainties (such count) themselves, population FMWT avail- might as the tion estimates there able, population even if a strict count was not. shifting accuracy population be levels of as Sometimes we have to read the shadows to change? certainly [It] levels ... true reality Nearly twenty discern the behind it: population growth appropri- rate is an [that years the after smelt were declared endan- population ate and reliable measure of the gered, population we know the smelt is con- year year], to increases and decreases tinuing imperiled, to decline and is but still if it is known without error but what are many no one knows how there are. It must is, sampling? assumptions That the about something tell us about the difficulties that scarce, increasingly as smelt become does trying migrating, in two-inch inhere to count 'thinner their overall distribution become fish. ‘patchy’, might all around' or is it how reliability changes the such influence refers 22. “Noise” is a statistical term that surveys data from different of abundance? unexplained the randomness or that variation general, ‘noisy’ make it diffi- data more sample. particular is found in a It is of patterns, underlying even if samples cult to detect concern when statistical are small. Quinn genuine. explained: patterns As Dr. the uncertainty population the number of larval and a smelt whose “Minimize

delta smelt entrained at facilities” threatened, existence is the FWS chose to controlling flow from March to June. OMR in the flow in setting be conservative limits analytical approach at 357. The 1, 2, Actions and 3. This choice was well and the preferred appellees district discretion; Supreme within its Court num- gauged court is best to measure the agency may has held that an choose to pumps ber of smelt entrained at the rela- in “counteract the uncertainties” inherent population may tive to the size. This be a analyses by “overestimat[ing]” its scientific more accurate reflection of the relative parameters being known without unrea- impact popula- flows on the smelt OMR sonable, Co., Baltimore Gas & Elec. tion, protect but it is not tailored to U.S. 103 S.Ct. and we have maximum absolute number of individual upheld agency’s an reliance on models that smelt, is; BiOp’s approach pro- as “yield conservative data because the mod- adjusting salvage cess of raw for the smelt incorporate higher po- els of [known population size results in num- normalized tential in assessing values] the overall bers, but it potential does so cost of Coal, Alts, Pesticides, risk,” Nw. minimizing impact of each individual for Likewise, give F.3d we the FWS smelt lost.23 great in deference its choice of scientific agency Our deference to determinations tools, and, circumstances, in these hold is at its greatest when that arbitrarily the FWS did not act or choosing models, between various scientific capriciously choosing analytical tool as the present FWS did instance. Alts, Coal, greater Pesticides, protections resulted for the See Nw. F.3d at Facing great imperiled population.24 1050. measurement smelt Appellee-Metropolitan implicates significantly differing management Water District uses - points "one of the data located at policies. about 5,000 Figure year cfs on B-13” for the example. unusually high as an substantially An smelt sal- 24. The district court relied 2000, but, vage Appellee was observed in testimony experts, Dr. Andre Punt indicates, population high- Quinn, smelt was also concluding and Dr. Thomas year. Appellee argues er than usual in that BiOp's Figures reliance on B-14 B-13 and expected that because is to be "[i]t more "scientifically unacceptable.” was San Luis salvaged year fish would be in a in which the -Mendota, & Delta at 891. But population extremely large,” the raw sal- experts present these a more nuanced view of vage number should be normalized for total Figures the FWS’s use of B-13 and B-14. The population. appellee’s Accepting argu- smelt already results confirm what the FWS had normalized, salvage ment that rates should be analyses using said: that additional normal- year actually represent- 2000 would have ized data from B-13 could have informed the (normalized) average ed a salvage. below experts FWS’s conclusions. The noted that Yet, undisputed extraordinary it is that an measure, proper although B-13 awas not the salvaged year. number of were smelt in that only, perhaps preferable, the not meas- analysis salvage A normalized of smelt experts ure. The testified the need also as to year below-average year, counts the 2000 as a parameters the for flows, FWS to set some on OMR analysis salvage while an of raw smelt counts difficulty figuring pre- and the in out year above-average year. the 2000 as an cisely parameters where the should be. analysis FWS’s choice of influences whether response question to the district court's it is the OMR flow’s relative or absolute im- rely pact whether it was "unreasonable for FWS to population prioritized. on smelt that is Thus, part represented in quality on the information in the of the statistical method is Quinn B-13,” only figure Dr. play, not the relevant answered that he did factor at as the erroneously regard district court "not concluded: the as unreasonable for FWS to BiOp’s part figure choice of one model over the other in have relied on this and the data -5,000 of OMR the cfs OMR flow limit on its find- BiOp’s 2. The determination ing “depend[s] heavily” that the limit so by influenced more flow limits was San Luis & Del- Figures B-14. B-13 and and B-14 Figures than B-13 tar-Mendota, F.Supp.2d at 968. After experts concluded The district court’s conducting independent review of the apparent exclusive BiOp’s that it was the record, however, BiOp’s we hold that B-13 and B-14 that Figures reliance on determination OMR flow limits was suf- they acknowledged, as problematic: ficiently popu- influenced several other could useful in tandem with figures be analyses lation-level Figures addition to analyses and data. The district other Together Figures B-13 and B-14. B-14, analyses its invalidation of B13 and provide court therefore based these sub- further, rely step recognizing it.” But he cautioned that Dr. Punt went one “[t]o behind however, B-13, would, neglected advantage entirely on it have of non-normalized data: complexities produced BiOp, justi- of the issue.... Both the as it is in the would be entrainment, salvaged proportion "any fish and the fied if no number of matter how small view, relevant, size, my population salvaged long- ... are as are relative to the total has consequences population kinds of Dr. Punt had a term for the other information.” size reaction: "it was unreasonable ... delta smelt.” similar only Figures responses experts question have relied on the information in These from the analysis expand analysis B-13 and B-14 rather than on an FWS’s failure to salvage expressed popula- underlying which relative to to normalize the data B-13. Had it so, added). concluded, (emphasis experts done the FWS tion size.” Moreover, although experts complete both had been would have had a more sense of the relationship salvage, critical of the FWS’s failure to run the addi- between OMR flow and numbers, although study were it is far from clear that tional both cautious about that shown, might what the normalized have would have affected the FWS’s conclusions. data any experts’ support whether hard conclusions could be drawn What the testimonies do not set, any even whether is the court’s overstated conclusion data normal- district preferable salvage that “use of raw in the ized data would be to the non- the FWS’s Quinn produced. analyses depicted Figures B-14 normalized data Dr. ac- B-13 and knowledged scientifically unacceptable” even if the FWS had normal- or that "such B-13, "plotting against meaningless management ized the data in flow metrics are as Delta-Mendota, salvage fully capture might index the risk tools.” San Luis & only evaluating population.” respect B- Were we With to what Quinn showed, experts’ opinion BiOp, we Dr. cautioned that it was on the would question continuing precision” as to "unwise to overestimate the face difficult BiOp: validity aspect of the it is clear data. He was not convinced that there was a of this data, experts "though believed the [his] "break” in the basis for that these two analysis, although saying than have fallen short in this as [was] so more intuitive statisti- indicates, point persuasive justifications ”emphasize[d] any cal.” He value Dr. Punt -5000, BiOp's has a measure of If exist for the reliance on non-normal- arbitrariness. clear, however, why many is less then not -4900 cfs? Given the ized data. It data, through eyes of Drs. sources of variation in the it strikes me as seen —even Quinn necessary though arbi- to set limits even there Punt and be rendered —would capricious by may strong partic- trary a sole reliance on not be statistical basis for a figure slightly Figures reach than a different B-13 and B-14. We need not ular rather question, accept opinions although this as we one.” He concluded that "the valid- experts only they ity specific regimes insofar as [FWS’s] [was] flow district court's informative, persuasive incomplete analyses and—as will be undermined data,” independently "ap- subsequently were done on the available it was described—we BiOp exclusively leeway propriate have in conclude that the does not for the FWS to some rely making setting Figures B-13 and B-14 in determin- decision and limits in their species.” ing protect to and recover the OMR flow limits. efforts listed that flows are more supports to show when OMR stantial evidence -5,000 than cfs—the “break” ob- -5,000 negative limit. flow BiOp’s cfs OMR in loss Figure population served B-13— that it First, expressly notes BiOp typically percent. BiOp exceeds 10 analyses determining in multiple employed E-4). Figure strong E-4 shows a (Figure limits, stating regards the OMR flow negative flows more correlation between analyses “recent 1 and to Actions -5,000 high population than cfs and smelt adult entrainment that cumulative indicate losses, Figure B-13’s supporting “break.” flows salvage are lower when OMR 250; at 213 BiOp BiOp see also See -5,000 in the negative than cfs are no more E^4).26 (discussing Figure period.” BiOp through December March added). Second, in BiOp explicit more Although the was (emphasis at 281 - exactly describing determining its bases for specify to which studies BiOp fails 5,000 on,25 stating cfs limit for Action at least one of these studies it relies study by pro- a flows associated with the contextually apparent: “[t]he OMR Kimmerer, Sacramento tectiveness criteria defined above have Losses Wim J. tracking model- particle and Delta Smelt to been derived from River Chinook Salmon Entrainment, Estuary ing27 input assumptions Francisco with the defined San (hereinafter BiOp “Kimmer- below.” at 360. Because Action 3 Science Watershed 2008”). juvenile study protects was referred larval and delta smelt er Kimmerer’s specifically, generally was which are smaller throughout BiOp, and his data (0.79 inches) mm estimating adult en- than 20 small for reproduced a chart —too BiOp salvage facility tracking particle tracking at 281. The Kimmerer trainment. — relationship necessary to Action study quantified modeling be- estimate BiOp flows 3 entrainment. at 282. The district population tween losses OMR 2006, concluding appears court to have misunderstood the between 1995 and —con- Figure analyses support- B-14—that as OMR distinction between the sistent with ing smelt the Action 1 and 2 limits and those negative, flow becomes more delta Moreover, in population supporting increase. the Action 3 limits. San Luis losses Delta-Mendota, E-4, Kimmerer 2008 at 922 Figure the FWS read & explicit significance. may some Kimmerer 2008 25. The should have been more That - provide justification describing exactly also a reasonable for a which studies it used 4,999 -5,001 analysis. similarly or a does not make its Had the been limit policy arbitrary capricious, spec- FWS's choice vague throughout analysis, lack its Here, especially given significant deference ificity may presented problem. have agency review. however, owed to the in our To do so provided has sufficient scrutiny arbitrary smacks more of strict than support for its conclusions. capricious review under the APA and ESA. Figure were less Even if E~4’s correlation apparent, analyses yield not all will clear range justified Even were the of the limit "change points,” and therefore the fact that substantially greater Kimmerer 2008 than immediately study apparent none is in a does this, is, likely appropriate and it not most support study mean that the cannot guess reviewing court second limits; agency’s specifically defined such a agency's judgments in scientific such matters holding effectively prohibit an would where, here, agency’s as is the case deter- setting any from exact limits circumstance data-justi- appears fall within the mination change point emerge in which a clear did not range. fied situation, present the data. In the Kim- Figure present tracking modeling merer 2008 and E-4 a reason- larval 27.Particle simulates -5,000 limit, justification by inserting neutrally buoyant particles able for a cfs which fish appear approximate point be an into a model with flow conditions. does domain added) -5,000 (finding (emphasis that the FWS derived The DWR information.” limit for Action 3 from the district cfs added that order, a previous court’s from non-linear DWR’s October 16 submittal pre- [i]n we model, B-13, Figure and from fail- DWR monthly analysis sented the of Old and ing particle tracking model- mention (OMR) Middle salvage River flows and tracking modeling re- ing). particle The of adult delta smelt that provid- was also in B- BiOp’s figure sults are.reflected to Judge Wanger ed in 2007. This anal- a and demonstrate low risk of entrain- ysis appears ignored by to have been long ment as as OMR flows be- remain though analysis FWS even is more -2,000 cfs, a increasing percent low highly predictive of adult delta smelt -3,500 at cfs. at The risk 366-67. salvage any than analysis present- of the high negative entrainment at more flows ed in the FWS the Draft BO. It potentially place percent

would at risk 80 clearly monthly shows that when the larval/juvenile period of all smelt over the OMR negative are more than flows approximately months that Ac- four -5,000 cfs., about the risk salvage tion is under effect. at 366. As dramatically. increases cfs, flow approaches-5,000 the OMR modeling predicts that smelt entrainment Given the lack of clear relationship be- percent multiple

will near 100 stations. effects, tween salvage population findings, at 367. Given such any act constraints unreasonably setting salvage FWS did not in needs to fo- limits; flow the particle tracking OMR avoiding peak cus on entrainment events modeling completed provides for Action 3 rather attempting than to eliminate sal- limits. independent support for the OMR Therefore, vage. again we suggest that this in analysis be included the BO Third, the DWR recommended to the -5,000 -5,000 that it set the OMR on a parameters H cfs flows FWS, cfs. In its informal comments to the day average period except be used justification recommended “[t]he rare circumstances where the data indi- using range of -5000 to -1250 cfs OMR negative that less flows are cates needed FWS, USBR, rather than the -5000 cfs the protect against peak entrainment proposed DWR to the District Court events. clearly last summer should be more ex- added). (emphasis plained.” In formal comments submitted Fourth, early its comments on an just to Reclamation two weeks before BiOp, independent peer draft review due, BiOp was the DWR criticized the urged relationship also consideration of the (Action 2) Component RPA because it salvage and flows: “[The between OMR for much more “call[ed] restrictive Old and relationship between OMR flows and sal- supported by Middle River Flows than vage] way sound and valuable to set best to reasonably available data minimize targets to reduce entrainment. entrainment of adult delta smelt.” As the presents regres- USFWS also a reasonable noted, the “proposed range DWR flows analysis break-point sion to determine the -1,250 -5,000 cfs to cfs. The current OMR-salvage relationship.... -5,000 set except should be used flow of cfs breakpoints analyses determined these specific relating circumstances to fish data, justify target were used to the selection of survey salvage because it better supported by arbitrary the available OMR It flows.”28 was scientific time, analysis appeared 28. At the same the DWR criticized the draft because "the to be are out of when fish rely to determine on DWR’s for the FWS capricious Delta. the appropri- views on South/Central expertise own limits, accept if it did not even flow for ate court criticized FWS The district order, en- provisional recommendation. court preferred relying DWR’s on “a pend- measure stopgap as a remedial tered last observation Finally, we have one analysis.” scientific ing comprehensive court re- the district this area. When Delta-Mendota, San Luis & arbitrary and BiOp as manded the 2005 the dis- at 896. We understand F!Supp.2d BiOp that concluded capricious—a agree with the court’s concern and trict would not and SWP operations of CVP that evidence general proposition court’s dis- delta smelt—the adversely affect the measure” is “stopgap support used to Reclamation DWR trict court ordered more, not, evidence without substantial to achieve the CVP and SWP “operate But also a final measure. we supporting (reverse) flow upstream daily average net easily applied not so principle think the [-]5,000 cfs on a not to exceed the OMR noted, court in this case. As the district average.” San Luis & seven-day running -5,000 cfs order—in which the its 2007 Delta-Mendota, at 863-64. oh an eviden- figure was found—was based BiOp, an inter- During drafting were tiary hearing in which two studies (known as the team agency delta smelt relationship introduced that considered Team”) Technical “Delta OCAP BO Smelt sal- OMR flows and delta smelt between -5,000 using OMR “discussed the merits court, Id. According to the district vage. -3,500 rather than per Wanger Order the basis for B13. one of these studies was proposed in the recommended OMR as hardly faulted for The FWS can be meeting state: action.” Notes of draft accep- court’s thinking that the district -5,000 established cap OMR was issuing and the of an tance of those studies discussed the bio- Wanger. The Team consequences with real-world for order to remove logical needs of the smelt at least a people might present and smelt modification. jeopardy and adverse -5,000 prima facie figure. case for the cfs year have might What worked agree with the district understand and We correctly in 2009 might not work if the district court’s interim court It has crashed. population because order was the sole basis for FWS’s amount suggested scaling was BiOp that it would not constitute substan- fall mid-water trawl protection But, was tial evidence. abundance) (FMWT) (i.e. an estimate of rely on the studies the at least entitled helpful. would be (albeit on an accepted court had district basis) interagency in 2007. The interim -5,000 cap protect If OMR is the to the “Wan- team’s infelicitous reference adults, Action # 3 cap then the for the surely a shorthand ger Order” that, than because the should be less court based evidence on which the district juvenile particles.... behave more like order, just and not the order its interim directly relate to zone of OMR rates Additionally, we can take notice of itself. really If out of entrainment. fish BiOp was fact that when the -5,000 perhaps issued, then year’s experience the central Delta the FWS had a order, as provide protec- sufficient the district court’s might living OMR under delta However, interagency members of the extremely it is difficult tion. ly poorly large- undocumented.” well done but was described and if undoubtedly Again, team knew. who monitor delta smelt smelt conditions and “Wanger Order” was the sole evidence the recommend OMR flow trig- levels—is Second, Compo- in the record for the RPA gered. FWS’s Id. the ITS “functions as (Action 3), would not hesitate to nent we an action that operations influences under Delta-Mendota, had not relied on the find the RPA.” San Luis & available,” data “best scientific but where at 929. Pertinent to the -5,000 discussion, cfs in the figure Wanger present the Order the ITS is used to set more for point agency, was one data actual flow levels under the RPA in real agency’s 352, 354, we cannot find that the reference time. BiOp at 357. The real- improper. is irrelevant or time actual flow levels are set adjusted and input

based on the of the Working Smelt 3. The OMR flow limits exist as one Group Project operators group in- —a part dynamic monitoring system in a by 352, 354, formed BiOp ITS. at popula-

that accounts for the smelt Thus, the actual highly OMR flow is tion as a whole29 ITS, by influenced as provi- these two described, As we have task of FWS’s together sions work part complex as of the monitoring popula- OMR flow smelt dynamic system by established the BiOp. tion is a daunting BiOp one. accounts This is relevant to present discussion: challenges ways, for these a number of limits, even while the OMR flow viewed in including choosing isolation, conservative models so may not account for total smelt obligations. as to best meet its ESA One population in acceptable a manner way in BiOp court, other which the addresses district the ITS—and therefore the practical by integrating these difficulties is actual OMR flows—takes the total smelt lim- protections; its various the OMR flow population into account. exist part complicated as but one a The ITS take establishes limits that dynamic system. significant part Another vary year each based on the preceding system of this is a limit on the total allow- (FMWT), Fall Midwater Trawl index an able take of delta smelt: the incidental proxy pop- abundance for the delta smelt’s (ITS). limit BiOp take proxy population ulation size—and the suggested

The ITS influences OMR levels in flow both and the court- First, primary ways. two appointed experts they ITS is used nor- considered to establish a “Concern malizing Level” estimate to the data in B-13.

“help guide implementation of the RPA.” by project- 383-86. ITS is calculated By Id. “indicating salvage ing salvage salvage when] levels future from historic approach^] threshold,” RPA, years comparable the take the Con- with flows cern Level oper- BiOp scaling acts “as indicator that and then that number ations need to be more using prior year’s constrained to to overall abundance FMWT, exceeding avoid the incidental take.” procedure Id. at 385. This reached, actually “yields If the Concern Level is salvage discrete value for take as meeting Working Group adaptive process operate Smelt so that the can —a group comprised interagency biologists to an relative estimate of the absolute BiOp’s 29. We do not believe the OMR flow flow limits—it did not—and even if such total arbitrary capricious arbitrary limits to be for sever- reliance would otherwise be previously capricious alternative al reasons described. would not—we hold that the —it But, entirely BiOp's population even had the FWS relied use of whole numbers in Figures setting fully supports B-14 in B-13 and the OMR the ITS its OMR flow limits. 5,000 captured arbitrary the sense system.” in the San fish extant number of Delta-Mendota, APA; not the result of F.Supp.2d at its decision was &Luis such, are ad- The APA does caprice. take limits arbitrariness or 926-27. As scrutiny. The FWS is best estimate not demand strict justed reflect smelt, at 354 the delta charged protecting existing population. smelt’s figure. size as a reasonable salvage population and chose (noting both flow). setting “important variables” X2 BiOp’s B. The 2008 Determination

Thus, rec- operational in real-time making Arbitrary Capricious and Was Not RPA, the the implement ommendations salvage numbers on raw BiOp relies both that Reclamation and The FWS found BiOp at population. and the whole smelt likely “are proposed operations DWR’s 352. of delta negatively affect the abundance decreas[ing] the conclusion, by “substantially smelt” agree that the FWS

In we delta amount of suitable abiotic habitat for prepared graph a should have at least address the BiOp smelt.” at 236-37. To on normalized data similar to B-13 based habitat, proposed the in RPA not. loss of FWS why it could Neverthe- explained or 4) (Action in overall, sup- Component September less, use of OMR flows is its October, years preceding in the in the rec- and when by substantial evidence ported is, runoff was defined BiOp precipitation period we can and As convoluted as the ord. normal,” arrive as “wet or above Reclamation path took to discern -5,000 provide RPA must sufficient Delta figure used in Com- DWR at the cfs 2) (Actions X2 monthly average to maintain no 1 and and RPA Com- outflow ponent 1 3). (Action more eastward than 74 km from the Gold- See Bowman ponent years en Gate in wet 81 km “above Freight Inc. v. Transp., Arkansas-Best Inc., 281, 286, years. BiOp immediate water 95 S.Ct. normal” Sys., 419 U.S. (1974) (“[W]e previously had found uphold will 369. FWS L.Ed.2d quality spawning that the amount and than ideal if the clarity decision of less available to delta smelt is linked to may reasonably be dis- habitat agency’s path cerned.”). BiOp Dr. of X2. at 239-40. As we explained In the sense the location discussed, -5,000 represents X2 previously choice of cfs was Quinn, BiOp’s is, Bay-Delta estuary where the point had to choose arbitrary; the FWS thousand, salinity per and is range per- parts from a broad is two some number — -6,000 LSZ, -4,000 range— point the center which is con- haps in the cfs to cfs for the likely give spawning sidered suitable habitat graphs and no or charts were turn, X2, depends smelt. on delta out- precise number. this FWS flow, sense, by the -5,000 arbitrary largely is because no which is determined cfs identified, inflow from number can be difference between the total precise more Rivers range Joaquin must the Sacramento and San and a number from be selected; -5,000 exported num- amount of arbitrary and the total water thus through pumping cho- the Banks and Jones could also have ber because stations, annually -4,999 -5,001 changes which both or or some other number sen -5,000 arbitrary seasonally.30 BiOp BiOp at 236. As the range. That within - found, operations control the not make the choice of this sense does “CVP/SWP cycles, moving range strongly by tidal think X2 as a influenced It is more correct to daily up 6-10 km for single, point in twice and downstream points rather than a fixed out, average daily BiOp points location.” estuary. "X2 is its As the position primary of X2 and therefore are a BiOp as a “historic” baseline for X2. BiOp suitability.” smelt habitat driver of delta at 207. CALSIM II is a computer sim- BiOp at 234. Because the location of X2 ulation model that uses Valley Central directly affects how much water can be hydrology from 1922-2003 to simulate exported agri- to southern California for Project operations. Despite the fact that purposes, cultural and domestic the deter- regulation environmental non-Project X2 mination of where is located was criti- historically water demands have changed, parties. cal to the the CALSIM II model assumes The district court found the these factors are fixed in modeling arbitrary capricious respect scenarios. at 207. This is in con- First, X2 grounds. the location of on two model, trast which, with the DAYFLOW objected court district to the data the largely because relies report- actual FWS used to locate X2. San Luis & Del- data, ed changes does account for in reg- ta-Mendota, F.Supp.2d 760 at 903-09. ulations that impact water demand. Second, it found that actual location of DAYFLOW and CALSIM II also differ X2 support lacked in the record. Id. at ways: in other example, DAYFLOW 910-13, 922-23. will We address each uses provide historical data to daily sim- point turn. ulations of operation whereas CALSIM II Project models operations on a monthly 1. The not arbitrary FWS was and ca- basis, and DAYFLOW X2 calculates loca- pricious comparing DAYFLOW using tion a mathematical method known II CALSIM as the “KM” method whereas CALSIM sophisticated used two FWS II uses the “ANN” mathematical method. impacts models to determine the of past, BiOp, analyzing predicted present, operations and future of the Pro X2, location of estimated that median X2 ject on X2. The computer FWS chose a percent would move 10 to 15 DAYFLOW, up- model farther developed by called DWR, proposed to measure the historical stream under the action environ relative baseline, mental a computer sim to the historic X2 median baseline. II, ulation model known as 265; CALSIM de at see at Appellees also 235. veloped jointly Reclamation, by DWR and assert, however, that impossible operations. to measure future Id. at 896- change discern whether “this was due to 98, 907. These models were used in a project operations continued into fu- ways BiOp, number of in the and were change ture or whether the was due to the important BiOp’s determination of modeling differences between the ‘historic’ impacts Project operation on the Dayflow-derived baseline and the Calsim location of X2. at 145.31 Appellees argue II studies.” that a CAL- comparison II to II SIM CALSIM would computer The DAYFLOW model uses preferable. have been The district court flow, past export pumping, precip- river agreed, concluding although itation, the FWS agricultural and estimated diver- sions had discretion to use historical baseline from 1967-2007 to estimate the out- DAYFLOW, Bay-Delta flow from the model such as to the San Bay. compared Francisco This data was used abused its discretion when it 4, management 31. Action which involves the these models. X2, significantly was also influenced 618 (“[T]he at Diversity, 215 F.3d discussing logical or models-without

two different in- resulting obligation bias. See San has no to conduct Secretary accounting for the Deltar-Mendota, at F.Supp.2d studies.”), these dependent Luis & but whether court ac- Although the district 902-05. represent the “best scientific models superior “no set of mod- knowledged that available,” 16 [currently] data commercial identified,” that a it held els have been 1536(a)(2); Empire Inland Pub- U.S.C. DAY- by using created problem calibration Serv., Forest lic Lands Council U.S. II with CALSIM made FWS’s FLOW Cir.1996). (9th Because we F.3d arbitrary capri- methodology choice of the district court’s conclusion agree with explana- or required cious and correction superior “no set of models have been 877, 899, recognize tion. Id. We Delta-Mendota, identified,” Luis & San II to DAYFLOW com- that the CALSIM only question limitations, but is not without parison whether the failure to calibrate FWS’s decision to use these hold that the FWS’s conclusions drawn two models renders together, even without further two models “weak,” merely not from their evidence calibration, arbitrary capri- “arbitrary capricious.” Green- but cious.32 Franklin, peace Action v. to use the Because the FWS’s decision Cir.1992) (9th (noting that “the fact II models to- DAYFLOW and CALSIM relies the evidence [that ” determination,” is a “scientific Bal- gether dispositive); see also is “weak’ is not on] Co., 103, at Gas & Elec. 462 U.S. timore Bldg. Superior Indus. Ass’n Cal. v. of 2246, “requires high that a level S.Ct. (D.C.Cir. Norton, 1241, 1246-47 Marsh, 490 at expertise,” of technical U.S. 2001) (holding that the fact “stud- 1851, at our 109 S.Ct. we must be imperfect ies the Service relied on were in reviewing provi- this most deferential to undermine ... alone is insufficient BiOp, of Baltimore Gas & Elec. sion scien- those authorities’ status as the ‘best Co., at 103 S.Ct. 2246. 462 U.S. ”). ... tific data available’ Therefore, question, under the ESA is why it chose to use explained The FWS should have con- not whether FWS “a tools and DAY- combination available independent studies in lieu of ducted Delta-Mendota, II, San Luis & or Sw. Ctr. Bio- data.”33 FLOW CALSIM BiOp 33. The makes clear that the FWS did 32. It is not clear from the record that concluding rely exclusively that district court was correct not on the CALSIM II/DAY- comparison the FWS’s reliance on a of CAL- comparison, previous FLOW but looked studies, II and DAYFLOW models taints the SIM government peer-re- and several impacts BiOp’s regarding the conclusions viewed scientific studies: X2, Project operations on the location of or analysis proposed the effects of CVP This regarding Action 4. See San Luis & Delta- operations and SWP on the delta smelt and Mendota, F.Supp.2d at 922-23. The its critical habitat uses a combination of possible FWS has indicated bases for its other data, including the CAL- available tools and example, FWS refers to conclusions. For ap- outputs provided SIM II model in the results, data, modeling sampling and not field biological pendices of Reclamations' providing as the basis for its conclusions assessment, hydrologic pro- historical data conditions, X2 at least in the fall. about database, vided in the DAYFLOW statistical BiOp also relies on historical data to demon- unique 90-day derived from 936 time, summaries upstream strate the shift of fall X2 over tracking published by 237, 264, particle simulations BiOp well the extent to at as (2008), Nobriga Kimmerer and statisti- Project operations which the influenced the 179-82, X2, analyses BiOp cal summaries and derivative upstream at shift fall published hydrodynamic and fishers data Note 33 270. See also infra. 912; ... BiOp compare against at at 204-05. The FLOW summaries outputs.” CALSIM II BiOp it used the DAY- BiOp explained baseline, than a model as a rather FLOW acknowledged district court simulation, II CALSIM because different problems using “[t]he theoretical a compared it a CALSIM II sim- when Calsim II to II comparison Calsim were Project operations to a ulation of current manifest,” yet found that the “FWS’s Project II past op- CALSIM simulation of Dayflow decision use Calsim II to erations, it found that the two were “near- comparison ... attempting without to cali- identical,” ly the fact that and brate the two models ... despite past arbitrary and capricious ignored the- Project best avail- operations significant- current able science showing pres- a bias was ly BiOp According different. at 204-05. Delta-Mendota, ent.” San Luis & BiOp, expected” were “changes be- F.Supp.2d at 906-07. The evidence for the II BiOp tween the two CALSIM studies. modeling bias the CALSIM II/DAY- comparison at 204. Because that did not comparison clearly FLOW was not identi- differences, yield significant as the FWS fied the comments sent to the FWS anticipated, BiOp concluded that “the BiOp, before it issued the and the com- monthly model CALSIM simulation does largely ments correcting recommended capture precise operation.” Delta bias comparing one CALSIM II model 204. These “inaccuracies in BiOp at CAL- to another II precisely CALSIM model— actual [lead SIM to use data to FWS] comparison the FWS found flawed and develop empirical baseline.” San Luis & Del- example, inaccurate. For 206. The chose “the DAYFLOW Authority, ta-Mendota Water one of the ... database OMR data obtained here, plaintiffs-appellees submitted exten- from USGS.” at 206. The also sive and thoughtful portion comments on a noted that the II pro- CALSIM simulation BiOp peer of the draft analysis. review “an imperfect representation vided of the authority expressed, The water its concern pre-POD [pelagic organism envi- decline]” that it was “methodologically inappropriate ronmental supplemental situation and that to compare historical data to simulated analysis compensate was needed “to for data. Simulated data must be compared modeling BiOp this limitation.” at 205. to simulated data.” The DWR advised words, In other when the FWS used CAL- problems. FWS some of the same In II pre- SIM as both a baseline and as a portion comments a its draft dictor, appeared yield it inaccurate re- BiOp, pointed DWR out that the model sults, so the FWS used the next best assumptions might made that not be borne available baseline: the model. DAYFLOW out in the “[g]reat historical data and that Thus, explained: the FWS the “[CALSIM caution should taken comparing be when Study 7.0 run II] was the model that Rec- actual data to modeled data.” Like the thought lamation and repre- DWR best authority, water it too recommended operations, sented current and was thus CALSIM II modeling compare “be used to intended as ‘current baseline.’ Howev- one set of model runs to another.” er, due to limitations of CALSIM II to ac- comments, later repeated DWR its curately operations, model actual BiOp “compared [the concern that model FWS] also used the 1967-2007 DAY- runs of future operation with historical (2007), (2008),

by Feyrer et al. Kimmerer (accepted manuscript). and Grimaldo et al. Punt that the had articu- “[i]mpact Dr. noted FWS suggested conditions” explained why assumptions scenar- lated its compare only model analyses often per- should have thought The he the FWS problems.” ... to avoid these ios also ex- formed additional calibration: panel review independent. peer comparing some concern pressed recdgnized in the record that the It is II the CALSIM baseline with historical X2 not reflect the “histori- modeled does pointed It out results. E-28), simulated (BiOp Figure X2 and the cal” [CALSIM II] between “large difference compare does historical and CAL- baseline condi- the historical results and (Fig- X2 month SIM-predicted values can confuse the data E-26). tions defined with However, not does ure ... between sim- comparisons of metrics comparison comparable make this .for The study and historical baseline.” ulated years. attempt Failure to examination “[ijdeally, a model- suggested that panel necessary of whether it is to calibrate baseline should-be simulated available.that data and the CALSIM out- the historical It with the historical data.... is consistent normally not be considered put would model-generated base- unfortunate that practice in the appropriate scientific reliability were high degree with a lines field. analysis.” this

not made available for (internal omitted). citations experts of the court’s post-hoc views recognize that the FWS’s de We Quinn Dr. similar concerns. reflect some baseline, pro its chosen compare cision to models agreed comparison that' a between DAYFLOW, projec to its future duced assumptions with identical databases II, tion, produced by CALSIM was not long but advised that preferable, “[a]s was perfect everyone acknowledged. has —as in mind the fact that these are as we bear comparison nevertheless hold this We models, I do not see very two different arbitrary capricious. them.” After dis- why compare we cannot BiOp explained assumptions and ex him, provided of the data cussing some plained why rejected the CALSIM-to- he concluded that comparison suggested by the CALSIM cannot outputs from the two models the draft DWR and others who reviewed interchangeably estimating be used fact that the chose one BiOp. The (in km) (in cfs). X2 or either flow This over another model is flawed model flawed any not reflect criticism of either does judgment the kind to which we must workings appar- model. Their inner are defer. As said in Environmental De we different, ently quite as are their funda- Center, EPA: to an Inc. v. “We defer fense explained as to us. purposes, mental not to invest the resources agency decision any comparisons between However necessary perfect study, to conduct the explicitly them account for the dif- must and we defer to a decision to use available ferences. data unless there is no rational relation Similarly, Dr. Punt stated that ship [the use[d] between means FWS] the. any in its data principle, nothing wrong imperfections there is to account for

[i]n with, means are fitting using a model a set of and the situation which those (9th Cir.2003) 832, applied.” 872 values from one model and 344 F.3d OMR/X2 omitted). (citations using val- “The existence of a making predictions OMR/X2 require ... not us to hold that output ues which are based on the flaw does model, of the was arbi long agency’s a different as as the two the use model trary.” Am. Inst. v. sets of values are calibrated.... Iron & Steel EPA (D.C.Cir.1997) 979, (per gies, 115 F.3d cu- and weighing scientific evidence” to riam). Rather, “reject agency’s we will an agency’s ensure that “choices [are] ‘only choice of a scientific model supported by analysis.” when reasoned Ecology relationship model bears no rational to the Castaneda, (9th 652, Ctr. v. 574 F.3d of the data to characteristics which is Cir.2009). Particularly in an area as “un- ” EPA, applied.’ Natl Fed’n v. 286 wieldy this, and science-driven” as Wildlife (D.C.Cir.2002) (quoting Ap- F.3d FWS’s statistical modeling easily “does not EPA, v. palachian Power Co. 135 F.3d judicial lend itself to 'Appala- review.” (D.C.Cir.1998)); accord Envtl. Power, 802; chian 135 F.3d at see id. Ctr., 344 F.3d 872. That is not the (“Statistical Def. analysis perhaps the prime case here. The district court’s determina- example of those areas of technical wilder- contrary heavily tion to the was influenced judicial ness into which expeditions which, experts supplied parties best limited to ascertaining lay explained, the reasons that we have land.”). some or many [experts] “[T]hat inappropriate.' The record is less cer- disapprove would approach [of FWS’s] tain than the district court was willing to does not answer the question presented to appellees admit.34 That were pro- able to us. In reviewing BiOp], [the FWS’s we do post-hoc duce alluding pos- theories panel sit as a of referees on a profes- sibility comparison of bias from such a journal, sional [scientific] but as a panel of does not invalidate the FWS’s choices generalist judges obliged to defer to a rea- where there is no indication that' a failure judgment by sonable an agency acting to calibrate would do more than add pursuant congressionally delegated au- results, uncertainty factor to the as well as thority.” City L.A. Dep’t Transp., no indication that a de-biasing calibration (D.C.Cir.1999). In this technically feasible. Under our deferen- instance, provided FWS has a rea- review, tial standard of we-therefore hold soned analysis explaining why the DAY- that the FWS’s choice and use of models FLOW model was used baseline. was based on the best available science. Consequently, we hold that the did Ideally, the FWS would have thoroughly not act arbitrarily rely- capriciously reasoning regard discussed its to pos- ing the CALSIM II and DAYFLOW arising sible issues from the use of DAY- evaluating models in the impacts Project But, with CALSIM II. FLOW the fact operations X2, on the location of or in explanation FWS’s for its choices justifying 4. Action fully does not every possible address issue that flows from that choice does not ren- BiOp sufficiently explained der the FWS’s determination unreasonable fall X2 locations unsupported. or require agen- We do not The district court held that analyze every potential cies to conse- make; quence every they BiOp explain why of choice “fails to it is essential to to do km, put impossible so would an maintain X2 74 respec burden on at km and 81 Rather, agencies. agency tively, “we review all opposed any specific as to other models, respect Delta-Mendota, choices with to methodolo- location.” San Luis & ("Undis- testimony 34. The district court perts agree”), refers to from in this case id. at 912 parties’ experts ("the puted testimony”), the in terms expert that overstate id. at 922 See, e.g., Dayflow comparison what the record will bear. po- San Luis Calsim II to has the Delta-Mendota, ("un- & significant, 760 at 909 tential to introduce if not over- ("[A]ll disputed bias”). expert testimony”), whelming, id. ex- 622 a number of includes different BiOp Specifically, 923. F.Supp.2d at

760 general, have specific both explanations, the “Federal Defendants court found (Action 4)’s any record regulation evidence not identified Component RPA total This explanation. such an provides First, X2 location. as discussed in of APA[].” explanation violates lack of section, the FWS used CALSIM previous latter respect 922. With Id. at of give picture it a II and DAYFLOW contrary to the dis- the record is finding, pre- Those models where X2 located. sup- record finding. There trict court’s is values for the location dicted different BiOp’s proposal. Whether port for the DAY- differing assumptions. X2 based “explain why is essential record will data, historical relied on while FLOW km ... as X2 at 74 km and 81 maintain X2 be might II where predicted CALSIM ” location any specific other opposed Project operations. on future located based fact, it In question. a different of- found that the two models The FWS wrong question. location, X2 estimates of fered different explained, as previously have As we that, II general, “CALSIM modeled operations of the pumping combined SWP/ up- percent 10-15 further scenarios were gal- remove hundreds thousands CVP at BiOp actual historic X2.” stream than Bay-Delta, of fresh water lons Thus, historic fall X2 was 235. “[m]edian salinity-defined of the X2—the location km, for the while median values CAL- primary spawning habitat —shifts smelt’s ranged from 87 SIM II modeled scenarios BiOp at 373. towards delta. eastward II sce- to 91 km. The CALSIM modeled encroaches salty the ocean’s influence As at upper range all an of X2 narios had condi- upstream, drought it mimics further at 235. The FWS BiOp about 90 km.” Bay-Delta, regardless in the tions figures numerous and charts generated BiOp at previous precipitation. season’s models, taken from these with information X2 spring and fall (plotting 273-74 a many graphed of which X2 location as locations). Among things, other this has See, e.g., function another variable. divergence be- increasing in an resulted 282, 283, 260-61, 265, 270, 281, BiOp at fall X2 locations. The spring tween graphing range for 285. The FWS’s “long-term up- BiOp determined km, most data X2 location was 60-95 with during in X2 fall has caused stream shift km, significant points between 65-85 area avail- decrease habitat long-term smelt,” ability for and it set forth cluster 70-85 km. delta between mini- management program to adaptive Second, data from the models was Project pumping on X2: mize the effects of finding written consistent FWS’s 4. Action Action 374. RPA monthly “[d]uring years, past during X2 “wet” the fall location of targets varied from as far average X2 has down- years: years “the and “above normal” (45 km) Bay far stream as Pablo San signifi- have most project operations which on the Sacramento upstream as Rio Vista cantly adversely affected fall loca- [X2 (95 km).... general, delta River smelt Action 4 Specifically, tion].” quality great- and surface area are habitat that fall X2 be at a requires maintained Bay.” er when X2 is located Suisun greater upstream no than km location BiOp, in the BiOp at 191. Elsewhere following Bridge from the Golden Gate on an outside scientific FWS found based greater than years, water and no “wet” study “prior spawning entrainment following km “above normal” wa- upstream vulnerability of adult delta smelt increased years. BiOp ter at 282-83. (“the X2 up- E.g., BiOp when was at daily at SWP CVP fluctuation X2 BiOp upstream point at 219. around an stream 80km.” such as Brown’s Island popu- confines [smelt] Third, BiOp pointed out that X2 channels, lation to narrow where delta factors, response varies in to a number of may smelt exposed be to more stressors both natural and man-made. The natural (e.g., diversions, agricultural predation) spring factors include the tides and the X2”); relative to a BiOp downstream at BiOp inflow. The found that “X2 is (“[T]he eastward movement X2 of will strongly by cycles, moving influenced tidal shift the distribution of up- delta smelt daily up twice and downstream 6-10 km stream, provide environmental condi- average daily BiOp from its location.” at tions for nonnative fishes that thrive X2 spring 372. The runoff affects as well. (internal stable conditions.” citation omit- BiOp “very high spring found that ted)). Thus, the BiOp points out that al- always pushed outflows have X2 far down- though operations remain the CVP/SWP resulting stream in delta smelt distribu- important most factor affecting smelt habi- tions distant the influence Banks tat, 177-79, BiOp at single “there is no BiOp and Jones.” at 221. Aside from driver of population dynamics,” delta smelt operations, there' are other CVP/SWP’s 238; (refer- BiOp BiOp at see also at 202 man-made factors to consider. For exam- ring to “the multitude of factors that affect ple, Salinity the Suisun Marsh Control population dynamics”). delta smelt historically operational Gates have been complex dynamic smelt habitat is a anywhere days year. for from 10 to 120 system. BiOp salinity at 218. Operation Fifth, X2 only because varies not gates control X2 may shift as far as 3 km operations, but with natural CVP/SWP’s 218; upstream. BiOp at see also BiOp at phenomena' beyond the control of the 241, 243-44. The BiOp noted that the 3 agencies tides, seasons, as the —such upstream km shift preferable was and the annual rainfall —the FWS had a “10-20 km shifts that up have occurred for range of goal choices for X2. Its towas days per year during or more late away “move habitat from Delta im- summer through early winter due to South pacts and open into broader waters west of Delta diversions.” Sherman Island.”35 at Rather Fourth, points also out that trying X2, than single point to define a not all Bay-Delta habitats range, chose a and refined the smelt, equally suited to the so that X2 as range by year whether the or “wet” shifts, location may smelt be affected “above normal.” think We there was am- by other factors as For example, well. ple evidence in the support record to shifts, X2 location delta smelt encoun- km, choice of X2 between 74 and 81 de- runoff; changes agricultural ter changes pending on the season. The district in turbidity, which affects smelt feeding finding court’s that there was a “total lack predation; exposure to predators. explanation” is belied the record.36 35. Sherman Island (noting lies confluence of that a Joaquin the Sacramento and San Rivers. The survey of delta smelt abundance "increases range 74-81 km selected the FWS lies west dramatically whenever X2 is located between of Sherman Island. islands," Chipps and Roe and that “[w]hen Island, upstream Chipps X2 is located especially 36. The district court’s conclusion is smelt are vulnerable to entrainment and are puzzling accepted because it a similar FWS feeding located in an area that is not ideal for finding prior Kempthorne, in its decision. See Luis Deltar- San & specific on location.” focused the district court appears It Mendota, in the record support there was whether km, opposed of 74 as choice for FWS’s the the ESA thought we demanded Even if km, decision to hence its km or 75 to 73 by the precision insisted the kind adequately the FWS had question whether court, expla- believe there is an district we maintain to “why it is essential explained choice record for the FWS’s nation the opposed ... to at 74km and 81km X2 param- km km as its seasonal of 74 and 81 Luis & location.” specific San any other logical is both explanation the eters. And Delta-Mendota, F.Supp.2d at 923. 74 km there is a monitor- simple: At responsi- court’s not the district This was Bay-Delta Chipps ing station for the at that Reclama- provides The ESA bility. Island; monitoring a km there is at 81 to opinion the FWS’s to seek tion had BiOp at 132. at Collinsville. station likely were “not operations that its ensure FWS, the DWR In its comments existence of the continued jeopardize locating fall feasibility of questioned the spe- species or threatened any endangered X2, X2 location is “especially spring if the or adverse in the destruction cies or result Chicago.”37 of Port significantly west species.” of habitat of such modification agen- with how the was concerned DWR 1536(a)(2). § It is the FWS’s 16 U.S.C. any compliance cies would measure opinion forth its responsibility to set by pointed selected the FWS and standard prudent alternatives reasonable and “those difficult to measure out that “it would be Secretary of the be- Interior] which [the km, it be much an X2 at whereas would 1536(a)(2) and [§ ] not violate lieves would (81 km) or easier to measure Collinsville Id. by [Reclamation].” taken can be (92 km).” represents X2 Where Emmaton 1536(b)(3)(A). pru- § A “reasonable choices, that can range choosing an X2 flexible for is a standard dent alternative” is a perfectly be and enforced measured equiva- it consulting agency; is not Biolog- response. rational See Sw. Ctr. for alternative,” lent of the “least restrictive Diversity, ical 523. way the district court treated which is the the stakes are ex- recognize that We Biological Diver- Sw. Ctr. inquiry. case, we con- ceedingly high this but ESA, sity, 143 F.3d at 523. Under choice, given the clude the FWS’s required to ex- agency is “not consulting record, pru- a reasonable and represented RPA over anoth- why he chose one plain dent alternative. er,” “required pick the best is it nor that would most or the one alternative Take State- BiOp’s Incidental C. [species] jeop- effectively protect ment Is Not Flawed Rather, only the FWS “need ardy.” Id. of an implementation complied RPA adopted a final which When have spe take” RPA will cause “incidental jeopardy standard and which with the from, cies, “takings that agency.” Id. defined as result implemented by the could be of, carrying out purpose but are Accordingly, it was error for the district activity conducted “explain otherwise lawful to hold that the FWS must court agency applicant,” km or 50 C.F.R. at 74 Federal why is essential to maintain X2 402.02, provide required any other opposed km ... as target at fall X2 (citing 37. The FWS did not set its protection”) rec- or administrative Chicago, 74 km which is west ord). Port BiOp at also See proposed the FWS. marker

625 (ITS) an incidental take statement that on an average salvage cumulative index i.e., “[specifies impact, the the amount or Level,” when it a which, set “Concern extent, taking such incidental of the triggered, when requires the Smelt Work- 402.14(i)(l)(i); § species,” C.F.R. see 16 ing Group to make “an immediate specific 1536(b)(4)(C)(i). U.S.C. The FWS de- recommendation to the BiOp [FWS].” at termined that “take of the delta smelt is 387; Delta-Mendota, see San Luis & likely kill, to occur in form capture the of F.Supp.2.d at 928-29. We hold the (via wound, harm, salvage), harass ITS is not arbitrary and capricious be- operations result of the within CVP/SWP cause it adequate explanation includes area,” BiOp action at and prepared an support for its determinations. ITS, BiOp ITS, at In 285-310. provided separate ju- FWS take for limits 1. The reasonably ITS uses different smelt, noting

venile and adult that “indi- data juvenile sets for adult and take larval/juvenile viduals lifestage of limits demographically significant less than BiOp First, adults.” at 289. The FWS used objected the district court determining different data sets in these FWS’s choice of data. The ITS’s of use separate notes, limits. As the ITS “[t]he separate data sets to calculate the separate mean values from 2005-2008 were used as incidental juvenile take limits for and adult [juvenile] an estimate take under smelt is adequately explained in the ITS. RPA,” BiOp at average[ while “[t]he ] In discussing the use of “[t]he mean values years] value for 2006 to [water 2008” was juvenile 2005-2008” to calculate take prespawning used to calculate adult delta limits, explains the ITS that “[t]he reason smelt, BiOp words, at 287. other for selecting this span years is that the year was used in calculating juve- apparent abundance of delta smelt since nile, adult, but not the take limit. BiOp at 2005 as indexed the 20-mm Survey and 287, 289. The acknowledges ITS also its Survey] [Summer Townet is the lowest there limitations: are numerous uncertain- on record.” BiOp at 289. Accordingly, ties inherent in the of smelt measurement necessary separate “[i]t was out this (our take, “salvage data most defini- variable, abundance but also to account for tive endpoint) only measurement reflects a other poorly understood relating factors portion mortality of the total associated OMR, salvage distribution, ex- BiOp with entrainment.” at tant BiOp conditions.” at 289. As noted elsewhere BiOp, in the “[i]n contrast

The district court concluded that smelt, adult delta there is no well estab- BiOp’s arbitrary capricious ITS was or juvenile lished index of larval First, abun- two reasons: the district found court reliably dance to scale the take this explain the FWS failed to adequately lifestage to its decision abundance.... This to use 2005-2008 should be salvage data ” kept when mind.... setting the incidental take 389. There- limit for fore, smelt, juvenile faced only greater degree but to use 2006-2008 uncer- tainty in salvage setting calculating juvenile data when the incidental incidental take, chose, take limit for conservatively, adult smelt. Luis San & Delta-Mendota, incorporate year at 925-28. additional of data —a Second, the year district court found that which smelt abundance “the to explain FWS failed to rely lowest on record.”38 decision 389-90. 38. This determination came on the heels of the district court's invalidation of increasing entrainment.” without possible it used 2006-2008 is clear added). (emphasis at BiOp years “these within for adults because

data ex- approximate best dataset the historic average in to use an decision FWS’s 287, and 2005- BiOp at salvage,” pected is, like its take limits setting the incidental juvenile because juveniles 2008 data sets, to entitled data a choice choice of significant demographieally “are less smelt Lands See The difference. substantial 289, adults,” and therefore BiOp at than (9th McNair, 981, 991 v. Council preferable.39 be set would larger data Cir.2008) it is “well-estab- (noting a more conservative to use a decision Such “to courts owe deference law” that lished exactly the set, necessary, is data when methodological their agencies discretion agencies afford that we choices”). note, sort ITS’s appellees As Finest, Inc. See, e.g., Fishermen’s make. an ‘esti- provides ... averaging “use of (9th 886, 893, Locke, 896-97 that, F.3d on the take based expected mate’ of Cir.2010) decision agency’s ap- (upholding salvage, record of when historical many other disregarding likely while be exceeded rely plied[,] on some data would words, object data). appellees years.” other choice of method- to the fact that the ITS’s of a man- in too restrictive ology results an aver- reasonably uses 2. The FWS district court agement regime. salvage index age cumulative not record does agreed, finding “[t]he approach was why ‘averaging’ an explain objected Second, court the district Delta-Mendota, 760 San Luis & used.” cumula- average decision to use FWS’s (“Based ad- at 929 on known “Concern to create its salvage tive index oper- supply consequences water verse The Con- at 389-90. BiOp Level.” man- Projects ‘constrained’ ating salvage ap- levels “indicate[s] cern Level ner, that the FWS did inexplicable is at take threshold.” proaching the explanation a clear and rational provide 387; (noting at 289 see also set.”). how the ITS is a concern represents estimated number high reached has level where “entrainment offers an ade- hold that the record We the need for numbers to indicate enough explanation. As with its decision quate restrictions.”). OMR protective more year of data when calcu- use an additional reached, two the Concern Level When limit, juvenile take the ITS’s use lating the Working Group follow: The Smelt actions “counter- averaging methodology of an make “an immediate must convene and in its uncertainties” inherent acts] [FWS],” specific recommendation pa- by “overestimating]” known analyses adjusted Co., to be may flows have and OMR Elec. Baltimore Gas & rameters. See BiOp at level.” 387. “to a more restrictive 2246. In the at 103 S.Ct. 462 U.S. trigger auto- repeated recognition Level does not The Concern context of the ITS’s “may “difficulty] definitively pro- It restrictions in OMR flows. [in] matic of take of “specific data level jecting]” a reduction “unless available require” pump- delta smelt at the exports adult greater level indicate some CVP/SWP disagree we with the reasons consider 39. For similar BiOp’s ITS based its failure to why explain must court that the FWS district setting population abundance in record-low it calculated include 2006 when it decided to Delta-Mendota, Luis & take limits. See San San Luis & larval/juvenile abundance. See F.Supp.2d at 918. Delta-Mendota, F.Supp.2d at 917-18. ... ing facilities due to inherent uncertain- Project cluded that operations were rea- ties,” BiOp at (1) FWS’s use of an sonably certain to limit delta smelt food methodology that, averaging (2) its na- supply, and pollution increase harmful — ture, yields conservative limits other- (3) It contaminants. also noted wise would have been exceeded in eleven harmful indirect effects that likely would past years, sixteen San Luis & spring from three “other pre- stressors”: Delta-Mendota, 760 F.Supp.2d 929-is a dation, aquatic macrophytes, and micro- discernable, justified path. See Bow- cystis. 182-88, See 285-86, Transp., man at U.S. 95 S.Ct. BiOp found that a “multitude of Coal, 438; also Nw. see Alternatives to factors ... affect delta population smelt Pesticides, 544 F.3d at 1050 (upholding an dynamics including predation, contami- agency’s discernable reliance models nants, species, entrainment, introduced' “yield that conservative data the [where] habitat suitability, supply, aquatic food ma- incorporate[d] higher models the of [the crophytes, microcystis.” and BiOp at 202. potential in assessing values] the overall It concluded that extent to “[t]he which risk”). Moreover, the establishment of a these factors adversely affect delta smelt norm, Level is an Concern enforcement is related to hydrodynamic conditions in policy and thus a decision. of “[Selection Delta, in which turn are controlled to a an action primarily level is a legislative large by extent CVP and opera- SWP policy that uphold decision we will so long BiOp tions.” at 202. It noted that there it reasonably as was drawn the rec- were other sources water diversion that ord.” Public Citizen Health Group Res. entrainment, affect through that, smelt but Labor, (3d Dep’t 557 F.3d Cir. even when together,” they “taken did not 2009). Ultimately, the appellees’ objection “approach! ] the influence of the Banks really not to the record support for the and Jones export facilities.” Level, Concern but to the Concern Level Although BiOp candidly assessed that itself. The choice of the Concern Level is there single was “no primary driver of quintessentially one within the discretion delta smelt population dynamics” and that agency, and we have basis for no factors,” there were it ulti- “non-CVP/SWP disturbing it here. mately that concluded “the CVP and SWP are a driver of delta primary smelt abiotic D. Supports BiOp’s The Record Con- health, and biotic suitability, habitat and Regarding clusions the Indirect Ef- mortality” that operations and CVP/SWP Project Operations fects of played “have also an indirect role in the is required The FWS ac- take into delta smelt’s creating decline an altered count both the “direct and indirect effects environment Delta that has fostered of an action species on the or critical habi- non-indigenous establishment of spe- tat” determining when whether an action cies and exacerbates these and other stres- likely jeopardy. cause 50 C.F.R. sors that adversely impacting are delta 402.02; 1536(a)(2). § § see also 16 U.S.C. smelt.” 202-03. effects” are “Indirect “those that are court proposed persuaded, caused district was not action and are time, later finding but still are that record does reasonably support cer- “[t]he Here, tain BiOp’s § to occur.” 50 C.F.R. 402.02. conclusion food web and identified a pollutants/contaminant number indirect impacts are indi- on the delta Project Project effects rect operations.” smelt from effects of San operations. Delta-Mendota, Among other it con- things, Luis & fall,” BiOp at from summer separately- also court The district Project’s effects BiOp concluded about conclusions BiOp’s that “the

noted indirectly population P. on the Project between the causal connections forbesi viability of delta smelt. threaten ambig- stressors’ and ‘other operations “un- the effects were uous,” thought analysis is whether this consider We ex- evidence by record supported and/or science, available by the best supported of these address each Id. We planation” arbitrarily acted the FWS and whether turn, conclude effects indirect concluding, to a “rea- capriciously sufficiently clear was BiOp analysis Project operations certainty,” that sonable arbitrary and not to be so as thorough through its indirectly affect delta smelt on the based capricious, As 402.02. supply. See C.F.R. food Kern, See available science. peer best and the re- the district court both indicated, analysis at 1080-81. the FWS’s panel view limita- methodological from some suffers indirectly affect Project operations 1. tions; analyses underlying some supply food smelt extrapo- involved agency’s conclusions the sites, and sampling from a few lating First, likely BiOp explores the scarci- may have overstated therefore food Project operations on smelt impact of The district court seized ty of P. “entrainment BiOp The describes supply. forbesi comments, panel’s review peer on the (P. forbesi), Pseudodiaptomus forbesi “Rather than correct pointed out: which smelt,” as a prey of delta primary was to response problem, this FWS’s adversely affect effect that “will primary analysis, choos- quantitative abandon Plainly at BiOp 203-04. delta smelt.”40 same, potentially ing to advance stated, severely food appear delta smelt subjective, in a more conclusion flawed the time. P. is the limited much of forbesi sug- analysis. This conduct qualitative food source summer principal smelt’s unlawful, another results-driven gests fall, Project any impact so on early and choice, science.” ignoring best available represents P. a availability of forbesi Delta-Mendota, San Luis & BiOp 380. threat to the smelt. See think the district F.Supp.2d at 940. We that such a threat BiOp The concludes criticism is incorrect. court’s high exports water reduce flows present: P. transport otherwise that would com- peer panel review independent The forbesi habitat, thereby con the delta smelt’s into analysis directly mented FWS’s tributing mortality population to smelt on P. impact of CVP/SWP forbesi Moreover, 184-85, 228. BiOp declines. BiOp’s conclusion that the abundance Bay- pumped inversely from the south “may vary as water is of P. forbesi P. through September, during directly Delta June ... with outflow.” export flow stations, with this pumping [p]anel agrees are entrained in “The It stated: forbesi availability justification the overall with the thereby reducing conceptual model and elements, well-supported.” Be delta. which are P. of its forbesi suggestions suggests] offered technical panel “statistical evidence The cause commented, figures “the meant [prey and then co-occurrence of delta smelt are not convinc- analysis influ this strong support ... such as P. has ] forbesi “suggested] then ing.” panel The young delta smelt ence on the survival 8,000 in the Delta through acres of habitat ration of proposed to address this 40. The FWS BiOp at 381. and in Suisun Marsh. requires the creation or resto- Action which analysis this be with the redone above the BiOp’s Project conclusion that opera- in mind. If [the] considerations revised indirectly tions affect the delta smelt analysis (not does not show a substantial impacting the smelt supply food was “well necessarily statistically significant) pat- supported.” We need independently tern, analysis should he mentioned but conclude that this conclusion sup- is well the results dropped as quantitative metric ported to a certainty.” “reasonable It is added). from the EA.” (emphasis sufficient that we conclude that the BiOp’s exactly panel FWS did as the recom- conclusion sufficiently supported such mended. It omitted the analysis statistical that the FWS did not arbitrarily and capri- justification for its conclusion. ciously find this to be an indirect effect to certainty. reasonable cannot

We see the error in the FWS following the peer recommendation of the 2. Project operations indirectly affect Moreover, panel. review the FWS had the smelt through water contamina- reasons, other explained record, in the tion concluding that operations had CVP/SWP Second, an effect on P. and that the abun- the BiOp explores the forbesi dance of P. indirectly Project’s impact affects the on water contamination forbesi 184-85, smelt. therefore, delta See indirectly, on popula smelt That it omitted a study statistical viability. because tion first concludes it did not have data to justify sufficient that water contamination poses a threat to does strike us as a “results-driven the delta smelt. throughout Smelt the choice,” but responsible as exposed science. Noth- Delta are pesticides to various ing in the ESA compelled contaminants, the FWS to and which “may affect em particular conduct the study peer bryo the re- survival or prey production.” inhibit panel view thought inadequately supported BiOp The singles out ammonia released data, by the nothing peer the re- from waste processing facility in Sacra panel’s view comments even pesticides hints that the mento and agricultural from op statistics suggested a contrary 153, 187, conclu- erations. BiOp at 237. The sion—there is no evidence here to suggest BiOp that observes over “concern contami FWS, conspiracy peer with the nants the Delta is not new” and refers panel, review was trying mercury, selenium, to hide to pesticides, evidence. herbi cides, Even if thought we that “rigorous, large- concentrations, ammonium and undi study scale ... preferable,” would luted be we drainwater. BiOp 187. Reclama authority have no compel to one: “in the tion observes that the “delta smelt are of study,” absence such a highly even high “credible sensitive to levels of ammonia” anecdotal “represent evidence” can that the such ] contaminants are detrimen best scientific ... to data available.” Nw. tal the health of the smelt population FWS, Ecosystem Alliance v. they 475 F.3d because render susceptible smelt (9th Cir.2007) (internal quota- BiOp disease. at 187-88. omitted). marks tion BiOp Project concludes that opera- Like peer panel review in question, dangerously tions will increase the impact we conclude that although FWS’s anal- of contaminants on smelt. One reason ysis limitations, here is not without its it is topographical: Project is operations upon “based the best available science.” constrain the smelt habitat to smaller riv- Alliance, Ecosystem See Nw. estuaries, 475 F.3d at ers and overall expo- smelt’s Similarly, 1147. we persuaded are that contaminants, sure these which result effects those sonably certain that indirect runoff, increases. See land primarily the ac- will result from Project harmful and op- Relatedly, when at 153. BiOp habitat, to do so. We decline at issue. We suitable tions overall reduce erations BiOp if the smelt in that we returned on are confident contaminants impact of help intensified. become we could agency, habitat remaining in the this effect out errors analyzes by “pointing] latter improve BiOp importance “insisting] recognizing missing context of information” viability: The FWS Cnty. smelt v. Nor- X2 delta fall detail.” Churchill additional Cir.2001). (9th mechanisms “[potential two ton, described effect,” one which was role, the observed course.” Id. not our But is “[t]hat may increase distribution “a more confined Instead, has suffi- we hold in- events of stochastic impact relation be- the harmful ciently explained ... delta smelt rates of mortality crease contaminants, Project operations, tween such effects anthropogenic ... including] did not such that the FWS and delta smelt (citing at 234 as contaminants.” to a arbitrarily capriciously conclude 2007). other harmful al. One et Sommer Project opera- certainty that reasonable the FWS Project operations, effect of to harmful contaminant- tions contribute flow created notes, from the comes effects. related indirect “in- Projectrelated flow Project pumping: many pesticides exposure crease[s] indirectly affect Project operations This [during spawning].” “other stres- through the the smelt contaminants.” can “mobilize flows because macrophytes, and predation, sors” of BiOp at 240. microcystis that “[i]t concluded court The district im analysis of the part As of its any other docu- BiOp or not clear how the smelt, on the Project operations pacts impacts of ment in the record links of “other the effect BiOp discusses Project Operations.” San to contaminants that of factors multitude stressors”: “the Delta-Mendota, F.Supp.2d at Luis & dynamics in population affect delta smelt well-founded. criticism not 942. This macro- aquatic ... cluding predation, that “con- Although recognized the FWS BiOp at 202. microcystis.'” phytes, and ecosystem and its effects loading taminant fac to these extent which “[t]he Because understood,” not Delta are well within the is relat affect delta smelt adversely tors science must at the fact that BiOp Del conditions hydrodynamic ed to complicated further before advance large ta, to in turn are controlled which Bay-Delta interactions in ecosystem operations,” and SWP extent CVP necessarily fully understood are does analysis consid BiOp’s BiOp at rely on to mean the FWS failed that deter stressors” ers these “other when science, arbitrarily or that best available effects of mining proposed “the CVP/ there was concluded that capriciously smelt,” BiOp at delta operations on SWP Project opera- certainty that reasonable con BiOp’s hold first 203. We through indirectly smelt tions will affect conditions “hydrodynamic clusion that Appellees would water contamination. opera driven or influenced CVP/SWP impossi- it is have us hold that presumably delta dynamics of ... influence tions simultaneously recog- to ble for an these other stres- with smelt interaction indi- of an that some characteristics nize in the sufficiently supported sors” was or not well understood” rect effect “are explicit- BiOp 202. The at uncertain,” BiOp it is rea- record. “highly ly relied on outside scientific studies eggs smelt prey larvae were for inland Project concluding operations “af- BiOp silversides. at 183. The BiOp also or recognized fect!] control!]” some these other the risks presented from other impacting stressors delta smelt abun- potential predators of smelt eggs and lar- Second, BiOp dance. having areas, de- vae such including yellowfin goby, Project termined that operations likely centrarchids, af- and Chinook salmon.41 BiOp impact fect the of “other stressors” on at 183. smelt, BiOp considers the harmful After concluding that predation poses a effects of these stressors on the other harmful indirect effect delta smelt popu- to population. smelt lation viability, the discusses BiOp the re- lation Project district court faulted the between operations FWS for

failing information,” predation. consider smelt By shifting “available X2 further Delta-Mendota, San Luis & upstream, Project operations move smelt 934-35, F.Supp.2d at or habitat “make a rational include more these littoral (areas connection between areas Bay-Delta the facts in the rec- close to shore), id. conclusions,” ord and its thereby increasing 936. We smelt exposure disagree, predators. BiOp and find the sufficiently BiOp at 153. As the location X2 thorough and hold shifts that the FWS did not towards Bay-Delta new re- arbitrarily and capriciously gions reach are rife predators this Moreover, conclusion. smelt follow. BiOp tion the smelt “should be considered smelt,” ty for delta part Delta-Mendota, in because the San Luis & significant.” scarcity of smelt means that Here, smelt have at 934. the district recently gone undetected recent studies very specific: court was It faulted the predator stomach contents. BiOp BiOp failing for any “include estimates 183. The BiOp speculated that “!d]elta of predation effect on the delta may experience smelt high predation mor- population” smelt when “[s]uch informa- tality around water diversion where smelt tion was available” was “decidedly and are entrained predators and aggregate,” contrary to BiOp findings.” Id. In sup- citing to an outside study showing that port, the district court referenced a 1999 BiOp 41. The pirical considered whether delta smelt evidence support the conclusion fish, competition are affected from other competition species that these between is a study and cited one suggesting they that were. factor that influences the abundance of delta BiOp The competition did not rely on aas BiOp smelt in the wild.” at 183. stressor, however, because "there is no em- broadly, we de- smelt. More to the delta & Game of Fish Department California comb a fine-toothed review with of an cline to part to the FWS report submitted relied the FWS (citing Califor- on which Id. the studies permit. take incidental Game, Conservation conclusions. reaching of Fish Dep’t nia Department Plan The California for Macrophytes Manage- Aquatic b. Bass Striped Fish and Game 1999)). (November We Program ment BiOp is in the noted Another stressor report Fish & Game not sure what Macrophytes— macrophytes. aquatic is report that adds, cannot see and we wa- in or near grow that aquatic plants conclu- contrary” to the FWS’s “decidedly interi- extensively colonized ter —have sions. BiOp decades. past two or Delta over that stated report Fish & Game notes, sug- research BiOp As im- yields information “the best available have “al- macrophytes that these gests estimates of loosely constrained precise, dynamics in the community fish tered It smelt.” on delta striped predation bass for Delta, increasing habitat including overlap considerable that there was said including largemouth eentrarchid fishes bass, striped smelt and the delta between fishes, bass, native reducing habitat for rarely ate delta bass “striped that but for pathway a food web supporting are sur- smelt,” “delta smelt likely because fishes.” other littoral eentrarchids tend to striped bass while face oriented (citations studies to scientific BiOp at &Fish Game near the bottom.” forage omitted). both impact smelt These effects responsi- bass were striped that estimated Submerged indirectly. directly consumption annual for “an ble estimated littoral vegetation can overwhelm aquatic popula- of the delta smelt about 5.3% habitats, shoals such as inter-tidal between not see conflict tion.” We do beaches, spawn- as delta smelt that serve BiOp. report and the Game the Fish & un- locations, rendering them thereby ing unknown “[i]t concluded is BiOp at 182. spawning. for suitable by striped predation incidental whether suspended Moreover, trap macrophytes source of ... a substantial represents bass tur- water and therefore reduce sediment smelt.” mortality delta to a de- contributed bidity, which has is consistent with report & The Fish Game and adult smelt juvenile crease both anything, If conclusion. the FWS’s hab- increasing available while habitat suggest report might Fish & Game BiOp at on smelt. prey for fish that itat if the study required further see and No- Feyrer et (citing al. 182-83 of mor- a substantial source striped bass is This de- support). briga et al. 2008 Although the the smelt. tality for may facilitate turbidity also creased showing studies post-1999 cited hampering delta smelt predation of while predator significant not a striped bass was thereby further feeding, own smelt’s smelt, if Fish & Game delta viability. BiOp harming population smelt current) estimates, (and a in its correct at 183. percent mor- for a 5.3 predator responsible notes, hydrologic conditions BiOp As the might endangered species tality rate anof significant a play temperature water event, we fail to see any significant. be of the Del- colonization macrophyte role in conclu- respect to the significance with that The concludes BiOp ta. at 182. BiOp concluded BiOp. of sions the im- Project operations’ that “likely” it is than fish predators primarily other — conditions Bay-Delta hydrologic pact threat potential a striped the bass—were of flushing reduction seasonal flows c. Microcystis the spread exacerbate of macrophytes Microcystis aeruginosa cyanobac- is a Bay-Delta. the BiOp at 277. Flushing terium that produces toxins throughout its flows known to lead to “abrupt are cycle, life with toxin concentrations sharply changes turbidity.” in flow and BiOp at increasing when the bacteria population Indeed, 146. the FWS found that “[avail- dies, usually September or October. able information is regarding inconclusive BiOp at 372. high These toxin pres- levels extent, magnitude pathways by ent a threat and, to the delta smelt as the which may delta smelt be affected these BiOp recognizes, high microcystis toxin independent stressors opera- CVP/SWP levels have been associated with low delta tions.” BiOp at 277. smelt BiOp abundances. at 372. Micro- The district court cystis was also not can persuad- directly “pose animal and hu- analysis, ed finding this man health risks if “[al- contacted or ingested though a Project directly,” connection op- although [between it does not appear that exist, erations macrophytes] may current concentrations are sufficiently se- discussion, record any does reflect vere nor to threaten BiOp smelt. at 186. parties Rather, have the pointed to any study, microcystis’s primary to threat connecting flushing indirect, ‘seasonal ... smelt is appears flows as “it that M. natural frequency upstream aeruginosa and down- toxic to copepods that delta LSZ, stream length- movement smelt eat.” BiOp at 186 (citing an outside upstream study). ened] shifts of the scientific LSZ’ There is also concern presence any aquatic macrophyte.” microcystis out-compete “could dia- Delta-Mendota, toms[, San Luis & rich food zooplankton,] source for F.Supp.2d at Again, light 935-86. we disagree. and nutrients.” BiOp at 186. As BiOp notes, set forth plausible BiOp however, several expla- more studies are Project needed, nations for how operations and, fact, will in- underway “are crease the detrimental impact zooplankton of macro- determine if production is phytes on delta compromised smelt viability, during and cited M. aeruginosa in support. studies BiOp See blooms to 182- an extent that likely ad- not, 277. That the BiOp versely did as the affect delta smelt.” BiOp at 186. requests, district point court study discusses, As also CVP/SWP directly addressing Project’s effect on operations likely are to increase the harm- Bay-Delta macrophytes does not render ful impact microcystis on delta smelt the FWS’s conclusions unreasonable or un- because “[l]ow flow conditions among supported: The FWS has drawn rational the factors associated Microcystis science, from the conclusions best available By flows, blooms.” at 372. reducing and, consequently, we hold that BiOp’s Project operations would cause “larval and determination that it is certain reasonably juvenile smelt ... delta remain [to] that macrophytes indirectly will affect del- Delta, Central and they South where could ta smelt is not arbitrary and capricious. It ... to predation microcystis succumb or job is not our to task the filling FWS with Overall, blooms.” Microcys- gaps in the scientific evidence. We tis suitability.” BiOp habitat “reduce[s] must respect agency’s judgment even “in the face of uncertainty.” Cattle Am. Salazar, Growers v. Ass’n The district court found that (9th Cir.2010). 1164 “makes no connection whatsoever between consistent with in a manner mented CVP and continued ... and

microcystis SWP operation” and that “[g]iven that the intended purpose of the action, ^[2] that Project consistent with the Operations implemented can be impacts regulating at- unsupported legal au- agency’s such Federal consequential, of the scope are so rationale) (a of a in search result that eco- thority jurisdiction, [3] and tributions feasible, & Detta- San Luis technologically are unconscionable.” nomically and Mendota, Again, and ^[4] Director believes would sev- proposed has FWS disagree. we The jeopardizing the likelihood of avoid the hypotheses evidence-based plausible eral species or listed existence of continued Project-related im- the harmful explaining or adverse in the destruction resulting pact of inherent not namic force hold instance, highly sonable. ations, science or their CVP/SWP BiOp candidly al sions were studies microcystis render recognizing the interdependent Bay-Delta microcystis the evidence uncertainty associated in this We should the FWS’s operations not acting on the acknowledges harms is sufficient BiOp’s recognition arbitrary area are own hydrologic on delta smelt. linking knowledge. are limitations of either microcystis conclu- not deter ecosystem underway. not the Bay-Delta does Project conditions, capricious. with the addition- agencies only that we unrea- oper- That The this dy- factors. the definition that would C.F.R. modification, but such “jeopardy” 50 meet one of the alternative alternative,’ clude that certain book ^[3] consultation show was § 402.02 is modification are explains it referred to as the The factor; commonly referred the Services should process.” U.S. avoid considered § the other three elements further: “If the services FWS’s of ‘reasonable of critical in the 402.02. Element alternatives jeopardy elements Consultation Hand- biological alternatives during the formal habitat. “non Fish & and are available ^[1] opinion to jeopardy” to as the document through prudent adverse ^[4] fail Wild- con- Support Required Serv., Is Not E. Fisheries & Nat’l Marine Serv. life “Non-Jeopardy” Elements (March 1998), Section Con- at 4-41 ESA RPA Handbook, http:// available sultation www.fws.gov/endangered/esa-library/pdf/ that an Secretary determines When the 2013) (second (last July visited CH4.pdf to, jeopardy or action cause agency will added). Thus, according emphasis of, an endan- modification adverse habitat Handbook, if a draft alterna- Consultation Secretary species, gered or threatened jeopardy the non to meet one tive pru- those reasonable suggest “shall fails RPA, the Service of a valid “elements” he would alternatives which believes dent to show that provide documentation should adversely or species [jeopardize during consulta- it alternatives by the considered and can be taken modify its habitat] previously have afforded tion. Id. We implement- applicant or agency Federal deference FWS’s Consul- U.S.C. Skidmore action.” ing Cattle 1536(b)(3)(A); Handbook. See Ariz. Grow- also C.F.R. tation see Ass’n, 402.14(h)(3), (concluding FWS’s 606 F.3d at 402.14(g)(5). §§ ers’ ap- duty under in the handbook explain its definition regulations “[t]he further agency’s “reason- the result of regulations pears the ESA. The define to be per- ... we prudent judgment able and alternatives” —RPAs—as considered entitled one reasonable during [that for- suaded identified alternative actions mal consultation ^[1] that can be imple- deference]”); see also Skidmore Swift Co., & 323 U.S. S.Ct. 89 L.Ed. non jeopardy factors. San Luis & Delta- (1944). Mendota, 760 F.Supp.2d at 957. Nothing §in obligates 402.02 FWS address

Referring to the non jeopardy fac jeopardy non factors when proposes it tors, court the district found that the FWS RPAs. Section 402.02 is a definitional sec- “has articulated absolutely no connection tion; it is defining what constitutes between the facts in the record and the RPA, (1) required setting hoops out that conclusion that the RPA is FWS jump consistent must purpose through. with the underly See 50 C.F.R. (2) action; ing 402.02; § (the consistent with the § action see also id. at 402.14(g)(5) (3) agency’s authority; economically and FWS shall “discuss an agency] [with ... technologically feasible.” San Luis & the availability of prudent reasonable and Delta-Mendota, at 956-57. alternatives”); 302.02(h)(3) (“A ‘jeopardy’ The court continued: biological opinion shall include reasonable

the APA requires, and public is enti prudent any.”). alternatives if More- receive, tled under the law to ex some over, the Consultation implies Handbook position the record of why that no such necessary. discussion is As (if all) concluded it did so at that all four the Handbook notes: regulatory requirements for a valid RPA [Although] it imperative that the opin- were satisfied. The RPA Actions mani ion contain a thorough explanation of festly interdict supply the water for do how component each of the [reasonable mestic consumption human agricul prudent] alternative is essential to tural use for over twenty people million jeopardy avoid adverse and/or modifica- depend who Projects on the for their [,] tion ... [i]f the Services conclude supply. water “Trust us” is not accept that certain alternatives are available able. FWS has shown no inclination to would, avoid jeopardy and adverse fully honestly address water supply modification, but such alternatives fail beyond species needs despite the to meet one of the other three elements fact regulation its own requires in the definition of pru- ‘reasonable and such consideration. alternative,’ dent the Services should Delta-Mendota, San Luis & document the alternative the biologi- F.Supp.2d at 957. Because the had FWS cal opinion to it show was considered explain why RPAs, failed to it chose its “to during the process. formal consultation the exclusion of implementing less harmful alternatives,” the district court remanded Handbook, Section 7 Consultation avail- Id. simply, FWS. Put more able http://www.fws.gov/endangered/esa- district court found that both the FWS’s (last library/pdf/CH4.pdf visited July regulation and APA required the FWS 2013); 402.14(h)(3) see also 50 C.F.R. engage exposition a record of the non (“If the Service is to develop unable such factors, jeopardy and that the FWS did prudent] alternatives, [reasonable and not do so. indicate [must] to the best of its disagree We both with the district knowledge there are no reasonable and legal analysis court’s and with its reading alternatives.”). prudent words, In other of the record. “thorough” jeopardy/ad- documentation of

First, contrary to the district verse modification in always court’s conclusion, the regulation” required, FWS’s “own whereas documentation of the require does not jeopardy only address the non required factors is when 636 itself. Section ESA manded the jeopardy- a non to meet fails RPA

the agen- federal 1536(a)(2) that each requires factor. ... is not any action that “insure cy shall that the FWS anywhere fail to see We habi- species or its the likely jeopardize” to explana- an provide to itself required has 1536(a)(2). includes § This tat. 16 U.S.C. factors jeopardy when non tion of the that its FWS, warrant which must the “impose on may not an RPA. We lays out 1536(a)(2)].” [§ violate RPA “would not pro- of which notion own agency [our] the is, 1536(b)(3)(a). the That 16 U.S.C. likely to further or most are ‘best’ cedures RPA, an proposing FWS, in the course Nor good. public undefined vague, some jeop- not RPA does insure must requirements procedural may impose we can its habitat. We species or the ardize pertinent in the enumerated explicitly [not] ESA the requirement no similar find (inter- McNair, at F.3d 993 statutes.” remaining three address FWS that omit- citations marks and quotation nal does If ESA nonjeopardy factors. ted).42 to extremely reluctant it, we are require not, dis- Second, as the APA does APA. into the requirement a such read held, to ad- require the FWS trict court Moreover, persuaded we are case. factors this non-jeopardy dress the economic what the Dettar-Mendota, court misread 760 district Luis & See San The court factor addresses. APA, feasibility F.Supp.2d Under accounting for for not the FWS deci- faulted agency has held Supreme Court for supply water “interdiet[ing] the cost of consider an to “entirely fail[] sions that consumption agricul- and human are arbi- domestic problem” aspect of the important twenty people million use for tural over Motor Vehicle trary capricious. and Mfrs. Projects their water on the depend who U.S., Mut. Inc. State Farm Ass’n of Delta-Mendota, 760 Luis & 29, 43, supply.” San Co., S.Ct. 463 U.S. Auto. Ins. the ESA (1983). This misreads have We 77 L.Ed.2d regulations. Section implementing its prevent RPA will an held that whether the economic with only concerned critical 402.02 or adverse modification jeopardy feasibility of the RPA. technological aspect important is “an habitat is, whether must consider See, That FWS e.g., Fish Conservan- problem.” Wild financially and (9th alternative is proposed its Salazar, 522-23 628 F.3d cy v. consid- Those two technologically possible. Cir.2010) that failed (finding BiOp technology jeopardy erations —economics avoided how the RPA explain —are the FWS measures what Coast constraints capricious); arbitrary and Pacific alter- as an to the Bu- can recommend v. U.S. Ass’ns Fed’n Fishermen’s entirely. To activity Reclamation, ceasing native reau of case: Recla- in this Cir.2005) (same). perspective it into (9th jeopardy put But the be- has consulted independently de- mation RPA factor in the any requirement without cially recently onerous Circuit 42. We Fourth note Id. at thought it was feasible. for whether FWS for failure remanded (RPA applica- prohibit pesticide would techno- RPA for economic and an evaluate (for ground applica- "within 500 feet AgroSciences, 707 feasibility. tions logical Dow tions) 1,000 (for applications) feet aerial Dow to We do not read 474-75. F.3d at connected, directly or any waterway that is require to address economic the FWS any year, any indirectly, at time of the feasibility procedural mat- technological as a might be Dow, body which salmonids con- water the court was ter. we read As original)). (emphases point." imposed espe- some found that the FWS had cerned legitímate law, cause it has whether concerns eral including but not limited to the may jeopard- continued CVP Act, activities Federal Endangered Species 16 ize or § the smelt its habitat. When the U.S.C. 1531 et seq”); Valley Tenn. Auth., 185, (hold- FWS concludes that Reclamation’s contin- U.S. at 98 S.Ct. 2279 ued jeopardize ing activities will that the smelt ESA reflects “a the conscious then, by decision presumptively, may Congress Reclamation not to give endangered species priority take or continue over ‘primary such activities. the See Nat’l missions’ Assn, of agencies”). federal Home Builders v. Accordingly, the Defenders of duty FWS’s Wildlife, 644, 652, 2518, is to opine viability 551 U.S. on the S.Ct. (2007) the smelt and “to halt 168 L.Ed.2d 467 and reverse (“Following the the trend toward species extinction, ‘jeopardy’ issuance of a agency the whatever opinion, (em- the cost.” Id. at action, must either S.Ct. 2279 imple- terminate the added). phasis proposed alternative, ment the or an seek exemption from the Cabinet-Level Endan- if Even the APA require did the FWS to ”). gered Species Committee.... In this factors, consider the nonjeopardy the rec- case, course, terminating Reclamations’ ord shows that the FWS has sufficiently CVP-related activities is unthinkable. The considered them. See Motor Vehicle Mfrs. point of the prudent whole “reasonable and U.S., Inc., Ass’n 463 U.S. at alternative” is for the suggest FWS to 2856 (holding S.Ct. that insufficient consid- what Reclamation can todo avoid such a eration purposes of APA arbitrary and regulation result. The identifies “econom- capricious review is an “entire[] failure] ic technological and feasibility” as factors consider”). Although the FWS’s consid- go because these the RPA whether “can eration of the non-jeopardy factors could be by taken the Federal' ... in certainly exhaustive, have been even more implementing action,” the agency 16 or stated expressly, more its determination 1536(b)(3)(A) § added), (emphasis U.S.C. that the RPA satisfied the non-jeopardy not to whether restricting CVP activities “may factors reasonably be discerned” will affect its consumers.43 The “economic record, from the and therefore should be technological feasibility” factor does upheld.- Id. (holding even “a decision not address downstream economic im- of less than clarity” should be upheld .ideal pacts of being circumstances). Reclamation to con- unable in such Application of the tinue operations its CVP as it has done jeopardy non really factors this case is past. important As consequential quite straightforward. See 50 C.F.R. question is, as the respon- First, § is not 402.02. the record shows that the sible for balancing the life of the delta RPA is purpose consistent with against smelt impact of restrictions on underlying action. The document that was operations. That balance has prepared consultation, CVP/SWP as a basis for Recla- already been struck Congress in (BA), the mation’s biological assessment iden- ESA and the Valley Project Central Im- purpose tified the of this action to be provement 3406(b), Act. See CVPIA “operat[ing] [Projects] divert, store, 102-575, Pub.L. No. redivert, 106 Stat. and convey CVP and ... SWP (stating that Secretary of the Interior water consistent applicable law.” “operate Valley the Central Project to The largely RPAs —which deal with regu- to. meet all obligations under State and Fed- lating the water export CVP/SWP parties Actions) Neither proposed nor district court economically not argue (and that the RPAs technologically themselves their feasible. discretion- separating not the ESA any ed require Delta-Bay

from the —do nondiscretionary actions when way ary Reclamation in the major changes Second, similarly, baseline; Rec- operations. runs its the environmental set RPA can be record indicates arbitrarily capriciously acted lamation scope of with the “consistent implemented BiOp; and that accepted the when it authority.” legal agency’s the Federal prepare failed to Reclamation FWS and BiOp discuss and the the BA Both statement, as re- impact environmental See, authority. extent Reclamation’s each NEPA. We consider quired (Reclamation’s obli- 21-25 e.g., at turn. Operations its Coordinated gations under DWR). is Finally, there Agreement with Discretionary From Segregating A. con- record for FWS’s in the support Nondiscretionary Actions technological- is both RPA clusion that the determining whether think feasible. We When economically ly and close- jeopardize The RPA likely nearly agency’s operations self-evident. this re- in the interim measures ly species resembles a listed existence of the continued order, feasibility of which medial or adverse in the destruction or result through 2007 in mid-December proven habitat, critical the FWS of a modification implementation. 2008 December of the action must effects “[e]valuate incorporates RPA Additionally, 327-28. species listed effects on the and cumulative that were comments feasibility-related habitat.” 50 C.F.R. or critical from RPA Reclamation made on a draft requires § This evaluation 402.14(g)(3). regu- RPAs propose Again, and DWR. baseline,” determining an “environmental does in what Reclamation latory changes include effects “does not which basis, not the RPAs do but day-to-day on a review,” adding the and then action under affecting Reclama- changes require major proposed effects of the and indirect direct technological- or ability financially tion’s — whether that to determine federal action RPAs. comply with the ly—to species. See jeopardize listed action will sum, district disagree with the In we 402.02; § Section 7 Consultation 50 C.F.R. the FWS’s own that court’s determination Handbook, http://www.fws. available the FWS require APA regulation and the gov/endangered/esalibrary/pdf/CH4.pdf; § 402.02 the RPA satisfies explain that Cooperation Interangency see also —En Alternatively, we jeopardy non factors. Amended; Act of dangered Species of these that the FWS’s consideration hold 19,926, 19,932 Rule, Fed.Reg. Final reasonably discerned may factors be (June 1986) (noting the environ any explanation re- satisfy the record to baseline as the “serve[s] mental baseline quirements.44 on of the action determining the effects for habitat”). or critical species CROSS-APPEAL V. Builders, Supreme Court In Home claims of error raise three appellees requirement the consultation violat- held that that the in the court: FWS district another, justify decision based his over or not ex- We need also hold F.3d at 523. solely apolitical over factors.” 143 plain why the RPA measures it chose & Del- agree San Luis “less harmful alternatives.” not with the dis- Consequently, we do ta-Mendota, F.Supp.2d at 957. In South- demands re- such a trict failure court we held Biological Diversity, Center west agency. & Delta-Men- San Luis mand to ESA, Secretary was not that "under the dota, F.Supp.2d at 957. explain why he one RPA required to chose 7(a)(2), 1536(a)(2), § § ESA U.S.C. does obligation consultation attaches to a partic- “impliedly repeal[] nondiscretionary agency ular at all.” San Luis & Deltar- mandates, statutory they might even when Mendota, 760 F.Supp.2d at 948. It did agency result in some action.” Nat’l Ass’n not, however, whether, “address once sec- Builders, Home 551 U.S. at 127 tion 7 consultation triggered, jeopar- case, S.Ct. 2518. the Clean Water dy analysis must separately identify and required Act the EPA to transfer certain segregate discretionary from nondiscre- permitting powers to state authorities if actions, tionary relegating the non-discre- nine criteria were satisfied. 33 U.S.C. tionary actions to the environmental base- 1342(b). § In accordance with the ESA removed). line.” (emphasis Id. 7(a)(2), EPA consulted with the agree We with the analy- district court’s FWS. The found that Service the transfer sis that Home Builders does not require would any not have direct effect on any to segregate discretionary from species Arizona listed under the ESA. non-discretionary actions when it considers FWS, however, The was concerned that the environmental baseline. Home Build- once EPA permitting transferred authori- 7(a)(2) only ers dealt § with whether ap- Arizona, ty to the ESA would no longer plies. question The real after Home decisions, apply permitting to and Arizona Builders is what counts as a non-discre- permits could issue without regard to list- action, 7(a)(2) tionary § to which does not species. ed The FWS concluded that the apply. might transfer have indirect effects on the species, although no decision or action of question We addressed this in National EPA in question, was other than the trans- Federation v. National Marine Wildlife fer of permitting authority. 653-54, Id. Service, (9th Fisheries 524 F.3d 917 Cir. 127 S.Ct. 2518. The Court held that the 2008) (“NWF”). NWF, the National ESA did not function as an additional con- (NMFS) Marine Fisheries Service had is- straint on the duty per- EPA’s to transfer impacts sued a on the of the opera- mitting authority to Arizona: “[Section] tion of the Federal Columbia River Power 7(a)(2)’s no jeopardy duty only covers dis- (FCRPS) System on a species. listed Id. cretionary agency actions and does not attach to actions ... that an agency is The 2004 BiOp’s jeopardy analysis in- required by statute to undertake once cer- cluded the environmental baseline for specified tain triggering events have oc- proposed existing action the 669,127 curred.” Id. at S.Ct. 2518. FCRPS, supposedly various nondiscre- Builders, Relying on Home San Luis & tionary operations, dam past and all Authority Delta-Mendota Water asserts present impacts discretionary op- 7(a)(2)] part of this analysis, “[a]s [§ found, .... erations NMFS also though, must, among things, other distin- that certain aspects opera- of FCRPS guish discretionary between the and non- tions—such operations relating as ir- discretionary actions of an operation, so rigation, control, power flood gener- only discretionary actions are con- nondiscretionary, ation —were given the sidered as agency effects of the action.” existence, dams’ and that aspects those argument This reject- considered and part should not he considered by ed the district court. See San Luis & action under ESA review. Delta-Mendota, 760 947-48. added). The district court (emphasis reasoned that “Home Id. at agency The Builders addressed whether the section “segregated analysis, its evaluating first (9th 1024, 1032 action—con- Coal. agency proposed

whether EPA Cir.2005)). discretionary only proposed sisting of have the FCRPS —would operation of Authority pointed us has Water species.” a Id. net effect appreciable statutory obligation Congress any assessing this “instead agency did that is both imposed on Reclamation has jeopar- be fish would whether the listed inconsistent with its obli- mandatory and proposed aggregate of dized Like the FCRPS gations under ESA.45 baseline, action, the environmental NWF, very broad Reclamation has effects, and current status cumulative Moreover, stated, has Congress mandate. species.” Id. can, that Reclamation is to clearly as it con- obligations CVP administer held, was incor approach, This we of the ESA. with the mandates sistent determining may not avoid rect: “NMFS (stat- 3406(b), CVPIA, § at 4714 106 Stat. agencies’ action discretion the limits of the Secretary of the Interior is to ing that the sweep using operation reference Project Valley “operate Central ‘nondiscretionary’ operations into so-called and Fed- obligations all under State meet baseline, thereby ex the environmental law, including but not limited to eral jeop requisite from the ESA cluding them Act, Endangered Species Federal analysis.” at 929. We distin ardy Id. seq.”). 1531 et U.S.C. on the guished from Home Builders NWF of the mandate in specificity basis B. Reclamation Did Not Violate the ESA Congress case question. present the “[I]n Accepting BiOp the 2008 mandates, rather than imposed has broad Authority argues that Recla- The Water ac directing agency specific the to take independent violation mation committed tions, agencies perfectly capa and the by relying BiOp. on the See of the ESA simultaneously obeying Section ble Wildlife, 420 F.3d at 976 and those mandates.” Id. 928. We Defenders of “[arbitrarily capriciously and (noting that that “in contrast Home Build repeated [to vio- relying faulty Biological Opinion on a ers,] Congress imposed broad man has agency’s independent lates action [the agencies per do not direct to dates which added)). Be- duty” (emphasis substantive] actions, nondiscretionary any specific form to BiOp do not believe the 2008 cause we rather, characterized as di but are better arbitrary join we the capricious, be recting agencies particular the to achieve declining to find that Rec- district court Thus, goals “while the goals.” Id. at 928. reliance on the was arbi- lamation’s mandatory, agen may themselves be trary capricious. See San Luis & cies retain considerable discretion Delta-Mendota, at 966-67. actions to take in choosing specific what them.” Id. at 929. implement order to “ NEPA Application C. to the FWS escape obligation agency cannot its ‘[A]n Reclamation merely because it comply to with the ESA that the fullest ex- requires another statute NEPA “to comply to with bound consistent, objec agencies ... all of the Fed- complementary possible tent that has ” an envi- Washington complete eral Government shall” (quoting tives.’ Id. Toxics pointed State Water Re- Authority us to wa- decision of California 45. The Water has ap- Board. These do not Reclamation and wild- sources Control ter contracts between contractors, statutory refuge exchange proach the mandate the Court water con- life holders, under in Home Builders. rights water and a found EPA was tracts with senior (EIS) impact ronmental statement in moved summary judgment for on their “every connection with claims recommendation the FWS and Reclamation failed report comply or on with proposals legislation and NEPA. The defen- dant-intervenors, NRDC, major including other significantly Federal actions filed a cross-motion for summary judgment, affecting quality argu- of the human envi- ing that 4332(2X0. agencies obligated were not ronment.” 42 U.S.C. adhere to NEPA in issuing implement- or agency may begin by preparing an envi- ing BiOp. November (EA). dis- ronmental assessment 40 C.F.R. trict issued court par- decision 1501.4(a)-(c); §§ 1508.9. It may then is- ties’ summary cross-motions for judgment sue a finding significant of no impact on the NEPA issues. San Luis & Delta- (FONSI) rather than completing an EIS Salazar, Mendota Water Auth. v. if EA reveals that the ques- action in (E.D.Cal.2009). F.Supp.2d 1026 The dis- tion “will not a significant have effect on trict court observed that is a “[i]t close call the human environment.” 40 C.F.R. whether FWS’s issuance of 1501.4(e); §§ 1508.13. its RPA under these circumstances” re- case, In this neither the nor quires the preparation of an pursuant EIS Reclamation an EA or prepared an EIS. to NEPA. Id. at 1044. It concluded that We are thus with two questions confronted made, call need “[t]his not be because Rec- concerning scope require NEPA’s lamation, with the ultimate au- and, agencies produce ment that an EA if thority RPA, implement is now necessary, First, an EIS. was the FWS’s joined party, as a whose actions must be “major issuance of the Federal evaluated under NEPA.” Id. The court significantly affecting action[ ] the quality granted summary judgment in favor of the of the human environment” imposed plaintiffs against on their claim Reclama- obligation on the FWS an comply tion, concluding “that Reclamation violated second, NEPA? And was Reclamation’s failing NEPA to by perform any NEPA provisional adoption implementation and of analysis prior to provisionally adopting and “major BiOp the a signifi Federal action[ ] implementing the 2008 and its RPA.” cantly affecting quality the of the human Id. at 1051. imposed environment” that on Reclama In a subsequent motion summary for tion an obligation to comply with NEPA?46 judgment, plaintiffs the again asserted court, Before the plaintiffs district the by FWS violated NEPA not issu- initially alleged that the FWS violated ing along an BiOp. EIS with the The dis- by NEPA failing prepare to an EIS when trict court remarked that was an “[t]his issuing the BiOp. In their first attempt to re-argue and argu- re-frame complaint, amended the San plaintiffs Luis previously ments decided” because the dis- further asserted that Reclamation violated trict “prior court’s NEPA rulings deter- by accepting NEPA and implementing the mined that Reclamation bears the NEPA BiOp without completing an EIS. In responsibility three in this case.” San Luis & cases, of the Delta-Mendota, five plaintiffs consolidated the agency’s 46. We review contending an decision that EIS ... or EA that NEPA does not prepare need not an EIS for apply_We "reasonable- hold the less deferential stan- Glickman, ness.” See Northcoast Envtl. Ctr. v. applies dard of ‘reasonableness’ to threshold (9th Cir.1998) ("Here, we agency that certain decisions activities are not have a question applicabil- threshold of NEPA subject procedures.”). NEPA's to ity. prepared The Secretaries have not an “[m]ajor federal action includes and an EIS. A opposition an filed The defendants major and may be urg- effects that summary judgment, actions with cross-motion the subject that to to reaffirm Federal potentially court which are ing the district NEPA comply with need to did not C.F.R. responsibility.” FWS control and court The district BiOp. issuing the regulations offer several § 1508.18. The granted and motion plaintiffs’ the actions, denied includ- major federal categories of respect motion with the defendants’ as plans, formal such “[a]doption of ing Id. against FWS. NEPA claims approved prepared or official documents guide pre- or agencies which argue that both federal appeal, plaintiffs On re- comply must of Federal Reclamation uses scribe alternative the FWS con- federal defendants actions sources, agency NEPA. The future upon with which comply with need not that the FWS specific tend “[ajpproval will be based” complete will Reclamation because NEPA manage- or construction projects, as such position takes EIS. And NRDC geo- located a defined ment activities by the of the the issuance neither (4). 1508.18(b)(2), §at graphic area.” Id. implementa- acceptance and nor the FWS argue defendants The federal triggers by Reclamation tion of the agen- FWS, consulting capacity its under NEPA. obligations ESA, merely cy 7 of the under Section judgment the district court’s We affirm suggestions offering opinions its First, the NEPA respect to claims. with which, Reclamation, agency, action as the circumstances, that, under these hold we ap- adopt whether to or ultimately decides pre- require does NEPA sup- well plan. This view is prove the the issu- conjunction pare an EIS statute, and our regulations, ported Second, we hold that BiOp. ance of 7(b) explains law. case Section adoption provisional Reclamation’s agency the Federal provide to FWS “shall BiOp triggered its implementation ... a written statement [e.g., Reclamation] We with NEPA. obligation comply Secretary’s opinion, and setting forth court’s order affirm the district therefore the information on which summary that it can remanding Reclamation so based, detailing how opinion evaluating the effects complete an EIS or its critical species action affects the adoption implementation (em- 1536(b)(3)(A) 16 U.S.C. habitat.” BiOp.47 added). or jeopardy “If adverse

phasis to the FWS Application 1. of NEPA found, Secretary shall the is modification al- prudent reasonable and suggest those the whether first consider We would vio- ternatives which he believes not “major a was FWS’s issuance (a)(2) of this section lated subsection affecting the significantly action[ ] Federal [e.g., agency the Federal can be taken such human environment” quality of the the implementing ... complete Reclamation] to obligated the that 2012). to court continues active- intent The district provided of its Reclamation notice 28, 2012. See deadline for com- prepare ly manage an EIS on March Reclamation's to Biological Opinions on the Coordi- process. Remanded pleting EIS See Memorandum Long-Term Operation of the Central nated Regarding Ex- Motion to and Order Decision Project: No- Valley Project and State Water Schedule, Delta Consolidated tend Remand Prepare an Environmental of Intent to tice Cases, l:09-cv-00407(E.D.Cal. Apr. No. Smelt Scoping of Impact and Notice Statement 9, 2013), EOF No. 1106. 28, (March Meetings, Fed.Reg. 18858-02 added). agency (emphasis action.” Id. person’ knowingly who ‘takes’ an endan- ordinarily not an “opin- We would consider gered or species subject threatened to “major a “suggestion]” ion” or Federal substantial civil and penalties, criminal in- regulations provide further action[ ].” cluding imprisonment.” Id. at “[f]ollowing biologi- the issuance of a S.Ct. 1154.48 But “powerful coercive opinion, agency cal the Federal [e.g., Rec- effect” of a BiOp on an agency action like shall determine whether and in lamation] Reclamation does not render it akin to the proceed what manner to with the action in “[a]doption of plans” formal or “[a]pproval light obligations of its section 7 and the specific projects,” which tend trigger biological opinion.” Service’s 50 C.F.R. requirements. NEPA’s 40 C.F.R. 402.15(a). Our cases confirm that an (4). 1508.18(b)(2), Reclamation, Unlike agency action like Reclamation has some for, responsible FWS is not and will not discretion to deviate and its implement, the RPAs. And even if Recla- Pyramid RPAs. See Lake Paiute Tribe of compelled mation felt implement Dep’t Navy, Indians v. U.S. 898 F.2d proposal, FWS’s we must bear in mind (9th Cir.1990) (“We 1410, 1418 recog- have complete Reclamation will an EIS Secretary nized that the is to be afforded evaluating the effects of implementing the ascertaining some discretion in how best to BiOp. See section IV.B.2. infra fulfill the mandate to conserve under sec- Reclamation, The fact that 7(a)(1) and not the tion example, [of ESA].... For FWS, responsibility bears for implement- agency given [an action] discretion to ing implement decide whether to alternative com- conservation —or plies with FWS.”); Section 7’s put recommendations forth mandate —distin- Hodel, guishes Kantor, Tribal Vill. Akutan v. this case from Ramsey (9th Cir.1988) (“The (9th Cir.1996), 96 F.3d 434 agency is where we held required adopt sug- alternatives issuing BiOp and ITS gested biological opinion.... required [The was to comply with NEPA. In 7(a)(2) Secretary] satisfied section if he Ramsey, the National Marine Fisheries alternative, took reasonably adequate (NMFS) produced BiOp Service and ITS steps insure the continued existence of in connection with the Columbia River any endangered or species.”). threatened Plan, Management Fish which is a “unique, created, judicially federal-state-

We are mindful of the fact that “while compact” tribal that “apportions the fish- the Service’s Biological Opinion theoreti- ing rights to the state and tribal mem- function,’ cally ‘advisory serves an in reali- Notably, bers.” Id. at 438. “The states ty it powerful has a coercive effect on the regulations governing fishing then enact agency.” action Spear, Bennett v. River, 154, 169, although they Columbia must do U.S. 117 S.Ct. 137 L.Ed.2d (1997) (citation omitted). so in compliance with the “The terms action agency River technically disregard Management fi-ee to Columbia Fish Plan.” the Biological Opinion and proceed completed with its Id. After the NMFS action, ITS, proposed but it does Washington so its own and the states of (and peril employees), that of its ‘any Oregon regulations, for issued which “would Bennett, question In "The for decision Act.” Id. at 117 S.Ct. 1154. The case did petitioners any questions concerning [was] whether the ... [had] stand- not raise or resolve ing judicial biological application agency to seek review of the of NEPA action to ESA, opinion provision under the citizen-suit under Section of the which is at issue ESA ... and the Administrative Procedure in this case. issued a Commission Energy Regulatory take] [incidental .not for that illegal, if

be hydro- a to construct permit preliminary amount statement,” specified a permitting at 1508. plant. power electric Id. River. fishing in Columbia of salmon argued that the Commission plaintiffs “con- Ramsey court at 444. Id. an before issu- have conducted EIS should take incidental statement that the clude[d] disa- at 1509. We permit. Id. ing the to a functionally equivalent in this case is “can applicants explaining greed, already estab- had cases permit.” Id. Our and conduct federal land only enter prereq- a permit a federal lished that “if obtaining after activities ground-breaking on impact adverse project a with uisite for per- use special and BLM Service Forest environment, permit of that issuance Thus, the Com- agencies, these not mits. action and major federal does constitute evaluating mission, responsible for will be conduct must agency involved the federal au- impact of activities the environmental grant- an EIS before EA and possibly an permits.” use by special, their thorized Gordon, 792 (citing v. ing it.” Id. Jones F.2d Id.; Burford, v. also Conner see (9th Cir.1986); Port F.2d 827-29 Cir.1988) (9th that the sale (holding (9th Hodel, 595 F.2d 478-79 v. Astoria did not gas oil and leases particular Cir.1979)). reason, the court held this For complete to an the Forest require Service take] of [the “that the issuance incidental governmental “absent EIS because further major action constitutes federal statement absolutely prohibit leases approval, [ ] Ramsey, 96 F.3d purposes of NEPA.” for activity.” (emphasis surface-disturbing added)). Washington Ramsey, the states consulting reason to a require We see no position typically Oregon occupied an complete EIS like FWS agency like agency action by a federal inhabited agency like Reclamation when an action and ITS Reclamation because (1) it EIS when prepare will either an of a part issued federal-state-tribal were (2) reject proposal or implements FWS’s only to applies NEPA compact. Because prepare an EIS proposal FWS’s 4332(2)(C), actions,” § 42 U.S.C. “federal implements. We alternative whatever no in that there was downstream case to use NEPA as have efforts condemned an If the agency complete EIS. federal Bay tactic.” See Drakes an “obstructionist NMFS, agency, did consulting (9th Jewell, 967, 984 Oyster Co. requirement Ram- comply with the EIS Cir.2013); Douglas Cnty., 48 F.3d at 1508. evaded then action would have sey, completing fact that an EIS Of course though the altogether even NEPA review costly or does consuming be time might was, substance, action identical complying with agency not excuse an permit, a would process issuing which NEPA; struck Con- that is balance an issuing agency to require the prepare courts. But the statute gress, not the EIS. in connec- completion of an EIS requires Here, need to comparable is no actions.” 42 “major there all tion with Federal added). 4332(2)(C) be- It prepare (emphasis an EIS require U.S.C. ready to do so. of an EIS require completion stands cause Reclamation does not need not that make agency agencies have all recommendations We held participate an action or respect another complete an EIS where for action. As the formulating proposal implement or action will authorize observed, FERC, “it makes little court v. district triggers NEPA. Sierra Club sepa- (9th Cir.1985), agencies prepare have two sense to the Federal 754 F.2d 1506 645 agen- rate NEPA documents for the same an is action that will trigger Reclamation’s obligation Deltar-Mendota, complete to an EIS. cy action.” San Luis & F.Supp.2d 686 Application 2. of NEPA to Reclamation that, We hold under these circum- Having explained why the FWS stances, the required FWS was not to required was not produce to an EIS when comply issuing BiOp. NEPA it issued the BiOp, we now why address There are situations where a consulting the district court correctly concluded that adoption Reclamation’s implementa required like the FWS is to com- tion of the BiOp requires preparation plete conjunction an prep- EIS with the of an EIS. The federal defendants do not aration of a BiOp under Section 7 of the contest the district court’s decision that already ESA. We have identified one such Reclamation completed should have an EA Ramsey. scenario in But neither the stat- and, if necessary, conjunction an EIS ute nor our case supports proposi- law implementation with its BiOp. But tion that production of a BiOp FWS’s appeals NRDC the district court’s order “major constitutes a Federal action[ ]” granting summary judgment in favor of implementation when its contingent is on plaintiffs on their claim that Reclama adoption BiOp, Reclamation’s of the which tion violated NEPA.49 demonstrated, 49. Because the federal defendants have not NRDC has for Article III appealed the district court’s decision that standing purposes, the district court’s adoption Reclamation's subject of the judgment is requiring complete Reclamation to NEPA, we must first confirm that NRDC poses an EIS "a credible threat of harm” to standing challenge has the decision. See the delta smelt. The 2008 we —which Charles, 54, 68, Diamond v. 476 U.S. 106 today arbitrary hold capricious— is not 1697, (1986) (”[A]n S.Ct. 90 L.Ed.2d 48 inter- project operations jeopardized concluded that right a venor's to continue suit in the absence is, therefore, the delta smelt. There "a credi- party on whose side intervention was ble threat of pro- harm” to the smelt if delta permitted contingent upon showing by a ject operations Although continue. the dis- requirements the intervenor that he fulfills the vacatur, trict court remanded without III.”). of Art. plaintiffs potentially judgment can use the enjoin that Reclamation violated NEPA to im- standing We hold that NRDC appeal has plementation long doing of the RPAs so so the district court's decision. "To determine does not violate Section 7 of the ESA. See In may appeal whether an intervenor from a Cases, re Consolidated Delta Smelt being appealed by decision not one of the 1133, (E.D.Cal.2011) (judg- court, parties in the district the test is wheth- by ment vacated San Luis & Delta-Mendota er the intervenor's interests have been ad- Salazar, 11-17143, Water Auth. v. No. versely judgment.” affected Didrick- 23, (9th 2012)) ("A Aug. WL 6929161 Cir. Intenor, Dep't son v. U.S. 982 F.2d may injunction court issue an under (9th Cir.1992). "To invoke this NEPA that would cause a violation of other jurisdiction court's injury the basis of an statutory requirements, such as those found in judgment, related to the Intervenors must es- However, Section 7 ESA.... where the tablish that judgment the district court's evidence indicates that the will ESA not be particu- causes their members a concrete and by injunctive violated relief injury issued under larized that is actual or imminent and NEPA, presence permits of a NEPA claim likely to be redressed deci- favorable evidence.”). Project consideration of economic Kraayenbrink, sion.” W. Watersheds harm (9th Cir.2010). grant injunc- The fact that a court cannot credi- "[A] judgment ble threat of harm is tion based on the sufficient to constitute Reclamation injury standing purposes, doing actual violated NEPA if whether so would cause a statutory or not violation violation has occurred.” of Section 7 of the ESA does not States, Agency preclude Cent. Delta having standing ap- Water v. United NRDC from peal F.3d judgment because "a credible threat 646 analyses pub- of BiOp uncontroverted agree with noting that we begin

We Projects’ water de- ... the [that] that Recla- lic data conclusion court’s district materially is a must be livery of implementation operations mation’s to af- water flows project restrict “major significantly changed action[] Federal to Delta- of the human environ- San Luis & fecting quality protect smelt.” di- ment,” Mendota, does not though F.Supp.2d NRDC even 686 the court’s portion this of rectly challenge of Second, implementation Reclamation’s First, 4332(2)(C), § 42 decision. U.S.C. human BiOp “significantly affect[s] “major is a of the implementation that “[a]n We have held environment.” “that have held action.” We Federal EIS required prepare to an agency is would not federal action proposed a where questions are substantial where there not nec- quo, an EIS is change the status signifi- may cause project about whether Chapter River essary.” Upper Snake of environ- degradation of the human cant Hodel, 921 F.2d v. Trout Unlimited Ecosystems Native Council ment.” (9th Cir.1990); Burbank see also 235 (9th Serv., 1233, 1239 428 F.3d US. Forest Goldschmidt, 623 Group v. Anti-Noise Cir.2005) original). The dis- (emphasis (9th Cir.1980) (“An need EIS F.2d “dis- correctly trict concluded that court of the environmental effects not discuss can be made without positive conclusions facility.”). operation continued of mere looking record].” the [administrative River, the court held Upper Snake Delta-Mendota, & San Luis the flow decision reduce Reclamation’s example, the fed- at 1050. For 1,000 cfs was not a from a dam to of water “De- answer states that eral defendants’ River, Snake major Upper federal action. in exports aver that ‘reductions fendants already Reclamation had 921 F.2d at 233. may ‘place greater de- from the Delta’ 1,000 during cfs the flow rate below set water, sources of mands alternative upon days that the dam had of the total 4.75% ” Id. The district including groundwater.’ the court to operation, which been led correctly concluded that court Reclama- they prior did in “[w]hat conclude BiOp a “ma- implementation of the is tion’s they doing during the years and what were significantly affecting jor action[ ] Federal no more under consideration period were human environment.” quality regu- managerial the routine actions than 4332(2)(C). 42 U.S.C. larly carried on from the without outset that, new, argument appeal NRDC’s nothing change they doing ... circumstances, con- NEPA extensive, under these nor other than that nor more goal preserving the ESA’s was first flicts with contemplated project when by im- like the delta smelt species Id. at 234-35. The district listed operational.” require- posing procedural additional correctly distinguished Upper Snake court reason, this not ment on Reclamation. For by observing does River ap- NEPA does adjustment merely involve a “routine” NRDC contends Rather, imple- adoption project. ply of the Reclamation’s operation BiOp, if the even action from the face of mentation “can be determined ESA, statutory 7 of which injury violation of Section sufficient to actual harm is constitute standing. id. at for Article III See standing purposes, or not a statu- is sufficient whether (" environmental harm is tory Cent. Delta Wa- 'Threatened violation has occurred.” ” added). (quoting (emphasis probabilistic.' Friends Agency, nature ter Earth, Copper Corp., probabil- Recycling judgment v. Gaston thus introduces some Inc. (4th Cir.2000) (en banc))). harm of a chance of environmental short F.3d istic

647 of an “major significantly “Endangered Species a Federal af- action[ ] Committee” quality of the human environ- fecting empowered grant that exemptions ordinarily trigger ment” that would NEPA general from the prohibition agency review. We therefore consider whether actions that “jeopardize the continued exis- requirement set aside the EIS we should tence” of species listed or “result in the agency implements BiOp an a when destruction or adverse modification of that designed RPAs to ensure its action “is 1536(e). § [their] habitat.” 16 U.S.C. likely jeopardize the continued exis- specifically provides statute that a decision any endangered species or threat- tence by Endangered Species Committee to species ened or result the destruction or exempt agency an action from the ESA’s adverse modification of habitat of such subject prohibitions is not to NEPA if the 1536(a)(2). § species.” 16 U.S.C. We find agency already completed an EIS concern- no basis the statute or our case law for ing the effects of the action. Id. at excusing Reclamation from its NEPA obli- 1536(k). § these gations under circumstances. very This could well be the end of our face, permit the statute does not On inquiry. Congress repeatedly has demon- case-by-case exceptions that assess how strated that it exempt par- knows how to NEPA interacts with the substantive stat- ticular substantive statutes from the EIS simply It that requires ute issue. “to the requirement when it wishes to do so. possible extent ... all agencies fullest of Moreover, Congress expressly exempt- has complete the Federal Government shall” particular ed a subset of actions under and, EA if necessary, an an EIS for all Section of the ESA—decisions “major significantly Federal actions affect- Endangered Species Committee an where ing quality of the human environment.” already completed by EIS was the action 4332(2)(C). § Congress U.S.C. has ex- agency. Congress But did not all exempt pressly provided that NEPA ap- does not jeopardizing efforts to avoid the survival of ply statutory to certain schemes. For ex- species requirement. listed from the EIS ample, “No action taken under the Clean fact, In Congress’s decision to exempt cer- major Air Act shall be deemed Federal tain Endangered Species decisions significantly affecting quality action requirement Committee from the EIS human environment within the mean- NEPA applies reaffirms that to other ac- ing Policy of the National Environmental ESA, tions under Section 7 of the includ- 793(c)(1). § Act of 1969.” 15 U.S.C. ing implementation Reclamation’s words, other no action taken under the BiOp. trigger Air Act Clean will NEPA’s re- statutory There is additional evidence quirement agency produce that adoption imple- Reclamation’s § EIS. See also 42 (stating U.S.C. BiOp triggers mentation of the its obli- requirement apply that the EIS does not gations under NEPA. Section 7 of the exemptions power- certain for electric provides biological ESA that a assessment § plants); (providing U.S.C. “may part be undertaken as of a Federal restoring actions particular facilities to agency’s compliance with requirements prior their condition to a disaster or emer- of section 102 the National gency exempt Environmen- require- the EIS (42 ment). 4332),” Policy § tal Act of 1969 notably, Most Section 7 of the U.S.C. provision governs prep- which is the section that ESA—the issue here —carves 1536(c)(1). exception require- out a narrow to the EIS aration of an 16 U.S.C. EIS. Congress ment. The authorizes the specifically ESA formation This is evidence that of a critical habi- apply designation dis- an action contemplated tat.”). 7 of under Section its duties charging NEPA comply with would also ESA First, Ridge Supreme Court in Flint and, necessary, if an EA completing an not requirement did held that the EIS acknowledge also regulations EIS. pre- to requiring agency apply because expected to concur- agencies are “would create an irreconcil- pare an EIS 7 of the with both Section rently comply with the fundamental conflict able and § 402.06 50 and NEPA. See C.F.R. ESA Secretary’s [substantive duties under conference, (“Consultation, biological Ridge, 426 U.S. statute at Flint issue].” under section procedures assessment There, 788, the substan- at 96 S.Ct. 2430. interagency with co- consolidated may be that a document filed provided tive statute by other required operation procedures automatically agency would be- with the statutes, Environ- as the National such days under certain thirty come effective in (NEPA).”). Policy Act mental 788, circumstances. Id. at 96 S.Ct. single not cite a case where

NRDC does inconceiv- explained that “[i]t The Court agency’s any held that an action court has impact state- able that an environmental of the ESA obligations under Section 7 drafted, could, days, be circu- ment with NEPA. complying excuse it from lated, and then reviewed upon, commented however, are, a number of cases There Id. light and revised the comments.” holding that other substantive statutes 789, at 96 S.Ct. 2430. But Jones Gor- requirement, even exempt from the EIS (9th don, Cir.1986), we ob- 792 F.2d 821 expressly provid- has not though Congress only Ridge applies “Flint served that would have us exemption. an NRDC ed unavoidable’ when a conflict is ‘clear and ” those cases to these extend the rationale of Id. and ‘irreconcilable and fundamental.’ circumstances. court ex- 792 F.2d 821. Jones that, Ridge, unlike in Flint plained circum recognized have two We agency publication long “could withhold need not com agency stances where an any require- NEPA enough comply with despite an absence of an plete an EIS an preparation ment for environmental First, an statutory exemption. express Id. impact statement.” complying with agency is excused from and funda create an There is no “irreconcilable doing NEPA where so “would conflict” mental conflict” between NEPA and Sec irreconcilable and fundamental Although the statute issue. tion 7 of the ESA. with the substantive statute a timetable for the consultation Ridge Dev. v. Scenic Rivers sets out Flint Co. 776, 788, to accommo process, enough it is flexible ofOkla., Ass’n 426 U.S. S.Ct. (1976). Second, preparation date the of an EIS. See 49 L.Ed.2d 205 we 1536(b)(1)(A) (“Consultation § un of in have identified a limited number U.S.C. (a)(2) has der subsection of this section stances where a substantive statute any action shall be con requirements, respect agency even “displaced” NEPA’s 90-day period beginning not “an irreconcilable” con cluded within the though there is or, subject initiated statute and on the date on which flict between substantive (B), period such other Douglas Cnty., paragraph within requirement. the EIS See (“[The agreeable to the argues mutually time as is plaintiff] 48 F.3d at 1502 (emphasis statutory Secretary agency.” that without this ‘irreconcilable’ federal 1536(b)(1)(B) (“The added)); disagree, 16 U.S.C. apply. conflict NEPA must We may Secretary Federal ... NEPA and the we hold that does tively accomplished a consultation all mutually agree goals extend of NEPA’s EIS, preceding requiring thereby established under without period Secretary, “mak[ing] procedure if the before the close the NEPA sentence seem ‘su- ” perfluous.’ the consent of the period, of such obtains Id. at 1503. Section 4 of the extension.”); application compels Secretary to the see also ESA “the consider [to] NEPA, impacts Water Dist. v. United States that concern Westlands extent Interior, Dep’t F.Supp. designation 1423 that the critical habitat has a (“Section (E.D.Cal.1994) positive 7 of the ESA environmental spe- effect on the Furthermore, gives agencies question.” control over the time within cies in Id. “The designation process which consultation is to be concluded.... critical provides also notice, provides public goal 7 for the inclusion for [And] ESA another of NEPA.” pro within the Id. ‘applicants’ of consultation cess, which demonstrates access to the But the cannot same be said for Section process by parties.... ESA interested ESA, 7 the of which is at issue in this case. timing secrecy nor concerns bar Neither Block, In Save the Yaak Committee v. 840 NEPA”). ability comply

the with (9th Cir.1988), F.2d 714 explained we the

Second, agency we have held that an biological difference between a assessment (BA) produced from NEPA might exempt pursuant action be even to Section 7 of the ESA,, statutory “without this ‘irreconcilable’ con- and an EA or in prepared EIS ac- in Ridge. Douglas flict” identified Flint cordance with NEPA. The Save the Yaak 1502; Cnty., argument 48 F.3d at see also Drakes court considered an that if “even 984; Merrell, EA Bay Oyster, inadequate, 729 F.3d at 807 the was it supple- (BA)” Douglas County, biological F.2d at 778. In we held mented the assessment Secretary that completed the the Interior need under Section 7 of the Id. ESA. complete designating rejected reasoning, EIS when the at 718. We this ex- species pursuant plaining analyzes critical habitat of a listed that a BA the “[w]hile Cnty., impact to Section 4 of the of a Douglas proposed upon ESA. 48 action endan- Douglas County gered species, at 1507.50 court an EA analyzes impact F.3d “Congress concluded that to dis- of all proposed intended action on facets of the Thus, place” procedures only pre- NEPA’s when authoriz- environment. if a BA is ing agency pared may in designate gaps agency’s critical habitat there be analysis.” under 4 of the ESA. Id. at 1504 n. environmental Id. Section 10. But none of the factors relied on Courts have offered several other exam- Douglas County this de- reaching court ples of the differences between the Section where, apply cision the same force with prescribed by 7 and the one process here, agency action at issue is the instance, 7 NEPA. For “the ESA’s Section implementation of a under Section 7 process provide consultation fails to for of the ESA. public way comment the same Hall, Douglas County NEPA does.” Fund court reasoned Animals for (D.D.C.2006). 127,136 This process designating F.Supp.2d for critical 448 particularly important “[p]ubli- habitat under Section 4 of the ESA effec- because subsequently disagreed 50. The Cnty. Tenth Circuit Bd.. v. U.S. See Catron Commissioners 1429, Serv., F.3d 1436 with the result that we reached in 75 Douglas Fish & Wildlife (10th Cir.1996); County, creating split concerning a circuit Cape see also Hatteras Access Interior, designation applies whether NEPA Dep’t Pres. Alliance v. U.S. (D.D.C.2004). critical habitat under Section of the ESA. 133-36 EIS, designation extent that the critical habitat both in draft and final cation of an form, informational larger a positive also serves has a environmental effect on assurance public gives role. It species question. designa- The critical ‘has indeed considered notice, provides process public tion also for in the decision- concerns environmental Merrell, NEPA.”); goal of another and, perhaps signif- more making process,’ (“[W]hen Congress F.2d at 779 revised icantly, springboard public provides 1972, designed registration FIFRA Valley Robertson v. Methow comment.” public public notice and procedure Council, 332, 350, 109 490 U.S. Citizens provisions that differ materi- participation (1989) (quot- S.Ct. 104 L.Ed.2d ally that NEPA would re- from those Co., Elec. ing Baltimore Gas & U.S. quire.”). 2246) (internal citation omit- 103 S.Ct. ted). only Additionally, requires Although already acknowledged “the ESA we have im- agencies to consider the cumulative the differences between Section 7 of the actions, while NEPA pacts of non-federal NEPA, we do not think that the ESA requires agencies to consider the cumula- pronounced as distinctions those are impacts tive of all actions.” Fund Ani- Merrell, where the court concluded that mals, 136; at 50 448 see also apply registration NEPA to FIFRA’s “[t]o (describing § 402.02 the cumulative C.F.R. sabotage would delicate ma- process ESA); analysis under the 40 effects C.F.R. chinery Congress designed that to register (describing § the cumulative im- 1508.7 Merrell, pesticides.” new 807 F.2d NEPA). pacts analysis under We cannot observed, Congress specifically As have we say that Section 7 of the ESA renders contemplated agency that an could comply “superfluous” NEPA when the statutes NEPA discharging while its duties types evaluate different of environmental under Section 7 of the ESA. See 16 U.S.C. impacts through processes that involve 1536(c)(1) § (explaining biological varying degrees public participation. of required by assessment Section 7 of the Merrell, that we held the EIS re- “may part be undertaken as of a ESA quirement apply not does FIFRA’s agency’s compliance Federal with the re- registering pesticides scheme for because quirements of section 102 the National of processes markedly the two are different. (42 Policy Environmental Act of 1969 Merrell, (“The F.2d at See 807 778 differ- 4332)”). case, § very In this U.S.C. registration pro- ences between FIFRA’s agencies acknowledge federal that Recla- requirements cedure and NEPA’s indicate complete conjunction mation will an EIS in that NEPA Congress did not intend adoption implementation with its apply.”). Although Doug- both Merrell and BiOp, which undercuts the notion that County las conclude that NEPA does not processes the two are incompatible. Un- statute, apply particular to a substantive circumstances, der these we cannot con- they opposite Douglas do so for reasons. process by clude that the set out Section 7 County that 4 of the holds Section ESA of the ESA clashes with NEPA to such an superfluous renders NEPA because the that requiring pro- extent Reclamation to similar, processes sufficiently while sabotage duce an EIS “would the delicate Merrell holds that FIFRA renders NEPA machinery Congress designed.” Mer- superfluous processes because are suf- rell, Instead, 779. we find that ficiently Douglas Cnty., different. See (“[T]he Section 7 of the ESA fits within the broad Secretary F.3d at 1503 will consid- NEPA, impacts er that concern swath of statutes coexist with complete will an Reclamation EIS evaluat- NEPA.51 ing implementing BiOp. the effects of Douglas County court also noted (“Consultation, See 50 C.F.R. 402.06 con- to decisions Congress acquiesced ference, biological assessment proce- compliance and courts that agencies may dures under section be consolidated required desig- when with NEPA was not with interagency cooperation procedures nating critical under habitat Section statutes, required other such as the noted that in 1983 the Secre- the ESA. We Policy National Environmental Act tary stopped preparing of the Interior EAs (NEPA).”). designating And, critical habi- knowledge, and EISs before to our there tat, yet Congress did not address stating are no cases that an need interpretation of the statute when agency’s comply with NEPA because of its obli- Douglas it amended the ESA 1988. gations under 7 of Section the ESA. Cnty., 48 F.3d at 1504. The Sixth Circuit Next, Douglas County court stated might that it suggested had also dicta prepare that there is no reason an EIS necessary prepare not be an EIS before “when the action issue does not alter critical the 1988 designating habitat before natural, physical untouched environ- (citing amendments to the ESA. Id. Pac. Douglas Cnty., ment at all.” 48 F.3d at Found, Andrus, F.2d Legal explained that “the purpose We Cir.1981)). (6th here, But neither NEPA protect physical is to environ- interpreted agencies nor the courts have ment, purpose preparing and the permit noncompli- 7 of the ESA to Section agencies public EIS is to alert and the noted, NEPA. As the relevant ance with land, potential consequences adverse regulations agency indicate that the action Therefore, designation sea or air.” Id. complete carrying will an EIS while out its habitat, of which results in the land 7, critical under and the federal duties Section alone, being require does not defendants this case have assumed that issue left equivalent' argument 'functional that one 51. A number of other circuits have held that agency produce another.”); EIS need not an where process requires steps the same proce- the substantive statute at issue offers a Merrell, ("While at 781 we 807 F.2d hesitate equivalent” of the dure that is the "functional equivalence' adopt ratio- the 'functional See, process. e.g., EIS State Ala. ex rel. nale, of Congress that did not we are confident EPA, (11th Siegelman v. 504 F.2d registra- apply intend NEPA to to FIFRA Action, Cir.1990); Ecology Inc. v. Limerick tions.”). Comm'n, Regulatory U.S. Nuclear 869 F.2d Douglas County NRDC and Mer- relies (3d Cir.1989); 729 n. Walton Izaak rell, "displacement” which are cases rather Marsh, League Am. v. 367 n. equivalent” our than "functional cases. To (D.C.Cir.1981). knowledge, none of the circuits that have skeptical We have been "functional adopted equivalent” the "functional test have equivalent” approach and have not used this procedures set out Section 7 held Douglas Cnty., language in our cases. See require- equivalent to the EIS of the ESA are ("Courts have used a F.3d 1504 n. 10 urge Although ment. does not us to NRDC exempt equivalent' agency 'functional test adopt equivalent” approach, the "functional requirements.... NEPA The de- action from considered in the we note that factors fendants here do not advance the functional paragraphs preceding are the same ones that argument, equivalent so we do not address it. analysis. Re- we would address under that [plaintiff] would have us believe that the gardless language conduct the of the used to 'displacement' argument defendants make is regulations analysis, reveal the statutes and equivalent’ the same as the 'functional test. and NEPA involve that Section 7 of the ESA argu- agree. 'displacement' We do not processes different that measure different Congress ment asserts that intended to dis- impacts. procedure kinds of environmental place one with another. The Here, pressure does far more one form of on the natural envi- an EIS. ronment, have than leave nature alone. Humans implemen- such as Reclamation’s Bay-Delta ecosys- dramatically altered BiOp, distinguishable tation of the implement- if we assume that tem. Even portion a decision to continue to leave a returning ing step is a toward altogether. nature untouched state, ecosystem natural there is to its point, opinion At its broadest our project operations no will con- doubt Douglas County implied agency’s environment, physical tinue to alter the designation trig- of critical habitat did not ways. already albeit in different We have ger NEPA review because it was an envi- interpreted portion Douglas County this preservation ronmental effort. We ex- narrowly. quite See Kootenai Tribe of plained “[b]y designating critical Veneman, (9th Idaho v. 313 F.3d 1094 endangered habitats or threatened Cir.2002) (abrogated grounds by on other species, Secretary working pre- ‘is Serv., Soc’y v. Wilderness U.S. Forest prevent serve the ir- environment (9th Cir.2011)). F.3d 1173 In Kootenai retrievable loss of a natural resource.’ Tribe, announced a “Roadless Secretary desig- Thus the action of the prohibited Rule” that construction of new nating pur- a critical furthers the habitat Tribe, in certain roads areas. Kootenai pose of NEPA. EPA Requiring the to file argu- 313 F.3d at 1105. We considered an only an EIS “would hinder its efforts at ment that NEPA should not apply to the attaining goal of improving the envi- agency’s in light Douglas action ” Douglas Cnty., ronment.’ 48 F.3d at County court’s statement that “an EIS is recently applied principle We this required to leave nature alone.” Id. at Bay Oyster, in Drakes where we wrote Douglas 1114 (quoting Cnty., 48 F.3d at Secretary’s that “[t]he decision is essen- 1505). “[bjecause We held that human tially an environmental conservation ef- intervention, in the form manage- of forest fort, triggered which has not NEPA in ment, part has been of the fabric of our past.” Bay Oyster, Drakes 729 F.3d long, national forests for so we conclude (citing Douglas Cnty., at 984 48 F.3d at that, case, in the context of this unusual 1505-06). the reduction in human intervention that Douglas County We do not read would result from the Roadless either Rule actu- ally Bay Oyster or does alter the environmental status Drakes to stand for the quo.... proposition preserve Forest Service’s Roadless ini- that efforts to tiative an required per thus EIS under natural environment are exempt se noted, NEPA.” Id. at 1115. An action to Douglas lessen from NEPA.52 As recently 52. We observed that our court has ronmental assessment.... final environ- yet agency already to hold that an that has finding mental assessment resulted in a of no produced produce an EA need not an EIS NEPA.”). significant impact under In Hu- question only when the in will action have Society, question we mane did not resolve the impacts beneficial on environment. See agency produced whether an that has an EA Locke, Soc’y Humane v.U.S. of showing significant beneficial environmental (9th Cir.2010) ("As matter, a threshold impacts and no adverse environmental im- plaintiffs’ argument appears to raise an issue pacts complete must still an EIS. See id. at impression of first in this circuit: whether 1056. requires prepare NEPA an an EIS In order to hold that Reclamation need not significant when an action has a beneficial NEPA, comply with we impact significant would need to take impact but not adverse environment.”); because, ("[T]o steps see two substantial in this also id. at 1046 forward NEPA, case, comply prepared complete NMFS an envi- Reclamation did not even an implementation on the observation that Reclamation’s of the County court relied critical habitat under Sec- in- designating BiOp important has effects on human the ESA is an “action[] tion teraction with the natural environment. physical natural nothing to alter the people do[es] We know that millions of and vast Cnty., 48 F.3d at Douglas environment.” produc- areas of some of America’s most Similarly, Bay Oyster, in Drakes impacted by tive farmland will be Recla- only purported “adverse environmental impacts mation’s actions. Those not were designating the area consequences” of sum, BiOp. the focus of the we cannot a wilderness were “short-term question as reach an informed decision about the ex- harms, heavy such as noise associated with implementation tent to which Bay’s Drakes machinery needed to remove an preservation is environmental action the area to structures” order return Douglas County the vein of and Drakes Bay Oyster, 729 its natural state. Drakes Bay Oyster because we do not know how noted that “such relative- F.3d 984. We impact action will the broader natural ‘sig- ly harms do not themselves minor environment. find no We basis ex- nificantly the environment such affeet[ ]’ empting Reclamation from the EIS re- But way implicate as to NEPA.” Id. quirement. Valley Methow See Citizens here, does not even contest NRDC Council, 490 U.S. S.Ct. 1835 implementa- court’s conclusion that district (“NEPA important ensures that effects BiOp “significantly affect[s] tion of the will not be overlooked or underestimated Luis human environment.” See San & only to .be discovered after resources have Deltar-Mendota, been committed or the die otherwise implementing effects Whatever cast.”). recognize preparation We environment, it have on the human might of an EIS will not alter Reclamation’s obli- they complex are more apparent gations may under the ESA. But the EIS wide-ranging than the removal of a well inform Reclamation of the overall buildings Bay Oyster. few in Drakes including the human costs—of fur- costs— informed, only speculate thering can the ESA. So Reclama- point, At this we seeking exemp- significant option effects will tion has the about what kind Endangered from the eventually implementation result from tion from the ESA yet Species See 16 U.S.C. Reclamation has Committee. because 1536(e).53 beyond dispute completed its EIS. But may exemption holding agen- an obtain an from the ESA's EA. We see no basis for that prohibitions. See 16 U.S.C. cy altogether when it substantive can avoid NEPA review 1536(a)(2) (“Each agency § shall ... agency federal that an action will have bene- believes authorized, funded, any insure that action or impacts we ficial on the environment when likely agency ... is not to carried out such agency pro- have not even excused an from any jeopardize of en- the continued existence ducing that its an EIS when its EA shows dangered species species or re- or threatened exclusively impacts will have beneficial action modification words, sult in the destruction or adverse even if on the environment. In other species ... unless such of habitat of such assuming that Recla- we had some basis for exemption agency granted for has been an implementation would mation’s pursuant [16 such action the Committee exclusively impacts the en- have beneficial 1536(h) ].”); § see also Portland Audu- U.S.C. vironment, founda- we would still lack a firm Comm., Soc’y Endangered Species bon v. holding need not tion for that Reclamation Cir.1993) ("The (9th Com- F.2d and, necessary, prepare an EA if an EIS. Endangered Spe- by the mittee was created making purpose final Endangered Species cies Act for the sole Committee is exemptions applications from through applicant decisions on avenue which an exclusive argues pertinent that Reclama- not Finally, NRDC issue also to our hold- tion’s issuance of the 2004 OCAP is the implementation ing that Reclamation’s “major action[ Federal should have ]” BiOp requires agency prepare subject to NEPA review instead of been an if implemen- EIS. Even Reclamation’s implementation BiOp. its of the 2008 In a tation of the 2004 OCAP was a final deci- case, alleged different NRDC that Recla- changed sion that quo status of the “approval implementation” mation’s of project operations way in a signifi- triggered obligation the 2004 its OCAP environment, cantly affected the that does Pac. complete EIS. See Coast Fed’n of not mean that implementa- Reclamation’s Fishermen’s Res. Fisheries Ass’n/Inst. for tion change of the 2008 did not also Gutierrez, l:06-cv-00245, No. 2007 WL quo way significantly the status (E.D.Cal. 2007). *4 June environment, thereby alters the requiring Gutierrez, the district court held that an additional EIS. subject the 2004 OCAP was not to NEPA NRDC bolsters its contention that agency because its was not a “final action.” apply NEPA should to the 2004 OCAP (“[The See id. *12-13 OCAP] do[es] BiOp by referencing lieu of the 2008 Sec- implement any They actions or inactions. regulations. tion 7 of the ESA its As any proposed informational. If above, regula- noted the statute and its changes are initiated that will have the explain tions that an action agency like environment, requisite effect on the such Reclamation can coordinate prepara- agency subject changes will be action biological tion of its assessment with its purpose NEPA review. The of the OCAP obligations is ‘to serve as a under NEPA. description baseline See 16 U.S.C. 1586(c)(1); operating § facilities and environment 50 C.F.R. 402.06. NRDC ”). CVP SWP.’ provisions reasons that these suggest that NEPA review of the BiOp required is not NRDC contends that Gutierrez was because complete Reclamation could not wrongly decided and that the district court its EIS until after its Section 7 consulta- required should have Reclamation to com- tion with the FWS rather than at the same plete an EIS on the 2004 OCAP. Yet obligations. time as its consultation But and the plaintiffs NRDC other in Gutier- fully even if we credit this line of reason- appeal rez did not the district court’s deci- ing, it does not affect our concerning sion its conclusion that NEPA claims. Not only collaterally NEPA estopped applies implemen- NRDC to Reclamation’s Gutierrez, relitigating the decision in but tation of BiOp. We need not locate Committee, be, the Act....” it Because is the ultimate arbiter unlikely may cies however endangered species, of the fate of an the Com- preparation renders the of an EIS more than Squad.”). mittee is known as "The God involving a mere academic exercise in cases only Committee has convened aon handful of Although Section 7 of the ESA. only granted exemp- occasions and has two ignore obligations cannot under ESA Soc'y, tions. See Portland Audubon 984 F.2d impacts aspects because of the on other at 1537. Commentators have discussed environment, might human the Committee whether action from the Committee would be wish to making consider these factors in an appropriate under these circumstances. See And, noted, exemption previously decision. as Yuknis, Note, generally Eric M. Would a "God contemplates agency the statute that the will Squad" Exemption Endangered Spe- under the completed by providing have an EIS that the Crisis?, cies Act Solve the Water 38 California agency Committee itself need not do so if the (2011). B.C. Envtl. Aff.L.Rev. 567 already prepared respect has an EIS with purposes, noting subject For our it is worth that the the action to the Committee’s review. possibility 1536(k). Endangered Spe- § of review the See 16 U.S.C. simply present are here. confirmation those cases unequivocal affirmative of our particu- cognizant commitment to statute that We every substantive review; “mak[ing] NEPA more of an ‘ob- requires NEPA avoid agency lar action prevent never mention structionist tactic’ to environmen- substantive statutes most protection may already not need to be- tal than it have They at all. do NEPA 1508; Cnty., that “to the at provides Douglas NEPA itself become.” cause ... possible agencies Bay Oyster, all see also Drakes F.3d fullest extent noted, But, complete shall” the district court re- the Federal Government and, necessary, if an EIS for all and RPAs without vaca- an EA manded tur, significantly actively manage affect- it continues to “major Federal actions ing quality completing of the human environment.” Reclamation’s deadline 4332(2)(C). process. The fact Sec- EIS We conclude that Reclama- U.S.C. NEPA, expressly obligated comply the ESA mentions tion is tion 7 of than re- requirement supports, judgment rather and we affirm the of the district EIS buts, applies respect notion that NEPA court with to the NEPA claims. at issue here. action VI. CONCLUSION ap- that NEPA starting point was

Our implementation to Reclamation’s plies reasons, forgoing judgment For the “major Federal because part of the district court is reversed in affecting quality significantly action[] part. affirmed in The matter is remanded acknowl- of the human environment.” We proceed- to the district court for further edge previously that we have held ings opinion. with this Each consistent apply partic- does not requirement EIS appeal. its own costs on party shall bear in the absence of ular actions even IN PART AND AF- REVERSED statutory exemption. See express FIRMED IN PART. 984; Oyster, 729 Bay Drakes F.3d Merrell, 1507; Cnty., Douglas 48 F.3d at GLOSSARY OF TERMS F.2d at 781. But the factors identified

APA Administrative Procedure Act Biological BA Bay-Delta Assessment n Bay Joaquin Delta Francisco and Sacramento-San San biologicalopinion Bay-Delta Program CALFED CALFED cubicfeet Central Central California per cfs second Valley Project Project Valley CVP CVPIA Act Improvement Department Resources DWR EIS ESA FCRPS FIFRA FMWT FWS ITS of Water Impact Environmental Statement Endangered Species Act System Federal ColumbiaRiver Power Federal Fall Midwater Trawl index U.S. Insecticide, Fungicide, RodenticideAct Fish and WildlifeService incidental take statement National Marine Fisheries Service NMFS Act Environmental Protection NEPA National Operating Criteria and Plan OCAP OMR RPA Middle Rivers Old and prudent alternatives reasonable Bureau of Reclamation Reclamation U.S. SWP Project State Water ARNOLD, Judge, concurring by Circuit in Deriso that should salvage be scaled population some measure of dissenting part: abundance. part and issue, As for the merits of this Appel- III, respectfully I dissent Parts lants do not contend that the use of raw IV.A, IV.B, IV.E, and V.B court’s salvage scientifically acceptable; data was opinion and concur the rest of it. I they pre- maintain instead that the flow address with in Part III the issue dealt scription supported also relied on and was considering it arises in the merits of the by my other information. on Based review challenges BiOp. to the information, however, of this did all, not connect it to flow limits at or there 1. I do not believe that the district explanation why yielded was no for it in holding BiOp’s court erred prescription BiOp specified. flow arbitrarily OMR flow limits were set toAs FWS’s use of normalized data in the all, I capriciously. First of discern no ITS, I am not convinced that this is rele- admitting portion error in of the declara- scientifically vant to whether it was sound Deriso, tion from Dr. Richard who holds only salvage FWS to use raw data to degrees advanced in mathematics and prescription. set the flow While certain biomathematics, discussing the use of raw DWR comments support prescrip- the flow salvage justify prescrip- data to the flow tion, parties dispute do not that these tion. A decision to include evidence that is comments arose from the district court’s outside the administrative record is re- previous remedial imposition pre- of such a discretion, viewed for abuse of see Lands which, scription, as the district court not- Council, 395 F.3d at 1030 n. and admit- ed, occurred before the court became ting this evidence fell within one of the using aware that salvage raw data was not exceptions general narrow rule accepted scientific methodology. Because evidence, against extra-record because FWS based prescription solely its flow necessary was to explain technical terms unexplained data, salvage use of raw I subject matter, complex or see Nw. Envtl. believe that its expertise methodological Serv., Advocates v. Nat’l Marine Fisheries deference, matters is not entitled to since (9th Cir.2006). 460 F.3d Fur- rationally use was not connected to thermore, Dr. Deriso’s declarations were science, the best available see W. Water- consistent with advice indepen- offered Project Kraayenbrink, sheds peer dent reviewers and draft notes of a (9th Cir.2011); and because FWS delta smelt evaluation team at FWS as- did not consider all relevant factors or issued, sembled before final BiOp articulate a rational connection between testimony with the experts made, of Rule 706 Dr. facts found and the I choices Quinn, agree Punt and with the recognized Dr. who district court that its deter- Dr. mination as to prescription the flow Deriso’s was declarations and stated that the arbitrary and capricious, see id. at 496. validity regimes of the flow specified in the incomplete undermined I also concur the district court’s analysis. Nor any was there “battle of the BiOp’s conclusion that determination here, experts” maintains, as the court be- arbitrary of X2 was capricious. First cause the responses to Dr. Deriso’s decla- all, there was no abuse of discretion in rations from FWS’s mathematical statisti- the district court’s decision to admit the cian, Newman, mostly Dr. Ken Miller, were declarations of Aaron a DWR tech- vague, and generally agreed he with Dr. nical engineer closely who worked

657 II, conducted comparison II FWS had which develop to Calsim Reclamation II-Dayflow expect- differences that were to the Calsim did not show relevant were Council, noted, F.3d Lands 895 court the record See ed. As the district comparison. was experts considered, of the n. 11. No battle much at 1030 did not reflect that FWS assessment by doing so: Miller’s created Fur- recognized, less the sources of bias. II-Dayflow validity of the Calsim thermore, significant there are differences with the testi- was consistent comparison models, X2 including two how between the Quinn Drs. experts Rule 706 mony of would positions are determined. FWS assessment; Punt, recognized Miller’s who significant differ- have to address these hydrologist the declarations FWS on way ences in some to obtain information expe- Hilts —whose credentials Derek rely reasonably it could to base the which Miller’s, and who similar to rience are conclusions, BiOp’s including Action on unre- mostly helped draft —were management of X2’s location. FWS declarations on sponsive to Miller’s required provide to some evidence was comparison that the sources of bias several its conclusions to ensure that supporting view, court my the district introduced. judgment rendered its no clear error simply extra-record evidence relied on this arbitrary capricious. See actions had considered to determine whether FWS League Wilderness Defenders-Blue factors, here, the sources of all relevant Project Biodiversity Mountains v. U.S. Advocates, bias, F.3d Nw. Envtl. see (9th Serv., 1211, Forest 549 F.3d Cir. comparison. relying on the before 2008). dispute not Appellants do Projects analyze proposed the effects of to existed, or that the biases sources of bias habitat, includ- on smelt and its operations material; or and the clear significant were with- Doing so was well ing X2’s location. requiring to use the best purpose of highly role. Because techni- in the court’s is to ensure scientific evidence available involved, were it was difficult cal matters implemented haphaz- is not that the ESA if all rele- FWS considered determine speculation. on surmise or ardly or based looking factors without outside vant Fed’n Fishermen’s See Pac. Coast should have record to see what matters Ass’ns, 426 F.3d at 1094-95. considered, not. Inland Em- but were See Forest pire Finally, agree Pub. Lands Council v. U.S. I with the district court (9th Cir.1996). Serv., why F.3d 760 n. 5 sufficiently explain not that FWS did properly dis- The district court could km critical 74 km and 81 were selected as “in charge duty engage ‘substantial habitat. preserve for X2 to smelt points ” by taking word inquiry’ simply FWS’s not therefore re- The district court was considered all relevant matters. had this mat- give FWS deference on quired SeeAsarco, 616 F.2d at 1160. ter, reasoning could not be rea- as FWS’s San sonably from the record. See inferred II-Day-

FWS’s choice to use Calsim Auth., 672 Luis & Delta-Mendota Water by the unsupported comparison flow justifying As to the other F.3d at 700. Ecology analysis. reasoned See requisite evidence, not relate to the (9th it either does Castaneda, Ctr. v. undisputed points or it is Cir.2009). choice of critical from DWR received Comments time in for the first it was offered echoed and other entities —which were I am reluc- proceedings. post-judgment testimony experts’ the Rule —alerted of an ade- bias, over the absence pass tant yet sources of FWS to the several in the administrative quate explanation using the only explanation in the post-hoc relying Appellants’ II record that a Calsim to Calsim comparison was 658 Soc’y

rationalizations. See Humane considered challenged FWS which Locke, conclusions, 1040, BiOp’s Lake, United States v. Pyramid F.3d see (9th Cir.2010). 1049-50 F.2d at and that there is no indica- tion that this occurred here. But the dis- authority I find no requiring FWS to trict court failed to consider another basis specifically analyze, address in the finding for action independently record, BiOp or ques- administrative liable, namely, legally reliance on a flawed tion of whether RPA meets the non- BiOp. Discerning such flaws involves no jeopardy question elements. I also technical or expertise, scientific so failure district Greenpeace court’s reliance on may to do so result action based on Serv., Nat’l Marine Fisheries reasoning not in accordance with the law (W.D.Wash.1999), F.Supp.2d 1248 for the rendering thus arbitrary the action proposition that expla- there must be some capricious. See Wild Fish Conservancy v. nation in the administrative record as to Salazar, (9th Cir.2010). 628 F.3d why FWS that all concluded four elements legal district court’s conclusions neces- for a valid RPA were satisfied. Nonethe- sarily arose from fact-finding, but less, I would affirm district court on clearly, court and I believe correctly, con- issue, this because the record belies Appel- cluded that FWS had not used the best lants’ contention that DWR and Reclama- available science or considered relevant tion raised no about nonjeo- concerns factors, and had arbitrarily acted and ca- pardy elements. The record shows that because, priciously, among things, other it concerns were relating raised to RPA feas- relied on II-Dayflow the Calsim compari- ability relationship and its to the action’s son, and did not use salvage normalized purpose intended (providing water for var- data to set the flow prescription. The uses), ious possibly to DWR’s and district court therefore should have found authority Reclamation’s implement the Reclamation independently liable under Thus, RPA. under FWS’s own interpreta- § ESA 7 for accepting legally flawed 402.02, tion of required it was to consid- BiOp and immediately beginning imple- er and address these elements specifically mentation of the RPA modifying opera- in the instant or administrative rec- tions. ord. sum, I find no abuse of discretion in Finally, as agency, the action Recla- L the district court’s limited admission rely mation could not solely on FWS’s evidence outside the administrative record BiOp to establish conclusively compli- as relevant to the OMR flow limits and the ance with its obligations substantive under X2, determination of including the use of 7,§ ESA because it could not delegate its II-Dayflow the Calsim comparison. I be- responsibility to see that its actions would lieve that determining whether FWS’s smelt, jeopardize Lake, see Pyramid decisions on these matters in the BiOp 1415; 898 F.2d at and as the action agency arbitrary, were or capricious, otherwise blindly it could not adopt FWS’s conclu- law, not in accordance the district sions because it ultimately responsible court’s analysis thorough and well- for compliance, Tacoma, ESA City see reasoned. I disagree While with the basis Wash. v. Fed. Energy Regulatory for the district court’s conclusion that Comm’n, (D.C.Cir.2006). I non jeopardy elements must be addressed agree with the district court that Reclama- in the BiOp record, or administrative I tion subject would be independent ESA nonetheless believe that affirmance is war- liability if possessed new information not ranted on this Finally, issue. I believe the reau”) impact prepare environmental Reclama- found have court should district (“EIS”) ad- periodically before ESA statements liable under independently tion the Pali- of water from the flow BiOp. justing flawed legally relying on a ‘We responded: panel Dam?” The sades ” RAWLINSON, Judge, Circuit is Id. at answer ‘No.’ that the are clear in dissenting part: part in and concurring majority bulk of the the I concur in Upper Snake reasoning of apply To the rationale only with the disagree I opinion. River, name of only change the we need of Recla- Bureau conclusion and me, the posed. To in the question the dam of implementation and adoption mation’s of The Bureau equally clear. answer is obli- triggered Biological Opinion the prepare to required not was Reclamation Envi- National with the comply gation to when Impact Environmental Statement (NEPA) by prepar- Act Policy ronmental al- prudent and the reasonable adopted Impact Statement an Environmental ing to Biological Opinion the ternatives from the En- under required generally is that by the water controlled flow limits for set Act. Species dangered project. Valley Central water the in that keep to mind important It is in supply the Cen- like the water Much imple- and adopted of Reclamation Bureau in the of water Valley, amount tral the Biological Opinion the mented detailed considerably from River “fluctuates Snake to threatened harm case to this alleviate amount of year, depending the year criti- their species endangered and and/or Id. Wa- mountains....” pack in the snow the fed- of by operation habitat caused cal in reservoirs captured from the river ter Pro- Water Valley erally operated Central regu- and is controlled water flow and the State Water operated and the state ject Plaintiffs id. The by dams. See lated Project. of Reclamation’s the Bureau challenged matter, agree that I do not an initial As flow rate below the to reduce decision prudent of “reasonable adoption the of less 1,000 per second times cubic feet threat- harm to alleviate alternatives” to rejecting In id. at 234. precipitation. See their species endangered ened and and/or applicable, that NEPA was contention the operations of ongoing from critical habitat held: district court “major constituted projects water Dam, the the Palisades In case of requirement triggering action” Federal actions are routine fluctuating flows Impact Statement. of an Environmental Na- upon Mother contingent which are 4332(2)(C) (requiring 42 U.S.C. See runoff, snowpack, precipitation ture for Impact Environmental of an preparation its routine part As carryover. actions”). I “major Federal Statement Bureau ongoing operations, our view toward that persuaded am flows de- fluctuates the Reclamation similar with facts cases rationale two past weather conditions upon pending Upper The first in this case. those Overall, views the Court and future. Unlimited Chapter Trout River Snake Palisades of flows below the fluctuation Cir.1990). (9th Hodel, 921 F.2d 232 do not which operations,” “ongoing River, posed panel Snake Upper ... NEPA. comply have Court “Did the District following question: added). omitted) (alterations (emphasis Id. Envi- the National concluding that err equivocation, and without (“NEPA”) again not Once did Policy Act ronmental (“Bu- clear panel “[W]e concluded: Bureau of require the Reclamation the district court’s conclusion was correct.” During period, [NEPA]. have they Id. panel explicitly determined from to time depending time on the “major level, water flow fluctuation was adjusted river’s up flow or down Federal under NEPA. action” Id. As with the volume of water released Valley operations, Central the Palisades Dam. they prior years did in What project completed well before the ef- they doing what were during period fective NEPA. According date of See id. under consideration were no more than panel, post-construction managerial fluctuation the routine actions regularly operation flow constituted water routine carried on from the outset without major of the dam rather than a action change. They simply operating the triggering requirements. NEPA facility See id. in the manner intended. short, at 234-35. they new, are doing nothing nor extensive, more nor other than that con- panel heavily relied the reason- templated project when the was first ing Andrus, in Trinity County v. operational.... (E.D.Cal.1977). F.Supp. 1368 In Trinity 921 F.2d at 235. County, sought injunction Plaintiffs prohibiting the Bureau of Reclamation Because the Bureau of sim Reclamation lowering the water level in the Cen- ilarly continued to do nothing more than tral Valley project water during a drought “monitor[ and control ... the flow ] rate to year. See id. at 1371. The district court practicable ensure that the most conserva the framed issue as “not whether the ac- achieved,” tion 235-36, of water id. at tions are of magnitude require sufficient panel the concluded that no Environmental the of an preparation [Environmental Im- Impact Statement was required. id. See pact Statement], but rather whether 236; see also Canyon Grand v. Trust apply NEPA was intended to at all to the Reclamation, United States Bureau of continuing operations of completed facili- (9th Cir.2012), F.3d 1021-22 ” (citation omitted). ties .... Id. at 1388 (citing Upper amended Snake River and similarly concluding that execution of an The district court distinguished water nual operating plans regulating fluctua flow determinations from cases “when a tions in water flow was “not major feder project takes place stages incremental al requiring action compliance with major of proportions or when a revision or NEPA”); Burbank Anti-Noise Group expansion of original the facilities is con- Goldschmidt, (9th 623 F.2d Cir. (citations omitted). templated ...” Id. 1980) (“An EIS need not discuss envi the The district Trinity court in County also ronmental effects of mere continued opera noted that the Bureau of Reclamation had (citation omitted). tion of a facility-”) not enlarged its diversion capacity, or re- procedures vised its or standards for re- The second case with similar facts of lease water or draws from reservoirs. Douglas Babbitt, County v. F.3d See id. at 1388-89. Under these circum- (9th Cir.1995). In Douglas the County, stances, requirements the of simply NEPA panel considered whether designation triggered. were not See id. Taking its of critical habitat under Endangered cue from Trinity County, panel in Species required Act compliance with Upper Snake River reasoned that: NEPA. id. at 1497. In process See

The Federal defendants in issue, this case deciding panel focused on been operating had upwards the dam for between requirements difference years ten before the effective date of purpose of NEPA and require- evaluating] all reasonable alternatives to a Endangered purpose ments and action ...” Southeast Alaska [Endangered Species proposed “The Act. Species Highway NEPA without v. Federal Ad goals of Conserv. Council furthers the Act] (9th Cir.2011) Impact min., 1050, 1056 demanding an 649 F.3d [Environmental 1502.14(a)). (emphasis at 1506 ...” Id. An 40 C.F.R. (quoting Statement] added). Indeed, Spe- [Endangered “[t]he does not violate NEPA declin statute whose is a substantive Act] cies alternatives that have ing to re-examine By desig- ... extinction prevent See, is to goal e.g., Ho previously been evaluated. Secretary is ... nating critical habitats Ad v. Federal Transit noluluTraffic.com ...” the environment working preserve 1222, 1231, min., 13-15277, No. (citations quota- and internal Id. at 1506 (9th February 607320 at *6 Cir. 2014 WL omitted). panel concluded tion marks 2014). “fur- critical habitat designation undisputed Biological Opin- It is NEPA purpose [and] [re- ther[ed] objectively “rigorously explore[d] ion would to file an EIS quiring [agency] alternatives” all reasonable evaluate[d] attaining goal only hinder its efforts pages. over hundreds of C.F.R. (cita- Id. the environment.” improving of 1502.14(a). Indeed, express purpose § omitted). quotation marks tion and internal Biological Opinion develop was to nature of to the substantive contrast prudent alternatives to the reasonable Act, Species “NEPA Endangered that were harmful to the existing activities designed essentially a statute procedural (set- 402.14 environment. See C.F.R. are issues to insure that environmental procedures forth formal consultation ting consideration in the decision- given proper proposed ac- impact ensure ” Alaska Trustees making process.... their critical species tions on listed (9th 1378, Hodel, Cir. v. 806 F.2d considered); see also Save fully are habitat (citation omitted). 1986) “NEPA does not Block, 840 F.2d Yaak v. Committee results, simply pro but particular mandate (not- (9th Cir.1988), 714, amended necessary process to ensure vides the an envi- requirement ing the NEPA take a hard look at agencies federal “include dis- assessment brief ronmental ac consequences of their environmental impacts of of the environmental cussions Tribe Idaho v. Vene tion.” Kootenai ”) action and alternatives proposed (9th man, 1094, 1115-16 Cir. 313 F.3d added). (emphasis 2002), grounds, part on other overruled Douglas listing As with the habitat Forest Soc’y v. States United Wilderness in this case Biological Opinion County, the (9th 1173, Serv., 1178-79 Cir. an Environ- function as the same served 2011) (citation quotation and internal 48 F.3d at Impact See mental Statement. omitted). NEPA purpose “The marks (“The obligation procedural is a EIS or a mechanism enhance provide is to give prop- agencies to assure that designed fur prevent the environment improve con- to the environmental er consideration Bay damage.” Drakes irreparable ther actions....”) (citation of their sequences Jewell, 13-15227, 747 v. No. Oyster Co. omitted). marks quotation internal *12 1073, 1090, 114699 at 2014 WL F.3d are de- (cita requirements NEPA Because the (9th 2014), as amended Cir. Jan. to en- a mechanism “provide signed omit quotation marks tion and internal improve the environment ted). hance or Impact State An Environmental damage” to irreparable further regulations prevent the NEPA implements ment environment, Im- Environmental no objectively the exploring] and by “rigorously pact Statement is needed “for federal ac- Quijada CORONADO, Petitioner, Raul the environment...

tions conserve (citation Id. at 1505 and footnote reference omitted). circumstance, In this as in HOLDER, Jr., Attorney Eric H.

Douglas County, procedure “the NEPA General, Respondent. (internal superfluous.” Id. at 1503 seem[s] omitted). quotation panel marks As the No. 11-72121. recognized Bay, in Drakes if the federal essentially Appeals, United States Court of decision “is an environmental effort,” obligations conservation NEPA Ninth Circuit. triggered. 2014 WL 114699 at *12. Argued July and Submitted 2013. panel explained that the Endangered Species goals Act NEPA Filed March “farthers demanding without (quot- EIS....” Id. 1505,1506)

ing Douglas County, 48 F.3d at

(emphasis Bay). in Drakes We should be

similarly “reluctant to make NEPA more

of an prevent obstructionist tactic to envi- protection

ronmental than it may already

have become.” Id. at (quoting Doug- *13 1508). County,

las 48 F.3d at At this

point, imposing overlay of NEPA re-

quirements only “would hinder Bu- [the

reau of efforts at attaining Reclamation’s] goal of improving the environment.” (citation

Douglas County, 48 F.3d at 1506

omitted). Majority Opinion, See p. 652

([T]he preparation of an EIS will not alter obligations

Reclamation’s under

ESA....). require There is no need to

the Bureau of engage Reclamation to busywork.

scientific simply “This case present

does not type of situation

NEPA was intended address.” Bur- Group,

bank Anti-Noise 623 F.2d at 117.

I respectfully dissent. notes confounding temporal trends this decision was motivated a concern during period.... the baseline for the absolute number of smelt entrained experts pumps, district acknowl- in the not the relative number of As the court’s edged, significant practical population faces smelt: “The current cannot to- challenges setting mortality through flow rates to lerate direct adult en- OMR approaching For trainment levels even minimize delta smelt entrainment. example, day-to-day through variations OMR ‘moderate’ take as observed sampling22 and “noise” in smelt used historic record of recent decades.” flows Thus, designed the RPA is to establish abundance and distribution of significant confounding pre-spawning “reduce entrainment of adult the delta smelt are

Notes

notes that “[t]he Delta-wide increase in water a. Predation transparency may have preda- intensified pressures smelt,” tion on delta citing the One stressor to predation. the smelt is documenta[tion],” “[w]ide including several As the acknowledged, there is much specific studies, indicating that water clari- here that is unknown. It is known that in ty significantly predation pe- influences 1960s, when the delta smelt were more fishes, lagic including many species. smelt plentiful, they prey bass, were for striped BiOp at 183. crappie, black white catfish. 183. The observed that un- “[i]t district court analysis found this known whether predation incidental by lacking, specifically citing the FWS’s fail- (and striped bass other lesser predators) explain striped ure whether preda- bass represents a substantial source of mortali-

Case Details

Case Name: San Luis & Delta-Mendota Water v. Natural Resources Defense Coun

Court Name: Court of Appeals for the Ninth Circuit

Date Published: Mar 13, 2014

Citation: 747 F.3d 581

Docket Number: 11-15871, 11-16623, 11-16662, 11-16617, 11-16624, 11-16621, 11-16660

Court Abbreviation: 9th Cir.