OPINION

This case arises from a dispute over scientific methodology. Cascadia Wild-lands and other environmental groups (Cascadia) appeal from the district court’s denial of their motion seeking to enjoin the Douglas Fire Complex Recovery Project (Recovery Project) in the southern Oregon Klamath Mountains. Cascadia specifically challenges the biological opinion issued by the United States Fish and Wildlife Service (the Service). Because Cascadia has not demonstrated a likelihood of success on the merits, we affirm the district court’s denial of a preliminary injunction.

I. BACKGROUND

A. The Recovery Project

The Douglas Complex Fire burned approximately 48,000 acres of federal and non-federal land in Oregon’s Klamath Mountains. In response, the Medford District of the Bureau of Land Management (Bureau) initiated the Recovery Project aimed at salvaging burned acreage. The Bureau completed a Douglas Complex Fire Recovery Project Environmental Assessment, and solicited public comment, which Cascadia timely provided. Subsequently, the Bureau issued a Record of Decision and Finding of No Significant Impact approving the Douglas Complex Fire Recovery Project. This finding authorized the salvage logging of approximately 1,600 acres of fire-killed or injured trees, including hazard tree removal (to which Casca-dia does not object), and logging of interior forests for economic benefit. The Bureau also consulted with the Service, after conducting a biological assessment and determining that the Recovery Project “may affect and is likely to adversely affect” the Northern Spotted Owl and its critical habitat. The Service in turn issued a biological opinion, which concluded in part:

Nevertheless, the biological opinion concluded that the Recovery Project was “not likely to result in jeopardy to the species or destruction or adverse modification of critical habitat.”

B. District Court Decision

The district court denied Cascadia’s motion for a preliminary injunction, finding that Cascadia failed to adequately establish that it was likely to succeed on the merits, that there were “serious questions” going to the merits, or that irreparable harm to the spotted owl was likely.

1. Barred Owls’ Ejfect on the Northern Spotted Owl

Cascadia’s primary argument focused on barred owls, which are predators of the Northern Spotted Owl. Cascadia contended that when barred owls are present, Northern Spotted Owls are less likely to respond to survey calls. Cascadia submitted that the Service failed to account for this possibility and “underestimated the number of spotted owls sites by relying on false no occupancy determinations.” Cascadia Wildlands v. Thrailkill, 49 F.Supp.3d 774, 779 (D.Or.2014) (internal quotation marks omitted).

The court disagreed with Cascadia’s contention that the Service’s no jeopardy conclusion was arbitrary and capricious, finding that the Service adequately “acknowledge[d] and accounted] for the potential impact of barred owls on [spotted owl] detectability-” Id. The court ob served that the biological opinion specifically referenced the presence of barred owls and the effect of barred owls’ presence.

The court found that the Service utilized the best available scientific information — a series of long-term and uniform Bureau surveys. The consistent nature of the surveys provided surveyors with accurate site locations and movement patterns of the owl in the area of the proposed action. The Service also recommended that the Bureau continue the survey process during the upcoming survey season to further inform project planning. Cascadia did not assert that the Service failed to identify the best available scientific information, and did not cite alternative occupancy data. The court found that the Service’s approval of the salvage project was not arbitrary, capricious, or otherwise in violation of the law.

2. Wildfire’s Effect on the Northern Spotted Owl Habitat

Cascadia asserted that the best available scientific data suggests that following a wildfire, the Northern Spotted Owl expands its core areas and home ranges, including roosting and foraging sites. As a result, the required habitat area would also increase.

The district court found that the record did not support Cascadia’s contention that the Service failed to consider this data. The court determined that the Service “fully considered the possibility” that the spotted owl may have expanded or shifted its core-use areas and home ranges post-fire. Although Cascadia failed to produce any evidence that the home ranges actually did expand, the Service examined and analyzed a variety of data from the study area to locate, to its best estimation, the sites where shifts may have occurred.

The court concluded that the one study Cascadia cited did not support Cascadia’s argument. Indeed, that study found no significant differences between core-use areas and home ranges in and around fire-affected areas. The court observed that the Service examined home range circles that were twice as large as those described in the cited study, and core-use areas that were four times the size of those in the study. Thus, the court found the Service’s methodology appropriate to evaluate any potential shifts in home range due to wildfires.

The court was not swayed by Cascadia’s other documentary evidence. The court concluded that an email from a Bureau official discussing movement of a single owl due to the presence of a barred owl was not evidence of expansion or shifting of ranges due to the Douglas Complex Fire. Moreover, the Bureau was aware of and tracked the single owl, completely accounting for its movement. Similarly, an internal Service memorandum suggesting that some owls could move because of the new conditions created by the Douglas Complex Fire did not undermine the Service’s analysis. The Service anticipated and accounted for these potential shifts by surveying larger home range and core-use areas and by using long-term, intensive spotted owl demographic studies on the action area. With this data, the Service identified spotted owls at eight of fourteen sites that could potentially shift. Because the Service “adequately and lawfully accounted for the effect of wildfire on spotted owl site locations,” the court concluded that the Service’s decisions regarding owl habitat were not arbitrary, capricious, or an abuse of discretion.

8. Endangered Species Act Procedural Requirements

Cascadia argued that the Service failed to comply with the procedural requirements of the Endangered Species Act because it did not: 1) assess the effects of the proposed action on six known spotted owl sites that overlap the planning area, but not the salvage units; 2) explain why it used two different methods for assessing effects to the species; and 3) consistently apply its effects analysis methodologies.

The district court rejected Cascadia’s arguments, finding that the Service appropriately assessed the sites around the spotted owl nesting areas and appropriately defined the action area to include the home ranges of known spotted owl sites that could be impacted. Of the 45 historical nest sites within the action area, only 39 would be affected by any salvage treatment or road/landing construction work. Of the six sites that overlapped the action area, none of them would be affected by' the habitat modifications in Recovery Project areas. Cascadia did not contend otherwise. Rather, Cascadia argued that the Service miscalculated the nesting, roosting, and foraging coverage at the home range and core-use areas. However, as the court noted, the calculations urged by Cascadia were guidelines rather than fixed formulas.

In addition, the Service took into consideration other site-specific factors, beyond the amount of remaining nesting, roosting, and foraging habitat, including: 1) the post-fire condition of the habitat; 2) the amount of post-fire foraging habitat left or slated for removal and its proximity to the critical habitat; 3) owl occupancy in the action area; and 4) other abiotic factors such as stream distance, elevations, and slope positions. The Service also evaluated the relative habitat suitability to determine owl occupancy/viability of an area. Areas that were unlikely to be affected by the Recovery Project were to be left relatively intact, thereby supporting a higher level of site occupancy and habitat fitness potential. Despite Cascadia’s contentions that several sites were occupied by the Northern Spotted Owl, the record surveys proved otherwise.

The court rejected Cascadia’s argument that the Recovery Project will remove more than minimal post-fire foraging at two sites. It also rejected Cascadia’s contention that the Service’s “not likely to adversely affect” determination was arbitrary and capricious. Mindful of the deference due to the Service on scientific matters, the court declined to substitute Cascadia’s definition of “minimal” for that employed by the Service. The court explained that post-fire habitat must be evaluated in relationship to the remaining habitat. For example, one site was slated to have a total of 44 acres of post-fire foraging land removed from the home range, but the areas most impacted were in a low relative habitat suitability area. Only approximately 0.7 acres of the total post-fire foraging land would be removed from the core-use area. For another site, only 25.5 acres of post-fire foraging land would be removed from the home range(s), and none from the core-use area(s). The court concluded that the Service credibly determined that the proposed post-fire foraging removals were “minimal” in nature and that the Recovery Project was not likely to adversely affect the spotted owl at the respective sites.

Finally, the court was not convinced by Cascadia’s contention that the Service inconsistently applied its methodology to determine when a “take” had transpired at a respective spotted owl site. The Service had concluded that spotted owls were not likely to occupy the site where the “ho take” determination was made, but were likely to occupy the site where the “take” determination was made. The court concluded that the Service’s rationale was not arbitrary, capricious, or an abuse of discretion.

A Recovery Plan for the Northern Spotted Owl

Cascadia’s final argument regarded what it perceives to be aii inconsistency between the Service’s determination that the Project would not jeopardize the spotted owl and the recovery measures identified in the Service’s 2011 Northern Spotted Owl Recovery Plan (Recovery Plan).

In 2011, the Service devised the Recovery Plan for the spotted owl population. The Recovery Plan included various Recovery Actions. Recovery Actions are “near-term recommendations to guide the activities needed to accomplish the recovery objectives and achieve the recovery criteria such that a species may be delisted from [Endangered Species Act] protection.” Id. at 786 (citation and internal quotation marks omitted).

Recovery Action 10 instructs agencies to “[cjonserve spotted owl sites and high value spotted owl habitat to provide additional demographic support to the spotted owl population. The intent of this recovery action is to protect, enhance, and develop habitat in the quantity and distribution necessary to provide for the long-term recovery of spotted owls.... ” Id. (citation omitted).

Recovery Action 12 directs that “[i]n lands where management is focused on development of spotted owl habitat, post-fire silvicultural activities should concentrate on conserving and restoring habitat elements that take a long time to develop (e.g., large trees, medium and large snags, downed wood)_” Id. (citation omitted).

The district court initially reasoned that Cascadia’s claims regarding Recovery Actions 10 and 12 must fail because recovery plans do not have the force of law, a point Cascadia conceded at oral argument in the district court. Pursuant to the Endangered Species Act, the Service’s jeopardy analysis should consider whether a particular action “is reasonably likely to appreciably reduce the likelihood of both survival and recovery of a listed species....” Id. at 787 (citing 50 C.F.R. § 402.02 ). The court explained that the jeopardy analysis is not focused on whether the “federal action would itself implement or bring about recovery[,]” and thus, the court rejected Cas-cadia’s attempt to conflate jeopardy with recovery. Id.

The court also found that the biological opinion was consistent with Recovery Actions 10 and 12. Consistent with Recovery Action 10, the Bureau and the Service examined the 39 known affected sites and for the top tier of the identified sites, avoided and mitigated Recovery Project impacts by excluding over 800 acres of post-fire foraging land from salvage activity. The Bureau avoided road and landing construction within the high priority spoN ted owl core-use areas. Rather, the Bureau concentrated its salvage activities on spotted owl sites with established non-occupancy results for several years prior to the fire. In sum, the Bureau mitigated the potential adverse effects to the spotted owl and its habitat, consistent with Recovery Action 10.

The court similarly concluded that the biological opinion discussed and properly implemented Recovery Action 12. The Bureau excluded any acres that were subject to low severity fire in the areas allo cated for salvage activity, as these sites were likely to be frequented by the spotted owl. This exclusion left a significant portion of action area land with both burned and “green legacy features” (e.g., snags, downed wood, etc.), which is important to the spotted owl’s development and future nesting, roosting, and foraging habitat. Id. at 788 . Conversely, approximately 25 percent of the area within the fire perimeter was burned by medium to high severity fire, and within this smaller area only eight percent is subject to harvest. The Bureau also “took special snag[-]related precautions” with regard to the areas set for harvest. Id.

The court denied Cascadia’s motion for a preliminary injunction after considering all arguments presented by the parties. Cas-cadia filed a timely appeal, and a motion in the district court for an injunction pending appeal. Despite the court’s concern that it lacked jurisdiction to grant the relief requested by Cascadia, it proceeded to address the motion on its merits. The court explained that Cascadia was asking the court to reconsider its previous order denying Cascadia’s motion for a preliminary injunction, as Cascadia reiterated many of the same arguments that were previously submitted to the court. Ultimately, the court found that Cascadia failed to satisfy its burden to prove that: 1) it was likely to succeed on the merits; 2) it was likely to suffer irreparable harm in the absence of injunctive relief; 3) the balance of equities tipped in its favor; and 4) an injunction was in the public interest. As in its previous order, the court found that the biological opinion, which was at the center of Cascadia’s argument, was “thorough, reasoned, and reflective of an in-depth analysis of the issues pertaining to the impact of the Recovery Project on the spotted owl.” Order Denying Pl. Mot. for Injunction Pending Appeal, p. 4, Oct. 14, 2014, ECF No. 33.

II. STANDARD OF REVIEW

We review a district court’s denial of a preliminary injunction for an abuse of discretion. See Alliance for the Wild Rockies v. Cottrell, 632 F.3d 1127, 1131 (9th Cir.2011). Our review is “limited and deferential.” Nat’l Wildlife Fed’n v. Nat’l Marine Fisheries Serv., 422 F.3d 782, 793 (9th Cir.2005) (per curiam) (citation omitted). We will reverse only if the district court committed an abuse of discretion, or an error of law, or made a clearly erroneous factual finding. See id. We defer to the expertise of the consulting, agency on matters of science. See Conservation Cong. v. Finley, 774 F.3d 611, 620 (9th Cir.2014).

III. DISCUSSION

Section 7 of the Endangered Species Act directs that federal agencies conserve species listed as endangered or threatened, and whenever a federal action could affect an endangered or threatened species, the agency involved must consult the service with jurisdiction over the relevant listed species. See 16 U.S.C. § 1536 (a)(3). Here, the Fish and Wildlife Service was the appropriate consulting agency because it is responsible for administering the statute with respect to terrestrial wildlife. See 50 C.F.R. § 402.01 (b); see also Pyramid Lake Paiute Tribe of Indians v. U.S. Dep’t of the Navy, 898 F.2d 1410 , 1415 n. 10 (9th Cir.1990) (noting that the National Marine Fisheries Service administers the Endangered Species Act only as to marine species and anadromous fish listed in 50 C.F.R. § 222.203 (a) and 40 C.F.R. § 227.4 ); Building Industry Ass’n of the Bay Area v. U.S. Dep’t of Commerce, F.3d 1027, 1029 (9th Cir.2015). Pursuant to this directive, the Service issued a biological opinion examining the required factors, such as the current status of the listed species, its critical habitat, and effects of the proposed action. The Service employed several methodologies to examine the effect of the proposed action on the spotted owl, and ultimately determined that there was no jeopardy and no adverse modification to the spotted owl’s critical habitat from the proposed action.

Cascadia concedes that the Service identified the relevant scientific data but argues that compliance with the Endangered Species Act requires more than merely restating the scientific data. Cascadia contends that the Service did not actually apply the scientific data in preparing its biological opinion, thereby violating the requirements of the Endangered Species Act. As outlined above, Cascadia contends that the Service failed to apply the best available science regarding: l).the effect of barred owls on detecting the presence of spotted owls; 2) the effect of wildfires on spotted owl habitat and home range; and 3) Recovery Actions 10 and 12. The district court acted within its discretion when it determined that Cascadia failed to show a likelihood of succeeding on the merits of these contentions.

A. Barred Owls’ Effect on the Spotted Owl

The record suggests that the predatory barred owl is a threat to the spotted owl, and that when it is present there is a potential adverse impact on the detectability of the spotted owl.

The Service used several “long term and consistent” Bureau surveys to delineate the study area, which included the respective' action area, to ascertain the locations) of the spotted owl. Cascadia Wildlands, 49 F.Supp.3d at 779 . These surveys recognized the potential impact barred owls have on the efficacy of spotted owl surveying, and the Service acknowledged and incorporated the survey findings into the biological opinion. The biological opinion observed: “Evidence that northern spotted owls were responding less frequently during surveys led the Service and .its many research partners to update the northern spotted owl survey protocol.... ” Cascadia does not dispute that the Service cited the best available science. . Rather, Cascadia contends that the Service did not actually use this science to reach its conclusion regarding the lack of jeopardy. However, Cascadia’s mere disagreement with the result of the biological opinion does not mean that the Service failed to use this scientific data. See United States v. Lewis, 611 F.3d 1172, 1180 (9th Cir.2010). The record reflects that the Service indeed relied upon the data of several surveys from an array of surveyors regarding the effect that barred owls have on the spotted owl.

We give wide latitude to an agency to determine what constitutes the best scientific and commercial data available, as “[t]he determination of what constitutes the best scientific data available belongs to the agency’s special expertise, and thus when examining such a determination, a reviewing court must generally be at its most deferential....” Conservation Cong., 774 F.3d at 620 (citation and internal quotation marks omitted) (emphasis in the original). In view of the deference owed to the agency’s determination, and the record evidence of reliable data, the district court’s rejection of Cascadia’s challenge was not an abuse of discretion, legally erroneous, or factually erroneous. See Nat’l Wildlife Fed’n 422 F.3d at 793 .

B. Wildfíre’s Effect on the Spotted Owl

As with Cascadia’s previous argument, the record does not support a finding that the Service failed to use the best available scientific information regarding the effect the wildfire had on the spotted owl’s habitat use, or a finding that the Serviced conclusions were arbitrary. The Service considered the possibility that spotted owls shifted habitat locations post-fire, specifically their core-use areas and home ranges. The Service simultaneously evaluated the long-studied pre-fire habitat conditions of the spotted owl sites to establish what post-fire habitat conditions could support spotted owl-occupied core-use areas. The biological opinion expressly addressed the effects of wildfire, observing:

(Italics omitted).

Contrary to Cascadia’s contentions, the Service referenced scientific reports that it consulted to help inform its conclusion.

Although the record evidence establishes the possibility of expansion of the spotted owl’s home range post-fire, the biological opinion explicitly acknowledged that because post-fire conditions examined in the scientific literature were “highly variable ... and not directly comparable to one another,” these studies could not be used in a singular fashion to determine post-fire spotted owl occupancy. Accordingly, the Service “relie[d] on professional judgment and interpretation of [the] best available information, including pre-and post-fire habitat conditions in the action area, data in the literature on spotted owl habitat use and occupancy following ... post-fire forest management practices, and ... abiotic factors such as distance to streams, slope position, elevation and aspect. ...”

Nevertheless, three scientific studies from 1998, 2007, and 2011 all indicated that when spotted owl site fidelity has been affected by fire, the spotted owl may increase its home range or shift locations for better nesting, roosting, and foraging opportunities. In the 1998 study, two spotted owl home ranges were surveyed before and up to three years after the 1994 wildfires in the respective areas. For both home ranges, spotted owl habitat use shifted toward unburned habitat, but some lightly/moderately burned habitat was also used.

As observed by the district court, the Service analyzed a 1.3 mile radius home range and was able to evaluate previous spotted owl response and movement patterns in concert with post-fire habitat conditions. The Service’s own evaluation and the other available scientific data amply support the conclusion that the spotted owl may shift or expand its habitat post-fire, thus supporting the Service’s no jeopardy determination.

The biological assessment documented approximately 45 owl sites within the action area, with 39 sites slated for salvage treatment in their home ranges. The Service adequately addressed the 39 sites that may be affected by the Recovery Project, explaining that the home ranges of six of these sites overlap with the action area, but the record suggests that none of these sites would be adversely affected by the Recovery Project.

Based on the spotted owl home range and core-use areas, the Service regularly evaluated recovery projects based on the potential to modify the respective habitat and the degree of potential modification. Given the best available science, the Service explained that it would evaluate a recovery project based on a 40 percent nesting, roosting and foraging home range and a 50 percent nesting, roosting and foraging core-use area. Post-Recovery Project habitat results indicate that these estimates were reliable. These percentages represent estimates, and are just one factor that the Service factored into its analysis. The Service also examined several other site-specific factors, including pre-and post-fire habitat conditions, habitat suitability, and abiotic factors. Despite Cascadia’s contention otherwise, the Service consulted and applied a multitude of scientific data to conclude that proceeding with the salvage project would result in no jeopardy to the spotted owl.

Importantly, the salvage project is slated to affect less than 10 percent of acreage located on federal land. The Douglas Complex Fire burned approximately 48,-000 acres of federal and non-federal land, with a salvage harvest of around 1,276-1,612 acres of fire-affected trees. Further, post-fire nesting, roosting and foraging habitat disturbance is limited, such that less than 20 percent is affected from the home range and core-use areas, and scientific data suggests that there was non-occupancy in several of those sites before the fire. To ensure that the habitat is minimally disturbed, the Bureau implemented restrictions on salvage damage by: 1) precluding harvest on any of the low severity burned areas; 2) limiting salvaging in core-use areas; 3) retaining large trees, snags and downed wood; and 4) reforestation of the burned units. Given the Service’s cautious, conservative and data-guided approach to salvaging, although the spotted owl may increase its range post-fire, the Service’s no jeopardy determination complied with both the Endangered Species Act and the Administrative Procedure Act. See Conservation Cong., 774 F.3d at 620 ; see also Ariz. Cattle Growers’ Ass’n v. U.S. Fish & Wildlife, 273 F.3d 1229, 1236 (9th Cir.2001) (explaining that an administrative decision withstands scrutiny under the Administrative Procedure Act if the decision is not “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.”) (quoting 5 U.S.C. § 706 (2)(A)).

Finally, Cascadia’s reliance on Blue Mountains Biodiversity Project v. Blackwood, 161 F.3d 1208, 1214 (9th Cir.1998) is unavailing. In that case, we faulted the Service for advancing an environmental assessment that “contain[ed] virtually no references to any material in support of or in opposition to its conclusions....” Id. Conversely, in this case the Service cited to several lengthy scientific reports to support its conclusions regarding the potential shift patterns of the spotted owl in a post-fire landscape. As previously noted, it is not within the province of a reviewing court to substitute its judgment for that of the respective agency as long as the agency used adequate and reliable data. See Conservation Cong., 774 F.3d at 620 . We affirm the district court’s ruling rejecting Cascadia’s argument to the contrary. See Nat’l Wildlife Fed’n, 422 F.3d at 793 .

C. Recovery Actions 10 and 12

In an argument similar to that rejected in Conservation Congress, Cascadia contends that the Service failed to adequately utilize the best scientific information necessary to ensure spotted owl recovery when evaluating the Project and rendering its jeopardy determination. According to Cascadia, the Recovery Plan represents the best available science, which the Service was either obligated to follow or explain its departure from when rendering the jeopardy determination. Like the district court, we disagree.

Pursuant to Section 4 of the Endangered Species Act, the Service created a Recovery Plan for the spotted owl with the avowed function “to protect, enhance and develop habitat in the quantity and distribution necessary to provide for the long-term recovery of spotted owls.... ” The purpose of the Recovery Plan is evident— promote recovery of the spotted owl. Although they are not necessarily mutually exclusive, recovery and jeopardy are two distinct concepts.

The biological opinion properly focused on jeopardy rather than monitoring for perfect compliance with the recovery plans. See Conservation Cong., 774 F.3d at 620 . “[DJeclining to adopt particular recommendations in a recovery plan or a study — neither of which is binding on an agency — does not constitute failing to consider them under 50 C.F.R. § 402.16 .... ” Id. In any event, the biological opinion is consistent with Recovery Action 10 because the focus in both is on conserving spotted owl sites and habitat to support the spotted owl population.

The same is true for Recovery Action 12, with its focus on the development of post-fire habitat and restoration of habitat elements that require significant time to develop. The Recovery Project will provide high retention of snags and coarse woody debris in the spotted owl’s critical habitat. Thus, the largest snags, the richest woody debris, and other green legacy features will be retained and aggregated to provide short and long-term benefits to the spotted owl. We affirm the district court’s rejection of Cascadia’s claim that the Service’s jeopardy determination was inconsistent with the Recovery Plan and therefore deficient. See Nat’l Wildlife Fed’n, 422 F.3d at 793 .

In sum, the district court’s conclusion that Cascadia failed to prove a likelihood of success on the merits was supported legally and factually. Because the district court acted within its discretion in reaching that conclusion, we need not consider the remaining preliminary injunction factors. See DISH Network Corp. v. F.C.C., 653 F.3d 771 , 776-77 (9th Cir.2011).

TV. CONCLUSION

The Douglas Complex Fire destroyed approximately 48,000 acres of forest in Oregon. In response, the Bureau created the Recovery Project to salvage remaining spotted owl habitat and optimize the growth of new habitat. Finding that the Service’s conclusions were supported by the best available science and were not arbitrary or capricious, the district court denied Cascadia’s motion for a preliminary injunction. Given our limited review of the denial of a preliminary injunction, we conclude that the district court acted within its discretion. The denial of injunctive relief was legally sound and supported by the record.

AFFIRMED.

. The referenced timber sales include Rogue Cow, Burnt Rattler, and Rock Star. . The logging operations commenced nearly one year ago. During oral argument, the intervenors estimated that approximately 33 acres have already been felled. .Survey calls are attempts made by scientific surveyors to observe/locate the Northern Spotted Owl in its habitat. Typically, survey calls are verbal in nature, with an anticipated audible response from any owls that are present. . Home range is the area covered in the normal activities of feeding, mating, and nurturing. Areas subject to concentrated use, usually around the nest site and favored foraging areas, are core areas. . A “take” is defined under the Endangered Species Act as “to harass, harm, pursue, hunt, shoot, wound, kill, trap, capture, or collect, or to attempt to engage in any such 'conduct.” 16U.S.C. § 1532(19). . It is worth noting that recovery plans are not binding, and Cascadia conceded this point at oral argument before the district court. See Conservation Cong., 774 F.3d at 620 .