OPINION

The western Distinct Population Segment of the Steller sea lions (“wDPS”) live in the great northern Pacific Ocean region off Alaska, and they were declared endangered in 1997. More recently, in two of the seven sub-regions they inhabit, they have been experiencing population declines because they have been showing signs of nutritional stress. In 2010, the National Marine Fisheries Service (“NMFS” or “the agency”) therefore limited commercial fishing in those waters, causing representatives of the fishing industry and the State of Alaska (“Plaintiffs”) to file this action challenging the limitations.

The plaintiffs’ principal argument is that the NMFS violated the Endangered Species Act (“ESA”) because it based the fishing restrictions on declines in sub-regions rather than in the entire population of the endangered species. Plaintiffs also contend the agency utilized the wrong standards in measuring the effects of continued fishing and failed to find a sufficient causal link between authorizing fisheries and the population decline. We hold that use of subregions did not violate the ESA and that the agency utilized appropriate standards to find that continuing previous fishing levels in those sub-regions would adversely modify the critical habitat and jeopardize the continued existence of the entire population. We therefore affirm the district court’s judgment rejecting plaintiffs’ claims.

I. BACKGROUND

A. Statutory Framework

This case involves the interaction of three statutes: the Magnuson-Stevens Fisheries Conservation Act (“MSA”), the ESA, and the National Environmental Policy Act (“NEPA”). The first concerns management of fisheries, and the others concern more generally the environmental consequences of government actions. Plaintiffs claim NMFS violated all three in its promulgation of a 2010 Biological Opinion (“BiOp”) reducing commercial fishing in wDPS habitat.

The MSA governs the federal management of fisheries in various waters off the United States and establishes regional councils that are responsible for the sustainable management of fisheries. 16 U.S.C. § 1852 (h). These councils create fishery management plans, which are prepared using scientific evidence and are geared toward ensuring conservation of the fisheries. Id. § 1853. The Secretary of Commerce must approve the management plans, which can include, among other things, limitations on or closure of fishing in designated zones. Id.

The ESA requires the Secretaries of Interior and Commerce to list endangered species and designate their critical habitats. 16 U.S.C. § 1533 (c). Section 4(f) of the ESA requires the Secretary of Commerce to design and carry out “recovery plans” and to implement programs to conserve the species under section 7(a)(1). 16 U.S.C. §§ 1533 (f), 1536(a)(1). Section 7(a)(2) of the ESA mandates that federal agencies ensure that actions they take will not “jeopardize the continued existence of any endangered species or threatened species or result in the destruction or adverse modification of habitat of such species.” 16 U.S.C. § 1536 (a)(2).

Under the ESA, when a governmental entity plans to take action that may impact an endangered species, it must consult with the agency that has authority over the species. The consulted agency must then prepare a BiOp to determine whether the planned action will either likely jeopardize the species’s continued existence or adversely modify its critical habitat. See id; Nat’l Wildlife Fed’n v. Nat’l Marine Fisheries Serv., 524 F.3d 917, 924 (9th Cir.2008). If either of those criteria is met, the agency may suggest a “reasonable and prudent alternative” (“RPA”), which is designed to avoid jeopardy or adverse habitat modification. 16 U.S.C. § 1536 (b)(3)(A); Nat’l Wildlife Fed’n, 524 F.3d at 925 .

The NMFS in this case stands at the intersection of the MSA and the ESA. Its duty is to ensure that actions taken by the regional councils, including fishery management plans, do not jeopardize the continued existence of a threatened or endangered species or adversely modify critical habitat of an endangered species. See Greenpeace v. Nat’l Marine Fisheries Serv. (“Greenpeace IV”), 237 F.Supp.2d 1181, 1185 (W.D.Wash.2002). In this case it must evaluate the effect of the fishing plans on the wDPS and its habitat.

The third statute, NEPA, requires study of the environmental effects of any government action. 42 U.S.C. § 4332 (C). Under NEPA, an agency planning an action may use an Environmental Assessment (“EA”) to determine whether the proposed action, including an RPA, would have a significant environmental impact. 40 C.F.R. § 1501.4 . If the action would have such an effect, the agency must prepare an Environmental Impact Statement (“EIS”). Id. If it would not, no EIS is required. An EIS must describe the environmental consequences of the governmental action and the alternatives that were considered by the agency. When an agency makes a decision after preparation of an EIS, it must prepare a Record of Decision (“ROD”), stating its decision, the alternatives it considered, and how its decision minimizes environmental harms. 40 C.F.R. § 1505.2 .

In this case, the operation of these statutes has a profound impact on both the Steller sea lions and on fishing interests because the statutes deal with the habitat in which sea lions live. This habitat includes their food sources, and, as a result, the resources available for commercial fishing. Environmental groups, the State of Alaska, and fishing industries, therefore, have all long been concerned about how the federal government applies these statutes in the waters off Alaska, and their concerns have led to a decade of litigation concerning the scope of federal regulation of these fisheries.

B. History of Regulation of Fisheries and wDPS

The regulation at issue in this case has centered on the Bering Sea/Aleutian Islands and Gulf of Alaska regions, which are home to Steller sea lions and to commercial fisheries that are governed by the MSA. See Greenpeace v. Nat’l Marine Fisheries Serv. (“Greenpeace I ”), 55 F.Supp.2d 1248, 1252 (W.D.Wash.1999). By 1990, there had been a decline in the sea lion population over a thirty-year period, so the NMFS listed the entire Steller sea lion population as “threatened” under the ESA. Id. at 1254 . Seven years later, because of new research, the agency divided the population of the Steller sea lions into the eastern Distinct Population Seg ment and the wDPS. Id. The new research revealed genetic differences between the two populations. Id. There were also markedly different survival rates in the two populations. Therefore, the agency changed the wDPS’s status to “endangered,” because of the “precipitous, large population decline” among this segment, contrasted with the more stable population trend among the eastern population segment. 62 Fed.Reg. 24,345, 24,354 (May 5, 1997).

The regional council responsible under the MSA for the areas affected by the agency actions in this case is the North Pacific Fishery Management Council. It has been preparing fishery management plans on an annual basis since the MSA was passed in 1976. These plans set limits on the total amount of fishing permitted as well as times and areas of fishery closure, and they allocated the allowable catch among vessels. Pursuant to the ESA, the Council consulted with NMFS on the plans, and NMFS prepared BiOps. Up until 1998, NMFS had found that the Council’s fishery management plans adequately took into account the needs of the Steller sea lions and that there was no jeopardy to the species or adverse modification of its habitat as a result of the fisheries.

In 1998, however, NMFS issued a BiOp that concluded that the fishing was causing some jeopardy to the species and adversely modifying its critical habitat. This led to a first round of litigation concerning regulation of fishing in the Alaskan waters. Environmental groups, including groups who are intervenor-defendants in this case, and fishing industry interests all challenged NMFS’s conclusions contained in the 1998 BiOp. That earlier litigation resulted in several district court opinions, Greenpeace I, 55 F.Supp.2d at 1276 ; Greenpeace v. Nat’l Marine Fisheries Serv. (“Greenpeace II ”), 80 F.Supp.2d 1137, 1152 (W.D.Wash.2000); Greenpeace v. Nat’l Marine Fisheries Serv. (“Greenpeace III’’), 106 F.Supp.2d 1066, 1080 (W.D.Wash.2000); Greenpeace IV, 237 F.Supp.2d at 1204 , which led to revisions of the BiOp with respect to fishery management.

In addition to reviewing fishery management pursuant to the ESA and MSA, NMFS was also required by the ESA to plan for sea lion population recovery. In 1992, NMFS promulgated a recovery plan for Steller sea lions pursuant to section 4 of the ESA. The agency did so using scientific experts working with interested parties, including fishing interests.

When, in 1997, NMFS divided the sea lion population into eastern and western distinct population segments, separate recovery plans were required for each of the two species. The most recent Recovery Plan for the wDPS was published in 2008 and divided the wDPS into seven subregions to monitor the species’s progress. NMFS concluded in the Plan that “it was important to consider sub-population vital rates and demographic characteristics when considering the status of recovery of the western DPS” because subregional impacts “could indicate that extinction risk may still be high.”

The Plan further established specific criteria that had to be met before the wDPS could be downlisted or delisted. To down-list the wDPS to “threatened” or to delist it entirely, the Plan required progress in growth of the wDPS population as a whole, and stable or increasing populations in five of the seven sub-regions. The Plan further mandated that the species could not be delisted if the wDPS population in any two adjacent sub-regions declined “significantly” or if the population in a single sub-region declined more than 50%. The Recovery Plan itself is not challenged here, nor is there any dispute that the wDPS as a whole is not meeting the recovery criteria set forth in the Plan and therefore remains endangered.

The decline in the wDPS population and the lack of robust recovery, despite the extensive efforts since 1990, has led to concerns about the health of the species and hence to federally funded research studying possible causes of the decline, including fishing. The North Pacific Fishery Management Council, in 2005, cited this new research in requesting that the NMFS initiate a new consultation pursuant to the ESA to determine the impact of continued fishing at the rates authorized by the then-existing fishery management plans. The Protected Resources Division of NMFS conducted the study using the new research on the impact of commercial fishing and other factors on the wDPS population. During its study, it took into account the 2008 Recovery Plan to determine whether continued fishing would jeopardize the wDPS or adversely modify the species’s critical habitat.

This consultation process culminated in NMFS’s promulgation of a new BiOp in 2010. That BiOp concluded that continued fishing would have deleterious impacts on the species and the habitat in which it lives and that, therefore, greater limitations on commercial fishing were required. The plaintiffs challenge that BiOp in this action.

C. The Challenged Agency Actions: The 2010 BiOp, EA and Interim Final Rule

In November 2010, after surveying the new research on the causes of wDPS’s lack of recovery, NMFS issued the BiOp pursuant to the ESA and an EA pursuant to NEPA. The BiOp was promulgated after review by the Council and public comment, including feedback from the fishing industry plaintiffs here. The BiOp concluded that continuing to authorize fisheries at the levels previously authorized in the fishery management plans would both jeopardize the continued existence of the wDPS and adversely modify its critical habitat. Relying on the Plan’s geographic divisions, the BiOp found that the wDPS was experiencing population declines in two of the sub-regions, although the population was increasing in four others. The BiOp concluded that these sub-regional declines caused the entire wDPS population to fall short of the Recovery Plan’s goal of a statistically significant increase of the species’s population as a whole.

The agency’s critical finding was that natality rates were lower for the wDPS than for the eastern population, and that the most likely explanation for the difference was that portions of the wDPS were experiencing nutritional stress. The BiOp concluded that nutritional stress had directly or indirectly contributed to the reduction in population growth. It observed that although the nutritional stress hypothesis had been questioned by some experts, fishery presence in the two wDPS sub-regions was nevertheless negatively correlated with population numbers. In other words, as fishing increased, the wDPS population fell.

With respect to the standard for measuring or determining adverse habitat modification, the BiOp noted that we had previously held part of the agency’s regulatory definition of adverse modification conflicted with the statute. See Gifford Pinchot Task Force v. U.S. Fish & Wildlife Serv., 378 F.3d 1059 , 1070 (9th Cir.2004). The agency therefore relied on the statutory language of the ESA, 16 U.S.C. § 1536 (a)(2), framing the critical question to be whether “the action reduces the value of critical habitat for the conservation of the species.” Applying this standard, the BiOp found the weight of evidence suggested that continued authorization of the fisheries was likely to adversely modify the critical habitat of the wDPS and jeopardize its continued existence. The BiOp acknowledged there were potentially multiple factors contributing to the decline in the wDPS population, but it found nutritional stress was one contributing cause. Thus the agency concluded it was mandated by the ESA to take steps to prevent continued fishing from likely reducing or negatively affecting the survival and recovery of the wDPS.

The BiOp therefore proposed an RPA that restricted the activity of several fisheries and closed other fisheries in the Aleutian Islands region. The RPA recommended closure of all mackerel and cod fishing in one part of the region and reducing the catch allowed in others. The RPA recommended swift implementation to “support the recovery of’ the wDPS population, which so far had not met the criteria outlined in the Recovery Plan.

Pursuant to NEPA, NMFS also prepared an EA, which considered various alternatives. NMFS found that of the various legally permissible alternatives, the RPA was the least likely to disrupt fishing or reduce the number of jobs and that the RPA’s fishing restrictions would have no significant impact on the environment. NMFS therefore promulgated an interim final rule, implementing the new restrictions, as described in the RPA. The likelihood of adverse economic consequences to the fishing industry, however, prompted this litigation.

D. District Court Proceedings

The plaintiffs challenged the agency’s actions under the Administrative Procedure Act (“APA”), contending that the BiOp, EA, and finding of no significant impact ran afoul of the MSA, ESA, and NEPA. The district court granted a motion by non-profit environmental groups Oceana, Inc. and Greenpeace, Inc., who had participated in the previous litigation, to intervene as defendants.

The district court held that the agency applied the correct standards under the ESA. It ruled that the agency’s consideration of population declines in two sub-regions was permissible because these declines implicated the survival and recovery of the species as a whole. It further held that the agency properly relied on the statutory definition of adverse modification and that the agency’s consideration of the impact of continued fishing on the wDPS’s recovery was appropriate. The district court rejected plaintiffs’ other ESA contentions as well, concluding that the RPA was supported by the record and that the agency was not required to show a definitive causal connection between authorization of fisheries and harm to the wDPS.

With respect to NEPA, however, the district court found that the agency should not have summarily concluded the restrictions would cause no significant impact, and it ordered the agency to prepare an EIS. The court refused to enter a broader injunction that would have also required the agency to suspend the new restrictions or prepare a ROD.

In this appeal, plaintiffs argue the agency violated the ESA and that the district court’s NEPA injunction was too narrow. Plaintiffs contend that the BiOp violated the ESA in relying on trends in two sub-regions rather than considering population trends of the whole species. They further contend that the agency should have applied the regulatory definition of adverse habitat modification that this court has partially rejected. They additionally argue the agency should have limited its consideration to possible affirmative harm to the existing population rather than focusing on the adverse effect of fishing on the species’s recovery. Plaintiffs also contend there must be a definitive causal link shown between authorization of commercial fishing and injury to the wDPS population or habitat. Plaintiffs finally contend the court abused its discretion under NEPA by not requiring the agency to prepare a ROD in addition to requiring an EIS.

Our review of agency actions, including the promulgation of a BiOp, is narrow. Under the APA, we must determine only if the agency’s action was arbitrary or capricious — that is, whether the agency based its decision on the relevant factors and stated a rational connection between its factual findings and conclusions. 5 U.S.C. § 706 (2)(A); Gifford Pinchot Task Force, 378 F.3d at 1065.

II. DISCUSSION

A. The Agency’s Reliance on Data in Two of the Recovery Plan’s Sub-regions Was Not Contrary to the ESA Because That Data Implicated Recovery and Survival of the Species as a Whole.

NMFS found in the BiOp that the wDPS population was declining in two of the Recovery Plan’s seven sub-regions and that while the population as a whole had experienced some increase between 2000 and 2008, that increase was not statistically significant. Plaintiffs challenge the agency’s reliance on population trends in two sub-regions rather than in the species as a whole.

We have consistently held that the ESA permits agencies to consider the impact of actions on sub-populations, as long as such impact would affect the population as a whole. Wild Fish Conservancy v. Salazar, 628 F.3d 513, 529 (9th Cir.2010); Gifford Pinchot Task Force, 378 F.3d at 1075. We recognized that trends with regard to a subset of a species can provide important indicators about the health of the entire species. Wild Fish Conservancy, 628 F.3d at 529 . Our law thus supports the agency’s looking to sub-regional declines in reaching its conclusions that continued fishing would jeopardize the survival of the wDPS and adversely modify the wDPS’s critical habitat.

Wild Fish Conservancy illustrates the importance of considering declining sub-populations. We there held that a Fish and Wildlife Service BiOp that did not consider the impact of sub-populational decline on a species as a whole was inadequate, and therefore arbitrary. The BiOp in that case dealt with the impact of the Leavenworth National Fish Hatchery on endangered bull trout population. Id. at 516 . We held that the agency contradicted itself when it suggested that the declining Icicle Creek sub-population was of particular importance due to its separation from other parts of the species but nevertheless found that the decline of this subpopulation would not jeopardize the survival or recovery of the species as a whole. Id. at 529 . Where trends in a subpopulation may affect the entire population, the ESA requires the agency to consider the effects of the declining subpopulation. Id.

We also stressed the importance of analysis of the subregional impacts of agency actions in Gifford Pinchot. In that case, environmental groups objected to three Fish and Wildlife Service BiOps analyzing the effects of timber harvests on the Northern Spotted Owl, arguing that the BiOps did not sufficiently consider local impacts. Gifford Pinchot Task Force, 378 F.3d at 1075. We held the agency had adequately taken the local impacts into account. Id. Far from suggesting that local impacts were irrelevant, we emphasized the importance of considering such impacts when they are relevant to the species or habitat as a whole. We stated, “[Hocusing solely on a vast scale can mask multiple site-specific impacts that, when aggregated, do pose a significant risk to a species.” Id. Plaintiffs’ contention that the agency should not have considered sub-regional declines of wDPS is fundamentally at odds with our holding in Gifford Pinchot

Plaintiffs nevertheless contend that the agency did not establish any nexus between the two sub-regions in which there was population decline and the species as a whole to justify analysis based on sub-regions. Yet, agency findings in the BiOp and Recovery Plan provide a clear basis for the conclusion that sub-regional declines have a profound effect on the future of the entire species. The Plan explained that the wDPS’s viability depended on the maintenance of healthy populations in each of the seven sub-regions in order to prevent concentration in a smaller area where a single isolated destructive force could endanger the entire population. The Recovery Plan concluded that maintaining independently viable populations is vital because it ensures that the species will “persist through normal population variations, as well as through unexpected catastrophic events.” The Recovery Plan additionally stressed the importance of monitoring on a sub-regional basis because a declining sub-population could indicate an unpredicted threat to the species that could spread to other sub-regions. The Recovery Plan therefore established a goal of no more than 50% decline in any single sub-region before the species could be delisted.

The analysis of sub-regions in the BiOp yielded significant information that, in light of the Recovery Plan’s concerns, led to the conclusion that sub-regional declines indicated that the entire species was in jeopardy. The agency found that between 2001 and 2009, there had been a 7% decline in wDPS’s pup population in the central Aleutian Islands sub-region and a 43% decline in the western subregion. These declines in two sub-regions explained the lack of a statistically significant increase in the overall wDPS population and led to the BiOp’s finding that if current declines in pup population were to continue, wDPS would be extirpated in these sub-regions. Such extirpation would render the Recovery Plan’s goals unattainable.

The Recovery Plan thus set a goal of a stable population to be accomplished through sub-regional monitoring. The BiOp furnished the supporting statistical analysis, thereby establishing a nexus between population trends in the sub-regions and the health of the species as a whole. Therefore the agency was not arbitrary or capricious in relying on sub-regional declines to determine whether continued fishing would jeopardize the species as a whole or adversely modify its critical habitat.

B. The Agency Was Nat Required to Utilize a Partially Discredited Regulatory Standard.

Plaintiffs contend that the agency used the wrong definition of adverse habitat modification and should have applied its regulation stating that there is an adverse habitat modification where there is a “direct or indirect alteration that appreciably diminishes the value of critical habitat for both the survival and recovery of a listed species.” 50 C.F.R. § 402.02 .

The agency did not apply the regulatory definition, however, because in Gifford Pinchot Task Force, 378 F.3d at 1069-72, we held that a portion of the regulation was invalid as contrary to the statutory language. NMFS stopped using the regulatory definition after our opinion in 2004 and began relying on the statute itself. As the agency explained in the BiOp, “NMFS does not rely on the regulatory definition of ‘destruction or adverse modification’ of critical habitat at 50 CFR 402.02. Instead, we have relied upon the statutory provi sions of the ESA to complete the analysis with respect to the critical habitat.” We have previously considered and specifically upheld an agency’s reliance on the statutory language of the ESA itself when faced with uncertainty about the vitality of the regulatory definition in light of Gifford Pinchot. See Butte Envtl. Council v. U.S. Army Corps of Eng’rs, 620 F.3d 936, 947-48 (9th Cir.2010). There is no reason to hold the agency should have gone back to the questionable regulation here. NMFS asked and answered the critical question under both the statute and the portion of the regulation that was not at issue in Gifford Pinchot: whether reauthorization of the fisheries would adversely modify the habitat.

C. The Agency Properly Considered the Prospects for the Species’s Recovery in Determining Whether the wDPS Would be Jeopardized by Continued Fishing.

Plaintiffs contend that the agency, in the BiOp and RPA, should not have considered the impact of the proposed action on the wDPS’s prospects of recovery, and instead should have focused exclusively on whether continued fishing would affirmatively harm the existing species and its likelihood of survival. Under the ESA, the agency must ensure against government action jeopardizing the continued existence of an endangered species or harming its habitat. 16 U.S.C. § 1536 (a)(2). We have held that recovery considerations are an important component of both the jeopardy and adverse habitat modification determinations. See Gifford Pinchot Task Force, 378 F.3d at 1070; Nat’l Wildlife Fed’n, 524 F.3d at 931-33 . The goal of the ESA is not just to ensure survival, but to ensure that the species recovers to the point that it can be delisted. Gifford Pinchot Task Force, 378 F.3d at 1070. Survival and recovery are intertwined and are the complementary goals of the consultation process. Id.; Nat’l Wildlife Fed’n, 524 F.3d at 932 .

NMFS therefore had to consider whether the proposed action, continued fishing, could prevent the species from achieving the Recovery Plan’s goals for delisting. That is what NMFS did. Relying on the Recovery Plan, the agency concluded that the fishery reauthorizations would appreciably diminish the wDPS’s chances of recovery as the fishery could fully extirpate the species in at least one sub-region. The agency found that “the extirpation of Steller sea lions in the western Aleutians would be significant to the western DPS, and is expected to appreciably reduce the likelihood of both their survival and recovery in the wild.”

D. Plaintiffs’ Specific Criticisms of the Agency’s Analysis Lack Support in the Record.

Plaintiffs raise a number of criticisms with the agency’s analytical methodology, claiming each demonstrates the agency conclusions in the RPA lacked a rational basis. None of these criticisms is supported by the record.

The first criticism is that the agency used forage ratios — the ratio of prey required by sea lions to available fish — to determine the areas in which fishing should be closed. The use of forage ratios would have been problematic, because such ratios were already higher in the areas that were to be affected by the RPA. The agency, however, did not rely on such ratios, finding that they were “very difficult to interpret.” The agency instead determined fishery closures and limitations based on actual population decline in two sub-regions.

Plaintiffs’ second criticism is that the agency failed to account for the fact that restrictions on cod fishing could decrease the number of mackerel, since cod prey on mackerel, and mackerel are a food source for the wDPS. In essence, plaintiffs contend the agency should have used a “multi-species” model in its analysis. The agency, however, looked to studies that relied on multi-species models, which took into account this inter-species predation, and singlespecies models, which did not. The agency concluded that the multi-species models were theoretically more accurate but that they introduced more variables leading to uncertainty. Faced with competing interests of theoretical accuracy and analytical uncertainty, the agency made a rational choice. This court will not second guess the agency’s determination, which is supported by the record. See Ariz. Cattle Growers’Ass’n v. Salazar, 606 F.3d 1160 , 1164 (9th Cir.2010).

Third, plaintiffs contend that the fact that the areas with more fishing have had slower wDPS growth does not show that further fishery restrictions can cause growth to increase. Yet, the agency did not point to the slower growth in fishing areas in order to establish causation, but instead used it as an illustrative example of the correlation between fishing and population decline. The plaintiffs nevertheless suggest that even if there is a correlation between increased fishing and slowing wDPS growth, there is no correlation between the authorization of increased fishing and slower wDPS growth. The agency, however, appropriately pointed to the correlation between areas with more fishing and slower wDPS population growth to illustrate its larger theory that the fisheries and the wDPS were in competition for the same prey.

Plaintiffs’ final criticism is that the agency did not properly account for killer whale predation. The agency, however, considered the effect of killer whales, but did not find it strong enough to outweigh the effect of fishing. Citing a number of studies regarding killer whales, the agency specifically looked to the long-term historical trends and concluded that the studies “argue against the hypothesis that killer whale predation alone was responsible for the decline” in population.

Plaintiffs’ specific criticisms of the RPA are therefore not well taken. They do not undermine the RPA’s analysis in any way.

E. NMFS Was Not Arbitrary or Capricious in Concluding the Removal of Sea Lion Prey was an Indirect Effect of Fishery Reauthorization.

Plaintiffs fault the agency denying the reauthorizations without finding that fisheries were the direct cause of wDPS nutritional stress. They point to the BiOp’s statement that the agency did not have the data to demonstrate the extent of nutritional stress or that such stress was caused by the fisheries.

The agency was required to ensure that its actions would not have direct or indirect effects that would jeopardize the wDPS or adversely modify its critical habitat. 50 C.F.R. § 402.02 . Applying its regulation, the agency indisputably found that the fisheries were removing prey species of the wDPS. It also found evidence of nutritional stress. While the agency admitted it could not find a direct link between the fisheries and the species’s decline, it found that the indirect effect of the fisheries was the removal of wDPS’s food. The agency was not required to find that the fisheries were the direct cause of the species’s decline.

F. The District Court Did Not Need to Require a ROD In Its NEPA Injunction.

The district court granted summary judgment to the plaintiffs on their claim that the agency violated NEPA by not issuing an EIS. The district court en tered an injunction requiring the agency to prepare an EIS. On appeal, plaintiffs seek a broader injunction ordering the agency to also prepare a ROD. Plaintiffs argue that the ROD is necessary to ensure that the agency actually considers the information in the EIS.

The problem is that because the agency has not yet prepared the EIS, we do not yet know what final action, if any, it will propose to take. An injunction ordering a ROD would be premature in the absence of the agency’s proposed action based on the EIS record it develops. 40 C.F.R. § 1505.2 ; Stormans, Inc. v. Selecky, 586 F.3d 1109, 1140 (9th Cir.2009) (requiring injunctions to be narrowly tailored). The district court did not err in refusing to order a ROD at this stage.

CONCLUSION

The district court’s order granting summary judgment to the defendants on the ESA claims and injunction on the NEPA claims are AFFIRMED.