FILED
                                                                   United States Court of Appeals
                                       PUBLISH                             Tenth Circuit

                      UNITED STATES COURT OF APPEALS                      April 15, 2019

                                                                       Elisabeth A. Shumaker
                             FOR THE TENTH CIRCUIT                         Clerk of Court
                         _________________________________

 WILDEARTH GUARDIANS,

       Plaintiff - Appellant,

 v.
                                                             No. 17-1334
 TAMARA CONNER, in her official
 capacity as District Ranger, Leadville
 Ranger District, San Isabel National Forest,
 United States Forest Service; UNITED
 STATES FOREST SERVICE, a federal
 agency of the United States Department of
 Agriculture,

       Defendants - Appellees.
                      _________________________________

                     Appeal from the United States District Court
                             for the District of Colorado
                          (D.C. No. 1:15-CV-00858-CMA)
                       _________________________________

John R. Mellgren, (Peter M.K. Frost, with him on the briefs), Western Environmental
Law Center, Eugene, Oregon, for Plaintiff-Appellant.

Sommer H. Engels (Jeffrey H. Wood, Acting Assistant Attorney General, Eric Grant,
Deputy Assistant Attorney General, Andrew C. Mergen, Allen M. Brabender, and
Barclay T. Samford, Attorneys, United States Department of Justice, Environment &
Natural Resources Division, Washington, D.C., and Tyler Clarkson, Deputy General
Counsel, and Kenneth Capps, Office of General Counsel, United States Department of
Agriculture, with her on the brief), United States Department of Justice, Environment &
Natural Resources Division, Washington, D.C for Defendants-Appellees.
                        _________________________________

Before HARTZ, SEYMOUR, and EID, Circuit Judges.
                  _________________________________
HARTZ, Circuit Judge.
                         _________________________________

       The dispute before us concerns the Tennessee Creek Project (the Project), an effort

of the United States Forest Service (the Service) in the San Isabel and White River

National Forests to protect the forest from insects, disease, and fire; improve wildlife

habitat; and maintain watershed conditions. In 2014 the Service published an

environmental assessment (EA) of the Project, followed by a Decision Notice (DN) and

Finding of No Significant Impact (FONSI). One of the EA’s many conclusions was that

the Project was unlikely to adversely affect Canada lynx, and the DN/FONSI declared

that the Project would not significantly impact the human environment.

       WildEarth Guardians sought review in the United States District Court for the

District of Colorado, arguing that the Service had violated the National Environmental

Policy Act (NEPA), 42 U.S.C. § 4321 et seq., by failing in its EA to adequately assess the

Project’s effects on lynx and by failing to prepare an environmental impact statement

(EIS). The district court upheld the agency action. Exercising jurisdiction under 28

U.S.C. § 1291, we affirm. The Service satisfied its NEPA obligations when it reasonably

concluded in its EA that under a worst-case scenario the lynx would not be adversely

affected by the Project and reasonably concluded that an EIS was not necessary.

       I.     BACKGROUND

              A.     NEPA framework

       We have called NEPA the “centerpiece of environmental regulation in the United

States.” New Mexico ex rel. Richardson v. Bureau of Land Mgmt., 

, 703



                                             2
(10th Cir. 2009). The statute’s “twin aims” are to ensure that agencies consider the

environmental effects of their actions and inform the public of having done so. Baltimore

Gas & Elec. Co. v. Nat. Res. Def. Council, 

, 97 (1983). It does not compel

“agencies to elevate environmental concerns over other appropriate considerations.” 

mandate is that agencies “pause before committing resources to a

project and consider the likely environmental impacts of the preferred course of action as

well as reasonable alternatives.” 

; see also Marsh v. Oregon

Nat. Res.’s Council, 

, 371 (1989). It “merely prohibits uninformed—rather

than unwise—agency action.” Robertson v. Methow Valley Citizens Council, 

, 351 (1989).

      If an agency is considering an action that might affect the environment, it must

follow a process prescribed by NEPA and its implementing regulations. See 42 U.S.C. §

4321 et seq.; 40 C.F.R. § 1500 et seq. First, “the agency must determine whether the

proposed action will significantly affect the environment.” Western Watersheds Project

v. Bureau of Land Mgmt., 

, 1269 (10th Cir. 2013) (emphasis added).

Unless the answer is “immediately apparent,” the agency must prepare an EA, which is

“a concise public document that briefly provides sufficient evidence and analysis for

determining the appropriate next step.” Western Watersheds 

(internal quotation marks omitted); see 40 C.F.R. § 1508.9. “If the EA concludes that the

proposed action will have no significant effect on the environment, the agency may issue

a [FONSI] and move forward with the proposed action.” Western Watersheds 

; see 40 C.F.R. §§ 1501.4(e), 1508.13. Otherwise, the agency must


                                            3
prepare an EIS—a more extensive analysis assessing all the predicted impacts on the

environment and comparing the proposed action to all reasonable alternatives. See

–04; 40 C.F.R. § 1502.14.

      Because NEPA provides no private cause of action, see Utah Envtl. Congress v.

Russell, 

, 823 (10th Cir. 2008), challenges to an EA or FONSI must be

brought under the Administrative Procedure Act (APA), which instructs us to review

whether an agency’s action was “arbitrary, capricious, an abuse of discretion, or

otherwise not in accordance with law,” 5 U.S.C. § 706(2)(A). See 

; Utah Shared Access All. v. United States Forest Serv., 

, 1213 (10th

Cir. 2002).

              B.     Factual Background

                     1.    The Canada lynx

       The Canada lynx is native to the snowy, high-altitude coniferous forests of

Colorado’s Southern Rockies. These mountains provide the conditions necessary for

lynx habitat: elevated forests dominated by spruce-fir, lodgepole pine, and aspen-conifer

mix, and populated by snowshoe hare for lynx to prey on. The U.S. Fish and Wildlife

Service (FWS) designated the lynx as a threatened species in 2000. See 65 Fed. Reg.

16052 (March 24, 2000). A year earlier the Colorado Division of Wildlife began

releasing lynx into the wild to augment the very small population. By 2007, it had

released 218 lynx.

       From 1998 to 2000, biologists from the Service, Bureau of Land Management,

National Park Service, and the FWS jointly compiled information on lynx in the


                                            4
contiguous United States, culminating in the 2000 publication of the Canada Lynx

Conservation Assessment and Strategy (LCAS). The LCAS instructed agencies to map

Lynx Analysis Units (LAUs), which are geographic areas approximating the size of the

home range of a female lynx to be used “to begin the analysis of potential direct and

indirect effects of projects or activities on individual lynx, and to monitor habitat

changes.” Bill Ruediger et al., USDA Forest Service, USDI Fish and Wildlife Service,

USDI Bureau of Land Management, and USDI National Park Service, Canada Lynx

Conservation Assessment and Strategy, 6-2 (2d ed. 2000). It also recommended various

conservation measures to protect lynx habitat on federal lands. For example, if an agency

intended to take management action within an LAU, the LCAS instructed it to map lynx

denning and foraging habitat and to ensure that at least 10% of the lynx habitat in the

LAU would remain denning habitat. See LCAS at 7-4. And because the lynx population

in the Southern Rockies is limited by the availability of snowshoe hare (the primary prey

for lynx), the LCAS also recommended various measures to maintain the horizontal

cover (e.g., shrubs, understory trees, and low limbs) necessary for snowshoe-hare habitat.

See LCAS at pp. 7-4–7-6. A month after the LCAS was published, regional managers of

the Service and the FWS in the Southern Rockies signed the Lynx Conservation

Agreement, committing themselves to consider the LCAS’s recommendations before

undertaking new actions in lynx habitat. The agreement was revised and extended in

2005, and amended in 2006.

       In 2008 the Service adopted the Southern Rockies Lynx Amendment (SRLA).

This document superseded the Lynx Conservation Agreement, and it amended the Land


                                              5
and Resource Management Plans of eight National Forests in the Southern Rockies. Its

purpose was to strike “a reasonable balance in providing for the conservation of lynx

habitat while also allowing appropriate levels of human uses to occur.” Aplt. App. at

227. The SRLA imposes seven standards on agencies, such as a standard stating that

“[t]imber management projects shall not regenerate[1] more than 15 percent of lynx

habitat on [National Forest System] lands within an LAU in a ten-year period.” Aplee.

App. at GA7 (footnotes omitted). The SRLA also includes nonmandatory guidelines that

recommend “actions that will normally be taken to meet [SRLA] objectives.” Aplt. App.

at 227.

          The standards and guidelines of the SRLA were adopted only after completion of

an EIS. The draft EIS for the SRLA (issued in 2004) and the supplemental draft EIS

(issued in 2006) received nearly 300 comments. The final EIS explored five different

alternatives, at least one of which—Alternative B—would have included greater

protection for lynx denning habitat than Alternative F, which the Forest Service

ultimately adopted. For one thing, Alternative B would have included a standard similar

to the one in the LCAS that 10% of lynx habitat in each LAU must be maintained as

denning habitat. The Service explained in its EIS that such a standard was probably

unnecessary because most LAUs already have between 20% and 40% denning habitat, in

which case the availability of denning habitat would not be a limiting factor for lynx.


1
  The SRLA defines regeneration harvesting as “[t]he cutting of trees and creating an
entire new age class; an even-age harvest. The major methods are clearcutting, seed tree,
shelterwood, and group selective cuts.” Aplt. App. at 237.


                                             6
Indeed, research after the study that led to the 10% standard in the LCAS concluded that

lynx use “a greater variety of habitat for denning” than previously thought. Aplt. App. at

232. The research showed that lynx den sites “are found in both mature and younger

forests that have a large amount of cover and downed, large woody debris. . . . [L]ynx

have used all kinds of deadfall for den sites, so it is likely almost any forest does supply

denning habitat. . . . The research does not indicate a certain minimum amount of denning

habitat is required for lynx.” Aplee. Fed. R. App. P. 28(j) Letter of 11/17/2018,

attachment at T01678.

       Yet the SRLA still protected denning habitat. Guideline VEG G11 states that “[i]f

denning habitat appears to be lacking in the LAU, then projects should be designed to

retain some coarse woody debris, piles, or residual trees to provide denning habitat in the

future.” Aplee. App. at GA9 (footnotes omitted). The SRLA also advises agencies to

protect certain types of vegetation that are beneficial to lynx and their prey, and to ensure

that lynx denning habitat is near snowshoe-hare winter habitat:

       Guideline VEG G1: Vegetation management projects should be planned to
       recruit a high density of conifers, hardwoods, and shrubs where such habitat
       is scarce or not available. Priority for treatment should be given to stem-
       exclusion, closed-canopy structural stage stands to enhance habitat
       conditions for lynx or their prey (e.g. mesic, monotypic lodgepole stands).
       Winter snowshoe hare habitat should be near denning habitat.

Aplee. App. at GA9 (footnotes omitted).

       The SRLA’s standards and guidelines also aim to protect the winter habitat of

snowshoe hare, which are the primary source of food for lynx. The SRLA describes

snowshoe-hare winter habitat as “places where young trees or shrubs grow densely—



                                              7
thousands of woody stems per acre—and tall enough to protrude above the snow during

winter, so snowshoe hare can browse on the bark and small twigs.” Aplt. App. at 239.

Several studies have identified spruce-fir stands as providing the highest snowshoe-hare

densities of forest types in the region, but snowshoe hare can also populate aspen and

lodgepole-pine stands, so long as the stands provide enough horizontal cover, which the

SRLA defines as “visual obscurity provided by vegetation that extends to the ground or

snow surface.” Aplt. App. at 655.

       Further, the SRLA Implementation Guide provides clarification and guidance on

how agencies should remap LAUs. An LAU should approximate the size of the home

range of a female lynx in the Southern Rockies (one study reported that this averaged

about 18,500 acres). It also must contain at least 6,400 acres of “primary vegetation,”

Aplee. App. at GA45, which typically consists of spruce-fir, Douglas-fir, aspen-mix, and

seral lodgepole-pine stands—forest types that can support lynx denning, rearing, and

foraging. See Aplee. App. at GA43–45.

                     2.     The Tennessee Creek Project

                            a.      Scope

       The Project is a response to a mountain-pine-beetle epidemic that impacted forest

stands on the White River and San Isabel National Forests, and an associated threat to

headwaters that serve communities along Colorado’s Front Range. It will be

implemented over 10 to 15 years “to create forest conditions that are more resilient to

outbreaks of insects, disease and wildfire; to improve habitat for threatened, endangered

and sensitive species and other important wildlife species; and to provide for sustainable


                                             8
watershed conditions.” Aplt. App. at 506. Planned action involves a mix of clearcutting,

thinning, and prescribed burns.

       The dominant forest type in the 16,450-acre Project area is lodgepole pine (11,096

acres), although there are also significant spruce-fir stands (2,177 acres) and aspen stands

(564 acres). The lodgepole pines are currently vulnerable to beetle infestations and the

spread of dwarf mistletoe. In 2014, 40% of the lodgepole-pine stands were already

infected by dwarf mistletoe. The spruce-fir and aspen stands currently have a low

incidence of insects and disease, but they will become more susceptible as they age.

                            b.     Lynx habitat in the project area

       Each of the three main forest types in the Project area—lodgepole-pine, spruce-fir,

and aspen—can provide lynx habitat (although lodgepole-pine stands are not considered

habitat once they develop into mature monocultures). About 9,480 acres of the Project

area are mapped lynx habitat falling within one of two LAUs—the Tennessee Pass LAU

and the Massive LAU. Most of the Project area also falls within the 67,500-acre

Tennessee Pass Linkage Area, an area connecting blocks of lynx habitat. During a 2013

two-month research project, three lynx were trapped and released within the spruce-fir

and lodgepole forests of the LAUs that overlap with the Project area.

       The Project will include conservation measures to protect this lynx habitat and to

follow the standards and guidelines of the SRLA:

    The Service will not treat spruce-fir stands (which often provide lynx denning

       habitat and high snowshoe-hare densities) in mapped lynx habitat, except that dead




                                             9
      trees will be removed from areas affected by beetles or disease (always leaving

      10% for denning habitat).

    The Service will exclude from treatment any tree stands with greater than 35%

      dense horizontal cover. Such stands provide lynx denning habitat and the highest-

      quality winter habitat for snowshoe hare.

    Most clearcutting will take place in climax lodgepole stands and in mature

      lodgepole monocultures, which are either low-quality habitat or uninhabitable for

      lynx and snowshoe hare.

    Any adjacent clear-cuts will be kept at least 200 feet apart to allow travel corridors

      for wildlife.

    Thinning will be “in a mosaic fashion that would mimic natural disturbances,”

      Aplt. App. at 587.

    The Service will track the acreage of lynx habitat treated by the Project and report

      that data to the FWS as required by the SRLA.

    In the event that mechanical trampling or other harvest or salvage activity

      damages lynx or snowshoe-hare winter habitat, the Service will track that damage

      and count it toward any limits imposed by the SRLA.

                           c.     The Project’s EA

      In 2013 the Service issued a draft EA, as well as a draft biological assessment

(BA) that primarily analyzed the Project’s effects on lynx. See 16 U.S.C. § 1536 (a)(2),

(c)(1) (providing that agencies should conduct biological assessments to determine if




                                            10
actions will adversely affect threatened or endangered species). In early 2014 the Service

released the final EA and final BA together with a draft DN/FONSI. The final EA

examined three alternatives for treating the Project area, including a no-action alternative.

The Service’s chosen alternative will involve 2,370 acres of clearcutting, 6,765 acres of

thinning, 345 acres of precommercial thinning, and 6,040 acres of prescribed burns (some

of which will overlap with the clearcutting and thinning), as well as the creation of about

21 miles of temporary roads. The EA also describes how much those treatments—spaced

out over 10 to 15 years—will impact each forest type. Of the 9,480 acres of lodgepole-

pine stands that are treatable (not limited by slope, accessibility, or other factors

preventing treatment), 6,765 will be targeted for thinning and 345 will be targeted for

precommercial thinning (a process of thinning stands that were clear-cut 20-30 years

earlier, so that growth can be concentrated on the more commercially valuable trees).

And the Service will use clearcutting and prescribed burns to create openings in

lodgepole-pine stands, but on no more than 25% of the 9,480 acres of treatable pine, and

with clear-cuts limited to irregularly shaped 40-acre patches. Those clear-cuts will

“essentially eliminate” the risk of beetle infestation in treated stands and will allow new

stands to regenerate “mistletoe free.” Aplt. App. at 555. As for the 455 acres of treatable

aspen stands, 115 acres will be treated through a combination of cutting and burning,

which will promote younger aspen that are less susceptible to various forest pathogens.

And, as mentioned above, the 2,177 acres of spruce-fir stands (which are prime lynx

habitat) will not be treated unless infested by insects or infected by disease.




                                              11
       Although the EA quantified the amount of each type of treatment, it did not

specify the treatment locations. Rather, the Service intends to identify 300 to 500 acres

for thinning and clearcutting each year over the next 10 to 15 years. It has asserted that

this flexible approach is necessary for reacting to on-the-ground conditions, such as a

beetle infestation or fire risk.

       The EA includes nine pages analyzing the proposed action’s possible effects on

lynx, as well as an appendix assessing its adherence to each SRLA objective, standard,

and guideline. Because the Service does not yet know precisely which of the 9,480 acres

of mapped lynx habitat will be treated, it took the conservative approach of assuming that

all lynx habitat in the Project area will be treated.

       In that worst-case scenario, the Service found, clearcutting would temporarily

convert 6% of lynx habitat in each LAU to nonhabitat—well below the SRLA’s 15%

limit on how much lynx habitat within an LAU may be regenerated in a 10-year period.

Further, the clear-cut area would become habitat in 15 to 30 years, when trees would

have grown enough to be available to snowshoe hare above the snow level, although

denning habitat that is clear-cut would take more than 150 years to regenerate fully. The

Service also found that thinning treatments could cause lynx habitat to be temporarily

degraded but not rendered unsuitable, and within 30 years those thinning treatments will

have created more snowshoe-hare winter habitat by allowing light to reach the forest

floor and nurture new vegetation to provide horizontal cover. And precommercial

thinning—for which targeted stands have been preidentified—would affect only 65 acres

of lynx habitat, or no more than 0.2% of lynx habitat in each LAU (far short of the


                                              12
SRLA’s 1% cap on how much lynx habitat within an LAU may be subject to

precommercial thinning). The Service concluded in its EA that even in this worst-case

scenario the Project would satisfy all SRLA standards and guidelines, and the FWS

concurred. Moreover, the actual effects will be less severe than those depicted in the

worst-case-scenario analysis because of the Project’s many measures to protect lynx

habitat discussed above. The EA determined that the proposed action was consistent with

the SRLA and that it “may affect, [but is] not likely to adversely affect,” lynx. Aplt. App.

at 595.

          After reviewing objections to the EA and the draft DN/FONSI, Leadville District

Ranger Tamara Conner issued a final (slightly revised) DN/FONSI in November 2014. It

announced that an EIS was not necessary and that the Service would proceed with the

EA’s proposed alternative.

                       3.     Procedural History

          WildEarth brought this action in the United States District Court for Colorado

against the Service and Tamara Conner (in her official capacity as the Leadville District

Ranger) for allegedly violating NEPA. WildEarth claimed that the Service’s EA and

FONSI were inadequate because they did not “disclose, analyze, and otherwise take a

hard look” at the Project’s environmental effects, particularly its effects on lynx. Aplt.

App. at 051. The district court granted judgment in favor of the agency.




                                              13
       II.    DISCUSSION

       On appeal WildEarth raises two claims: (1) the Service violated NEPA because

its EA did not adequately assess the Project’s effects on lynx; and (2) the Service violated

NEPA by not producing an EIS. We address each in turn.

              A.     Standard of Review

       WildEarth brings its challenges to the Service’s actions under the APA. Our task

is therefore to determine whether those actions were “arbitrary, capricious, an abuse of

discretion, or otherwise not in accordance with law.” 5 U.S.C. § 706(2)(a). See

(“[W]e review an agency’s NEPA compliance to see

whether it is arbitrary, capricious, an abuse of discretion, or otherwise not in accordance

with law.” (internal quotation marks omitted)). An agency’s decision under NEPA is

arbitrary and capricious if the agency “(1) entirely failed to consider an important aspect

of the problem, (2) offered an explanation for its decision that runs counter to the

evidence before the agency, or is so implausible that it could not be ascribed to a

difference in view or the product of agency expertise, (3) failed to base its decision on

consideration of the relevant factors, or (4) made a clear error of judgment.” 

marks omitted). We have described this standard of review as assessing

whether the agency took a “hard look” at the issue before it.2 See 

upon to review factual determinations made by an agency as part of its NEPA process,


2
  Of course, it would be improper to view “hard look” as a requirement going beyond the
APA standard of review. As the Supreme Court explained in FCC v. Fox Television
Stations, Inc., 

, 513–16 (2009), courts should not impose a heightened
standard of review not found in the text of the APA.

                                             14
short of a clear error of judgment we ask only whether the agency took a ‘hard look’ at

information relevant to the decision.” (further internal quotation marks omitted)).

       In accordance with this standard, we assess the adequacy of the Service’s action

by asking whether its method of analyzing environmental effects “had a rational basis

and took into consideration the relevant factors.” Utah Shared Access 

–13. The analysis need not have used the best possible methodology, nor do we

ask whether it could have discussed environmental impacts in more detail. See 

Instead, we review whether the agency’s decision was reasoned, and we defer

to the agency’s expertise and discretion. See 

see also Utah Envtl. Cong. v.

Russell, 

, 824 (10th Cir. 2008) (“Deference to the agency is especially

strong where the challenged decisions involve technical or scientific matters within the

agency’s area of expertise.” (internal quotation marks omitted)).

              B.     Environmental Assessment

       WildEarth argues that the EA was inadequate because it did not sufficiently

evaluate the Project’s effects on lynx. We disagree.

       Before even beginning work on the EA, the Service had devoted a great deal of

attention to the lynx, assessing what type of habitat was needed and what actions would

imperil them. The LCAS, published in 2000, presented a number of recommendations.

It was incorporated in the Lynx Conservation Agreement of regional managers of the

Service and the FWS in the Southern Rockies, and the agreement was updated in 2005

and 2006. Then, after considerable input from the public, the Service in 2008

promulgated the SRLA, which superseded the Agreement and amended Land and


                                            15
Resource Management Plans for eight national forests in the southern Rockies. Based on

an EIS, it contained a number of standards and guidelines. With this scientific

knowledge in hand, the Service could reasonably assess the maximum impact that the

Project could have on the lynx and conclude it was unlikely to adversely affect them. In

particular, it could conclude that the Project will not violate any SRLA standards even in

a worst-case scenario in which every acre of mapped lynx habitat in the Project area is

treated.

       WildEarth nonetheless argues that the Service needed to collect and disclose

several additional types of data. We address each of its arguments.

                     1.     Need to specify treatment locations

       First, WildEarth argues that the Service was obligated to specify the sizes,

locations, and treatment planned for each of the treatment units and the locations of the

21 miles of temporary road expected to be built. According to WildEarth, our decision in

Richardson holds that an environmental assessment must include such “site-specific”

detail about a project area so that a proper analysis can be performed. Aplt. Reply Br. at

9. Sometimes. But that depends on the circumstances.

       In Richardson the federal Bureau of Land Management (BLM) produced a draft

EIS proposing a land-use plan that would allow drilling on New Mexico’s Otero Mesa.

, 690. The plan limited drilling to within 492 feet of existing

roadways to protect desert grassland from habitat fragmentation. See 

Three

years later, the BLM issued a final EIS adopting a modified plan that did not limit drilling

to areas close to existing roadways. See 

Instead, this alternative opened most


                                            16
of the Mesa to drilling, so long as only 5% of the surface area of the Mesa was subject to

drilling at any one time. See 

major change to the plan, the BLM barely

updated the sections of the draft EIS assessing the effects on vegetation and wildlife. See

We held that the EIS was inadequate, reasoning that “the location of

development greatly influences the likelihood and extent of habitat preservation.

Disturbances on the same total surface acreage may produce wildly different impacts on

plants and wildlife depending on the amount of contiguous habitat between them.” 

We analogized the difference between the two plans to the difference between

“building a dirt road along the edge of an ecosystem” and “building a four-lane highway

straight down the middle.” 

According to WildEarth, the lesson of Richardson

is that the “‘location, not merely total surface disturbance, affects habitat fragmentation,’”

so the EA’s analysis must identify specific treatment locations and assess their

importance to lynx. Aplt. Reply Br. at 9 (quoting 

).

       But Richardson did not hold that an agency’s EA or EIS always must specify the

precise locations within a project area that will be affected. The problem in Richardson

was simply that there had been no environmental assessment of the ultimate plan. The

earlier assessment contemplated a significantly different project from what was later

selected. That is hardly the case here. The EA analyzed what could happen whatever

sites were eventually chosen for treatment by the Project, so long as the Project

restrictions were satisfied. The Service’s analysis accounted for the uncertainty about

treatment locations by evaluating the Project’s effects on lynx in a worst-case scenario in

which all the mapped lynx habitat in the Project area is treated, and by including


                                             17
conservation measures to protect high-quality lynx habitat, such as not treating healthy

spruce-fir stands or any stands with greater than 35% dense horizontal cover. Moreover,

the Service had a valid reason for not identifying specific treatment sites in its EA: it

intends to select treatment units based on changing on-the-ground conditions over the 10

to 15 years of the Project. NEPA leaves “substantial discretion to an agency to determine

how best to gather and assess information” about a project’s environmental impacts.

Biodiversity Conservation Alliance v. U.S. Forest Serv., 

, 1270 (10th Cir.

2014). The Service used that discretion reasonably, assessing the Project’s maximum

possible effect on lynx habitat while also conserving agency resources and retaining

flexibility to respond to changing conditions. See Utah Shared Access 

(“By conducting an EA, an agency considers environmental concerns yet

reserves its resources for instances where a full EIS is appropriate.” (internal quotation

marks omitted)). We note that the Service was not postponing the requisite

environmental analysis until it picks the specific sites for treatment under the Project;

rather, it was saying that such future analysis would be unnecessary because, in its expert

opinion, whatever sites it ultimately chooses (within the constraints imposed by the

Project), there would not be a negative impact on the lynx.3




3
  WildEarth quotes various comments from Forest Service personnel in its argument that
greater location specificity was required in the EA. We are not sure that the comments
support WildEarth’s argument. But in any event, unanimity within an agency is not
required, so long as the agency analysis is reasonable.

                                             18
                     2.     Disclosure of precommercial thinning

       Repeating the same error just discussed, WildEarth specifically argues that the

Service violated NEPA by not disclosing the locations of its preidentified precommercial

thinning units. But such disclosure was not material to determining whether the Project

would adversely affect the lynx. The EA specifies that precommercial thinning will

affect no more than 0.2% of lynx habitat in each LAU—far below the SRLA’s 1% cap on

how much lynx habitat within an LAU may be subject to precommercial thinning.

       Relatedly, WildEarth contends in its reply brief that the Service failed to adhere to

the SRLA’s VEG S5 standard governing precommercial thinning, because it did not

specify the amount of precommercial thinning that would affect snowshoe-hare winter

habitat. But this argument is waived because it was not raised in the opening brief. See

Silverton Snowmobile Club v. U.S. Forest Serv., 

, 783 (10th Cir. 2006)

(“The failure to raise an issue in an opening brief waives that issue.” (brackets and

internal quotation marks omitted)).

                     3.     Need to identify denning habitat

       WildEarth also argues that the Service could not truly understand the Project’s

impact on lynx without knowing how much affected habitat would be denning habitat. It

ignores, however, the expert opinion relied on by the Service that denning habitat is not a

constraint on the lynx in the Project area. In particular, a study conducted shortly before

issuance of the EIS prepared for the SRLA noted how adept lynx are in creating dens:

“[L]ynx have used all kinds of deadfall for den sites, so it is likely almost any forest does

supply denning habitat. . . . The research does not indicate a certain minimum amount of


                                             19
denning habitat is required for lynx.” Aplee. Fed. R. App. P. 28(j) Letter of 11/17/2018,

attachment at T01678. Lynx denning habitat is therefore not expected to be a “limiting

factor” for lynx in the Southern Rockies. Aplt. App. at 232 (SRLA). Moreover, to

further ensure that lynx in the Project area will not suffer from any shortage of denning

habitat, the Project will avoid treating healthy spruce-fir stands or any tree stands with

greater than 35% dense horizontal cover. In light of those conservation measures and the

low likelihood of LAUs having insufficient denning habitat to begin with, the Service did

not need to quantify denning habitat to conclude that the Project will not adversely affect

lynx. See Utah Shared Access 

–13 (the Service must use a

methodology with a “rational basis” but does not need to use the best possible

methodology or create the most detailed EA possible).

                     4.     Need to quantify winter habitat

       WildEarth contends that the Service should have quantified the amount of winter

lynx habitat that will be affected. But such habitat analysis in the EA was not necessary,

because the Service reasonably found that the Project will preserve existing high-quality

winter habitat, target stands that provide poor or no winter habitat, and even generate new

winter habitat in those treated areas. As discussed above, lynx winter habitat is closely

correlated with snowshoe-hare winter habitat, which exists in areas with horizontal cover

above the snowline. Such conditions tend to exist in healthy spruce-fir stands as well as

in other forest stands with greater than 35% dense horizontal cover—all of which the

Service will exclude from treatment. The Service will instead target stands considered

nonhabitat or poor habitat for lynx and snowshoe hare. Clearcutting will generally occur


                                             20
in climax lodgepole stands and in mature lodgepole monocultures, which provide only

“marginal to poor horizontal cover for snowshoe hare.” Aplt. App. at 588. The Service

therefore anticipates that this clearcutting will “not have an effect on snowshoe hare”

other than improving winter foraging conditions within 15 to 30 years by allowing new

horizontal cover to develop. Aplt. App. at 588. Likewise, thinning treatments will

generally be restricted to low-quality stands, and treatments of those stands will improve

snowshoe-hare winter habitat in the course of 15 to 30 years. Given this project design,

the Service concluded that the Project satisfied SRLA objectives, standards, and

guidelines designed to protect winter habitat.

       WildEarth argues that the Service’s approach to winter habitat is nonetheless

inadequate because it treats snowshoe-hare winter habitat as a proxy for lynx winter

habitat instead of accounting for the distinct needs of lynx in winter. To show that

snowshoe-hare winter habitat is not a proxy for lynx winter habitat, WildEarth points to a

single sentence in the third edition of the LCAS that states, “Winter habitat may be more

limiting for lynx.” Aplt. App. at 374. But that sentence is not conveying that lynx winter

habitat may be more limiting than snowshoe-hare winter habitat; it is conveying that lynx

winter habitat may be more limiting than lynx habitat in other seasons. Indeed, the

article cited by that sentence specifically ties lynx winter habitat to snowshoe-hare winter

habitat, explaining that lynx in winter are particularly dependent on the availability of

snowshoe hare. WildEarth has failed to point to any source contradicting the statement

in the EA that “[l]ynx winter habitat is not exactly the same as snowshoe hare winter

habitat though the two are very closely associated.” Aplt. App. 494. We are not


                                             21
persuaded that it was unreasonable for the EA to treat snowshoe-hare availability as the

key factor for lynx winter habitat.

                            5.        Need for baseline data

       Next, WildEarth makes two arguments that the EA was inadequate for failing to

include “baseline data” regarding lynx denning and winter habitat in the Project area.

Aplt. Br. at 25. First, relying on two Ninth Circuit cases, it contends that agencies have a

general duty to ascertain baseline data during the NEPA process to determine the likely

impacts of their actions. See N. Plains Res. Council, Inc. v. Surface Transp. Bd., 

(9th Cir. 2011) and Nat’l Parks & Conservation Ass’n v. Babbitt, 

(9th Cir. 2001), abrogated on other grounds by Monsanto Co. v. Geertson Seed Farms,

, 157 (2010). Those cases, however, are readily distinguishable. In

Northern Plains Resource Council there was no dispute that the EIS prepared for the

Surface Transportation Board failed to include studies necessary to determine the impact

of the planned railroad on, among other things, the pallid sturgeon (an endangered fish)

and the sage grouse; rather, what was planned was to study the effects as part of

mitigation measures during execution of the project. See N. Plains Res. Council, 

–85. The circuit court held that these delayed studies were not

“sufficient to meet the Board’s NEPA obligations to determine the projected extent of the

environmental harm to enumerated resources before a project is approved.” 

It pointed out that “an agency must support its conclusions with studies that the agency

deems reliable,” and “[s]uch analyses must occur before the proposed action is approved,

not afterward.” 

In the case before us, in contrast, the Service determined,


                                              22
based on data and studies it deemed reliable, that the Project would not have an adverse

impact on the lynx. Similarly, in National Parks the circuit court held that the EA was

inadequate because it described various environmental effects as “unknown” but

proposed only future research and monitoring to determine those 

–33. The court said, “That is precisely the information and understanding that is

required before a decision that may have a significant adverse impact on the environment

is made, and precisely why an EIS must be prepared in this case.” 

Again, that

is not the situation here. The Service reasonably determined that it had sufficient

information to conclude that the lynx would not be adversely affected by the Project.

       WildEarth’s second argument is that more baseline data about the Project area is

necessary to monitor (what WildEarth calls “ground-truth”) the Service’s commitment

not to treat areas of mapped lynx habitat with greater than 35% dense horizontal cover.

Aplt. Br. at 22. In essence, WildEarth is saying that it does not trust the Service to do

what it promises and needs additional information at this time so that it can later

investigate whether the Service has lived up to its commitments. WildEarth does not

provide any authority supporting its argument, and we are aware of none. We generally

presume that government agencies comply with the law and NEPA creates no exception

to this presumption. See Pit River Tribe v. U.S. Forest Service, 

, 1082 (9th

Cir. 2010) (“[W]e presume that agencies will follow the law.”); cf. Poe v. Gerstein, 

, 281–82 (1974) (it was appropriate for district court to issue a declaratory

judgment but not an injunction against the State because court would not assume that




                                             23
State would decline to acquiesce in the decision). We decline to require additional steps

to facilitate monitoring of whether the Service complies with what it proposes to do.

       In sum, the record shows that the Service made a reasoned evaluation of how the

Project will affect lynx. WildEarth contends that the Service needed to state in the EA

precisely where the Project would do what and then evaluate the specific effects of those

actions on the lynx. But the nature of the Project, which requires responding to

conditions on the ground as they develop over the course of 10 to 15 years, makes such

precision impracticable. And the Service’s long study of the lynx and the requirements

for its habitat enabled it to reasonably conclude that even in the worst-case scenario, the

Project would not adversely affect that animal.

              C.      The FONSI

       WildEarth’s second claim is that the Service erred by issuing a FONSI instead of

conducting an EIS. An agency may issue a FONSI only if, after reviewing the direct and

indirect effects of a proposed action, it concludes that the action “will not have a

significant effect on the human environment.” 40 C.F.R. § 1508.13; see also 40 C.F.R.

§ 1508.8. To determine whether the effects of a proposed action on the human

environment are significant enough to require an EIS instead of a FONSI, an agency must

consider the “context and intensity” of the action. 40 C.F.R. § 1508.27. Factors affecting

the “intensity” of an action include effects “that may be both beneficial and adverse,”

effects that are “individually insignificant but cumulatively significant,” effects on

“unique characteristics” of the project area such as “cultural resources” and “ecologically

critical areas,” the “degree to which the effects . . . are likely to be highly controversial,”


                                              24
the “degree to which the possible effects . . . are highly uncertain or involve unique and

unknown risks,” and the degree to which endangered and threatened species will be

affected. 40 C.F.R. § 1508.27(b)(1), (3)–(5), (7), (9). The obligation to conduct an EIS

can be triggered by an effect on one of those significance factors, but the simple existence

of an effect does not trigger that obligation—the “relevant analysis is the degree to which

the proposed action affects” a listed factor. Hillsdale Envtl. Loss Prevention, Inc. v. U.S.

Army Corps of Eng’rs, 

, 1180 (10th Cir. 2012) (emphasis added).

       WildEarth argues that, contrary to the Service’s DN/FONSI, the Project will have

a significant effect in several respects listed in § 1508.27(b). As in our review of an EA,

we apply the arbitrary-and-capricious standard in assessing an agency’s decision to issue

a DN/FONSI instead of preparing an EIS. See Utah Shared Access 

. We address each of WildEarth’s arguments about these significance factors and

conclude that the Service was not arbitrary or capricious in deciding that they do not

individually or cumulatively demonstrate the need for an EIS.

       First, WildEarth contends that the sheer size of the Project—over 2,000 acres of

clearcutting and 7,000 acres of thinning—bears on two significance factors. See 40

C.F.R. § 1508.27(b)(1) (agency should consider both “beneficial and adverse” impacts

and a “significant effect may exist even if the Federal agency believes that on balance the

effect will be beneficial”); 

(7) (“Significance exists if it is reasonable to

anticipate a cumulatively significant impact on the environment”). Citing Colorado

Environmental Coal. v. Dombeck, 

(10th Cir. 1999), WildEarth argues that

a project affecting this much acreage of National Forest requires an EIS. To be sure,


                                              25
Dombeck considered an EIS for a project involving significantly less acreage than is

involved here; it analyzed the impacts of expanding a ski resort by roughly 2,000 acres in

the White River National Forest. See 

But that opinion does not address

whether an EIS was required. The issue before us was whether the EIS was adequate.

The plaintiffs argued that the EIS was inadequate for several reasons, including that it did

not include sufficient data about how the expansion would affect lynx or adequate

analysis of the socioeconomic impact of 200,000 additional skier visits each year. See 

1176. We rejected those arguments. See 

Nothing in the opinion

suggests that a project of more than 2,000 acres necessarily requires an EIS, as WildEarth

seems to argue here. Size in itself does not establish significance. As the D.C. Circuit

stated in TOMAC v. Norton, 

, 862 (D.C. Cir. 2006):

       TOMAC offers no support for the proposition that an EIS is required when
       a project reaches a certain size. The relevant benchmark is whether the
       federal action “significantly affect[s] the quality of the human
       environment.” 42 U.S.C. § 4332(2)(C). Large federal projects may, on the
       average, be more likely to meet this threshold. But there is no categorical
       rule that sizable federal undertakings always have a significant effect on the
       quality of the human environment.

Context is an important consideration. See 40 C.F.R. § 1508.27(a). Here, suffice it to

say that treatment by the Project will encompass less than .1% of the Holy Cross Ranger

District and only 5.4% of the Leadville Ranger District (slightly more than 1% of the San

Isabel National Forest). Thus, we must turn to the other factors.

       WildEarth next argues that the Project’s direct and cumulative impacts on lynx

and lynx habitat will be significant because the Project will destroy some denning habitat

for 150 years, degrade other winter and denning habitat, degrade linkage area, and render


                                            26
some lynx habitat unsuitable for up to 25 years. See 40 C.F.R. § 1508.27(b)(1), (b)(7).

But, as explained in our treatment of the EA, the Service reasonably determined that this

worst-case scenario would not significantly hurt the lynx, and WildEarth ignores how the

Project’s priorities and restrictions will limit the impact on denning and winter habitat

and eventually produce some new habitat.

       WildEarth further argues that the Project “would be implemented in and near areas

with ‘unique characteristics,’ including in and near areas with proximity to ‘ecologically

critical areas’ and historic resources.” Aplt. Br. at 38 (quoting 40 C.F.R.

§ 1508.27(b)(3)). The “ecologically critical areas,” according to WildEarth, include lynx,

wolverine, and elk habitat, as well as federally designated wilderness and various trails.

And the “historic resources” that might be affected are six 10th Mountain Division huts.

But the Service concluded that the effects on the lynx, wolverine, and elk would not be

significant. For reasons discussed above, we cannot reject the Service’s conclusions

about the unlikelihood of a significant detrimental impact on lynx habitat, and WildEarth

gives us no reason to doubt the Service’s conclusions about the wolverine and elk.

Regarding impact on wilderness areas, the Project is only adjacent to (not overlapping

with) wilderness and roadless areas, so the Service concluded that the sole anticipated

impact would be that wilderness visitors would be subjected to a short-term increase in

noise and visual disturbances. And as for the 10th Mountain Division huts, the Service

explained in its FONSI that it had designed the Project to ensure that there would be no

direct effect—and only slight risk of indirect effect—on cultural resources. The Service

reported that it consulted with the Colorado Historic Preservation Office to confirm that


                                             27
any adverse effect on heritage resources was unlikely. WildEarth has not attempted to

rebut any portion of the Service’s analysis on these matters.

       WildEarth also argues that the Project’s effects on lynx are “highly controversial”

and “highly uncertain”—two other significance factors under § 1508.27(b)(4)–(5). Even

in the absence of substantial public opposition, an action may be “highly controversial” if

there is “a substantial dispute as to the size, nature, or effect of the action.” Middle Rio

Grande Conservancy Dist. v. Norton, 

, 1229 (10th Cir. 2002). According

to WildEarth, the Project’s effects are controversial and uncertain because the Service has

not specified the size or location of the Project’s treatment units or their effects on lynx

habitat, and the Service could resolve this controversy and reduce uncertainty simply by

agreeing to conduct an EIS addressing these issues. But given that the Service

reasonably concluded that the Project was unlikely to harm lynx regardless of treatment

locations, it could properly conclude that there was no legitimate controversy.

       The final significance factor that WildEarth directs us to is the “degree to which

the action may adversely affect an endangered or threatened species or its habitat,” 40

C.F.R. § 1508.27(b)(9). WildEarth points out that even under the Service’s conclusion

that the Project is unlikely to adversely affect lynx, there is a possibility of some effect on

lynx. So even if the possibility of adversely affecting lynx is not independently sufficient

to compel an EIS, it contributes to the need for an EIS arising from cumulative

significance factors. See 40 C.F.R. § 1508.27(b)(7) (one factor affecting an action’s

“intensity” is whether “the action is related to other actions with individually insignificant

but cumulatively significant impacts”). We therefore turn to the ultimate issue, whether


                                              28
the cumulation of the factors set forth in the regulations shows that the Service was

arbitrary and capricious in concluding that the Project “will not have a significant effect

on the human environment.” 40 C.F.R. § 1508.13. Given that WildEarth has not

challenged any of the reasoning of the Service supporting its rejection of any of the non-

lynx factors as significant, we conclude that the Service’s conclusion must stand.

WildEarth has utterly failed to show what could be accomplished through an EIS that

would be material to whether the Project should proceed as planned. We see no “clear

error of judgment” in the Service’s conclusion that, based on the relevant factors, the

effects of the Project were not significant enough to require an EIS. Utah Shared Access

(internal quotation marks omitted).

       III.   CONCLUSION

       We AFFIRM the district court’s order rejecting WildEarth’s objections to the

Project.




                                             29