Pacific Coast Federation of Fishermen's Associations (Pacific Coast) and seven additional environmental and fishing organizations challenged a Bureau of Reclamation (BOR) and National Marine Fisheries Service (NMFS) plan to comply with Endangered Species Act[1] (ESA) obligations to protect threatened Southern Oregon and Northern California Coast (SONCC) coho salmon in the operation of a BOR irrigation project in the Klamath Basin. The district court invalidated as arbitrary and capricious the final phase of a three-phase plan, which required BOR to provide 57 percent of the Klamath's long-term flows, but upheld the plan's remaining short-term phases. Neither party appealed the district court's decision to invalidate the long-term plan, but Pacific Coast appealed the decision to uphold the short-term phases of the plan. The Ninth Circuit concluded the short-term plan was arbitrary and capricious, and remanded the case back to the district court for appropriate injunctive relief to be determined.

The Klamath Project, a federal irrigation project established in the early 1900s, is operated by BOR in the Klamath River Basin in Northern California and Southern Oregon. Water use in the Klamath Basin has been the source of much contentious litigation in recent years between irrigators and federal agencies, environmental groups, and tribes.[2] SONCC coho spawn and mature in the Klamath River main stem and tributaries, and as an anadromous species, also spend a portion of their lives in the ocean. Since the Klamath Project's construction, SONCC coho have continued to live in the river and its tributaries below Iron Gate Dam, the part of the Klamath Project closest to the Pacific Ocean. The Klamath River SONCC coho population has plummeted during the last century, from an estimated 50,000 to 125,000 wild coho in the 1940s to fewer than 6,000 in 1996. In 1997, SONCC coho were listed as a threatened species under the ESA.

SONCC coho have a three-year life cycle, half of which is spent in freshwater and half in the ocean. After hatching in March, the coho fry live in shallow areas near stream banks until for 15 months, when the fry become smolt and head out to sea between March and June. Three years later, the coho return to the same stream in which they were born to spawn, migrating between September and February and spawning between November and January. Thus, sufficient flows need to be available in the Klamath and its tributaries for coho to migrate upstream in the fall and downstream in the spring. NMFS considers it most critical to provide adequate instream flows March through June, when newly hatched fry need shaded habitat near stream banks and migrating older fish need enough water in the river to make it out to sea. During summer, there must be sufficient water in the river to protect the maturing fry from lethally high temperatures, and in the fall, there must be sufficient water to facilitate upstream migration and access to the Klamath's tributaries for spawning.

The ESA requires federal agencies to "insure" that "any action authorized, funded, or carried out" by the agency "is not likely to jeopardize the continued existence" of any threatened or endangered species, or to result in the destruction of critical habitat.[3] If there are threatened or endangered species present in the area of a proposed federal action, the agency must determine whether that action is likely to affect the species by preparing a biological assessment.[4] If the biological assessment reveals a fish species is likely to be affected by the proposed action, the agency must consult with NMFS, which will prepare a biological opinion (BiOp) to determine how the proposed action might affect the species or its critical habitat.[5] If the proposed action might cause jeopardy to a listed species or its critical habitat, NMFS must recommend "reasonable and prudent alternatives" to the proposed action to avoid jeopardy to species.[6]

To ensure compliance with the ESA in its Klamath Project operations, BOR first consulted with NMFS to prepare a BiOp for Klamath River SONCC coho in 1999. Following the severe drought in 2001, which resulted in BOR withholding water from Klamath Basin irrigators in order to remain in compliance with the 1999 BiOp, BOR prepared a biological assessment for revised operations in 2002. BOR proposed maintaining flows in the Klamath based on type of water year, providing less water in dry years and more water in wet years, using the historic average flows over the past ten years to establish its base flows. Any water over these flows would be available for irrigation, and BOR would establish a water bank to meet flow requirements. NMFS reviewed the BOR's operation plan in the BiOp at issue in this case, and concluded that BOR's plan of operation would cause jeopardy to SONCC coho and adversely modify its habitat. NMFS believed BOR's use of minimum flows over the past decade to establish its baseline flows would have the end result of driving down the average flows and lead to adverse modification of critical habitat, as well as harm to coho in the main stem. In addition, NMFS was concerned the decreased flows would reduce suitable fry habitat, impair downstream migration, and cut off access to tributaries when the coho migrated back upstream. NMFS proposed a "reasonable and prudent alternative" (RPA) to avoid the jeopardy finding in BOR's proposed operation plan.

The RPA, the subject of the appeal, covered Klamath Project operations from 2002 through 2012. It made BOR responsible only for the share of the water losses caused by the Klamath Project's operations. The Klamath Project irrigates 57% of the Klamath Basin, thus the required BOR to provide 57% of the water needed for the coho, establish an intergovernmental workgroup to address the remaining 43% of the flows, and allowed BOR to use a water bank to store water to meet its flow requirements. The RPA was divided into three phases. During the first phase, lasting from 2002 to 2005, BOR was required to provide at least the flows established in its biological assessment and supplement these flows in the spring and summer with water from a water bank. The second phase, spanning 2006 to 2010, required BOR to increase its storage in the water bank to 100,000 acre feet and to provide either 57% of the necessary minimum flows, or the flows in the biological assessment. In the final phase, NMFS anticipated that the full instream flow requirements for the salmon will be met through BOR's 57% share and the 43% contribution from unspecified sources. The district court overturned the final phase of the RPA because it found the requirement that BOR was responsible for only 57% of the minimum flows to be arbitrary and capricious.

This appeal challenges the first two phases of the RPA, which Pacific Coast alleges the district court erred in not finding them to be arbitrary and capricious. The standard of review for the district court's grant of summary judgment is de novo.[7] The BiOp is subject to review under the Administrative Procedure Act (APA)[8] as a final agency action, which may be set aside if it is "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law."[9] The court will sustain an agency action if the agency has provided adequate facts to demonstrate a rational connection between the record and its conclusion.[10] Although deferential to the agency decision, the court will evaluate whether the agency action was "based on a consideration of the relevant factors and whether there has been a clear error of judgment."[11] An agency decision will be overturned as unlawful if it relied on factors beyond the scope of those Congress intended it to consider, failed to consider an important factor, was supported by an explanation that runs counter to the available evidence, or is "so implausible that it could not be ascribed to a difference in view or the product of agency expertise."[12]

The Ninth Circuit found the short-term flow requirements in the RPA to be arbitrary and capricious because NMFS failed to provide evidence on the record in support of its decision to delay delivery of sufficient minimum flows for the first 8 years of the 10-year plan. While the district court assumed that NMFS had "implicitly considered" that all three phases of the RPA would avoid causing jeopardy to salmon, the Ninth Circuit refused to make the same assumption. "It is a basic principle of administrative law that the agency must articulate the reason for its decision."[13] When reviewing a BiOp, the court will consider only "what the agency actually said" and not what it claims to have "implicitly recognized."[14] Thus, the Ninth Circuit undertook a review of the entire BiOp to determine whether NFMS had sufficient evidence on the record to reach its no-jeopardy finding for the first two phases of the RPA.

The Ninth Circuit undertook a "searching and careful"[15] review of the BiOp to identify the basis for NMFS no-jeopardy conclusion for the initial phases of the RPA. It concluded that reasoning supporting NFMS conclusion "cannot be reasonably discerned."[16] In light of the agency's analysis of the benefits of the long-term flows and the absence of any analysis of the effects of the much-lower short-term flows, the court determined that the "agency decision appears to conflict with the analysis in the BiOp."[17] The BiOp failed to discuss the effects of each of the three phases of the plan on coho. While the Phase 1 plan discussed the water ban requirement, the discussion failed to demonstrate how water from the bank would affect instream flow or improve stream temperatures. In addition, the court found the Phase II discussion to be "even more conclusory"[18] because it failed to address how providing only 57 percent of the necessary flows would avoid jeopardy to salmon. Rather than using the Klamath Project's proportional share of water use to determine flows, the analysis should have focused on "what jeopardy might result from the agency's proposed actions in the present and future human and natural contexts."[19] Further, the BiOp discussed the benefits of increased flows for salmon in the final phase, but it failed to reconcile operations in the first two phases with this conclusion. Both Phase I and Phase II of the plan allow for much-lower flows during summer month in dry years, but the BiOp did not explain how coho would not be jeopardized by the reduced flows during the first 8 years of the plan. While the BiOp contained analysis in support of the long-term flow recommendations for the RPA's final phase, such a discussion is "conspicuously absent"[20] from the discussion of the first two phases.

The court was troubled by the failure of the BiOp to take into account the coho's three-year lifespan in the first two phases. It noted that five generations will complete their life-cycles during the first two phases of the plan, when lower flow requirements mean that there may be not be enough water for the coho to survive. The agency must take into account near-term habitat loss for species with short life cycles, and cannot impose flows with no explanation of the short-term effects they may have on the protected species.[21] By failing to consider the life-cycle of the coho, NMFS "entirely failed to consider an important aspect of the problem."[22] An agency cannot provide "only partial protection for a species for several generations without any analysis of how doing so will affect the species."[23] The agencies argued that NFMS was not required to provide a quantitative analysis because the analysis for the first two phases of the RPA reflect the agency's qualitative determination the plan would not cause jeopardy to salmon. The court rejected this argument because "while the [NMFS] can draw conclusions based on less than conclusive scientific evidence, it cannot base its conclusions on no evidence."[24] Thus, the Ninth Circuit concluded that the RPA was arbitrary and capricious because NMFS failed to provide explanation for how the first two phases of the RPA will not cause jeopardy to species.

The court remanded the case to the district court to determine appropriate relief, but it "emphasize[d]" the injunction should be tailored to reflect the short life cycle of the SONCC coho, stating it was "not enough to provide water for the coho to survive in five years, if in the meantime, the population has been weakened or destroyed by inadequate water flows."[25]

In conclusion, the Ninth Circuit invalidated the remainder of the RPA for operations of the Klamath Project as arbitrary and capricious because it failed to provide analysis of the effects of the proposed action on salmon or explain how the agency reached its no-jeopardy conclusion. The court remanded the case back to the district court to determine appropriate injunctive relief, with instructions that the injunction reflect the short life cycle of the SONCC coho.