In recent years, BLM regularly has had to remove, or
“gather,” horses from public lands to keep the population
within the appropriate management levels and to avoid the
degradation of rangeland resources and the suffering of wild
horses due to a lack of water and forage. Although the Act
allows excess horses to be adopted or euthanized, 16 U.S.C.
§ 1333(b)(2), adoption demand is insufficient to achieve
population control, and Congress effectively has prohibited
BLM from euthanizing healthy animals by prohibiting the
authorization of funds to do so. In Def. of 
Animals, 751 F.3d
at 1060 
n.6. Instead, BLM keeps removed horses in long-
term holding facilities for the remainder of their natural lives.

   About a decade ago, to assist it in improving the wild-
horses program, BLM commissioned a report from the
National Academy of Sciences (“NAS Report”) “to provide
8       AM. WILD HORSE CAMPAIGN V. BERNHARDT

BLM with a scientific evaluation of the program’s pressing
challenges.” The report, issued in 2013, evaluated various
methods of population control. It recommended that BLM
consider “changes in expression of sexual and social
behavior” when deciding whether to sterilize horses and
noted that the “ideal method would not eliminate sexual
behavior or change social structure substantially.” The report
concluded that “the methods judged most promising for
application to free-ranging horses or burros are PZP vaccines,
GonaCon vaccine,1 and chemical vasectomy.” It noted,
however, that “testing [of chemical vasectomy] in captive
horses would be needed before widespread application in the
field.” As for gelding some males in a herd, the report
concluded that the effects on reproduction and behavior
“could not be predicted at the time [the] report was prepared.”
The report stated that it was not clear how gelding “would be
better than [surgical] vasectomy, which does not affect
testosterone or male-type behaviors,” but the report
recognized that there are some risks to herds associated with
surgical vasectomy as well.

    To gain a better understanding of gelding as a population-
control method and of its effects on herd behavior, BLM
began a five-year study in 2016 (“Gelding Study”). The
Gelding Study, conducted as part of a gather plan
implemented in Utah, will evaluate whether gelding is “an
effective approach to slowing the annual population growth
rate,” the effects “of maintaining a population of gelded
males on the behavior and spatial ecology of the overall
population,” and the “health and short-term survival” of the


    1
      PZP and GonaCon are immune-contraceptive vaccines. PZP
vaccines are administered to mares; GonaCon vaccines can be
administered to either male or female horses.
       AM. WILD HORSE CAMPAIGN V. BERNHARDT                 9

geldings. This is the first study that has examined
specifically “the effect of gelding on the behavior of free-
roaming wild horses” and the way in which geldings interact
with intact wild stallions. Data from the study will not be
analyzed until October 2020 at the earliest.

   B. The Present Controversy

    The Antelope and Triple B Complexes comprise about
2.8 million acres of public lands in northeastern Nevada and
are home to thousands of wild horses. More than a decade
ago, BLM established that those areas could sustain a total of
between 899 to 1,678 wild horses. In 2017, based on a
current inventory of the lands, BLM determined that the
Complexes contained an excess population of about 8,600
wild horses and that action was necessary to remove them.
Once BLM made that determination, the Act required BLM
to “immediately remove excess animals.” 16 U.S.C.
§ 1333(b)(2).

    BLM developed the Antelope and Triple B Complexes
Gather Plan (“Gather Plan”) to address the excess, while also
keeping its management activities to “the minimal feasible
level.” Under the Gather Plan, BLM will remove the excess
wild horses over a ten-year period. Horses will be gathered
in phases as necessary to achieve a core breeding population
at the low range of the appropriate management level. The
plan also calls for adjusting sex ratios and administering
fertility-control treatments to mares in order to slow
population growth rates and increase intervals between
gathers.

   In an effort to reduce the number of horses that need to be
removed permanently from public lands and kept in long-
10     AM. WILD HORSE CAMPAIGN V. BERNHARDT

term holding facilities, BLM will geld some male horses and
release them back onto the range “where they can engage in
free-roaming behaviors.” By doing so, BLM can reduce the
breeding population to the low end of the appropriate
management level, but keep the total population of horses at
mid-range. The primary purpose of the gelding component
is not to slow population growth, but to allow more horses to
remain free-roaming than otherwise would be possible.

    Gelding horses is a centuries-old practice. It has been
studied extensively and is a routine procedure. It also is one
of the few permanent fertility-control options, resulting in
fewer “handling occasions” compared to other impermanent
methods. The only unknown aspects of gelding—as
identified in the NAS Report and acknowledged by BLM in
the Gelding Study—concern how useful it proves to be as a
population-control method and how wild horses will respond
when geldings are released back to the range.

    In 2017, BLM issued a preliminary environmental
assessment of the Gather Plan. 40 C.F.R. § 1508.9. BLM
received almost 5,000 public comments. Public comments
notified BLM of the concerns of experts on wild-horse
behavior about the negative effects of gelding wild horses. In
the final environmental assessment, BLM responded to those
comments and explained that the experts’ opinions were
speculative because none of them had conducted a study on
the topic. BLM concluded that the Gather Plan would not
significantly affect the human environment and, thus, that
BLM was not required to prepare an environmental impact
statement (“EIS”). In December 2017, BLM issued a finding
of no significant impact.
       AM. WILD HORSE CAMPAIGN V. BERNHARDT                 11

    In this action, Plaintiffs challenge the “geld and release”
portion of the Gather Plan. Plaintiffs do not deny that there
are too many wild horses in the Complexes, that BLM is
required to address that overpopulation, that removal of some
horses from the range is appropriate, or that BLM established
appropriate management levels of wild horses for the
Complexes. Plaintiffs challenge only BLM’s choice to geld
and release some of the male horses that would otherwise be
permanently removed, arguing that the agency acted
arbitrarily and capriciously. The district court granted
summary judgment to Defendants on all claims. Plaintiffs
timely appeal.

                STANDARDS OF REVIEW

   We review de novo a district court’s grant of summary
judgment. McGinest v. GTE Serv. Corp., 
360 F.3d 1103
,
1112 (9th Cir. 2004).

    We review agency decisions that allegedly violate NEPA
under the Administrative Procedure Act, and we set aside
those decisions only if they are “arbitrary, capricious, an
abuse of discretion, or otherwise not in accordance with law.”
Native Ecosystems Council v. U.S. Forest Serv., 
428 F.3d
1233
, 1238 (9th Cir. 2005) (quoting 5 U.S.C. § 706(2)(A)).

    In reviewing an agency’s decision not to prepare an EIS,
we “employ an arbitrary and capricious standard that requires
us to determine whether the agency has taken a ‘hard look’ at
the consequences of its actions, based [its decision] on a
consideration of the relevant factors, and provided a
convincing statement of reasons to explain why a project’s
impacts are insignificant.” Envtl. Prot. Info. Ctr. v. U.S.
12      AM. WILD HORSE CAMPAIGN V. BERNHARDT

Forest Serv. (EPIC), 
451 F.3d 1005
, 1009 (9th Cir. 2006)
(alteration in original) (internal quotation marks omitted).

                       DISCUSSION

    Plaintiffs argue that BLM must prepare an EIS for the
Gather Plan because five of NEPA’s intensity factors
demonstrate that gelding and release would have significant
effects on the environment. Plaintiffs also argue that BLM
acted arbitrarily and capriciously because it authorized
gelding and release without considering all the relevant
evidence.

     A. Intensity Factors

    NEPA requires agencies to prepare an EIS for “major
Federal actions significantly affecting the quality of the
human environment.” 42 U.S.C. § 4332(c). To determine
whether an action will have significant effects, agencies may
prepare an environmental assessment that “[b]riefly
provide[s] sufficient evidence and analysis for determining
whether to prepare an [EIS].” 40 C.F.R. § 1508.9(a)(1). An
EIS is required if that process raises “substantial questions”
about whether an agency action will have a significant effect.
Blue Mountains Biodiversity Project v. Blackwood, 
161 F.3d
1208
, 1212 (9th Cir. 1998). If the agency determines that an
EIS is unnecessary, it issues a finding of no significant
impact, which “briefly present[s] the reasons why an action
. . . will not have a significant effect on the human
environment.” 40 C.F.R. § 1508.13.

    To determine whether an action “significantly” affects the
environment, agencies must consider both the “context” and
“intensity” of the possible effects. 40 C.F.R. § 1508.27.
        AM. WILD HORSE CAMPAIGN V. BERNHARDT                  13

“Intensity” refers to the “severity of impact,” and NEPA
regulations include ten intensity factors that agencies must
consider.
 Id. An action 
may be, but is not necessarily,
“significant” if any one of those factors is met. Ctr. for
Biological Diversity v. Nat’l Highway Traffic Safety Admin.,
538 F.3d 1172
, 1220 (9th Cir. 2008).

    Plaintiffs argue that five intensity factors demonstrate that
the Gather Plan may have a significant impact: (1) the Plan
has highly uncertain effects; (2) the Plan has highly
controversial effects; (3) the area has unique characteristics;
(4) the decision establishes a precedent; and (5) the decision
threatens a violation of the Wild Free-Roaming Horses and
Burros Act. 40 C.F.R. § 1508.27(b)(3), (4), (5), (6), & (10).
We disagree.

        1. Highly Uncertain Effects

    When the possible effects of an agency’s action are so
“highly uncertain” that they raise “substantial questions”
about whether the action will have a significant impact on the
environment, the agency must prepare an EIS. 40 C.F.R.
§ 1508.27(b)(5); Blue 
Mountains, 161 F.3d at 1212
. NEPA
“regulations do not anticipate the need for an EIS anytime
there is some uncertainty, but only if the effects of the project
are ‘highly’ uncertain.” Ctr. for Biological Diversity v.
Kempthorne, 
588 F.3d 701
, 712 (9th Cir. 2009) (internal
quotation marks omitted).

    BLM’s plan to geld and release male horses to the range
does not meet that threshold. Gelding horses is not a new
practice, and its effects are well understood.          The
environmental assessment thoroughly reviewed the research
on the surgical procedure, on the effects of gelding on
14     AM. WILD HORSE CAMPAIGN V. BERNHARDT

domesticated and semi-feral horses, on the effects of
castration on other species, and on the natural social behavior
of wild horses. In that discussion, BLM acknowledged that
wild horses “are rarely gelded and released back into the
wild, resulting in few studies that have investigated their
behavior in free-roaming populations.” Thus, BLM used the
existing research to predict that those effects likely would be
insignificant.

    Some “quotient of uncertainty . . . is always present when
making predictions about the natural world.”
 Id. And we
have upheld agency predictions despite some uncertainty.
 Id.
In Center 
for Biological Diversity, for example, we held that
the Fish and Wildlife Service made reasonable predictions,
based on prior data, about the effects of their regulation on
polar bears even though “the specter of climate change made
the Service’s prediction less certain than it would be
otherwise.”
 Id. That the 
agency did not have perfect
information and had to extrapolate did not make the possible
effects “highly uncertain” and did not require the preparation
of an EIS.
 Id.

    Similarly, in 
EPIC, 
451 F.3d 1005
, and Native
Ecosystems Council, 
428 F.3d 1233
, we upheld the
reasonableness of agency predictions in the face of some
uncertainty. In EPIC, the precise effects on spotted owls
could not be “accurately described without additional
information,” but we held that the agency permissibly
declined to prepare an EIS because the agency provided
reasons, based on the available research, for the assumptions
that it made and for the conclusion that it drew: the effects
likely would be 
minor. 451 F.3d at 1011 
(internal quotation
marks omitted). We also did not fault the agency for not
preparing an EIS in Native Ecosystems despite imperfect
       AM. WILD HORSE CAMPAIGN V. BERNHARDT                 15

information and some evidence that the project would have
negative effects on wildlife. We explained that “the presence
of some negative effects [does not] necessarily rise[] to the
level of demonstrating a significant effect on the
environment.” 428 F.3d at 1240
.

    The same general analysis applies here.            While
acknowledging that available research was not perfectly
analogous, BLM used the existing evidence to assess the level
of uncertainty and made “reasonable predictions on the basis
of prior data” to conclude that there would be no significant
environmental impact. Ctr. for Biological 
Diversity, 588 F.3d
at 712
. BLM did not have to conclude that its project would
have no effect, but only that there were not substantial
questions as to whether gelding and release would have a
significant effect on the environment. Although the
environmental assessment did not always make BLM’s
reasoning explicit, we will “uphold a decision of less than
ideal clarity if the agency’s path may reasonably be
discerned.” Nat’l Ass’n of Home Builders v. Defs. of
Wildlife, 
551 U.S. 644
, 658 (2007) (internal quotation marks
omitted).

    Exactly what role geldings will play within the wild-horse
population when they return to the range is unknown, but
BLM reasonably concluded that there was no reason to
expect any behavioral change in individual geldings to be
significant. BLM explained that, although geldings might
exhibit some behavioral differences, they will still be free to
roam unhindered and will have “a number of biological
impulses” compelling them to do so. Male horses take on a
number of different roles during their lives, so a change in a
gelding’s behavior post-castration would not necessarily have
a significant or adverse effect because the gelding could fit a
16      AM. WILD HORSE CAMPAIGN V. BERNHARDT

natural role other than reproducing. BLM further concluded
that the effects of gelding on individual horses would not be
more significant than the effects of the alternative—
permanently removing those horses from public lands.

     BLM likewise considered the effects on family structures
among wild horses and reasonably concluded that there
would be no significant effects. BLM cited a study finding
that the presence of geldings did not disrupt relationships
between mares and foals. BLM further noted that the plan
will not geld a large enough number of males to substantially
reduce population growth rates. Because the number of
geldings will not be sufficient to affect reproductive patterns,
it is less likely that geldings would significantly affect other
family dynamics.

    BLM provided a scientific foundation for its assumptions
and predictions. We defer to an agency’s “scientific
prediction[s] within the scope of its technical expertise.” Ctr.
for Biological 
Diversity, 588 F.3d at 712
. Nothing in the
record, and none of Plaintiffs’ objections, undermines the
reasonableness of BLM’s conclusion.

    Nor do our prior cases require a different result. In Bark
v. United States Forest Service, 
958 F.3d 865
, 869–71 (9th
Cir. 2020), we held that the Forest Service’s finding of no
significant effects was arbitrary and capricious because the
plaintiffs “pointed to numerous expert sources,” including
expert studies and research directly contrary to the agency’s
conclusion, and the agency failed to engage with that body of
research. By contrast, Plaintiffs have not identified any
evidence affirmatively showing that returning geldings to the
range would affect herd behavior. And BLM engaged with
the currently available scientific evidence either by discussing
       AM. WILD HORSE CAMPAIGN V. BERNHARDT                 17

the studies expressly or by addressing the concerns that the
research raised. See In Def. of 
Animals, 751 F.3d at 1072
(explaining that an agency must address crucial factors, but
not necessarily specific evidence regarding that factor).

    Plaintiffs point to the NAS Report as evidence of
potentially significant effects. But that report acknowledged
the dearth of applicable scientific evidence and was
ultimately inconclusive. The report explained that “male-
type aggressive and sexual behaviors are usually reduced” in
geldings but that the response varies by individual horse. It
concluded that whether gelding and release would “increase
aggression and competition in herds or decrease it” was
uncertain.

    Only one study of domestic horses found increased
aggression in geldings. But, as BLM noted, wild male horses
naturally have varying levels of testosterone and behavior
depending on a variety of factors. Some change in behavior
among released geldings, then, would not necessarily be
disruptive. And a single study does not per se suffice to
demonstrate highly uncertain effects. See In Def. of 
Animals,
751 F.3d at 1071 
(finding that the effects of administering
PZP to horses were not “highly uncertain” despite two studies
discussing the possible negative effects of PZP on herd
behavior); Native 
Ecosystems, 428 F.3d at 1240 
(“[T]he
presence of some negative effects [does not] necessarily rise[]
to the level of demonstrating a significant effect on the
environment.”).

   Finally, there was no other available information that
BLM should, or could, have used to reduce the uncertainty
about the effects of gelding and release. Data from BLM’s
Gelding Study will not be analyzed until later this year. And
18     AM. WILD HORSE CAMPAIGN V. BERNHARDT

the mere fact that BLM chose to study the effects of gelding
and release is not evidence that the effects on herd behavior
are highly uncertain or might be significant. The study will
provide information not only about the effects on herd
behavior, but also about the effectiveness of gelding as a
population control method. Because the Gather Plan’s use of
gelding was not primarily intended as a population control
tool, any uncertainty regarding gelding’s effect on population
growth is not significant. The Gelding Study reveals BLM’s
opinion that the effects on herd behavior are not fully
understood but does not suggest that BLM expects the effects
to be significant—or even that there will be an effect at all.

    For that reason, our decision in National Parks &
Conservation Ass’n v. Babbitt, 
241 F.3d 722 
(9th Cir. 2001),
abrogated on other grounds as recognized by Monsanto Co.
v. Geertson Seed Farms, 
561 U.S. 139
, 157 (2010), is
inapposite. In that case, although the effects of the proposed
action were uncertain to some extent, the scientific evidence
revealed definite adverse effects.
 Id. at 732. 
The uncertainty
was over the intensity of the effects,
 id., and “[n]o 
new
scientific developments [were] required in order to obtain the
requisite information,”
 id. at 735. 
Under those circumstances,
the preparation of an EIS could help decision-makers better
understand the effects of the project.

    Similarly, in Ocean Advocates v. United States Army
Corps of Engineers, 
402 F.3d 846
, 867 (9th Cir. 2005), we
identified evidence in the record that the agency’s action
would have an “unquestionably severe” effect on the
environment. We noted that the agency failed to collect
available data, conduct projection analyses, or provide a
“justification regarding why more definitive information
could not be provided.”
 Id. at 870–71 
(internal quotation
       AM. WILD HORSE CAMPAIGN V. BERNHARDT                 19

marks omitted). Here, by contrast, the uncertainty is whether
gelding and release will have any effect at all on the behavior
of wild horses. BLM is not required to wait years before
taking action simply because an ongoing study might shine
light on an uncertainty that BLM reasonably predicts will be
minor or nonexistent.

    After a thorough review of the evidence that was
available, BLM made “reasonable predictions” about the
relevant area of uncertainty. Ctr. for Biological 
Diversity,
588 F.3d at 712
. Where there was no evidence that there
would be an effect, BLM was not required to conduct
additional research or wait for the results of the Gelding
Study before moving forward with its proposed action. The
NAS Report’s inconclusive statements and a single study
about domestic geldings does not undermine BLM’s
reasonable prediction that gelding and release will have
insignificant effects on herd behavior. BLM’s conclusion,
then, that the likelihood of significant effects of gelding was
not so uncertain as to require an EIS was not arbitrary or
capricious. 
EPIC, 451 F.3d at 1009
.

       2. Highly Controversial

    The effects of the Gather Plan are not “highly
controversial.” 40 C.F.R. § 1508.27(b)(4). The evidence that
Plaintiffs identify does not “cast[] serious doubt upon the
reasonableness of [the] agency’s conclusions.” Humane
Soc’y of U.S. v. Locke, 
626 F.3d 1040
, 1057 (9th Cir. 2010)
(internal quotation marks omitted). Plaintiffs have not
identified any evidence that contradicts BLM’s findings. The
NAS Report was inconclusive and reported no findings that
conflict directly with those in the environmental assessment.
The expert opinions that Plaintiffs cite were not based on
20     AM. WILD HORSE CAMPAIGN V. BERNHARDT

studies that those experts had conducted, and no existing
research supported their speculation. Mere opposition to an
action does not, by itself, create a controversy within the
meaning of NEPA regulations.
 Id. BLM considered 
and
addressed the existing literature in its environmental
assessment and provided reasoning for its conclusions. That
satisfies NEPA. See Native 
Ecosystems, 428 F.3d at 1240
(“Simply because a challenger can cherry pick information
and data out of the administrative record to support its
position does not mean that a project is highly controversial
. . . .”).

       3. Unique Characteristics

    BLM’s determination that the gather area is not in close
“proximity to historic or cultural resources” was not arbitrary
or capricious. 40 C.F.R. § 1508.27(b)(3). Wild horses are
not a cultural resource for purposes of NEPA. Congress,
through the Wild Free-Roaming Horses and Burros Act,
decided how wild horses should be managed and how the
effects of agency actions on those horses should be evaluated.
The Act states that wild horses will be considered “an integral
part of the natural system of the public lands,” 16 U.S.C.
§ 1331, and specifically instructs that they should be managed
“as components of the public lands” and as a part of a
“natural ecological balance.”