Ctr. for Biological Diversity v. David Bernhardt

FOR PUBLICATION

UNITED STATES COURT OF APPEALS
FOR THE NINTH CIRCUIT

CENTER FOR BIOLOGICAL DIVERSITY; No. 18-73400
DEFENDERS OF WILDLIFE; FRIENDS
OF THE EARTH; GREENPEACE USA;
PACIFIC ENVIRONMENT, OPINION
Petitioners,

v.

DAVID BERNHARDT; BUREAU OF
OCEAN ENERGY MANAGEMENT;
UNITED STATES FISH AND WILDLIFE
SERVICE,
Respondents,

HILCORP ALASKA LLC,
Respondent-Intervenor.

On Petition for Review of an Order of the
Bureau of Land Management, Interior

Argued and Submitted November 5, 2019
Portland, Oregon

Filed December 7, 2020
2 CTR. FOR BIO. DIVERSITY V. ZINKE

Before: Richard A. Paez and Johnnie B. Rawlinson,
Circuit Judges, and Leslie E. Kobayashi, * District Judge.

Opinion by Judge Paez

SUMMARY **

Appellate Jurisdiction / Environmental Law

The panel granted in part, and denied in part, a petition
for review brought by plaintiff conservation groups
challenging the U.S. Department of Interior’s Bureau of
Ocean Energy Management (“BOEM”)’s approval of the
Liberty project – an offshore drilling and production facility
along the coast of Alaska in the Beaufort Sea; vacated
BOEM’s approval of the project; and remanded to the
agency for further proceedings.

The site of the Liberty project is governed by the Outer
Continental Shelf Lands Act (“OCSLA”). Before Hillcorp
Alaska, LLC could begin drilling, it had to obtain approval
of the Liberty project from BOEM. Three environmental
statutes and their concomitant regulations governed
BOEM’s approval: the National Environmental Policy Act
(“NEPA”); the Endangered Species Act (“ESA”); and the
Marine Mammal Protection Act of 1973. Relying on a
biological opinion prepared by the U.S. Fish and Wildlife

*
The Honorable Leslie E. Kobayashi, United States District Judge
for the District of Hawaii, sitting by designation.
**
This summary constitutes no part of the opinion of the court. It
has been prepared by court staff for the convenience of the reader.
 CTR. FOR BIO. DIVERSITY V. ZINKE 3

Service and BOEM’s environmental impact statement
(“EIS”), BOEM’s Regional Supervisor of Leasing and Plans
signed a record of decision approving the Liberty project.

The panel held that it had original jurisdiction over
plaintiff’s challenge to BOEM’s approval of the Liberty
project under the OCSLA’s 43 U.S.C. § 1349(c)(2), which
included plaintiff’s challenge to the EIS prepared under
NEPA and the biological opinion prepared under the ESA.
The panel held that it also had jurisdiction over plaintiff’s
claims that BOEM’s conditional approval of the Liberty
project violated the ESA. The panel further held that the two
statutes relevant to plaintiff’s Section 7 ESA claim – the
OCSLA and the ESA - had conflicting jurisdictional
provisions, and it would follow the more specific statute –
the OCSLA. The OCSLA bifurcated jurisdiction between
the courts of appeal and district courts. The panel concluded
that under the OCSLA, it had jurisdiction to review whether
BOEM’s approval violated the ESA.

The panel concluded that BOEM acted arbitrarily and
capriciously by failing to quantify the emissions resulting
from foreign oil consumption in its EIS as required by the
NEPA, or, at least, explaining thoroughly why it could not
do so and summarizing the research upon which it relied.
The panel also held that the Fish and Wildlife Service
violated the ESA by (1) relying upon uncertain, nonbinding
mitigation measures in reaching its no-adverse-effect
conclusion in its biological opinion, and (2) failing to
estimate the Liberty project’s amount of nonlethal take of
polar bears. Because the panel concluded that Fish and
Wildlife Service’s biological opinion was flawed and
unlawful, the panel further concluded that BOEM’s reliance
on the Fish and Wildlife Service’s opinion was arbitrary and
4 CTR. FOR BIO. DIVERSITY V. ZINKE

capricious. In all other respects, the panel denied the petition
for review.

COUNSEL

Rebecca Noblin (argued) and Jeremy C. Lieb, Earthjustice,
Anchorage, Alaska; Eric P. Jorgensen, Earthjustice, Juneau,
Alaska; Kristen Monsell and Emily Jeffers, Center for
Biological Diversity, Oakland, California; for Petitioners.

James A. Maysonett (argued), Attorney, Appellate Section;
Eric Grant, Deputy Assistant Attorney General; Jeffrey
Bossert Clark, Assistant Attorney General; Environment &
Natural Resources Division, United States Department of
Justice, Washington, D.C.; for Respondents.

Svend A. Brandt-Erichsen (argued) and Linda R. Larson,
Nossaman LLP, Seattle, Washington, for Respondent-
Intervenor.

OPINION

PAEZ, Circuit Judge:

Hilcorp Alaska, LLC, is an energy management
company seeking to produce crude oil from Foggy Island
Bay, along the coast of Alaska in the Beaufort Sea. To
extract the oil from under the Beaufort Sea, Hilcorp will
need to construct an offshore drilling and production facility.
The facility—referred to as “the Liberty project,” or “the
Liberty prospect”—will be the first oil development project
fully submerged in federal waters. Hilcorp estimates that the
site contains about 120 million barrels of recoverable oil,
 CTR. FOR BIO. DIVERSITY V. ZINKE 5

which it hopes to extract over the course of fifteen to twenty
years.

The site of the Liberty project is within the outer
Continental Shelf of the United States and thus governed by
the Outer Continental Shelf Lands Act (“OCSLA”), 1
43 U.S.C. § 1331 et seq. OCSLA allows the Department of
Interior—which houses the Bureau of Ocean Energy
Management (“BOEM”)—to oversee the mineral
exploration and development of the outer Continental Shelf. 2
Administering the use of the Shelf under OCSLA may
include leasing federal land for oil and gas production to
entities like Hilcorp. See 43 U.S.C. §§ 1344; 1331(c), (k)–
(m). OCSLA requires BOEM to manage the outer Shelf in
“a manner which considers [the] economic, social, and
environmental values” of the Shelf’s natural resources.
43 U.S.C. § 1344(a)(1).

Before Hilcorp can begin drilling, it must obtain
approval of the project from BOEM. Three environmental
statutes and their concomitant regulations govern BOEM’s
approval. First, approval of the Liberty project is considered
a “major Federal action” under the National Environment
Policy Act (“NEPA”), 42 U.S.C. § 4321 et seq. See
42 U.S.C. § 4332(C); 40 C.F.R. § 1508.18. NEPA requires
BOEM to draft an “environmental impact statement”

1
We include a glossary of acronyms in an Appendix attached to this
opinion.
2
The outer Continental Shelf includes “all submerged lands lying
seaward of state coastal waters (3 miles offshore) which are under U.S.
jurisdiction.” OCS Lands Act History, U.S. Department of the Interior,
Bureau of Ocean Energy Management, http://www.boem.gov/oil-gas-
energy/leasing/ocs-lands-act-history (last visited Aug. 19, 2020); see
also 43 U.S.C. § 1331(a).
6 CTR. FOR BIO. DIVERSITY V. ZINKE

(“EIS”) evaluating the environmental consequences of the
drilling and oil extraction. 42 U.S.C. § 4332(C). The EIS
must contain, among other things, a statement of purpose, a
description of the project, and a comparison of the Liberty
project with other reasonable alternatives for extracting oil.
Id.; 40 C.F.R. § 1502.12–1502.14. It must also include a “no
action” alternative, in which BOEM evaluates the relative
consequences of not approving any drilling in the Beaufort
Sea. 40 C.F.R. § 1502.14(c). This comparative analysis is
“the heart” of the EIS. Id. § 1502.14.

Second, the remarkable biodiversity of the drilling site
implicates the Endangered Species Act of 1973 (“ESA”),
16 U.S.C. § 1531 et seq. The Liberty project requires the
construction of an offshore gravel island, wells, a pipeline to
transport the oil, gravel pads to support the intersections
between pipes, ice pads, a hovercraft shelter, a small boat
dock, a gravel mine, and additional ice roads and crossings.
The gravel island’s proposed site is in the middle of “the
Boulder Patch,” an isolated area of boulders and cobbles that
supports the only high arctic kelp forest in the Alaskan
Arctic and produces unusual species diversity and biomass.
The Bay is home to a wealth of threatened and endangered
marine mammals, including polar bears, six species of
whales, three species of seals, sea lions, sea otters, and
Pacific walruses. Seabirds, numerous species of fish, and
larger mammals all frequent the shallow waters around the
Bay.

The ESA requires BOEM to ensure that its approval of
the project does not jeopardize an endangered or threatened
species or destroy or adversely modify the species’s habitat.
16 U.S.C. § 1536(a)(2). BOEM must consult with either the
U.S. Fish and Wildlife Service (“FWS”) or the National
Marine Fisheries Service (“NMFS”), depending on the
 CTR. FOR BIO. DIVERSITY V. ZINKE 7

species at risk, and then either FWS or NMFS must prepare
a biological opinion to determine whether the agency’s
proposed action will jeopardize a species. Id. § 1536(b)–(c).
If BOEM concludes that the proposed action will not
jeopardize a species or adversely modify its critical habitat—
but that the project will result in the “incidental take” of the
members of a species—FWS or NMFS must provide an
“incidental take statement” authorizing such takings. A
“take” occurs under the ESA when an animal is harassed,
harmed, pursued, hunted, shot, wounded, killed, trapped,
captured, or collected, or when anyone attempts to engage in
such conduct. 16 U.S.C. § 1532(19).

Third and finally, the proposed project must comply with
the Marine Mammal Protection Act of 1972 (“MMPA”),
16 U.S.C. § 1361 et seq. The MMPA is narrower but more
restrictive than the ESA. It broadly prohibits the take of any
marine mammal. 16 U.S.C. § 1371(a). Under the MMPA,
the Department of Interior may promulgate incidental take
regulations that allow an agency to take marine mammals
where such take is “in accord with sound principles of
resource protection and conservation” as provided in the
MMPA. Id. § 1371(a)(3)(A).

Relying on a biological opinion prepared by FWS and
BOEM’s own EIS, BOEM’s Regional Supervisor of Leasing
and Plans signed a record of decision approving the Liberty
project. The Center for Biological Diversity and four other
conservation organizations (collectively, “CBD”), dispute
the legality of BOEM’s and FWS’s actions, arguing that the
agencies failed to comply adequately with the procedural
requirements imposed by NEPA, the ESA, and the MMPA.
Specifically, CBD claims that (1) BOEM violated NEPA by
arbitrarily and capriciously estimating the environmental
consequences of the alternatives included in the EIS;
8 CTR. FOR BIO. DIVERSITY V. ZINKE

(2) FWS violated the ESA and MMPA by producing a
legally inadequate biological opinion; and (3) BOEM
violated the ESA by relying on FWS’s unlawful biological
opinion to approve the Liberty project. Hilcorp intervened
on behalf of BOEM. We agree in part with CBD and vacate
BOEM’s approval of the project.

I. Court of Appeals Review

A. Jurisdiction

We have original jurisdiction over CBD’s challenge to
BOEM’s approval of the Liberty project under 43 U.S.C.
§ 1349(c)(2) (“Any action of the Secretary to approve . . .
any development and production plan under this subchapter
shall be subject to judicial review only in a United States
court of appeals for a circuit in which an affected State is
located.”). This includes CBD’s challenge to the EIS
prepared under NEPA and the biological opinion prepared
by FWS under the ESA. See 16 U.S.C. § 1531; Am. Bird
Conservancy v. F.C.C., 545 F.3d 1190, 1191 (9th Cir. 2008).

We also have jurisdiction over CBD’s claims that
BOEM’s conditional approval of the Liberty project violated
the ESA. “[W]hen a Section 7 claim challenges an agency
order issued pursuant to a substantive statute with a ‘more
specific’ judicial review scheme than the ESA, courts must
evaluate the plaintiff’s claims under the jurisdictional
provisions of that substantive statute.” Ctr. for Bio.
Diversity v. E.P.A., 847 F.3d 1075, 1089 (9th Cir. 2017)
(quoting Am. Bird. Conservancy, 545 F.3d at 1194). When
two claims are “inextricably intertwined between two
statutes,” “and those statutes contain conflicting
jurisdictional provisions,” we follow the more specific
statute. Id.
 CTR. FOR BIO. DIVERSITY V. ZINKE 9

The two statutes relevant to CBD’s Section 7 ESA claim
are OCSLA and the ESA, and they have conflicting
jurisdictional provisions. OCSLA grants standing to “any
person” to “compel compliance” with the Act. 43 U.S.C.
§ 1349(a)(1). A court of appeals has original jurisdiction
under OCSLA to review the Secretary of the Interior’s action
where that action is “to approve, require modification of, or
disapprove . . . any development and production plan” under
the Act. Id. § 1349(c)(2). If the agency action does not
“approve, require modification of, or disapprove” any plan,
but still arises from (1) “any operation . . . which involves
. . . development” or (2) “the cancellation, suspension, or
termination of a lease or permit,” then federal district courts
have jurisdiction to review the agency action. Id.
§ 1349(b)(1).

The ESA instead allows a citizen to “commence a civil
suit on his own behalf . . . to enjoin any person, including the
United States and any other governmental instrumentality or
agency . . . , who is alleged to be in violation of any
provision of this chapter or regulation issued under the
authority thereof[.]” 16 U.S.C. § 1540(g)(1)(A). The ESA
citizen-suit provision also provides, “The district courts shall
have jurisdiction, without regard to the amount in
controversy or the citizenship of the parties, to enforce any
such provision or regulation, or to order the Secretary to
perform such act or duty[.]” Id. § 1540(g)(1).

OCSLA is the more specific jurisdictional statute. It
bifurcates jurisdiction between the courts of appeal and
district courts, and it refers specifically to BOEM’s
“approv[al]” of development plans, like the one at issue here.
Additionally, OCSLA and the ESA are “inextricably
intertwined”: BOEM’s lawful approval under OCSLA is
contingent on whether it properly complies with the ESA.
10 CTR. FOR BIO. DIVERSITY V. ZINKE

Under OCSLA, then, we have jurisdiction to review whether
BOEM’s approval of the Liberty project violated the ESA.
43 U.S.C. § 1349(c)(2).

B. Standard of Review

NEPA, the ESA, and the MMPA all lack independent
judicial review provisions. Claims arising under all three are
therefore reviewed under the Administrative Procedure Act
(“APA”), 5 U.S.C. § 551 et seq., which authorizes courts to
set aside agency actions, findings, and conclusions if they
are “arbitrary, capricious, an abuse of discretion, or
otherwise not in accordance with law,” 5 U.S.C.
§ 706(2)(A); see also Klamath-Siskiyou Wildlands Ctr. v.
Bureau of Land Mgmt., 387 F.3d 989, 992 (9th Cir. 2004).

In reviewing the adequacy of an EIS under NEPA, we
employ “a rule of reason” analysis to determine whether the
discussion of the environmental consequences included in
the EIS is sufficiently thorough. Kern v. U.S. Bureau of
Land Mgmt., 284 F.3d 1062, 1071 (9th Cir. 2002) (internal
quotation marks omitted). The rule of reason analysis
requires evaluating whether the agency took a sufficiently
“hard look” at probable consequences; it is “essentially the
same” as an abuse of discretion analysis. Id. at 1071–72
(internal quotation marks omitted).

II. NEPA

A. The EIS Process

We begin with CBD’s challenge to BOEM’s NEPA
compliance. NEPA “is our basic national charter for
protection of the environment.” 40 C.F.R. § 1500.1(a). The
statute provides environmental protection not by mandating
“particular results,” but by prescribing the process that an
 CTR. FOR BIO. DIVERSITY V. ZINKE 11

agency must follow to evaluate and approve an action that
will have environmental consequences. Robertson v.
Methow Valley Citizens Council, 490 U.S. 332, 350 (1989).

The EIS is the linchpin of NEPA’s procedural
requirements. An EIS must be prepared for any and all
“major Federal actions significantly affecting the quality of
the human environment.” 3 42 U.S.C. § 4332(C); see also
Ctr. for Bio. Diversity v. U.S. Dep’t of Interior, 623 F.3d
633, 642 (9th Cir. 2010).

The purpose of the EIS is twofold: first and foremost, it
is an action-forcing device, ensuring that the goals of NEPA
are infused into the government’s actions. 40 C.F.R.
§ 1502.1. 4 NEPA’s requirements “are to be strictly
interpreted ‘to the fullest extent possible’ in accord with the
policies embodied in the Act.” State of Cal. v. Block,
690 F.2d 753, 769 (9th Cir. 1982) (quoting 42 U.S.C.
§ 4332(1)). Second, the EIS provides important information

3
A “[m]ajor Federal action” includes an action with “effects that
may be major” and is “potentially subject to Federal control and
responsibility.” 40 C.F.R. § 1508.18. The “[a]pproval of specific
projects, such as construction or management activities located in a
defined geographic area,” may be major federal actions. Id.
§ 1508.18(b)(4). “Projects” can “include actions approved by permit or
other regulatory decision as well as federal and federally assisted
activities.” Id.
4
We rely on two sets of NEPA regulations. The NEPA regulations
promulgated by the Council on Environmental Quality (CEQ), codified
at 40 C.F.R. §§ 1500.1–1508.28, provide NEPA guidance to all federal
agencies. The Department of Interior, like many other agencies, has also
promulgated its own NEPA regulations, codified at 43 C.F.R. §§ 46.10–
46.450, to be used alongside the CEQ regulations.
12 CTR. FOR BIO. DIVERSITY V. ZINKE

to the public and any party interested in the proposed
environmental action. See Robertson, 490 U.S. at 356.

Agencies prepare EISs in two stages. See 40 C.F.R.
§ 1502.9(a). First, the agency creates a draft EIS. See
40 C.F.R. § 1502.9(b). The draft examines the scope of the
federal action, evaluates the consequences of the action, and
includes viable alternatives for the project. Id.; see also
42 U.S.C. § 4332(c). The agency has discretion to develop
the alternatives it considers, see Citizens Against Burlington,
Inc. v. Busey, 938 F.2d 190, 195 (D.C. Cir. 1991), but a “no
action alternative”—in which the agency evaluates the
consequences of taking no action—must be considered in
every EIS, to provide a baseline against which every action
alternative is evaluated, see 40 C.F.R. § 1502.14(d). The no-
action alternative analysis should be “[i]nformed and
meaningful,” Bob Marshall All. v. Hodel, 852 F.2d 1223,
1228 (9th Cir. 1988), and the agency must not minimize
negative side effects, N. Alaska Envtl. Ctr. v. Kempthorne,
457 F.3d 969, 975 (9th Cir. 2006).

The discussion of environmental consequences must be
“reasonably thorough.” Kern, 284 F.3d at 1071. NEPA
emphasizes the early presentation of relevant information to
facilitate reaching fully informed decisions. See Blue
Mountains Biodiversity Project v. Blackwood, 161 F.3d
1208, 1216 (9th Cir. 1998). Drafting an EIS “necessarily
involves some degree of forecasting,” and the agency “must
use its best efforts to find out all that it reasonably can” when
predicting the environmental effects of the proposed action.
City of Davis v. Coleman, 521 F.2d 661, 676 (9th Cir. 1975).

After completing the draft, the agency must then “[m]ake
diligent efforts to involve the public in preparing and
implementing [its] NEPA procedures,” including soliciting
public comments where appropriate. 40 C.F.R. § 1506.6(a);
 CTR. FOR BIO. DIVERSITY V. ZINKE 13

see also id. § 1506.6(b)–(f). The substantive comments
received by the agency, and the agency’s responses to them,
are attached to the final EIS. See Block, 690 F.2d at 772–73
(quoting 40 C.F.R. § 1500.10(a)).

* * *

CBD argues that BOEM’s EIS is arbitrary and capricious
under the APA because BOEM improperly (1) relied on
different methodologies in calculating the lifecycle
greenhouse gas emissions produced by the no-action
alternative and the other project alternatives, thus making the
options incomparable, and (2) failed to include a key
variable (foreign oil consumption) in its analysis of the no-
action alternative. We consider each in turn.

B. Comparison of the Action and No-Action
Alternatives

CBD argues that BOEM unlawfully used different
methodologies to calculate the greenhouse gas emissions
resulting from the Liberty project and the no-action
alternative. We disagree.

CBD is correct that using different methodologies to
capture the emissions resulting from each alternative would
indeed prevent the agency from making an “informed and
meaningful” choice, see Bob Marshall All., 852 F.2d
at 1228, because the alternatives would be incomparable.
But the record indicates that BOEM did not apply different
methods in comparing the action and the no-action
alternatives.

In its final EIS, BOEM considered various alternatives:
the Proposed Action (the Liberty project), other action
alternatives (each of which propose different strategies,
14 CTR. FOR BIO. DIVERSITY V. ZINKE

locations, or other modifications of the Proposed Action),
and the no-action alternative, in which BOEM analyzed the
effects of not leasing the land at all. To calculate the
emissions for each of the action alternatives, BOEM
calculated both the “upstream” and the “downstream”
emissions. Upstream emissions are those that result directly
from the project itself (e.g., construction and operation), and
downstream emissions are those that result from the
consumption of the oil produced by the project (e.g., heating
homes or fueling cars). BOEM then summed the two types
of emissions, resulting in a “lifecycle greenhouse gas
emissions” estimate for each alternative. To facilitate
comparisons across the action alternatives, the total lifecycle
emissions for each proposed plan were converted to metric
tons of “carbon dioxide equivalents”—even though
emissions would include methane, nitrous oxide, and other
greenhouse gases.

The lifecycle greenhouse gas emissions for the no-action
alternative were not calculated by directly summing its
upstream and downstream emissions. The upstream
emissions for the no-action alternative are, clearly, zero. The
direct downstream emissions of the no-action alternative are
zero, but—as BOEM recognized—its indirect downstream
emissions may be much higher. Not drilling at the proposed
site may cause global oil supply to fall, demand to rise, and,
as a result, require drilling and oil extraction elsewhere. To
capture these indirect downstream emissions, BOEM used a
market-simulation model to predict the greenhouse gas
emissions for energy sources that would substitute for the oil
not produced at Liberty.

CBD argues that the use of this model renders the
choices incomparable. But, as the government notes, the
Proposed Action and action alternatives implicitly take this
 CTR. FOR BIO. DIVERSITY V. ZINKE 15

analysis into account: if the Liberty site is developed, none
of the emissions in other parts of the United States estimated
under the no-action alternative will result. In other words,
BOEM could have instead used the market simulation model
to offset the emissions calculated under each of the action
alternatives and then compared it to zero, the lifecycle
emissions produced by the no-action alternative. Summing
all emissions from the proposed project assumes that, if
Liberty is developed, there would be no need for the other
sites to satisfy demand under the no-action alternative. The
total numbers would be different, but the absolute
differences between them would be the same. Both methods
of calculation result in net—not gross—emissions. The
analysis is ultimately a relative comparison, sufficient for
making a “reasoned choice among alternatives.” 40 C.F.R.
§ 1502.22(a). We conclude BOEM did not arbitrarily and
capriciously apply a different method of calculation in
estimating the emissions from the action and no-action
alternatives.

C. Omission of Emissions Resulting from Foreign Oil
Consumption

But CBD’s second argument is persuasive. CBD argues
that BOEM arbitrarily failed to include emissions estimates
resulting from foreign oil consumption in its analysis of the
no-action alternative. In its EIS, BOEM concluded that the
Proposed Action and the action alternatives would each
produce about 64,570,000 metric tons of carbon dioxide
equivalents. It then estimated that the no-action alternative
would produce—somewhat perplexingly—89,940,000
metric tons of carbon dioxide equivalents, 25,370,000 more
metric tons than if the land were leased under any scenario.
The EIS explains that the no-action alternative will result in
more emissions because the oil substituted for the oil not
16 CTR. FOR BIO. DIVERSITY V. ZINKE

produced at Liberty will come from places with
“comparatively weaker environmental protection standards
associated with exploration and development of the
imported product and increased emissions from
transportation.” CBD explains that BOEM reached this
counterintuitive result by omitting a key variable in its
analysis: foreign oil consumption.

Understanding why foreign oil consumption is critical to
BOEM’s alternatives analysis requires some basic
economics principles. If oil is produced from Liberty, the
total supply of oil in the world will rise. Increasing global
supply will reduce prices. Once prices drop, foreign
consumers will buy and consume more oil. The model used
by BOEM assumes that foreign oil consumption will remain
static, whether or not oil is produced at Liberty.

This omission, according to CBD, makes BOEM’s
analysis “misleading” because it fails to capture the
emissions caused by increased global consumption in its
estimate of Liberty’s downstream emissions. BOEM
acknowledges that the no-action alternative will cause
foreign oil consumption to decline; the EIS estimates that the
no-action alternative will result in a reduction in oil
consumption of one, four, or six billion barrels of oil,
depending on the market price of oil. But the impacts on
greenhouse gas resulting from such reductions in oil
consumption “are not captured” in the EIS because BOEM
determined it did not have sufficiently “reliable information
on foreign emissions factors and consumption patterns.”
CBD replies that BOEM was both required and able to
estimate the variable and include its effect. We agree.

NEPA requires agencies to evaluate the direct and
indirect effects of the proposed action. 40 C.F.R. § 1502.16.
Indirect and direct effects are both “caused by the action,”
 CTR. FOR BIO. DIVERSITY V. ZINKE 17

but direct effects occur “at the same time and place” as the
proposed project, while indirect effects occur “later in time
or [are] farther removed in distance.” 40 C.F.R. § 1508.8(a),
(b). The agency need consider only indirect effects that are
“reasonably foreseeable,” id. § 1508.8(b); or those that “a
person of ordinary prudence would take [] into account in
reaching a decision.” EarthReports, Inc. v. F.E.R.C.,
828 F.3d 949, 955 (D.C. Cir. 2016) (internal quotation marks
omitted); see also 40 C.F.R. § 1502.22(b). An increased risk
of an oil spill caused by an increase in crude oil tanker traffic,
for example, is a reasonably foreseeable indirect effect of a
proposed dock extension. See Ocean Advocates v. U.S.
Army Corps. of Eng’rs, 402 F.3d 846, 867–70 (9th Cir.
2005). “[G]rowth inducing effects” to a forest that result
from a project that alters “pattern[s] of land use” are also
indirect impacts that must be considered. 40 C.F.R.
§ 1508.8.

An EIS that does not adequately consider the indirect
effects of a proposed action violates NEPA. In Sierra Club
v. Federal Energy Regulatory Comm’n, 867 F.3d 1357 (D.C.
Cir. 2017), for example, the D.C. Circuit concluded that the
Federal Energy Regulatory Commission had unlawfully
conducted its EIS for a natural gas pipeline project because
it failed to quantify the indirect greenhouse gas emissions
that would result from the burning of the natural gas
transported by the pipelines. Id. at 1374. The agency should
have “either given a quantitative estimate of the downstream
greenhouse emissions,” or “explained more specifically why
it could not have done so.” Id. Greenhouse gas emissions
were an indirect, reasonably foreseeable consequence of the
pipeline, and FERC’s justification for its omission—that
“emission estimates would be largely influenced by
assumptions rather than direct parameters about the
project”—was unsatisfactory. Id. The effects of the
18 CTR. FOR BIO. DIVERSITY V. ZINKE

agency’s assumptions on its estimates simply “can be
checked” by disclosing the estimates so that readers could
make informed decisions regarding the project and its
consequences. Id.; see also WildEarth Guardians v. Zinke,
368 F. Supp. 3d 41, 68 (D.D.C. 2019) (determining that an
agency’s assertion that “quantifying [greenhouse gas]
emissions . . .would be overly speculative” was “belied by
an administrative record replete with information on oil and
gas development and [greenhouse gas] emissions”).

BOEM refers to the omission of foreign oil consumption
in two separate pages of the final, 600-page EIS. The first is
in Appendix B of the EIS, in response to public comments
expressing concern over the omission of foreign oil
consumption. BOEM responds only that “[c]ontext suggests
that any change in foreign oil consumption resulting from
the pending decision on the Liberty DPP would be very
small,” 5 and because “Liberty DPP represents a very small
fraction of the amount of oil comprising the global market,”
it “could only have a negligible impact on worldwide oil
prices and, as a result, only a negligible impact on foreign
consumption and emissions levels.” It adds that “[e]ven if
BOEM could reliably estimate these marginal differences
(which it cannot, given the lack of reliable information on
foreign emissions factors and consumption patterns), such
estimates would not change the end results of BOEM’s
analysis to a meaningful extent.” BOEM cites to no
evidence in support of these conclusions and does not
provide any further explanation for the omission.

Appendix B then refers readers to a general report,
incorporated by reference into the EIS, that describes the
market-simulation model and its limitations. The relevant

5
“DPP” is shorthand for “development and production plan.”
 CTR. FOR BIO. DIVERSITY V. ZINKE 19

portion of that report explains that “[e]xcluding the foreign
oil and gas markets is reasonable” because “[o]il
consumption in each country is different, and BOEM does
not have information related to which countries would
consume less oil.” Again, BOEM does not cite any materials
in support of these statements nor describe the research it
relied upon to reach these conclusions.

This is insufficient to satisfy NEPA’s requirements.
Emissions resulting from the foreign consumption of oil are
surely a “reasonably foreseeable” indirect effect of drilling
at Liberty, just as foreseeable as the emissions resulting from
the consumption of oil produced at sites other than Liberty,
which the market-simulation model already considers. Even
if the extent of the emissions resulting from increased
foreign consumption is not foreseeable, the nature of the
effect is. Mid States Coal. for Progress v. Surface Transp.
Bd., 345 F.3d 520, 549 (8th Cir. 2003). This is sufficient to
require estimation or explanation under NEPA. Id.

The record belies BOEM’s contention that it could not
have summarized or estimated foreign emissions with
accurate or credible scientific evidence. See Seattle
Audubon Soc. v. Espy, 998 F.2d 699, 704–05 (9th Cir. 1993).
Various studies provided by CBD in the administrative
record confirm the effect of increasing domestic oil supply
on foreign consumption and the feasibility of its estimation.
In one study, the Stockholm Environment Institute—noting
that BOEM omitted the same calculation in its analysis of
the effects of the Keystone Pipeline—demonstrates how an
increase in foreign oil consumption translates into
greenhouse gas emissions. See Peter Erickson, U.S. Again
Overlooks Top CO2 Impact of Expanding Oil Supply, but
That Might Change, Stockholm Environment Institute (Apr.
30, 2016), http://www.sei.org/perspectives/us-co2-impact-
20 CTR. FOR BIO. DIVERSITY V. ZINKE

oil-supply. Using a “simple calculation,” relying on
parameters publicly provided in BOEM’s report, the
Institute calculates the expected resultant greenhouse gas
emissions from increased foreign consumption of oil. It
concludes that developing the Pipeline would cause an
increase in global oil consumption ten times greater than the
increase in domestic consumption forecasted by BOEM.
Other studies in the record confirm the same: domestic
consumption impacts foreign oil consumption, and increases
in foreign oil consumption can be translated into estimates
of greenhouse gas emissions. See Peter Erickson and
Michael Lazarus, Impact of the Keystone XL Pipeline on
Global Oil Markets and Greenhouse Gas Emissions, Nature
Climate Change 778, 778–80 (2014) (modeling increased
global oil consumption caused by the Keystone Pipeline and
finding an increase in greenhouse gas emissions four times
greater than that predicted by the model that did not account
for global oil market effects). Jason Bordoff and Trevor
Houser, Navigating the U.S. Oil Export Debate, Columbia
SIPA Center on Global Energy Policy, Jan. 2015, at 57
(assessing the net greenhouse gas impact of an increase in
global crude oil demand under different scenarios).

BOEM now explains that these studies rely on
“simplistic assumptions that [fall] well short of the detailed
model that BOEM used to analyze the U.S. energy market,”
but it is unclear from the record why these assumptions are
any more simplistic than those the market-simulation model
incorporates. The model assumes, for example, near
constant oil and gas demand over the next 40 to 70 years, an
unrestricted supply of foreign oil for substitution, and that all
oil and gas produced domestically is consumed
domestically. BOEM’s conclusion for the higher emissions
produced by the no-action alternative assumes that the
petroleum products substituted for oil not produced at
 CTR. FOR BIO. DIVERSITY V. ZINKE 21

Liberty will come from places with “comparatively weaker
environmental protection standards.” It is unclear from the
administrative record what justifies these assumptions and
not those needed to estimate foreign oil consumption.

Even if the nature of BOEM’s assumptions did not
sufficiently demonstrate the need for further explanation, the
result upon which the agency relied surely did. BOEM’s
conclusion that not drilling will result in more carbon
emissions than drilling is counterintuitive. An agency acts
arbitrarily and capriciously when it reaches a decision that 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 U.S., Inc. v. State Farm Mut. Auto. Ins. Co.,
463 U.S. 29, 43 (1983). Without further explanation, we
cannot ascribe the implausibility of the result to BOEM’s
expertise or rational decision-making. We will uphold a
decision “of less than ideal clarity if the agency’s path may
be reasonably discerned,” but we cannot “supply a reasoned
basis for the agency’s action that the agency itself has not
given.” Id. at 43, 57 (internal quotation marks omitted).

We “understand that in some cases quantification may
not be feasible.” Sierra Club, 867 F.3d at 1374. But even if
BOEM is unable to quantitatively evaluate the emissions
generated by foreign countries in the absence of the Liberty
project, it still must thoroughly explain why such an estimate
is impossible. The Department of Interior has promulgated
a regulation addressing such situations, where “incomplete
or unavailable information” impedes the agency’s ability to
evaluate a “reasonably foreseeable significant adverse
effect[]” of the project. 40 C.F.R. § 1502.22. The regulation
requires the agency to include a statement explaining that the
information is lacking, its relevance, a summary of any
existing credible evidence evaluating the foreseeable
22 CTR. FOR BIO. DIVERSITY V. ZINKE

adverse impacts, and the agency’s evaluation of the impacts
based upon “theoretical approaches or research methods
generally accepted in the scientific community.” 40 C.F.R.
§ 1502.22(b)(1). These requirements are read “in the
context of the more general requirements for preparation of
an EIS,” including the “rigorous evaluation” of the indirect,
direct, and cumulative effects of the selected alternatives.
National Environmental Policy Act Regulations, 50 FR
32,234, 32,237 (Aug. 9, 1985); see also 40 C.F.R.
§§ 1502.16(a)–(b), 1508.8(b).

The EIS’s two-page explanation of BOEM’s decision to
omit foreign oil emissions is insufficient to meet these
requirements. BOEM did not summarize existing research
addressing foreign oil emissions nor attempt to estimate the
magnitude of such emissions. It cannot ignore basic
economics principles and state—without citations or
discussion—that the impact of the Liberty project on foreign
oil consumption will be negligible. See WildEarth
Guardians v. Bureau of Land Mgmt., 870 F.3d 1222, 1237–
38 (10th Cir. 2017); Mont. Envtl. Info. Ctr. v. U.S. Off. of
Surface Mining, 274 F. Supp. 3d 1074, 1098 (D. Mont.
2017). Nor can it ignore this foreseeable effect entirely. EIS
estimates often involve some “[r]easonable forecasting and
speculation.” Scientists’ Inst. For Pub. Info., Inc. v. Atomic
Energy Comm’n, 481 F.2d 1079, 1092 (D.C. Cir. 1973).
Some “educated assumptions are inevitable in the NEPA
process,” and the “effects of assumptions on estimates can
be checked by disclosing those assumptions so that readers
can take the resulting estimates with the appropriate amount
of salt.” Sierra Club, 867 F.3d at 1374.

We note that we typically accord significant deference to
an agency’s decisions that require a “high level of technical
expertise.” Kleppe v. Sierra Club, 427 U.S. 390, 412 (1976).
 CTR. FOR BIO. DIVERSITY V. ZINKE 23

But such deference applies only when the agency is making
predictions “within its area of special expertise.” Baltimore
Gas & Elec. Co. v. Nat. Res. Def. Council, Inc., 462 U.S. 87,
103 (1983). BOEM’s area of expertise is the management
of “conventional (e.g., oil and gas) and renewable energy-
related” functions, including “activities involving resource
evaluation, planning, and leasing.” U.S. Dep’t of Interior,
Sec. Order No. 3299A2, Establishment of the Bureau of
Ocean Energy Management, the Bureau of Safety and
Environmental Enforcement, and the Office of Natural
Resources Revenue, § 2 (Aug. 29, 2011). The scope of its
expertise does not include the economic analysis of
greenhouse gas emissions. Therefore, we do not readily
defer to its decision to exclude a discussion of foreign oil
consumption, particularly in light of our conclusion that its
decision to do so was unreasonable. See The Lands Council
v. McNair, 537 F.3d 981, 993 (9th Cir. 2008), overruled in
part on other grounds by Winter v. Nat. Res. Def. Council,
Inc., 555 U.S. 7 (2008).

In short, the EIS “should have either given a quantitative
estimate of the downstream greenhouse gas emissions” that
will result from consuming oil abroad, or “explained more
specifically why it could not have done so,” and provided a
more thorough discussion of how foreign oil consumption
might change the carbon dioxide equivalents analysis.
Sierra Club, 867 F.3d at 1374. BOEM has the statutory
authority to act on the emissions resulting from foreign oil
consumption. If it later concludes that such emissions will
be significant, it may well approve another alternative
included in the EIS or deny the lease altogether. Cf. Dep’t
of Transp. v. Public Citizen, 541 U.S. 752, 766–68, 770
(2004). For these reasons, we agree with CBD that BOEM’s
alternatives analysis in the EIS was arbitrary and capricious.
24 CTR. FOR BIO. DIVERSITY V. ZINKE

III. ESA

A. Section 7 Consultation and Section 9 Take Regulation

CBD next challenges FWS’s compliance with the ESA.
In accordance with NEPA’s requirements, FWS prepared for
BOEM a biological opinion that discusses the effects of the
project on all threatened species and their habitats in the Bay.
In the opinion, FWS concluded that polar bears—which are
classified as threatened marine mammals—were present in
the project area, but that the project was unlikely to
jeopardize their continued existence or adversely modify
their habitat. CBD argues that FWS violated the ESA
because portions of its biological opinion and incidental take
statement were arbitrary and capricious.

The Department of the Interior and, by delegation, FWS,
is responsible for implementing the ESA. See 16 U.S.C.
§ 1531 et seq. Section 9 of the ESA regulates the “taking”
of a threatened or endangered species. It prohibits “any
person”—including an “instrumentality” of federal, state, or
municipal government, see id. § 1532(12), (13)—from,
among other things, “taking” endangered wildlife, fish, or
plants, id. § 1538(a). A “take” occurs under the ESA when
an animal is harassed, harmed, pursued, hunted, shot,
wounded, killed, trapped, captured, or collected, or when
anyone attempts to engage in such conduct. Id. § 1532(19).

FWS may issue a temporary permit approving conduct
normally barred by Section 9 if the taking is incidental to an
otherwise lawful activity. Id. § 1539(a)(1)(B). Before FWS
may issue such a permit, it must find that (1) the applicant
will minimize and mitigate the negative impacts of the
taking; (2) the applicant will ensure adequate funding for the
plan; and (3) the taking will not appreciably reduce the
 CTR. FOR BIO. DIVERSITY V. ZINKE 25

likelihood of the survival and recovery of the species in the
wild. Id. § 1539(a)(2)(B).

Section 7 of the ESA describes the process for agency
consultation. Unlike Section 9, it does not contain an
outright prohibition on take; it requires only that an agency
consult with FWS or NMFS before it takes any action that
may affect a species listed as threatened or endangered under
the ESA. See id. § 1536(a)(2), (4).

“Section 7 consultation” begins with an assessment of
the species affected by the action. If a threatened or
endangered species “may be present” in the area of the
proposed action, the agency must conduct a biological
assessment to determine whether the species will be
adversely affected by the project. Id. § 1536(c)(1); see also
50 C.F.R. § 402.14(a). If BOEM concludes that the species
is likely to be adversely affected, it must initiate formal
consultation with either FWS or NMFS (here, FWS). After
formal consultation, FWS issues a written opinion (a
“biological opinion,” or “BiOp”), concluding either that the
project is unlikely to adversely affect the species or that the
action will likely jeopardize the species or adversely modify
its critical habitat. See 16 U.S.C. § 1536(b)(3)(A). If FWS
determines that the proposed action is likely to jeopardize
the species or modify its habitat, then it must suggest
reasonable and prudent alternatives that could be taken by
the agency. Id; see also 50 C.F.R. § 402.14(g).

If, however, FWS determines that the proposed action
will neither harm the species nor adversely modify its
habitat, it may authorize the taking of a species incidental to
the proposed project. 16 U.S.C § 1536(b)(4). To determine
whether the action will ultimately jeopardize a listed species
or adversely modify its habitat, the agency may rely on
mitigation measures proposed by the project planners. See
26 CTR. FOR BIO. DIVERSITY V. ZINKE

Selkirk Conservation All. v. Forsgren, 336 F.3d 944, 955
(9th Cir. 2003).

When the agency authorizes the incidental taking of a
species, it must also issue an “incidental take statement” with
the biological opinion. 50 C.F.R. § 402.14(i); Ctr. for Bio.
Diversity v. Salazar, 695 F.3d 893, 909 (9th Cir. 2012). The
incidental take statement estimates the amount of the
project’s incidental take of the listed species, includes any
“reasonable and prudent measures” considered “necessary
or appropriate to minimize such impact,” and—in the case
of marine mammals like the polar bear—describes specific
measures necessary to comply with the aforementioned
provisions of the MMPA. 50 C.F.R. § 402.14(i)(1); see also
Salazar, 695 F.3d at 909; 16 U.S.C. § 1536(b)(4). The
statement also describes the terms that must be followed by
BOEM or the applicant to implement any mitigation
measures specified in the statement. 16 U.S.C. § 1536(b)(4).
A taking that complies with the terms and conditions of a
Section 7 incidental take statement is not prohibited by
Section 9. Salazar, 695 F.3d at 909; 16 U.S.C. § 1536(o)(2);
50 C.F.R. § 402.14(i)(5).

B. Coordination between the ESA and the MMPA

The MMPA prohibits the take or harassment of animals,
but its scope is narrower and its procedures distinct from
those of Sections 7 and 9 of the ESA. It entirely prohibits
the take of marine mammals in U.S. waters. “Take” in the
MMPA is similar to “take” under Section 9 of the ESA; the
MMPA defines it as encompassing, among other things,
“harassment,” “torment,” or “annoyance” which “has the
potential to injure . . . or . . . disturb a marine mammal . . . in
the wild by causing disruption of behavioral patterns,
including, but not limited to, migration, breathing, nursing,
 CTR. FOR BIO. DIVERSITY V. ZINKE 27

breeding, feeding, or sheltering.” 16 U.S.C. § 1362(13),
(18)(A)(i)–(ii); see also id. § 1371(a).

As under the ESA, the MMPA allows FWS to permit the
incidental take of “small numbers” of marine mammals
pursuant to a specified activity for a limited period. The total
incidental take must have a “negligible impact” on the
species and cannot have an “unmitigable adverse impact” on
the availability of the species for specified subsistence uses.
16 U.S.C. § 1371(a)(5)(A); see also 50 C.F.R. § 18.27(b). If
the incidental take meets these requirements, FWS may then
prescribe regulations setting forth permissible methods of
taking the species in question and describing methods of
effecting the least adverse impact possible on the species and
its habitat. See 50 C.F.R. § 18.27(b). The regulations are
subject to public notice-and-comment. 16 U.S.C.
§ 1371(a)(5)(D)(iii). Once the regulations are finalized and
promulgated, FWS issues individual letters of authorization
to the agency, authorizing the project and the take. Id.

Both the ESA and the MMPA apply when, as here, an
agency seeks approval for the incidental take of threatened
and endangered marine mammals. The MMPA is more
restrictive than the ESA; when the two statutes conflict, the
relevant MMPA provision applies. Id. § 1543. FWS cannot
issue an incidental take statement authorizing the take of an
endangered or threatened species under the ESA until the
take has been authorized under the MMPA. See id.
§ 1536(b)(4)(C); see also Incidental Take of Endangered,
Threatened, and Other Depleted Marine Mammals, 54 Fed.
Reg. 40,338, 40,346 (Sept. 29, 1989), codified at 50 C.F.R.
§§ 18.27, 228, 402.14. The incidental take statement must
incorporate any mitigation measures required under the
MMPA. 50 C.F.R. § 402.14(i)(1)(iii).
28 CTR. FOR BIO. DIVERSITY V. ZINKE

In consultation with BOEM, FWS issued a BiOp
authorizing the Liberty project’s incidental take of polar
bears. 16 U.S.C. § 1536(b)(4). The BiOp acknowledges that
Liberty may “adversely affect polar bears through
disturbance, an increase in polar bear-human interactions,
and habitat loss,” and concludes that denning polar bear
mothers and cubs are most likely to be affected, because they
are the most sensitive to the disturbance caused by the
project. The disturbance from the project is expected to
include (1) construction, drilling, production operations,
maintenance, and ancillary activities associated with the
project; (2) noise and disturbance caused by aircraft, vessel,
hovercraft, and vehicle traffic; and (3) drilling and
production activities. The BiOp concludes that the proposed
action is “not likely to jeopardize the continued existence of
polar bears by reducing appreciably the likelihood of
survival and recovery in the wild by reducing reproduction,
numbers, or distribution of this species.”

CBD argues that FWS violated the ESA by (1) relying
on uncertain, insufficiently specific mitigation measures in
reaching its no-jeopardy and no-adverse-modification
conclusions, and (2) failing to specify the amount and extent
of “take” in the incidental take statement included within the
BiOp.

C. Inadequacy of FWS’s Mitigation Measures

Throughout the BiOp, FWS describes mitigation
measures intended to alleviate the harm caused to polar bears
by the Liberty project. CBD argues that the mitigation
measures violate the ESA for two closely related reasons.
First, the measures themselves are not sufficiently specific,
binding, or certain to occur. Second, FWS relied on these
non-binding mitigation measures to reach both its “no
jeopardy and no adverse modification” conclusions for polar
 CTR. FOR BIO. DIVERSITY V. ZINKE 29

bears and their critical habitats, respectively. The
government responds that FWS did not rely on the
mitigation measures and, even if it did, the mitigation
methods are enforceable and sufficiently specific.

1. Enforceability of the BiOp’s Mitigation Measures

We first address whether the mitigation measures in
FWS’s BiOp are sufficiently binding or certain to occur.
Mitigation measures relied upon in a biological opinion must
constitute a “clear, definite commitment of resources,” and
be “under agency control or otherwise reasonably certain to
occur.” Nat’l Wildlife Fed. v. Nat’l Marine Fisheries Serv.,
524 F.3d 917, 936 & n.17 (9th Cir. 2008). A “sincere general
commitment to future improvements”—without more
specificity—is insufficient. Id. at 935–36. The measures
“must be subject to deadlines or otherwise-enforceable
obligations; and most important, they must address the
threats to the species in a way that satisfies the jeopardy and
adverse modification standards.” Ctr. for Bio. Diversity v.
Rumsfeld, 198 F. Supp. 2d 1139, 1152 (D. Ariz. 2002). 6
Binding mitigation measures cannot refer only to
generalized contingencies or gesture at hopeful plans; they
must describe, in detail, the action agency’s plan to offset the
environmental damage caused by the project.

If an action agency fails to carry out the mitigation
measures contained in a BiOp, it must re-initiate
consultation with FWS. See Ctr. for Bio. Diversity v. U.S.
Bureau of Land Mgmt., 698 F.3d 1101, 1115 (9th Cir. 2012)
6
District courts in this circuit follow the standard articulated by
Rumsfeld. See, e.g., AquAlliance v. U.S. Bureau of Reclamation,
287 F. Supp. 3d 969, 1071–72 (E.D. Cal. 2018) (“One district court
persuasively provided further guidance [regarding when mitigation
measures are sufficiently clear and definite].”) (citing to Rumsfeld).
30 CTR. FOR BIO. DIVERSITY V. ZINKE

(citing 50 C.F.R. § 402.16(c)). If the action agency does not
re-initiate consultation, the BiOp is invalid and “any person”
may bring suit and subject the action agency or the applicant
to “substantial civil and criminal penalties.” Id. (internal
quotation marks omitted). An indefinite mitigation measure
is less likely to trigger re-consultation because it will be
difficult to know at which point or whether the action agency
has failed to comply. For this reason, measures that are too
vague, or do not commit resources, or are otherwise
insufficiently integrated into the proposed action are
generally unenforceable under the ESA, and thus cannot be
properly relied upon. See id. at 1113–14. The measures can
be made enforceable in a variety of ways, including by
incorporation into the terms and conditions of an incidental
take statement. See id. at 1114 n.9.

CBD identifies four instances of planned mitigation in
FWS’s biological opinion. The first two state:

Available data indicate polar bears regularly
den at low densities in the action area. . . .
Den abandonment would be most likely to
occur during new construction activities
because ongoing activities during routine
operations would allow more sensitive bears
to select an alternative den site. However, the
applicant has indicated they would conduct
den detection surveys each winter in
compliance with [guidance issued under
MMPA incidental take regulations] and the
project’s polar bear interaction plan. These
surveys would be planned in cooperation
with [FWS]. If dens are detected within
1.6 km of the proposed locations of ice roads
 CTR. FOR BIO. DIVERSITY V. ZINKE 31

and pads, then [FWS] will be contacted for
guidance.[ 7]

...

As with denning polar bears, [FWS] expects
potential adverse effects to non-denning
polar bears would be reduced by the
applicant’s compliance with existing and
future authorizations issued under the
MMPA . . . . Disturbance that disrupts
behavioral patterns of polar bears is classified
as take under the MMPA. The MMPA
prohibits unpermitted incidental take of
marine mammals. Under the MMPA,
incidental take is only permitted provided the
total of such taking will have no more than a
negligible impact on the marine mammal
species . . . , and does not have an
unmitigable adverse impact on the
availability of these species for subsistence
uses. . .

Both measures rely principally on yet unapproved and
undefined mitigation measures under the MMPA. The
government argues that the BiOp’s reliance on these
measures is authorized under a 2013 Department of Interior
rule governing the conservation and protection of polar
bears. See Endangered and Threatened Wildlife and Plants;
Special Rule for the Polar Bear Under Section 4(d) of the
Endangered Species Act, 78 Fed. Reg. 11,766 (Feb. 20,
2013), codified at 50 C.F.R. § 17.40(q). The rule

7
“Den detection surveys” are used to evaluate (with infrared radar,
for example) where polar bear dens are located.
32 CTR. FOR BIO. DIVERSITY V. ZINKE

“synchronizes the management of the polar bear under the
ESA with management provisions under the MMPA[.]” Id.
at 11,768.

The rule does permit the agency to bypass Section 9
compliance under the ESA once it has obtained a letter of
authorization under the MMPA. It states that “if an activity
is authorized or exempted under the MMPA,” “no additional
authorization” under Section 9 of the ESA “for that activity
will be required.” Id.; see also 50 C.F.R. § 17.40(q)(2)
(“None of the prohibitions in § 17.31 of this part apply to
any activity that is authorized or exempted under the Marine
Mammal Protection Act (MMPA) . . . provided that the
person carrying out the activity has complied with all terms
and conditions that apply to that activity under the provisions
of the MMPA . . . and [its] implementing regulations.”).
MMPA protection is considered sufficient because the
definition of “take” under the MMPA is “more protective”
than take under the ESA. 78 Fed. Reg. at 11,770.

Therefore, “managing take of polar bears under the
MMPA adequately provides for the conservation of polar
bears.” Id. Obviously, if incidental take of a threatened
marine mammal is not authorized under the MMPA, “then
the general [ESA take prohibitions] would apply, and [the
Department of Interior] would require a permit for the
activity as specified in [its] ESA regulations.” Id. at 11,766.
But, as the rule repeatedly states, it “does not remove or alter
in any way the consultation requirements under section 7 of
the ESA.” Id. at 11,768. In other words, FWS’s BiOp
remains unaffected by the polar-bear rule because it is part
of the consultation process under Section 7 of the ESA.

We have already rejected a similar interpretation of the
rule, as applied to incidental take statements. Salazar,
695 F.3d at 910–11. In Salazar, we held that an agency
 CTR. FOR BIO. DIVERSITY V. ZINKE 33

acted unlawfully by failing to issue an incidental take
statement pursuant to Section 7 of the ESA, even though it
separately complied with the MMPA prohibitions on marine
mammal take. Id. at 910. The agency argued that the polar-
bear rule preempted any need to publish an incidental take
statement. Id. But, as we explained, Section 7 imposes a
separate requirement for an incidental take statement and
biological opinion in certain circumstances. Id. The rule
itself states that “[n]othing in this special rule affects the
issuance or contents of the biological opinions for polar
bears[.]” 73 Fed. Reg. 76,249, 76,252 (Dec. 16, 2008)
(emphasis added). In other words, compliance with
50 C.F.R. § 17.40(q) satisfies the ESA’s Section 9 take
requirement but does not fulfill the agency’s separate and
independent Section 7 obligations. Salazar, 695 F.3d at
910–11.

So too here. FWS must comply with both Section 7 and
Section 9 of the ESA, and approval of polar-bear take under
the MMPA will meet the agency’s obligations only under
Section 9. The rule does not preclude or preempt FWS’s
responsibility to include the mitigation measures that it relies
upon in a biological opinion under Section 7 of the ESA.
The agency cannot refer to future, unstated authorizations
under the MMPA to fulfill its obligations under Section 7. 8

8
FWS, NMFS, and two other federal agencies have also issued
guidance suggesting they did not contemplate that MMPA compliance
would automatically satisfy an action agency’s Section 7 obligations.
The timelines between the MMPA and ESA differ substantially; the ESA
process generally requires that Section 7 consultation be completed
within 90 days, 16 U.S.C. § 1536(b)(1)(A), (B), but the MMPA approval
process can take much longer because it requires public notice-and-
comment. The agencies recommend that action agencies handle timing
discrepancies in one of three ways. First, the action agency may
“consider initiating the MMPA [] process in advance of the ESA section
34 CTR. FOR BIO. DIVERSITY V. ZINKE

Reliance on future MMPA measures is particularly
inappropriate to satisfy the agency’s Section 7 obligations
here, where the authorizations under the MMPA last for only
five years, see 16 U.S.C. § 1371(a)(5)(A), and the Liberty
project is expected to last fifteen to twenty years.

7 process.” 54 Fed. Reg. at 40,346. The MMPA requirements can then
“be incorporated into the ESA incidental take statement when the
biological opinion is issued and subsequent revisions would not be
necessary.” Id. Second, FWS and the action agency may together agree
to extend the Section 7 consultation under the ESA “to accommodate
completion” of the MMPA regulations. Id. Or, third, the action agency
may begin “early consultation” with the ESA, and request a “preliminary
biological opinion.” Id. Once the MMPA process is completed, the
opinion “would be reviewed and the . . . incidental take statement
amended or added, as appropriate.” Id.

Whichever route the action agency chooses, the Department of
Interior “is expected to proceed with issuance of the biological opinion
and . . . incidental take statement in a timely manner” as Section 7
consultation requires. 99th Cong. 32,185 (1986) (statement of Rep.
Jones). The agency should “indicate that the findings and conditions
applicable to affected marine mammals are subject to final completion
of the MMPA” process and “that the statement would subsequently be
revised to reflect the outcome of that review.” Id. “In this situation,” as
the statute reflects, “incidental take of listed marine mammals would not
be authorized under the ESA” until after “the MMPA and the section 7[]
incidental take statement has been revised.” Id.; see also 16 U.S.C.
§ 1536(b)(4)(C).

In other words, even if the action agency obtains MMPA approval,
the take statement must be “subsequently revised” to reflect that
approval, and authorization will not occur under after the statement has
been revised. At no point does the guidance issued by FWS suggest that
Section 7 approval occurs automatically, upon MMPA approval. Indeed,
all three options suggested in the guidance recommend that the action
agency complete the MMPA approval process before seeking final
Section 7 consultation and approval.
 CTR. FOR BIO. DIVERSITY V. ZINKE 35

The third proposed mitigation measure states:

Additional information on possible
minimization measures that would reduce
effects to polar bears from oil and gas
industry activities can be found in the [2016
generalized list on mitigation measures used
in the Beaufort Sea].

This measure references “possible” strategies, without
selecting a mitigation measure from the incorporated list or
committing BOEM or Hilcorp to carrying out any specific
number of measures. These noncommittal assurances
cannot shoulder the government’s burden to identify a
“clear, definite commitment of resources.” Nat’l Wildlife
Fed’n, 524 F.3d at 936; see also Sierra Club v. Marsh,
816 F.2d 1376, 1388 (9th Cir. 1987) (concluding that “one
of several ‘reasonable and prudent alternatives’ that the FWS
found necessary to minimize the project’s effects” was not
reasonably certain to occur), abrogated on other grounds as
recognized in Cottonwood Envtl. Law Ctr. v. U.S. Forest
Serv., 789 F.3d 1075, 1088–91 (9th Cir. 2015).

The fourth and final measure states:

Mitigation measures applied to ensure least
practicable impacts include requirement of
site-specific plans of operation and site-
specific polar bear interaction plans. In
combination, these plans reduce attraction to
bears (e.g., through garbage disposal
procedures, snow management procedures)
and provide training and other measures to
eliminate the potential for injurious or lethal
take of bears in defense of human life in the
event that encounters occur. Other mitigation
36 CTR. FOR BIO. DIVERSITY V. ZINKE

measures may be required on a case-by-case
basis, such as use of infra-red thermal
technology or trained dogs to determine
presence or absence of dens in suitable
denning habitat; measures to protect pregnant
polar bears during denning activities (den
selection, birthing, and maturation of cubs);
and limiting industrial activities near barrier
islands . . . This incidental-take program and
the associated mitigation measures have
effectively limited human-bear interactions
and disturbance to bears, ensuring that, at
least to date, industry effects have had a
negligible impact on bears.

This contains the most concrete mitigation strategies found
in the BiOp, but even these suggestions do not truly commit
to the development of mitigation strategies. The few
concrete strategies provided are offered only as examples of
possible strategies that could be taken, “in the event that
encounters occur.” It is unclear what will constitute a polar
bear encounter or commit the action agency to carrying out
any of the mitigation measures listed in the examples
provided. It concludes that “[o]ther mitigation measures
may be required on a case-by-case basis,” a statement which,
alone, also does nothing to bind BOEM when the need for
those measures apply. See, e.g., Rumsfeld, 198 F. Supp. 2d
at 1153 (explaining that a “laundry list of possible mitigation
measures” is unenforceable). A mitigation strategy’s
eventual MMPA approval does not change this analysis
because, as we have held, MMPA authorization does not
alter the agency’s obligations under Section 7 of the ESA.

We agree with CBD that the mitigation measures
proposed by FWS are too vague to enforce. The
 CTR. FOR BIO. DIVERSITY V. ZINKE 37

administrative record does reflect a “general desire” to
impose mitigation strategies, but it does not reflect a definite
commitment to those improvements. The generality of the
mitigation measures makes it difficult to determine the point
at which the action agency may renege on its promise to
implement these measures. “[S]incere general
commitment[s] to future improvement” are insufficient
under Section 7. Nat’l Marine Fisheries Serv., 524 F.3d. at
935–36.

2. Reliance on Mitigation Measures

Our conclusion that the mitigation measures in the BiOp
are insufficiently specific to enforce has no legal
consequence unless we separately conclude that FWS relied
on those measures. The government and Hilcorp argue that
because the overall magnitude of the negative effect on polar
bears is estimated to be low, FWS did not rely on any of its
mitigation measures to reach its no-jeopardy and no-
adverse-modification findings.

Whether FWS relied on the proposed mitigation
measures in reaching its conclusion depends on the language
and structure of the BiOp. A BiOp that integrates mitigation
measures into its decision-making is more likely to have
relied upon those measures. Conversely, an opinion that
relies upon indefinite “background cumulative effects” and
uses those effects “as a basis for determining the likely
effects” of the proposed project, for example, renders the
agency’s reliance on that opinion arbitrary and capricious.
See Ctr. for Bio. Diversity, 698 F.3d at 1113. To be
enforceable, those effects “should properly have been part of
the project itself.” Id.; see also Nat’l Wildlife Fed’n v. Nat’l
Marine Fisheries Serv., 839 F. Supp. 2d 1117, 1125–26
(D. Or. 2011) (holding that the agency improperly relied on
habitat mitigation measures that “in some cases [were] not
38 CTR. FOR BIO. DIVERSITY V. ZINKE

even identified” and the agency had “assume[d] it w[ould]
be able to identify and implement the additional projects that
are necessary”); Ctr. for Bio. Diversity v. Salazar, 804
F. Supp. 2d 987, 1002 (D. Ariz. 2011) (concluding that a
biological opinion that relied on water saving mitigation
projects where the court could not ascertain the details of the
planned projects or the estimated water savings was
unlawful).

The portion of the BiOp describing FWS’s no-jeopardy
and no-adverse modification findings is brief. Its no-
jeopardy conclusion states:

A small number of polar bears may also be
adversely affected through disturbance or
polar bear-human interactions which may
include intentional take. These adverse
effects are expected to impact only small
numbers of individuals . . . and therefore, we
do not expect population-level impacts as a
result of the proposed Liberty DPP. After
reviewing the current status of the species,
environmental baseline, effects of the action,
and cumulative effects, [FWS] concludes the
proposed action is not likely to jeopardize the
continued existence of polar bears by
reducing appreciably the likelihood of
survival and recovery in the wild by reducing
reproduction, numbers, or distribution of this
species.

FWS appears to conclude that the Liberty project, as a
whole, will not significantly impact polar bears, with or
without the mitigation measures. We conclude FWS did not
 CTR. FOR BIO. DIVERSITY V. ZINKE 39

rely on any of the aforementioned mitigation measures in its
no-jeopardy determination.

But in concluding that the bears’ critical habitat will not
be adversely affected by the project, FWS relied on three
stated factors, the second of which incorporates the
mitigation measures. Specifically, the second basis for
FWS’s no-adverse-modification finding is that the “terms
and conditions associated with authorizations under the
MMPA would minimize the level of persistent disturbance
that may result from the Proposed Action[.]”

As discussed, unauthorized, future mitigation measures
under the MMPA cannot satisfy the FWS’s obligations
under Section 7 of the ESA. The mitigation measures
proposed in the BiOp are indefinite and do not constitute a
“clear, definite commitment of resources,” and FWS’s
reliance upon those measures to conclude that the polar
bear’s critical habitat would not be adversely modified by
the Liberty project was arbitrary and capricious. For these
reasons, we hold that FWS’s BiOp violated the ESA. We
further hold that FWS did not rely on its indefinite mitigation
measures in finding that the polar bear’s continued existence
would not be jeopardized by the project.

D. Incidental Take

We next evaluate whether FWS unlawfully failed to
specify the amount and extent of “take” in its incidental take
statement. The ESA requires an incidental take statement
where FWS concludes, as here, that a project will not
jeopardize a species or modify its critical habitat. The
purpose of the incidental take statement is, at least in part, to
specify the amount of take that may occur, and include
triggers that indicate non-compliance with the statement and
require re-consultation with FWS. See 16 U.S.C.
40 CTR. FOR BIO. DIVERSITY V. ZINKE

§ 1536(b)(4); 50 C.F.R. § 402.14(i)(l)(i). To “specify the
impact” of any incidental take, the statement should either
include a numerical cap on take or explain why it does not
include the cap. Ctr. for Bio. Diversity, 698 F.3d at 1127.
The numerical cap establishes a threshold that, when
exceeded, results in an unacceptable level of take and
requires parties to re-initiate Section 7 consultation. Ariz.
Cattle Growers Ass’n v. U.S. Fish and Wildlife, 273 F.3d
1229, 1249 (9th Cir. 2001).

The statement can use a proxy measure for take where
“no number may be practically obtained.” Ctr. for Bio. Div,
698 F.3d at 1126–27 (internal quotation marks omitted). For
example, where an agency is unable to quantify the number
of endangered or threatened fish that will be taken, it may
instead estimate the project’s impact on the number of eggs
laid by those fish. Id. (citing H.R. Rep. No. 97-567, at 27
(1982)). Take can also be expressed as a change in habitat
affecting the species (e.g., for “aquatic species, changes in
water temperature or chemistry, flows, or sediment loads”),
but “some detectable measures of effect should be
provided.” Ariz. Cattle Growers, 273 F.3d at 1250 (quoting
Final ESA Section 7 Consultation Handbook, March 1998
at 4-47–4-48). When it relies upon a proxy, the agency must
explain why it cannot directly quantify the animal’s expected
take. See Or. Nat. Resources Council v. Allen, 476 F.3d
1031, 1037–38 (9th Cir. 2007) (holding that FWS erred in
quantifying the expected take of northern spotted owls in
terms of habitat acreage without explaining why the agency
was unable to numerically estimate take).

CBD argues that FWS failed to quantify the amount of
nonlethal take in its incidental take statement. The
government argues that any nonlethal disturbance does not
rise to the level of take, and so FWS did not need to quantify
 CTR. FOR BIO. DIVERSITY V. ZINKE 41

any nonlethal take that may occur as a result of the project.
We agree that FWS contemplated that nonlethal harassment
of polar bears may rise to the level of “take” under the ESA
and should have quantified the nonlethal take of the bears.

In the BiOp, FWS does provide a numerical cap on the
amount of take that constitutes injury or death to polar bears;
injury or death to more than one polar bear triggers re-
consultation:

As provided in 50 C.F.R. 402.16, re-initiation
of formal consultation is required where
discretionary Federal agency involvement or
control over the action has been retained (or
is authorized by law), and re-initiation may
be required if:

1. The amount or extent of incidental
take for listed species is exceeded
over the life of the project;

a. . . .

b. If human-polar bear interactions
result in injury and/or death of
more than 1 polar bear over the
life of the project.

But FWS does not quantify the amount of other types of
incidental take that the Liberty project may cause. Take
under the ESA can occur via injury or death, as the BiOp
recognizes, but it can also occur via nonlethal harassment.
See 16 U.S.C. § 1532(19). FWS interprets “harassment” of
an animal to have occurred under the ESA when an entity,
either intentionally or negligently, “creates the likelihood of
injury to wildlife by annoying it to such an extent as to
42 CTR. FOR BIO. DIVERSITY V. ZINKE

significantly disrupt normal behavioral patterns which
include, but are not limited to, breeding, feeding, or
sheltering.” 50 C.F.R. § 17.3(c). The MMPA includes an
even broader view of “harassment” than the ESA—
“harassment” includes actions which “ha[ve] the potential to
disturb a marine mammal . . . by causing disruption of
behavioral patterns, including, but not limited to, migration,
breathing, nursing, breeding, feeding, or sheltering.”
16 U.S.C. § 1362(18)(A).

Though it now argues otherwise, it appears that FWS
contemplated these types of nonlethal take in its biological
opinion. It noted that polar bears might face disturbance
from “polar bear-human interactions,” and “ground-level
activities within the action area,” including “construction of
the LDPI, drilling activities, facility operations, pipeline
construction and maintenance, mine site development, ice
road construction and associated vehicle traffic, and air
traffic.” “In addition to disturbance from ground-level
activities within the action area, air traffic associated with
the Liberty DPP could potentially disturb polar bears,
affecting the success or likelihood of denning in the action
area.” “Denning females may also be more likely to
abandon their dens in the fall before cubs are born and
relocate if disturbed. . . . Den abandonment would be most
likely to occur during new construction activities because
ongoing activities during routine operations would allow
more sensitive bears to select an alternative den site.” These
disturbances implicate disruptions in behavioral patterns
contemplated in the ESA and MMPA, such as polar bears’
breeding and sheltering.

Two different portions of FWS’s biological opinion
suggest that FWS contemplated that these expected
disturbances rise to the level of nonlethal take. In its
 CTR. FOR BIO. DIVERSITY V. ZINKE 43

discussion on the effects of these disturbances on polar
bears, FWS noted that

The potential that disturbance will indirectly
reduce the value of polar bear critical habitat
would be significantly reduced by other
existing regulatory programs that directly
address the disturbance of polar bears. As
described previously, the MMPA allows for
incidental, non-intentional take from
harassment of small numbers of marine
mammals during specific activities[.]

The BiOp then lists the potential mitigating consequences of
future measures authorized under the MMPA. This list
suggests that FWS considered that such indirect harassment
would rise to the level of “incidental, non-intentional take”
under the MMPA, and that mitigation measures might
alleviate the severity of such take.

More pointedly, a later section of FWS’s BiOp states that
re-initiation of formal consultation may be required if:

New information reveals effects of the action
that may affect listed species in a manner or
to an extent not considered in this opinion
(e.g., if observations in the Liberty DPP
action area indicate levels of interaction with
polar bears, especially the need for hazing, is
increasing significantly over time, or is
resulting in chronic or repeated interference
with normal polar bear behavior).
44 CTR. FOR BIO. DIVERSITY V. ZINKE

FWS explains that the “levels of interaction with polar
bears, especially the need for hazing” 9 is itself a trigger for
further re-consultation: if interaction with the bears increases
significantly or results in chronic, repeated interference with
normal bear behavior, FWS requires re-consultation.
Considering “levels of interaction” as a trigger suggests that
this type of non-lethal harassment amounts to incidental take
and requires FWS to provide an estimate for such take. See
Ariz. Cattle Growers, 273 F.3d at 1249.

This trigger is particularly important here, where FWS
asserted that any take approved under the MMPA would take
effect without further action by FWS. As the incidental take
statement stands, there is no guarantee that these
“harassment” take provisions—once they are made
enforceable by authorization under the MMPA—will
contain the numerical triggers required by the ESA.

Because FWS contemplated that the harassment and
disturbances polar bears will suffer could trigger re-
consultation with FWS and did not quantify the nonlethal
take that polar bears are expected to face (or explain why it
could not do so), we hold that FWS’s incidental take
statement violated the ESA. It was therefore arbitrary and
capricious under the APA.

9
“Hazing” polar bears refers to actions taken to deter them from
entering a worksite. “Polar bears may need to be hazed if they approach
Liberty DPP infrastructure when humans are present (e.g., the work
surface of the LDPI). Although the partial sheet pile wall may prevent
some polar bears from accessing the LDPI, others may gain access to
areas occupied by humans and require hazing.” Bears can be hazed by
using loud noises (starting a car or revving an engine), or by using
stronger mechanisms (such as chemical repellants, electric fences, or
“firearm projectiles”).
 CTR. FOR BIO. DIVERSITY V. ZINKE 45

IV. BOEM’s Reliance on the Invalid BiOp

Finally, we evaluate whether BOEM’s reliance on
FWS’s biological opinion in its approval of the Liberty
project was arbitrary and capricious. Section 7 of the ESA
imposes a duty on BOEM to ensure that its actions are not
likely to jeopardize the continued existence of the listed
species or result in destruction or adverse modification of its
critical habitat. Ctr. for Bio. Diversity, 698 F.3d at 1127–28.
An agency cannot meet its Section 7 duties by relying on a
legally flawed biological opinion or failing to discuss
information that might undercut the opinion’s conclusions.
See id. Because we conclude that FWS’s biological opinion
is, at least in part, invalid, BOEM’s reliance on it is unlawful.

V. Relief

We vacate BOEM’s approval of the Liberty project. We
conclude that BOEM acted arbitrarily and capriciously by
failing to quantify the emissions resulting from foreign oil
consumption in its EIS as required by NEPA, or, at least,
explaining thoroughly why it cannot do so and summarizing
the research upon which it relied. We also hold that FWS
violated the ESA by (1) relying upon uncertain, nonbinding
mitigation measures in reaching its no-adverse-effect
conclusion in its biological opinion, and (2) failing to
estimate the Liberty project’s amount of nonlethal take of
polar bears. Because we conclude that FWS’s biological
opinion is flawed and unlawful, we conclude that BOEM’s
reliance on FWS’s opinion is arbitrary and capricious. In all
other respects, we deny the petition for review.

The petition for review is GRANTED in part and
DENIED in part. BOEM’s approval of the Liberty project
is VACATED and this action is REMANDED to the
46 CTR. FOR BIO. DIVERSITY V. ZINKE

agency for further proceedings consistent with this opinion.
CBD shall recover its costs.
 CTR. FOR BIO. DIVERSITY V. ZINKE 47

Appendix

APA Administrative Procedure Act

BOEM Bureau of Ocean Energy Management

BiOp Biological opinion

CBD Center for Biological Diversity

EIS Environmental Impact Statement

ESA Endangered Species Act

FWS U.S. Fish and Wildlife Service

MMPA Marine Mammal Protection Act

NEPA National Environmental Policy Act

NMFS National Marine Fisheries Service

OCSLA Outer Continental Shelf Lands Act