Defenders of Wildlife v. United States Forest Service

94 F.4th 1210

• Court: Court of Appeals for the Tenth Circuit
• Decided: March 11, 2024
• Docket: 23-1093
• Status: Published

Opinion

Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 1 FILED United States Court of Appeals PUBLISH Tenth Circuit

UNITED STATES COURT OF APPEALS March 11, 2024

Christopher M. Wolpert FOR THE TENTH CIRCUIT Clerk of Court _________________________________

DEFENDERS OF WILDLIFE,

Petitioner - Appellant,

v. No. 23-1093

UNITED STATES FOREST SERVICE; UNITED STATES FISH AND WILDLIFE SERVICE,

Respondents - Appellees. _________________________________

Appeal from the United States District Court for the District of Colorado (D.C. No. 1:21-CV-02992-RM) _________________________________

Ellen Medlin Richmond, Attorney (McCrystie Adams and W. Cory Haller, Attorneys, with her on the briefs) Defenders of Wildlife, Denver, Colorado for Appellant.

Jacob David Ecker (Todd Kim, Assistant Attorney General, and Katelin Shugart-Schmidt, Attorney, on the brief) United States Department of Justice, Denver, Colorado for Appellee. _________________________________

Before MATHESON, KELLY, and EID, Circuit Judges. _________________________________

MATHESON, Circuit Judge. _________________________________ Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 2

GLOSSARY

Term Definition

APA Administrative Procedure Act BA Biological Assessment BiOp Biological Opinion 2021 BiOp 2021 Biological Opinion for the 2020 Rio Grande National Forest Revised Land Management Plan 2019 BiOp 2019 Biological Opinion for the 2020 Rio Grande National Forest Revised Land Management Plan DPS Distinct Population Segment ESA Endangered Species Act FWS United States Fish & Wildlife Service Ivan Study Jake Ivan et al., Predictive Map of Canada Lynx Habitat Use in Colorado LAU Lynx Analysis Unit NEPA National Environmental Policy Act Plan 2020 Rio Grande National Forest Revised Land Management Plan RGNF Rio Grande National Forest SISS Stand Initiation Structural Stage SRLA Southern Rockies Lynx Amendment Squires Study John Squires et al., A Specialized Forest Carnivore Navigates Landscape-Level Disturbance: Canada Lynx in Spruce-Beetle Impacted Forests, 475 Forest Ecology & Mgmt. (2020), and related materials 2017 SSA FWS, Species Status Assessment for the Canada Lynx Contiguous United States Distinct Population Segment (2017) Theobald and David M. Theobald & Tanya M. Shenk, Areas of High Habitat Use Shenk Study from 1999-2010 for Radio-Collared Canada Lynx Reintroduced to Colorado (2011)

i Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 3

USFS United States Forest Service VEG S Vegetation Management Standard WUI Wildland Urban Interface

ii Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 4

INTRODUCTION

From about 2008 to 2017, a bark beetle epidemic killed nearly all the spruce trees

in the Rio Grande National Forest (“RGNF”) in Colorado. In response, the United States

Forest Service (“USFS”) revised its Land Management Plan (“the Plan”) for the RGNF.

The USFS consulted the United States Fish and Wildlife Service (“FWS”), as required by

the Endangered Species Act of 1973 (“ESA”),1 to consider the Plan’s effects on Canada

lynx in the contiguous United States. In 2021, the FWS issued a Biological Opinion

(“2021 BiOp”) concluding the Plan would not likely jeopardize the lynx’s continued

existence.

Defenders of Wildlife (“Defenders”) petitioned for review, arguing that the 2021

BiOp violated the ESA and the Administrative Procedure Act (“APA”)2 and that the

USFS improperly relied on the BiOp in preparing the Plan.3 The district court found the

2021 BiOp complied with the ESA and the APA and dismissed Defenders’ petition.

On appeal, Defenders renews its ESA and APA challenges to the 2021 BiOp. It

argues the FWS (A) failed to adequately address conclusions about the Canada lynx

subpopulation in Colorado from the agency’s 2017 Species Status Assessment

1 16 U.S.C. §§ 1531-1544. 2 5 U.S.C. §§ 551, 553-559, 701-706. 3 Defenders also petitioned for review of the Plan, the Plan’s environmental impact statement, and the accompanying Record of Decision under the National Environmental Policy Act of 1969 (“NEPA”), 42 U.S.C. §§ 4321-4370h. That claim is not at issue on appeal.

1 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 5

(“2017 SSA”), (B) acted arbitrarily and contrary to the best available science when it

described the northern part of the RGNF as “low-use” lynx habitat, and (C) inadequately

analyzed the Plan’s impact on lynx in “low-use,” and (D) “high-use” habitat. Defenders

also contends (E) the USFS improperly relied on the 2021 BiOp.

Exercising jurisdiction under 28 U.S.C. § 1291, we affirm. The FWS did not

violate the ESA or the APA, and the USFS appropriately relied on the FWS 2021 BiOp.

Canada lynx. App., Vol. 7 at 46.

2 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 6

I. BACKGROUND

A. Legal Background

National Forest Management Act

The USFS manages the national forest system under the National Forest

Management Act, 16 U.S.C. §§ 1600-1687, which prescribes “a two-step process” for

forest planning and management at the programmatic forest and individual project levels.

Biodiversity Conservation All. v. Jiron, 762 F.3d 1036, 1049 (10th Cir. 2014); see also

Ohio Forestry Ass’n v. Sierra Club, 523 U.S. 726, 729-30 (1998). At the programmatic

level, the USFS develops forest-wide planning goals in a Land and Resource

Management Plan, or forest plan. Utah Env’t Cong. v. Bosworth, 443 F.3d 732, 736-37

(10th Cir. 2006). In doing so, the USFS must “provide for multiple use and sustained

yield of the products and services,” including coordination of outdoor recreation, range,

timber, wildlife, and wilderness uses. See 16 U.S.C. § 1604(a), (e). The USFS then

implements forest plans through site-specific individual projects. Id. § 1604(a), (i);

see also Biodiversity Conservation All., 762 F.3d at 1049.

All agency actions, including site-specific projects, must comply with the forest

plan, Utah Env’t Cong., 443 F.3d at 737; 16 U.S.C. § 1604(i), and the National

Environmental Policy Act of 1969 (“NEPA”), 42 U.S.C. §§ 4321-4370h, see Silverton

Snowmobile Club v. USFS, 433 F.3d 772, 785 (10th Cir. 2006).

Endangered Species Act

Forest plans must comply with the ESA, which Congress enacted “to provide a

means whereby the ecosystems upon which endangered species and threatened species

3 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 7

depend may be conserved” and “to provide a program for the conservation of such

endangered species and threatened species.” 16 U.S.C. § 1531(b). The ESA “authorizes

the Secretary of the Interior to list domestic or foreign species as endangered or

threatened,” triggering certain protections. People for Ethical Treatment of Prop. Owners

v. FWS, 852 F.3d 990, 995 (10th Cir. 2017) (quotations omitted). The ESA defines

“species” to include subspecies, as well as “any distinct population segment” (“DPS”).

16 U.S.C. § 1532(16).4

Section 7 of the ESA requires federal agencies to consult with the relevant

Secretary, here the Secretary of the Interior, to “insure that any action authorized, funded,

or carried out by [an] agency . . . is not likely to jeopardize the continued existence of any

endangered species or threatened species.” Id. § 1536(a)(2). To “[j]eopardize the

continued existence” means “to engage in an action that reasonably would be expected,

directly or indirectly, to reduce appreciably the likelihood of both the survival and

recovery of a listed species in the wild by reducing the reproduction, numbers, or

distribution of that species.” 50 C.F.R. § 402.02 (2021).5

4 The ESA does not define DPS. The FWS and the National Marine Fisheries Service have interpreted the term to mean a population that is both discrete—either “markedly separated from other populations of the same taxon” or “delimited by” certain “international governmental boundaries”—and important “to the taxon to which it belongs.” Policy Regarding the Recognition of Distinct Vertebrate Population Segments Under the Endangered Species Act, 61 Fed. Reg. 4,722, 4,725 (Feb. 7, 1996). 5 We cite the ESA regulations that were applicable when the FWS issued the 2021 BiOp.

4 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 8

“The ESA duty to avoid jeopardy is policed by a procedural consultation

requirement,” which involves (1) an “action agency”—the agency taking an action that

could affect a listed species, and (2) a “consultant agency”—either the FWS or the

National Marine Fisheries Service, depending on the species involved. W. Watersheds

Project v. Haaland, 69 F.4th 689, 699 (10th Cir. 2023) (quotations omitted). Here, the

action agency is the USFS and the consultant agency is the FWS.

The agencies’ assessments of an action’s impact may trigger a formal consultation

requirement. If the action agency prepares a biological assessment (“BA”) that

determines the proposed action is not likely to adversely affect the listed species, formal

consultation is not necessary. See 50 C.F.R. § 402.14(b)(1). But if the BA concludes the

action “may affect listed species,” the action agency must formally consult with the

consultant agency. Id. § 402.14(a).

After formal consultation, the consultant agency must prepare a BiOp “as to

whether the action is likely to jeopardize the continued existence of [the] listed species.”

Id. § 402.14(g)(4). The consultant agency must (1) “[r]eview all relevant information

provided by the Federal agency or otherwise available”; (2) “[e]valuate the current status

. . . of the listed species”; (3) “[e]valuate the effects of the action and cumulative effects

on the listed species”; and (4) “use the best scientific and commercial data available.”

Id. § 402.14(g); see also 16 U.S.C. § 1536(a)(2) (in making a jeopardy determination,

“each agency shall use the best scientific and commercial data available”).

5 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 9

Administrative Procedure Act

Because the ESA does not provide a private right of action for Section 7 claims,

we review such claims under the APA. W. Watersheds Project, 69 F.4th at 700. Under

the APA, a court may overturn an agency’s decision only if it is “arbitrary, capricious, an

abuse of discretion, or otherwise not in accordance with law.” 5 U.S.C. § 706(2)(A);

see also W. Watersheds Project, 69 F.4th at 700. Agency action is arbitrary and

capricious if the agency “has relied on factors which Congress has not intended it to

consider, entirely failed to consider an important aspect of the problem, offered an

explanation for its decision that runs counter to the evidence before the agency,” or if the

action “is so implausible that it could not be ascribed to a difference in view or the

product of agency expertise.” Motor Vehicle Mfrs. Ass’n of the U.S. v. State Farm

Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983).

“We review a district court’s resolution of APA claims de novo, applying the same

deferential standard toward the agency’s decisions that the district court applies.”

W. Watersheds Project, 69 F.4th at 700 (quotations omitted). “[T]he burden is on the

petitioner to demonstrate that the action is arbitrary and capricious” and to overcome the

“presumption of validity” afforded to such action. Copar Pumice Co. v. Tidwell, 603 F.3d

780, 793 (10th Cir. 2010). “Our deference to the agency is especially strong where the

challenged decisions involve technical or scientific matters within the agency’s area of

expertise.” Diné Citizens Against Ruining Our Env’t v. Haaland, 59 F.4th 1016, 1029-30

(10th Cir. 2023) (quotations omitted).

6 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 10

Courts consider only the agency’s stated reasons for its decision. See State Farm,

463 U.S. at 43. We may not look to “post hoc rationalizations” offered by agency

officials or counsel. Dep’t of Homeland Sec. v. Regents of the Univ. of Cal., 140 S. Ct.

1891, 1909 (2020). Nor may we “attempt . . . to make up for [any] deficiencies” in the

agency’s reasoning by “supply[ing] a reasoned basis for the agency’s action that the

agency itself has not given.” State Farm, 463 U.S. at 43 (quotations omitted); see SEC v.

Chenery Corp., 318 U.S. 80, 95 (1943).

B. Factual Background

Canada Lynx, the DPS, and the RGNF

The Canada lynx is a medium-sized cat adapted for hunting in deep snow. It

primarily hunts snowshoe hare. The lynx lives almost exclusively in Alaska and Canada,

but also in the contiguous United States.

In 2000, the FWS listed the DPS of Canada lynx in the contiguous United States

as a threatened species. App., Vol. 4 at 2. The lynx DPS’s habitat is approximately

30 million acres, App., Vol. 1 at 236, across six separate geographic units, including

Western Colorado, App., Vol. 6 at 10. “Colorado represents the extreme southern edge of

the range of the lynx.” App., Vol. 4 at 9. The Colorado Division of Wildlife began

reintroducing lynx to the State in the late 1990s because few, if any, native lynx remained

then. Id.; App., Vol. 1 at 272.

Much of the reintroduced lynx progeny lives in the RGNF. The RGNF covers

more than 1.8 million acres of south-central Colorado. App., Vol. 1 at 197. According to

the FWS 2018 BA for the Plan, the RGNF “includes some of the most important lynx

7 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 11

habitat in Colorado”—“[o]f about 12 geographic locations in Colorado where lynx can

consistently be located, at least six (50%) occur on the RGNF.” Id. at 65. Lynx presence

in the RGNF “is associated with larger contiguous blocks of forest that [are] primarily

dominated by” a “specialized [spruce-fir] forest structure.” Id. at 62-63. The lynx habitat

in the RGNF, about 661,000 acres, amounts to about 2 percent of the DPS’s total habitat.

Id. at 236.6 The following map depicts the RGNF in Colorado.

Location of the RGNF in Colorado, shown in light green. App., Vol. 2 at 168.

6 The RGNF is not currently part of the lynx DPS’s designated critical habitat. App., Vol. 1 at 62. In 2016, the United States District Court for the District of Montana found that excluding Colorado from the lynx critical habitat designation was unlawful. App., Vol. 5 at 145-53 (WildEarth Guardians v. U.S. Dep’t of the Interior, No. 9:14-cv-00270 (D. Mont. Sept. 7, 2016) (order)). In April 2022, the court approved a settlement agreement under which the FWS will submit a proposed critical habitat rule to the Federal Register by November 21, 2024, and will submit a final rule within the statutory timeline. See WildEarth Guardians v. Williams, No. 9:20-cv-00097 (D. Mont. Apr. 25, 2022) (order approving settlement agreement); Joint Mot. to Approve Settlement Agmt. at 5, WildEarth Guardians, No. 9:20-cv-00097 (D. Mont. Apr. 15, 2022).

8 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 12

Southern Rockies Lynx Amendment

When the FWS listed the lynx DPS in 2000, it noted that the then-current USFS

forest plans for the RGNF and several other forests were inadequate to protect the lynx.

App., Vol. 4 at 2. In 2008, the USFS adopted the Southern Rockies Lynx Amendment

(“SRLA”), which applied to the RGNF and established goals, objectives, standards, and

guidelines for protecting lynx against several risk factors, including logging. Id. at 133,

137-41. Different vegetation management standards applied depending on the conditions

in a given lynx analysis unit (“LAU”). Id. at 142-45. An LAU is a 25-50 square-mile

area designed to represent the home range of a theoretical female lynx. Id. at 143, 186.

Spruce Beetle Epidemic

Although the SRLA considered the effects of a “large-scale mountain pine beetle

epidemic,” id. at 140-41, it did not anticipate the spruce beetle epidemic that swept

through the RGNF in the ensuing years. App., Vol. 1 at 228 (“[T]he SRLA BiOp did not

anticipate a landscape-level bark beetle epidemic, and the subsequent large-scale timber

salvage activities that may follow.”); App., Vol. 5 at 210 (showing spruce beetle replacing

mountain pine beetle). The spruce beetle outbreak devastated the spruce-fir ecosystem,

an essential lynx habitat. App., Vol. 1 at 66. The entire overstory—the uppermost layer

in the forest—died by 2017. Id. at 66, 225. Between 2011 and 2018, the proportion of

9 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 13

lynx habitat in the RGNF that was considered unsuitable grew from 0.6 percent to an

average of 24 percent across all LAUs.7 Id. at 65-66.

2017 Species Status Assessment, 2018 Biological Assessment, and 2019 BiOp

In response to the spruce beetle epidemic, the USFS began to revise the RGNF

forest plan. It conducted informal discussions with the FWS about lynx conservation

beginning in fall 2017. Around the same time, the FWS released the 2017 SSA for the

lynx DPS, which discussed the current and projected outlook for the species.

In September 2018, the USFS issued a BA concluding that the draft of the Plan

“may affect and is likely to adversely affect Canada lynx” in part because it was uncertain

how changes in management standards proposed in the draft would affect lynx and their

prey. Id. at 87-88. The USFS was thus required to formally consult with the FWS.

See id. at 170; 50 C.F.R. § 402.14(a).

In March 2019, the FWS issued a BiOp (“2019 BiOp”) concluding that the draft of

the Plan was “likely to adversely affect the Canada lynx,” App., Vol. 1 at 170, but “not

likely to jeopardize the continued existence” of the lynx DPS, id. at 189.

2020 Revised Forest Plan

In May 2020, the USFS finalized the new Plan. The Plan divided the RGNF into

“low-use” and “high-use” lynx habitat8 and applied new management standards to each

7 The USFS redrew the LAUs in 2018 and removed two of the original LAUs for lack of lynx persistence, so this comparison is not exact. See App., Vol. 1 at 182, 224, 226. 8 The Plan refers to “high probability lynx use area[s] (95 percent area[s]).” App., Vol. 3 at 38. The parties and the district court refer to these areas as “high-use” and all 10 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 14

category, maintaining some SRLA standards but altering others. See, e.g., App., Vol. 3

at 39.

The USFS based this classification largely on an RGNF-commissioned study by

Dr. John Squires (“Squires Study”) and other top lynx researchers that examined how

lynx were responding to the spruce beetle epidemic. Id. at 37-38; App., Vol. 5 at 165.

The Squires Study began in 2013 and culminated with a paper published in the Forest

Ecology and Management journal in 2020. App., Vol. 5 at 165; App., Vol. 1 at 242-56.

It analyzed a southern part of the RGNF and developed a model that predicted which

parts of the study area would be in the top 95 percent of any lynx’s use and which parts

would be in the bottom 5 percent. App., Vol. 1 at 247, 249-51 (explaining model);

compare id. at 244 (showing study area boundaries), with id. at 221 (showing RGNF

boundaries).

The USFS classified the northern part of the RGNF as low-use habitat. See App.,

Vol. 3 at 37-38. The Plan explained:

Dr. Squires consulted State and Forest-level biologists before initiating the field study to ensure that the designated study area captured all primary lynx use areas on the Forest. The northern area of the Forest was deemed to support minimal consistent lynx use, both currently and historically following the reintroduction [of lynx to Colorado]. Therefore, Dr. Squires and colleagues concluded that the northern portion of the Forest supported too few Canada lynx to capture enough individuals to inform reliable modeling and mapping products.

other areas as “low-use.” See, e.g., Aplt. Br. at 3; Aplee. Br. at 32, 50; App., Vol. 7 at 247. We use the parties’ terminology.

11 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 15

Id. at 38.

2021 BiOp

In September 2020, Defenders notified the two agencies of its intent to sue,

alleging the FWS 2019 BiOp and the USFS Plan violated the ESA. App., Vol. 1 at 195;

see 16 U.S.C. § 1540(g) (requiring such notice). In February 2021, the FWS issued the

2021 BiOp, which added analysis and reached the same no-jeopardy conclusion. App.,

Vol. 1 at 210-41. The USFS did not further revise the Plan.

The 2021 BiOp and some of the other materials mentioned above will be

described more fully as needed in our discussion of the issues raised by Defenders.

C. Procedural Background

Defenders sued the USFS and the FWS in the United States District Court for the

District of Colorado, alleging violations of NEPA, the ESA, and the APA. It argued the

FWS 2021 BiOp violated the ESA and the APA and that the USFS improperly relied on

the BiOp in developing the Plan. As noted above, Defenders also challenged the Plan,

the Plan’s environmental impact statement, and the accompanying Record of Decision

under NEPA. The district court found the agencies complied with all three statutes and

dismissed Defenders’ petition for review.

Defenders timely appealed. It reasserts only the ESA and APA claims on appeal

and does not challenge the district court’s NEPA decision.

II. DISCUSSION

Defenders challenges the FWS 2021 BiOp’s (A) consideration of the 2017 SSA,

(B) description of the northern part of the RGNF as “low-use” lynx habitat, and

12 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 16

(C) analysis of the Plan’s impact in “low-use” and (D) “high-use” habitat. It also

contends (E) the USFS improperly relied on the 2021 BiOp. We address these issues in

turn.

A. The 2021 BiOp’s Consideration of the 2017 Species Status Assessment

Defenders argues the FWS inadequately addressed the 2017 SSA in the 2021

BiOp. Aplt. Br. at 23-30. We disagree.

Additional Factual Background

a. 2017 SSA

The 2017 SSA for the lynx DPS “evaluate[d] the current and possible future

conditions for lynx in 6 geographic units”—Washington, Montana and Idaho, the Greater

Yellowstone area, Colorado, Minnesota, and Maine. App., Vol. 6 at 9-10. It was intended

to “inform a determination by [FWS] decision makers of whether (1) the DPS continues

to warrant protection under the ESA and (2) a recovery plan is needed to guide

conservation and recovery of the lynx DPS.” Id. at 19. It did not specifically evaluate

the RGNF, which is one of seven national forests in the Colorado geographic unit.

The 2017 SSA described Colorado as “currently” having “many more resident

lynx . . . than [it] likely [did] historically, and many more than were known or suspected

at the time the DPS was listed.” Id. at 17. It determined that “the current population in

Colorado” gives the DPS “greater” “redundancy . . . at least temporarily, now than it

[had] historically.” Id. at 18, 240. It predicted that “resident lynx in [Colorado] are likely

to persist at year 2025.” Id. at 17. But it also predicted that, compared to the projections

of surveyed experts, Colorado “is less likely . . . to support a resident population at 2050

13 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 17

[80 percent likelihood according to experts] or at 2100 [50 percent likelihood according

to experts].” Id. at 13, 17.

Evaluating the DPS as a whole, the 2017 SSA concluded that “[t]he current broad

distribution of resident lynx in large, geographically discrete areas (redundancy) makes

the DPS invulnerable to extirpation caused by a single catastrophic event.” Id. at 240.

It predicted that lynx in each geographic unit would likely “become smaller and more

patchily-distributed due largely to projected climate-driven losses in habitat quality and

quantity and related factors,” but “the timing, rate, and extent of habitat decline” was

“highly uncertain.” Id. at 241. It asserted that “[d]espite some reduced resiliency, . . .

resident lynx populations are very likely to persist in all 5 [geographical] units that

currently support them [including Colorado, but not the Greater Yellowstone area] in the

near-term (2025) and in all or most of those units at 2050.” Id. The 2017 SSA predicted

with “high[] uncertain[ty]” that “resident lynx populations could be functionally

extirpated from some [geographic] units by the end of the century,” which “would

indicate a loss of resiliency, reduced redundancy and representation, and an increased risk

of extirpation of the [entire] DPS.” Id.

b. 2021 BiOp

The purpose of the 2021 BiOp was to evaluate “whether the [Plan] is likely to

jeopardize the continued existence of [the] listed species,” here, the lynx DPS. 50 C.F.R.

§ 402.14(g)(4); 16 U.S.C. § 1536(a)(2).

The 2021 BiOp said the 2017 SSA “found no reliable information indicating a

substantial reduction of the current distribution and abundance of” lynx in the DPS “from

14 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 18

historical conditions.” App., Vol. 1 at 222. The risk was low because the lynx DPS is

“resilien[t]” and the DPS has “broad distribution” in “geographically discrete areas,”

making it “invulnerable to extirpation caused by a single catastrophic event.” Id. The

2021 BiOp also noted, based on the 2017 SSA, that the Colorado subpopulation provides

“increased redundancy in the DPS, at least temporarily,” compared to historical trends.

Id. at 223.

Citing other sources, the 2021 BiOp explained that “[t]he current size of the

resident lynx population in Colorado is unknown but thought to number between 100 and

250” and that “the long-term persistence of the introduced population remains uncertain.”

Id. Of the estimated 100 to 250 lynx in Colorado, it reported that only about 10 lynx

were thought to be present in the Squires Study area—the southern part of the RGNF. Id.

(citing id. at 244).

The 2021 BiOp concluded the Plan was “not likely to appreciably reduce the

likelihood of both survival and recovery of the Canada lynx” because “the RGNF . . . is

just one of seven national forests within the southern Rockies geographic area, and those

other forests would be unaffected by the [Plan],” and “the RGNF makes up just over two

percent of the total lynx habitat [for the DPS].” Id. at 236-37.

Analysis

Defenders contends the 2021 BiOp’s analysis of the Plan’s impact on the lynx

DPS failed to adequately consider the 2017 SSA. It argues the FWS “failed to analyze

the consequences for the lower-48 DPS if . . . the Colorado population deteriorates or it

disappears altogether,” Aplt. Br. at 25, including “ignor[ing] the [2017 SSA’s] extirpation

15 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 19

projection,” Aplt. Reply Br. at 3. It argues the FWS needed to analyze the risks to the

Colorado subpopulation because “subpopulation extirpation . . . must be considered as

part of jeopardy analysis” and failure to do so overlooks an “important aspect[] of the

problem.” Aplt. Br. at 23.

For two reasons, Defenders has failed to show an APA violation. First, the FWS’s

task under the ESA was to assess whether the Plan’s implementation in the RGNF would

likely jeopardize the lynx DPS. The FWS reasonably carried out this task in the

2021 BiOp. Second, Defenders relies on Wild Fish Conservancy v. Salazar, 628 F.3d 513

(9th Cir. 2010), which is not binding and both legally and factually inapposite.

a. The FWS’s no-jeopardy analysis

Defenders’ criticism of the 2021 BiOp misses the mark. The FWS’s focus in the

2021 BiOp was on the relationship between the RGNF and the lynx DPS. Section 7 of

the ESA required the FWS to consider whether the Plan was “likely to jeopardize the

continued existence of any endangered species or threatened species,” here, the lynx

DPS. 16 U.S.C. § 1536(a)(2). By contrast, the FWS’s focus in the 2017 SSA was on the

relationship between the six geographic units in the DPS and the overall lynx DPS.

To the extent the 2017 SSA’s analysis of the lynx subpopulation in Colorado was

relevant to the RGNF Plan’s impact on the lynx DPS, the 2021 BiOp considered it and

was not arbitrary or capricious in doing so. It explained:

 The 2017 SSA showed the DPS was at minimal risk and was “invulnerable to extirpation caused by a single catastrophic event.” App., Vol. 1 at 222.

16 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 20

 The Colorado and RGNF subpopulations were relatively small and nonessential to the DPS. See id. at 223. But it also recognized “the long-term persistence of the introduced [lynx] population [in Colorado] remains uncertain.” Id.

 The Plan would affect no more than 50,000 acres of RGNF lynx habitat, compared to 661,000 acres of lynx habitat in the RGNF, 7,000,000 acres in the southern Rockies, and 30,000,000 acres in the DPS. Id. at 236.

The 2021 BiOp concluded that because the Plan affected only 7.5 percent of a forest that

makes up only 2 percent of the DPS lynx habitat, the Plan would not likely jeopardize the

lynx DPS. Id. at 236-37.

Defenders argues the FWS failed to consider the Plan’s impact on “the Colorado

population,” which Defenders characterizes as “struggling” based on the 2017 SSA.

Aplt. Br. at 24-25. The FWS’s job was to consider the Plan’s impact on the DPS. When

the record shows a subpopulation is particularly important to the species, the FWS may

need to consider how the agency action affects that subpopulation to give a reasoned

jeopardy opinion. See W. Watersheds Project, 69 F.4th at 707-08; Wild Fish, 628 F.3d

at 528; Save Our Cabinets v. FWS, 255 F. Supp. 3d 1035, 1050-51 (D. Mont. 2017). But

as we further explain, the record does not support that (1) the Colorado subpopulation

plays such a role within the DPS or (2) the RGNF plays such a role within the Colorado

subpopulation.

First, the 2017 SSA showed the Colorado subpopulation was a fraction of the

estimated DPS population. See App., Vol. 6 at 15-17 (estimating “750-1,000” lynx in

Maine; “50 to 200” in Minnesota; “200-300” in Montana and Idaho; “30-35” in

Washington; “0-10” in the Greater Yellowstone area; and “100-250” in Colorado). It did

not identify a single DPS geographic unit or factor influencing lynx viability that was 17 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 21

necessary to its determination that by 2100 “resident lynx populations could be

functionally extirpated from some [geographic] units,” which “would indicate a loss of

resiliency, reduced redundancy and representation, and an increased risk of extirpation of

the DPS.” Id. at 241. Consistent with the foregoing information and the 2017 SSA’s

statement that “the current population in Colorado” gives the DPS “greater” “redundancy

. . . at least temporarily, now than it [had] historically,” id. at 240, the 2021 BiOp

discussed the historically small size of the Colorado subpopulation. It stated that “[t]he

best available information indicates that the lynx population in Colorado is, and likely has

always been low.” App., Vol. 1 at 223.9

Second, the 2021 BiOp refuted Defenders’ assertion that “the fate of the Colorado

population likely turns on the ability of the [RGNF] to support lynx.” Aplt. Br. at 25.

The BiOp pointed to the RGNF’s small lynx population relative to the Colorado

population. App., Vol. 1 at 223 (estimating 10 lynx in the Squires Study area but

“between 100 and 250” lynx in Colorado). Although the FWS 2018 BA for the Plan

stated that the RGNF “includes some of the most important lynx habitat in Colorado,”

id. at 65; accord id. at 244 (Squires Study), it did not conclude the RGNF subpopulation

was integral to the Colorado population. The 2021 BiOp concluded that the Plan would

not affect the six other “national forests within the southern Rockies geographic area.”

9 Defenders’ arguments about the Colorado subpopulation’s importance to the lynx in the face of climate change, Aplt. Br. at 22; Aplt. Reply Br. at 4, fail to account for the 2017 SSA’s recognition of “projected climate-driven losses in habitat quality and quantity” in all six lynx DPS geographic units, App., Vol. 6 at 241.

18 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 22

Id. at 236. And all seven Colorado forests are likely home to lynx. App., Vol. 6 at 120;

see also App., Vol. 1 at 65 (reporting that half of the “geographic locations in Colorado

where lynx can consistently be located” are outside the RGNF).

The 2021 BiOp’s consideration of the 2017 SSA’s analysis of the Colorado

subpopulation was reasonable.

b. Defenders’ reliance on Wild Fish

Defenders argues the 2021 BiOp was arbitrary or capricious in light of the Ninth

Circuit’s opinion in Wild Fish, but it again misses the mark.

In Wild Fish, the Ninth Circuit reviewed an FWS BiOp’s consideration of a

project’s impact on bull trout. The bull trout species, broken down into increasingly

smaller geographic units, consists of (1) “interim recovery units,” which include

(2) “geographic ‘core areas,’” which include (3) local populations. 628 F.3d at 519. The

project at issue in Wild Fish affected “one of seven” local populations in the Wenatchee

River core area, which itself was one of 90 core areas within the Columbia River interim

recovery unit. Id. A regulation required the FWS to “conduct[] its jeopardy

determinations for the bull trout at the interim recovery unit level” rather than the species

level. Id.; see Endangered and Threatened Wildlife and Plants; Determination of

Threatened Status for Bull Trout in the Coterminous United States, 64 Fed. Reg. 58,910,

58,930 (Nov. 1, 1999).

The BiOp in Wild Fish noted that “[t]he Wenatchee River core area is particularly

important to the recovery unit because it is a relative stronghold for bull trout” and “the

[affected] local population is important to the core area because its location . . . could

19 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 23

insulate th[e] population from disturbances . . . that would affect most of the other[] [local

populations].” 628 F.3d at 529 (quotations omitted). The BiOp then found that local

population reductions due to the project would “improve . . . the [population’s]

contribution . . . to the survival of the bull trout.” Id. at 528. The Ninth Circuit

concluded it was irrational for the BiOp to say reductions in one local population could

somehow improve the overall survival of the interim recovery unit. Id. at 528-29.

Wild Fish is nonbinding and legally and factually distinguishable.

First, Defenders’ contention that under Wild Fish, the ESA requires an agency to

consider “subpopulation extirpation . . . as part of jeopardy analysis,” Aplt. Br. at 23, fails

as a legal matter. It ignores that Wild Fish turned on a bull trout regulation that required

the FWS to “conduct[] its jeopardy determinations . . . at the interim recovery unit level.”

Wild Fish, 628 F.3d at 519. No comparable regulation applies here. As explained, the

FWS’s no-jeopardy determination properly addressed the Plan’s effect on lynx at the DPS

species level. See, e.g., App., Vol. 1 at 212, 222. Also, as discussed above, the 2021

BiOp reasonably considered the 2017 SSA’s analysis of the Colorado subpopulation.

Second, Wild Fish is factually inapposite. Unlike in Wild Fish, the FWS never

concluded in the 2017 SSA or the 2021 BiOp that reductions in the RGNF subpopulation,

let alone the Colorado subpopulation, would improve the survival or recovery of the lynx

DPS. More importantly, the FWS never concluded that Colorado, in contrast to the “core

area” in Wild Fish, was “particularly important” or a “relative stronghold” for the DPS.

Wild Fish, 628 F.3d at 529. Nor did it conclude that the RGNF, in contrast to the local

population in Wild Fish, could “insulate” the Colorado subpopulation “from disturbances

20 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 24

. . . that would affect” the other six national forests in the southern Rockies. Id.

(quotations omitted); see also Save Our Cabinets, 255 F. Supp. 3d at 1050-51 (relying on

the “essential role” and “importance of the local populations at issue” to conclude that the

FWS’s no-jeopardy conclusion was arbitrary). Although “particular small population

units” can in theory be significant to the FWS’s consideration of an “overall population,”

Aplt. Reply Br. at 5 (quoting Aplee. Br. at 27), the record does not show this was so with

the Colorado or RGNF subpopulations.

* * * *

The FWS did not arbitrarily fail to consider the 2017 SSA in the 2021 BiOp.10

B. The 2021 BiOp’s Description of Northern Part of the RGNF as “Low Use”

Defenders argues the FWS 2021 BiOp (a) arbitrarily assumed the northern part of

the RGNF was a low-use area because the Squires Study excluded the northern part of

the RGNF from its study area,11 and (b) did not rely on the best available science in

10 In one sentence, Defenders asserts that “the SSA’s conclusion that lynx would persist in Colorado until the end of the century was based on FWS’s assumption that the SRLA protections would remain in place—an assumption the Plan upends.” Aplt. Br. at 25. We agree with the Government that “Defenders does not actually support [this] claim.” Aplee. Br. at 29. Defenders’ discussion of this issue in its reply brief is no more illuminating. See Aplt. Reply Br. at 7. “[P]erfunctory” allegations of error that “fail[] to frame and develop an issue” are insufficient to “invoke appellate review.” Kelley v. City of Albuquerque, 542 F.3d 802, 819 (10th Cir. 2008) (quotations omitted); see also United States v. Cooper, 654 F.3d 1104, 1128 (10th Cir. 2011) (“It is well-settled that arguments inadequately briefed in the opening brief are waived.” (alterations and quotations omitted)). We decline to consider this issue. 11 Neither party defines the term “northern part of the RGNF,” but it appears to refer to the LAUs that do not overlap with the Squires Study area: 4-Mile to La Garita Creek, Bonanza, Carnero, Cochetopa, Creede, Embargo, Groundhog Park, La Garita 21 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 25

determining the area was low use. Aplt. Br. at 18-19, 31-39. Under the APA’s deferential

standard of review, the FWS did not act arbitrarily and did not err under the ESA’s best-

available-science standard.

a. 2021 BiOp

Like the Plan, the 2021 BiOp concluded the northern part of the RGNF is likely

low-use lynx habitat. App., Vol. 1 at 226, 230. It explained:

Dr. John Squires and [his] colleagues . . . focused their [study on] . . . the areas known to support Canada lynx on a consistent basis. Dr. Squires consulted State and Forest-level biologists before initiating the field study, to ensure that the designated study area captured all primary lynx use areas on the RGNF. Based on these discussions, the northern area of the RGNF was deemed to support little consistent to no lynx use . . . . Therefore, Dr. Squires and colleagues concluded that the northern portion of the RGNF supported too few Canada lynx to . . . inform reliable modeling and mapping products. Instead, they focused their resources on the southern portion of the RGNF . . . (P. McDonald pers comm. December 2020). We conclude, based on this information, that even though lynx habitat is present on the northern part of the Forest, it likely does not provide high quality lynx habitat, and likely never has, and is unlikely to develop into high quality lynx habitat.

Id. at 226. Dr. Squires helped write this text.12 Later in the 2021 BiOp, the FWS stated:

Wilderness, Saguache Park, Sangre De Cristo North, and Sangre De Cristo South. See App., Vol. 1 at 221, 230. 12 The administrative record includes the “(P. McDonald pers. comm. December 2020)” referenced in the quoted text. Dist. Ct. Doc. 16-1, at FWS00153-54. This refers to an email from a USFS employee to a FWS employee that says: “I put together the below paragraph [the same paragraph that was included in the 2021 BiOp] with [Dr.] Squires’ help to help respond to concerns raised . . . about failure to include lynx in 22 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 26

Based on Squires et al. 2020, we believe that habitat in . . . LAUs [not included in the Squires Study] is unlikely to provide sufficient high quality habitat to support high numbers of lynx. Mapping of lynx habitat on the RGNF, based upon Squires et al. (2020), indicate[s] that lynx occupancy of these LAUs is unlikely because the habitat conditions they prefer is lacking.

Id. at 230.

b. Squires Study

The Squires Study did not discuss how it determined the study area boundaries,

which did not cover the full RGNF. Within the study area, researchers captured 10 adult

lynx—“most individuals present”—and equipped them with GPS collars to gather hourly

location data. Id. at 244. The researchers used the location data to build a model that

predicted the relative probability that lynx used an area. They displayed the findings in

several maps, which were designed in part to guide forest management decisions about

timber salvage. Contrary to the researchers’ initial predictions, the study concluded that

lynx were not avoiding areas affected by spruce beetles, though areas with live trees

remained important. Id. at 251.

The Squires Study was the latest in a series of Canada lynx studies in Colorado.

Two others are relevant here.

the northern part of the forest in Squires Study or the habitat management direction in the new forest plan.” Id.

23 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 27

c. Theobald and Shenk Study

In 2011, David Theobald and Tanya Shenk released a report documenting “habitat

use for the 218 Canada lynx that were reintroduced to Colorado from 1999 to 2006.”

App., Vol. 4 at 209. Based on location data from a filtered set of 118 lynx, the

researchers produced distribution maps and models that determined where lynx were

most likely to be found and what habitat characteristics were associated with those

places. Id. at 210-27. The study concluded that its analysis was “consistent with

previous reports that the Canada lynx reintroduced to Colorado have primarily used high

elevation spruce-fir and aspen vegetation types as habitat.” Id. at 222.

d. Ivan Study

Jake Ivan and his coauthors—including Theobald and Shenk—built on the

Theobald and Shenk Study.13 Theobald and Shenk “provide[d] valuable information

regarding the types of areas that were known to be used by lynx from 1999 to 2010,” and

the Ivan Study “extend[ed] th[at] work . . . by producing a map of predicted lynx use on a

statewide scale.” Id. at 228. The model predicted lynx use based on habitat

characteristics and “[a]rbitrarily defin[ed] the top 20% of predictions as high quality lynx

13 Defenders argues the record version of the Ivan Study includes no date, but that the record also “makes clear that this study is from 2012.” Aplt. Br. at 37 n.10. Defenders also notes that the 2021 BiOp cited “‘Ivan et al. 2017,’” which Defenders says “seems to refer to the same (2012) study described here.” Id. (quoting App., Vol. 1 at 67, 69, 70). The Government argues the study is from 2011 but refers to the study as the “2012 Ivan study” to “avoid confusion.” Aplee. Br. at 39 n.13. The index to the administrative record filed in district court identified the study as “undated.” Dist. Ct. Doc. 16 at 8. We need not resolve the date of the study, which all parties agree was completed before the 2020 Squires Study.

24 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 28

habitat.” Id. at 233. The study acknowledged “shortcomings”—imprecise location data

for certain landscape features and a lack of “important vegetation data.” Id. at 233-34.

But “[d]espite these weaknesses, the predictive maps . . . ha[d] a distinct strength in that

they were constructed objectively from rigorous mathematical models based on empirical

data collected from wild lynx” and were “the first such maps for Colorado.” Id. at 234.

The maps showed a relatively high probability of observing lynx across Colorado,

including on the RGNF. See id. at 241-42.

a. Arbitrary and capricious challenge

Defenders argues the 2021 BiOp arbitrarily justified its “significant exclusion” of

the northern part of the RGNF from high-use classification based on “[i]nformal and

unexplained discussions.” Aplt. Br. at 33. Defenders has not met its burden.

i. Legal background

An agency “ha[s] discretion to rely on the reasonable opinions of its own qualified

experts,” but “[w]hen specialists express conflicting views,” we “carefully review[] the

record and satisfy[] [our]selves that the agency has made a reasoned decision.” Marsh v.

Or. Nat. Res. Council, 490 U.S. 360, 378 (1989). Even in the face of conflicting expert

views, an agency generally “has made a reasoned decision,” id., if other record evidence

supports the agency’s decision or if the record is consistent with the agency’s expert’s

assertions.14

14 See Colo. Wild v. USFS, 435 F.3d 1204, 1217 (10th Cir. 2006) (holding agency’s use of staff’s “personal observations” to conclude projects at issue “did not have 25 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 29

ii. Application

The record shows the agency made a reasoned decision to describe the northern

part of the RGNF as a “low-use” area.

1) Squires Study

In addition to the consultations involving Dr. Squires and other state and federal

government biologists, the habitat analysis in the Squires Study supported the “low-use”

description. The 2021 BiOp reasoned that the LAUs outside of the Squires Study area

(which include the northern RGNF LAUs) that are known to contain lynx habitat are

“unlikely to provide sufficient high quality habitat to support high numbers of lynx”

because “[m]apping of lynx habitat on the RGNF, based upon Squires et al. (2020),

indicate[s] . . . the habitat conditions they prefer [are] lacking” in these areas. App.,

Vol. 1 at 230; see also id. at 235 (stating the USFS used the 2020 published version of the

Squires Study and a 2018 interim report “to remap lynx habitat on the [RGNF],

identifying those areas where there is a high probability of lynx use and areas where lynx

are far less likely to use”); App., Vol. 3 at 37 (Plan explaining habitat characteristics of

significant environmental impacts” was permissible because “[t]he record fully demonstrate[d] the factors the [staff] considered . . . , the monitoring techniques they employed (personal observations, measurements, photo-point, etc.), and the resulting data”); Forest Guardians v. FWS, 611 F.3d 692, 709 (10th Cir. 2010) (upholding agency’s decision where petitioner questioned studies but did not “point to any [physical] evidence in the record indicating” the FWS’s conclusion was infirm); San Luis & Delta-Mendota Water Auth. v. Locke, 776 F.3d 971, 996 (9th Cir. 2014) (directing courts to “defer[] to the agency’s scientific conclusions when those conclusions are fairly traceable to the record”); Nw. Motorcycle Ass’n v. USDA, 18 F.3d 1468, 1475 (9th Cir. 1994) (upholding agency’s conclusion where personal “experiences of [USFS] personnel” were “not the only basis” for it).

26 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 30

high-use areas). The FWS thus “believe[d] lynx [are] unlikely to establish home ranges

in this area.” App., Vol. 1 at 230. The FWS reasonably read the Squires Study as

developing a set of preferential habitat qualities and determined these northern LAUs

lacked such qualities.

2) Administrative record

In determining whether the agency’s stated position is reasonable under the Marsh

and State Farm standards, we can and should look to the entire administrative record.

See, e.g., San Luis & Delta-Mendota Water Auth. v. Locke, 776 F.3d 971, 994, 996

(9th Cir. 2014). The administrative record is consistent with the FWS’s position.

a) Theobald and Shenk Study map

The 2021 BiOp’s description of the northern part of the RGNF as low use for lynx

is consistent with the Theobald and Shenk Study, as two maps show. The first map

appears in the 2021 BiOp. It depicts the low- and high-use area designations that the

USFS adopted in the BA and the Plan. The LAUs in the northern part of the RGNF are

considered low-use and labeled “Lynx Mapped Suitable Habitat.” App., Vol. 1 at 221;

see also id. at 230 (listing LAUs “outside of the 95 percent lynx use area . . . [that]

contain lynx habitat”).

27 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 31

2021 BiOp map showing low- and high-use lynx habitat determinations in the RGNF. App., Vol. 1 at 221 (2021 BiOp); id. at 74 (same map in USFS BA).

The second map is from the Theobald and Shenk Study, which tracked lynx across

most of Colorado. The color-coded areas on the map show intensity of lynx use—low

(yellow), moderate (orange), and high (blue). See App., Vol. 4 at 219. A faint outline of

the RGNF boundary is visible at the bottom right of the map. This map shows little or no

lynx use in the northern part of the RGNF from 1999-2010, which is consistent with the

FWS’s habitat determinations in the 2021 BiOp, as depicted in the first map.

28 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 32

Theobald and Shenk Study map with forest service administrative boundaries (including the RGNF) and lynx distribution (1999-2010), with low-intensity use shown in yellow, moderate in orange, and high in blue. App., Vol. 4 at 223.

29 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 33

A cropped version of the map showing only the RGNF makes this more clear:

Cropped version of Theobald and Shenk Study map from above.

Defenders argues the Government’s reference in its brief to the Theobald and

Shenk Study is an “improper post hoc” rationalization. Aplt. Reply Br. at 13. It is not.

“The basic rule here is clear: An agency must defend its actions based on the reasons it

gave when it acted.” Regents of the Univ. of Cal., 140 S. Ct. at 1909. “[The FWS]’s

stated rationale for its agency action,” Defs. of Wildlife v. Everson, 984 F.3d 918, 941

n.10 (10th Cir. 2020), was that the Squires Study researchers determined that areas

outside of their study area were not routinely used by lynx and that the Squires Study’s

habitat-preference analysis showed that habitat outside of the study area would not

support high lynx use. App., Vol. 1 at 226, 230; see also id. at 235. The Government’s

statement that the Theobald and Shenk Study “generally aligns with the Squires Study’s

delineation of the 95% high probability use area,” Aplee. Br. at 34, is not a new rationale.

It rebuts Defenders’ argument that the record is devoid of additional support for the

agency’s position, and “does not render the substance of the [Government]’s response a

post hoc rationalization.” Everson, 984 F.3d at 941 n.10.

30 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 34

b) Home ranges and linkage areas

The FWS’s “low-use” description is also consistent with record evidence about the

development of lynx home ranges and the importance of linkage areas—areas that

“provide[] landscape connectivity between blocks of lynx habitat . . . separated by

intervening areas of non-lynx habitat.” App., Vol. 4 at 186. The FWS reasoned that the

northern parts of the RGNF, while not high-use areas, could still be used for “exploratory

and dispersal movements,” App., Vol. 1 at 230—exactly the kind of activities that

characterize “linkage areas,” see, e.g., App., Vol. 3 at 202; App., Vol. 4 at 72.

Defenders argues the Ivan Study, which predicted habitat use across the RGNF,

and the Squires Study tracking data, which showed lynx traveling outside study area,

both conflict with describing the northern part of the RGNF as low-use habitat. Aplt. Br.

at 36. But this evidence, which shows some potential habitat or movement of lynx

beyond the Squires Study boundaries, reinforces the FWS’s conclusion that “even though

lynx habitat is present on the northern part of the [RGNF]” and lynx may travel there, the

northern part of the RGNF “likely does not provide high quality lynx habitat.” App.,

Vol. 1 at 226, 230. The nature of the habitat in the northern part of the RGNF, the FWS

observed, makes lynx unlikely to “establish [their] home ranges” there. Id. at 230;

see also id. at 236, 226. Defenders has not pointed to any data to show lynx home range

development in northern part of the RGNF. See Forest Guardians v. FWS, 611 F.3d 692,

709 (10th Cir. 2010) (deferring to the agency where petitioner advanced a contrary view

of the same science the agency reviewed).

31 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 35

Defenders has not identified record evidence that counters the Government’s

position, demonstrates the agency entirely failed to consider an important aspect of the

problem, or shows the FWS’s determination “is so implausible that it could not be

ascribed to a difference in view or the product of agency expertise.” State Farm,

463 U.S. at 43.

b. Best-available-science challenge

“The obvious purpose of the requirement that each agency ‘use the best scientific

and commercial data available’ is to ensure that the ESA not be implemented

haphazardly, on the basis of speculation or surmise.” Bennett v. Spear, 520 U.S. 154, 176

(1997) (quoting 16 U.S.C. § 1536(a)(2)).

This standard requires the agency to “seek out and consider all existing scientific

evidence relevant to the decision” and to not “ignore existing data.” Ecology Ctr., Inc. v.

USFS, 451 F.3d 1183, 1194 n.4 (10th Cir. 2006) (quotations omitted). We afford

especially strong deference to the agency’s choice of “which data are the most accurate,

reliable, and relevant, . . . but it still must [use] good science—that is reliable, peer-

reviewed, or otherwise complying with valid scientific methods.” Id. (quotations

omitted); San Juan Citizens All. v. Stiles, 654 F.3d 1038, 1045 (10th Cir. 2011) (“The

deference we give agency action is especially strong where the challenged decisions

involve technical or scientific matters within the agency’s area of expertise.” (quotations

omitted)); see also, e.g., Miccosukee Tribe of Indians of Fla. v. United States, 566 F.3d

32 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 36

1257, 1264-65 (11th Cir. 2009); Balt. Gas & Elec. Co. v. Nat. Res. Def. Council, Inc., 462

U.S. 87, 103 (1983).

An agency may “not disregard available scientific evidence that is in some way

better than the evidence it relies on.” Locke, 776 F.3d at 995 (alterations and quotations

omitted); accord Defs. of Wildlife v. U.S. Dep’t of the Interior, 931 F.3d 339, 346 (4th

Cir. 2019); see 16 U.S.C. § 1536(a)(2) (requiring use of the “best scientific and

commercial data available”). “Absent superior data, occasional imperfections do not

violate the ESA best available standard.” San Luis & Delta-Mendota Water Auth. v.

Jewell, 747 F.3d 581, 602 (9th Cir. 2014) (alterations and quotations omitted).

Further, the statute refers to the best “available” data. 16 U.S.C. § 1536(a)(2). An

agency does not need to “collect new data,” Ecology Ctr., 451 F.3d at 1194 n.4, or

conduct additional studies when a decision can be made based on existing evidence,

Sw. Ctr. for Biological Diversity v. Babbitt, 215 F.3d 58, 60 (D.C. Cir. 2000); accord

Heartwood, Inc. v. United States, 380 F.3d 428, 436 (8th Cir. 2004); Defs. of Wildlife, 931

F.3d at 345.

Defenders argues the Ivan Study expressly analyzed the northern part of the

RGNF. Aplt. Br. at 36. But that does not make the Ivan Study better. In Forest

Guardians, we said a litigant could not “rel[y] on the conclusion drawn from scholarly

articles . . . that [is] contrary to the conclusion drawn by the FWS” without “point[ing] to

any [contrary physical] evidence in the record.” 611 F.3d at 709. The Ivan Study does

not provide such data. It created a model based on the Theobald and Shenk Study data.

33 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 37

The Squires Study, by contrast, provided new tracking data (updated post-beetle-

epidemic) and new habitat modeling. The Squires Study did not track lynx outside the

study area, but the record shows that Dr. Squires and his colleagues considered that part

of the RGNF to be so low use as to be useless in “inform[ing] reliable modeling and

mapping products.” App., Vol. 1 at 226.15

The Ivan Study does not offer comparative advantages to the Squires Study. For

one, the Squires Study is peer-reviewed, and the Ivan Study is not. See Ecology Ctr., 451

F.3d at 1194 n.4. Although the FWS has cited the Ivan Study in some analyses,16 the

agency has not recognized it as high-quality science and has repeatedly said the study had

major shortcomings.

Defenders contends the “2017 SSA . . . used the Ivan study to create its habitat

map for the Colorado population unit.” Aplt. Br. at 36. But the 2017 SSA used the Ivan

Study only to inform its calculation of the total potential habitat in the Southern Rockies.

App., Vol. 6 at 169. It did not use the Ivan Study to determine high or low use of those

15 As the Government notes, “Dr. Ivan was a co-investigator of the later in time Squires study,” Aplee. Br. at 41 (citing App., Vol. 1 at 242), so it is arguably reasonable to conclude Dr. Ivan did not disagree with exclusion of the northern part of the RGNF for lack of lynx presence. 16 The Government argues the Ivan Study was cited in the “Biological Assessment that the [USFS] provided to [the] FWS,” along with several other papers on which Ivan was a coauthor. Aplee. Br. at 40. It was cited in the BA for the proposition that “hare densities in spruce-beetle killed forest on the RGNF are on average the highest in Colorado,” and that “[w]hether lynx can persist in such conditions has not been rigorously evaluated through science-based methods and was a primary research need identified on the Rio Grande National Forest.” App., Vol. 1 at 67; see also id. at 69-70.

34 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 38

areas, and it noted the study’s shortcomings. See id. One shortcoming was that the study

“[a]rbitrarily defin[ed] the top 20% of predictions as high quality lynx habitat,” App.,

Vol. 4 at 233, which excluded many of the “spruce-fir habitats” that are important to lynx,

App., Vol. 6 at 169. Similarly, the USFS 2013 Canada Lynx Conservation Assessment

and Strategy, which Defenders claims “attest[ed] to the importance of [the Ivan] stud[y],”

Aplt. Br. at 38, relied on the study only to say it “produc[ed] a statewide map of predicted

lynx use”—“the first such maps for Colorado”—and that it had several “shortcomings,”

App., Vol. 5 at 57-58.

The 2021 BiOp’s treatment of the Ivan Study is consistent with the study’s self-

imposed limitations. The Ivan Study stated it should be used only as a “compliment [sic]

to expert opinion and existing maps produced by other means.” App., Vol. 4 at 234. The

Squires Study, by contrast, stated without limitation that its “goal was to provide forest

managers with conservation insights and tools that distinguish forest stands essential to

the conservation of Canada lynx from those stands that were less important and therefore

available for timber salvage with little impact to this federally-listed species.” App.,

Vol. 1 at 243; see also id. at 251.

The FWS did not err in failing to rely on the Ivan Study for its analysis of lynx use

in the northern part of the RGNF.

C. The 2021 BiOp’s Assessment of the Plan’s Impact on Lynx in Low-Use Areas

Defenders argues the FWS 2021 BiOp “failed to adequately analyze the Plan

decision to abandon SRLA protections against logging for . . . ‘low-use’ habitat,” Aplt.

Br. at 19, including failing to analyze the effects of changes between the SRLA and the

35 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 39

Plan, id. at 41. The FWS’s only role was to analyze in the 2021 BiOp whether changes in

management standards would jeopardize the lynx DPS. It did so in a reasonable manner.

The following information provides relevant background for this issue (impacts in

low-use areas) and the next issue (impacts in high-use areas).

a. SRLA management standards

As noted above, in 2008 the SRLA established goals, objectives, standards, and

guidelines for protecting against several risk factors to the lynx. App., Vol. 4 at 139-41.

Four of the “standards”—“management requirements designed to meet the objectives”—

are relevant here: three “VEG” standards, which set requirements for vegetation

management, and one “ALL” standard, which set requirements for habitat connectivity.

Id. at 44-45, 141.

VEG S1 provided that if more than 30 percent of lynx habitat in an LAU was in

an “unsuitable condition,” certain activities were prohibited in that LAU. Id. at 143. An

“unsuitable condition” meant “in a stand initiation structural stage [(‘SISS’)] . . . not yet

tall enough to provide winter snowshoe hare habitat.” Id. An SISS “generally develops

after a . . . disturbance by fire, insects[,] or regeneration timber harvest,” and is

characterized by a “new single-story layer of shrubs, tree seedlings, and saplings” that

“reoccupy[] the site.” Id. at 188.

Prohibited activities in the affected LAUs included “vegetation management

projects,” id. at 143, e.g., “prescribed fire or timber harvest,” id. at 188-89, that “move

additional acres into a stand initiation stage,” id. at 143, or that “regenerate forested

36 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 40

stands,” id. at 176 (footnote omitted), by “cutting [including clearcutting] trees and

creating an entire new age class,” id. at 187. See also id. at 177 (prohibiting habitat

“regenerat[ion] by vegetation management projects”).

VEG S2 prohibited “[t]imber management projects” from “regenerat[ing] more

than 15 percent of lynx habitat . . . within an LAU in a ten-year period,” id. (footnotes

omitted), to “limit the rate of management-induced change in lynx habitat,” id. at 143.

This standard also provided that salvage harvest—“commercial timber sale of dead,

damaged, or dying trees,” id. at 188—“d[id] not add to the 15 percent, unless the harvest

treatment would cause the lynx habitat to change to an unsuitable condition.” Id. at 177.

VEG S6 was designed to protect the best lynx habitat. See id. at 145-46, 88. It

prohibited “[v]egetation management projects that reduce winter snowshoe hare habitat

in multi-story mature” forest stands, except near certain manmade structures, for

research, for incidental removal during salvage harvest, and where management was

designed to encourage “multi-story attributes.” Id. at 179 (footnotes omitted).

All three VEG standards included a limited exemption for fire management

actions in the wildland urban interface (“WUI”)—areas adjacent to at-risk communities.

Id. at 176-79, 189.

ALL S1 provided that “[n]ew or expanded permanent developments and

vegetation management projects must maintain habitat connectivity in an LAU and/or

linkage area.” Id. at 175 (footnotes omitted). As noted above, a “linkage area” “provides

landscape connectivity between blocks of lynx habitat . . . separated by intervening areas

of non-lynx habitat.” Id. at 186.

37 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 41

b. Plan management standards

The Plan made two relevant modifications to the SRLA.

First, it added a new VEG S7 standard. VEG S7 “applies to salvage harvest

activities conducted in conifer forests that have lynx habitat attributes, but no longer meet

the definition for standard VEG S6 due to tree mortality and associated forest structural

changes.” App., Vol. 3 at 37.17

Second, the Plan made VEG S1 and S2 applicable only to LAUs that overlap

“wholly or partially” with high-use areas—and thus removed them from low-use areas.

Id. at 39. The other SRLA standards continue to apply in low-use areas. Id.; App.,

Vol. 1 at 219.

a. The 2021 BiOp’s consideration of management standard changes for RGNF low-use areas

The FWS reasonably considered how changes in the management standards for

low-use areas would affect the lynx DPS. The 2021 BiOp acknowledged that removal of

VEG S1 and S2 in low-use areas could lessen protection for lynx. It said:

By removing these standards, forest management within these LAUs could result in SISS condition[s] that exceed 30 percent, or could result in more than 15 percent LAU conversion to SISS conditions within 10 years. In addition, removal of these standards could result in additional conversion to SISS conditions in LAUs that currently exceed these standards.

17 We will describe the specific VEG S7 standards for high-use areas below.

38 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 42

App., Vol. 1 at 230-31. The FWS then appropriately analyzed the effect of these changes

on lynx. It concluded the management standard changes for the low-use areas are not

likely to adversely affect the lynx because (1) as discussed previously, lynx do not often

use the low-use areas, see, e.g., id. at 231, 234-36; and (2) projects in these areas would

be small, id. at 231.

The FWS justified the latter prediction by “assum[ing]” that because “little forest

management occur[red] within the LAUs in the low lynx use area” between 2010 and

2020 and “the [RGNF] did not propose significant increased vegetation management on

this part of the forest,” “future vegetation management w[ould] be similar in scope and

scale to the preceding 10-year period.” Id. “Even if vegetation management exceeds

historic levels in the low use area[s],” the FWS concluded that “the consequences of such

actions” will not “appreciably reduce the likelihood of survival and recovery of lynx on

the RGNF.” Id. at 236. Defenders’ challenges to this conclusion are not persuasive.

First, Defenders argues the FWS could not justify its project-size prediction based

on 2010-2020 logging data because that data “predated peak tree mortality and the

increased incentive for salvage logging.” Aplt. Br. at 47. But as Defenders asserts, the

“spruce beetle outbreak swept through the [RGNF] in the years after the SRLA’s

adoption, peaking around 2014.” Id. at 13 (citing App., Vol. 5 at 210; App., Vol. 1 at 87).

And as the Government explains, “By 2017, overstory mortality in the [RGNF]’s

spruce-fir habitat had reached 100%.” Aplee. Br. at 6 (citing App., Vol. 1 at 66). So the

2010-2020 estimates overlap “peak tree mortality,” Aplt. Br. at 47, and they cover nearly

the entire beetle epidemic.

39 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 43

Second, Defenders argues logging data from the 2017 La Garita Hills Restoration

Project18 “undercut[s]” the FWS’s projection. Id. at 47-48. The record shows otherwise.

The project authorized approximately 94,000 acres of logging on the RGNF, including

38,000 acres of salvage logging. See App., Vol. 2 at 55. But as the Government explains,

Aplee. Br. at 50, the project minimized impacts to lynx habitat by leaving “[l]arge

percentages [of] the spruce-fir/spruce-mixed conifer vegetation zones in both the Carnero

and Four Mile LAUs [the two LAUs affected by the project] . . . unharvested (90 and

74 percent, respectively)” to “retain a mosaic of stand conditions across each LAU,”

App., Vol. 2 at 59; see also id. at 45 (identifying Carnero and Four Mile LAUs as the

only affected LAUs); id. at 64 (Record of Decision noting that “potentially not all the

acres proposed will be harvested in suitable lynx habitat” and that “[t]he best lynx/hare

habitat would be protected by [project design criteria], standards and guidelines, and

during unit layout”).

The project was designed to follow the SRLA, and its Record of Decision

explained that “[i]t was always intended that this project would be updated to be

consistent with the revised forest plan, as needed.” Id. at 54. A project that promised

future updates to comply with the Plan does not undermine the Plan. The FWS did not

18 The La Garita Hills Restoration Project proposed vegetation management projects across approximately 180,000 acres of the RGNF. App., Vol. 2 at 53. It was designed to “manag[e] vegetation and fuels to reduce risk . . . ; foster[] resilient and adaptive ecosystems and mitigate[e] wildfire risk; and address[] the widespread bark beetle epidemics considering human safety, recovery, and resiliency.” Id.; see also id. at 6.

40 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 44

“offer[] an explanation for its decision that runs counter to the evidence before the

agency.” State Farm, 463 U.S. at 43.

b. Defenders’ other arguments

First, Defenders asserts that the “FWS[’s] fail[ure] to acknowledge that the

removal of VEG S1 and VEG S2 protections for so-called ‘low-use’ areas represents a

fundamental shift in approach from the SRLA.” Aplt. Br. at 41; see also id. at 42-44. It

also disputes the Plan’s decision to use the ALL S1 standard in low-use areas. It argues

ALL S1 is “vague and difficult to enforce,” is “not as protective as VEG S1 and S2,” and

improperly puts off jeopardy analysis until the project level. Id. at 45. We agree with the

Government that these arguments are challenges to the USFS Plan and not to the FWS

2021 BiOp. Aplee. Br. at 45.

Second, Defenders also seems to challenge the FWS’s analysis of landscape

connectivity in low-use areas as arbitrary, arguing it deferred jeopardy analysis to specific

projects. Aplt. Br. at 45-46. We disagree. The 2021 BiOp stated that because ALL S1

requires maintenance of habitat connectivity and because of the low risk of “genetic

isolation,” the FWS “anticipate[d]” that the Plan would not likely “result in adverse

impacts to lynx connectivity.” App., Vol. 1 at 234. It thus did not defer analysis.

The FWS added that it would determine “site-specific effects of projects proposed

under the revised Forest Plan that could influence connectivity” later, “during project-

specific consultation.” Id. It was reasonable for the FWS to anticipate that development

of additional safeguards would be needed at a later stage, especially when nothing in the

record shows the FWS could have made a site-specific determination when evaluating the

41 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 45

Plan. 50 C.F.R. § 402.14(a) (requiring consideration of impacts at “the earliest possible

time”); see N.M. ex rel. Richardson v. Bureau of Land Mgmt., 565 F.3d 683, 717-19 (10th

Cir. 2009) (holding that NEPA’s direction to consider impacts at the earliest practicable

point required an analysis of site-specific impacts where “concrete [site-specific] plans”

were available at the time of the programmatic decision).

The FWS reasonably analyzed the Plan’s impacts on lynx in low-use areas.

D. The 2021 BiOp’s Assessment of the Plan’s Impact on Lynx in High-Use Areas

Defenders argues the FWS allowed an arbitrary amount of salvage logging in

high-use areas under VEG S7, failed to adequately analyze the effects of that logging

allowance or the hazard tree exemption to VEG S7, and failed to adequately analyze the

effects of VEG S7’s logging permissions outside the USFS’s suitable timber base. Aplt.

Br. at 19, 49-55. The first contention challenges the USFS Plan, not the FWS 2021 BiOp.

As to the latter challenges, the FWS did not act arbitrarily or capriciously in evaluating

the impact of the USFS’s changes in management standards on lynx in high-use areas.

The USFS’s Design of the VEG S7 Standard

Defenders argues that “[t]he record provides no scientific basis for the 7%

standard,” id. at 49, and that the FWS should have considered that VEG S7 “opens a far

bigger percentage of the best habitat to logging than [VEG S6] did,” id. at 52. It says the

“FWS failed to provide any rationale for choosing this threshold.” Id. at 49.

But the FWS did not choose the threshold, the USFS did. Defenders’ petition for

review did not challenge the USFS’s design of the VEG S7 standard. See App., Vol. 1

at 34-37. Defenders argues it can still challenge the 7-percent allowance because without

42 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 46

a reasoned standard, the “FWS would lack [a] scientific basis for its effects analysis.”

Aplt. Reply Br. at 22. But it cites no authority, and we fail to see any merit to the

argument.

The FWS’s Analysis of VEG S7’s Impact in High-Use Areas

For high-use areas, VEG S7 creates management rules for three types of habitat.

First, within the suitable timber base and in LAUs that do not exceed 30 percent SISS, it

permits salvage logging in up to 7 percent of the overlapping high-use and suitable timber

base areas, with exemptions for hazard and WUI tree removal. App., Vol. 1 at 229-32.

Second, within the USFS’s suitable timber base and in LAUs that exceed 30 percent

SISS, VEG S7 limits vegetation management to “WUI actions exempted by the SRLA

and the [VEG S7] proposed hazard tree exemption.” Id. at 231; see also id. at 229-30.

And third, outside the suitable timber base, it permits vegetation management to “proceed

consistent with the SRLA.” Id. at 232; see also id. at 229. Defenders asserts the FWS

acted arbitrarily in its assessment of each category’s impact. We disagree.

a. Salvage logging allowance

Defenders argues the “FWS failed to analyze the effects of the arbitrary 7%

standard on the lynx.” Aplt. Br. at 51. The FWS calculated the maximum amount of

salvage logging under the 7-percent allowance to be 4,265 acres. App., Vol. 1 at 231-32.

The 2021 BiOp explained that the Plan as a whole—the three VEG S7 provisions for

high-use areas and logging allowances in low-use areas—could lead to “approximately

50,000 acres” of vegetation management over the 15-year life of the Plan, id. at 236,

including the 4,265 acres from the 7-percent allowance. See id. at 232. The FWS did not

43 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 47

discern an overall adverse potential impact on lynx in high-use areas because (a) VEG S7

also prohibits any salvage logging in LAUs with greater than 30 percent SISS, allowing

those LAUs to “recover from the beetle epidemic” and “support continual lynx presence

within the action area,” and (b) 50,000 acres is not a significant percentage of the DPS’s

total habitat. Id. at 236-37. We fail to see how the FWS ignored the impact of the 7-

percent standard on the lynx DPS.

The FWS also explained that VEG S7 is more protective of lynx because it fills a

gap in salvage logging protections under the SRLA. Because VEG S7 avoided

“unlimited salvage potential” possible under the SRLA, the 2021 BiOp concluded that

the standard was “likely to reduce effects to lynx on the RGNF compared to the SRLA

alone.” Id. at 233. Defenders argues the FWS did not connect this statement to how the

standard would “affect the survival and recovery of the” lynx DPS and that “any limiting

standard” compared to the SRLA would not be the appropriate test because conditions in

the RGNF had changed. Aplt. Br. at 51. We disagree with Defenders’ description of the

FWS’s conclusion. The FWS concluded that even with VEG S7’s logging allowance—

which provided more protection than the SRLA—the Plan still would not cause jeopardy

for the reasons stated above.

b. Hazard tree removal exemption

Defenders argues the “FWS did not examine the effects of” the hazard tree

removal exemption (applicable in suitable timber base areas), nor did the agency justify

“its apparent conclusion that the uptick in logging under these [exemptions] will not

44 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 48

adversely affect lynx.”19 Id. at 53-54. But the FWS did not fail to consider the impact of

the hazard tree exemption. It analyzed the likely use of the exemption in detail,

concluded that adverse impacts were possible, stated it lacked the data to determine the

impact conclusively, and then set up future safeguards. That is hardly an APA violation.

The FWS acknowledged that (1) the potential acreage open to logging via the

exemption was up to 17,896 acres, (2) there might be “adverse consequences to lynx

where hazard tree treatments occur in those LAUs where existing conditions exceed

VEG S1 (30 percent SISS),” App., Vol. 1 at 232, and (3) there might be “additional

[habitat] fragmentation,” id. at 233. But the FWS had difficulty “quantify[ing]” the

effects on lynx “at this programmatic scale” because not all trees in open acres would be

logged and there was “uncertainty [in] the timing, location, size, and extent of future

actions.” Id. at 234; see also id. at 233. The timing of future projects would be

particularly important because whether the 7-percent salvage allowance applies depends

on the “habitat conditions within the LAU at the time” of the project—which varies based

on human-caused, as well as natural disturbances. Id. at 232. To account for this

uncertainty, the 2021 BiOp directed that “all subsequent actions that affect lynx are

subject to future section 7 analysis and consultation requirements.” Id. at 234.

19 Defenders uses the plural “exemptions” but seems to challenge only the hazard tree exemption and not the WUI exemption. Aplt. Br. at 53. The 2021 BiOp stated it calculated the acreage potentially affected by “[h]azard [t]ree [r]emoval r[oa]d & [s]ite [b]uffers,” App., Vol. 1 at 232, but this calculation appears to determine the impact of the hazard tree exemption alone, see id. at 230; App., Vol. 3 at 38, not multiple “exemptions for road buffers, ‘hazard trees,’ and the like,” as Defenders contends, Aplt. Br. at 53.

45 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 49

Acknowledging uncertainty does not make the FWS’s no-jeopardy conclusion

unreasonable. Swan View Coalition, Inc. v. Turner, 824 F. Supp. 923 (D. Mont. 1992), is

instructive. There, the court held the FWS did not need to determine whether “future

timber production equalling [sic] the levels proposed in the [land and resource

management plan’s] production objectives would jeopardize listed species.” Id. at 935.

It reasoned that the FWS must implement consultation at “later stages of development”

and “the safeguards already built into the plan guarantee[d] that such jeopardy would be

averted before such levels of production were ever reached.” Id. The same is true here.

Although an agency may not defer analysis to the project stage when analysis is

possible earlier, 50 C.F.R. § 402.14(a) (“Each Federal agency shall review its actions at

the earliest possible time to determine whether any action may affect listed species or

critical habitat.”), the FWS recognized that complete analysis is not possible at this time.

We owe deference to that careful consideration. See, e.g., Balt. Gas & Elec. Co., 462

U.S. at 103 (“[When] the [agency] is making predictions, within its area of special

expertise, at the frontiers of science,” we “must generally be at [our] most deferential.”);

Rural Cellular Ass’n v. FCC, 588 F.3d 1095, 1105 (D.C. Cir. 2009) (“In circumstances

involving agency predictions of uncertain future events, complete factual support in the

record for the [agency]’s judgment or prediction is not possible or required . . . . [W]e

require only that the agency acknowledge factual uncertainties and identify the

considerations it found persuasive.” (quotations and citations omitted)). The FWS’s

analysis of the hazard tree exemption was reasonable.

46 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 50

c. Analysis of areas outside the suitable timber base

Defenders also argues the FWS’s analysis of potential logging outside the suitable

timber base was arbitrary. Aplt. Br. at 52-53. The FWS reasoned that in high-use areas

outside the suitable timber base (and therefore not subject to any salvage-logging cap),

salvage harvest is likely to be minimal because data from 2005-2020 indicated that the

RGNF forest managers “usually focus[] salvage activity within the[] suitable timber

base.” App., Vol. 1 at 232. Defenders contends the 2005-2020 data are “not reliable”

because most of those years “predated the height of the beetle kill that made salvage

harvest a key [RGNF] objective.” Aplt. Br. at 53.

The FWS acted reasonably. The 2005-2020 range includes some years that

predated the beetle epidemic, but also includes some years during and post-peak-

epidemic. The record does not show significantly more logging in years after the height

of the beetle epidemic. See Suppl. App. at 2 (showing breakdown of 2005-2020 data by

year). Nor does it show that after the height of the beetle epidemic, a significant amount

of salvage harvest outside the suitable timber base was in high-use areas. Rather, of the

344,536 acres of high-use habitat, id. at 4, “a total of 118 acres have been salvage

harvested in areas not suitable for timber,” id. at 2. Further, this minimal logging took

place in “2007 (3 acres), 2008 (31 acres), and 2009 (84 acres)”—before or just as the

beetle epidemic was beginning but not after. Id.

In the 2021 BiOp, the FWS reasonably analyzed the Plan’s impacts on lynx in

high-use areas.

47 Appellate Case: 23-1093 Document: 010111013174 Date Filed: 03/11/2024 Page: 51

E. The USFS’s Reliance on the 2021 BiOp

Because the FWS did not violate the ESA or the APA, the USFS did not act

arbitrarily in relying on the FWS 2021 BiOp for its Plan.

III. CONCLUSION

We affirm the district court’s denial of the petition for review.