Turtle Island Restoration Network v. United States Department of Commerce

FOR PUBLICATION

UNITED STATES COURT OF APPEALS
FOR THE NINTH CIRCUIT

TURTLE ISLAND RESTORATION No. 13-17123
NETWORK; CENTER FOR BIOLOGICAL
DIVERSITY, D.C. No.
Plaintiffs-Appellants, 1:12-cv-00594-
SOM-RLP
v.

UNITED STATES DEPARTMENT OF OPINION
COMMERCE; NATIONAL MARINE
FISHERIES SERVICE; WILBUR L. ROSS,
in his official capacity as Secretary of
Commerce; U.S. DEPARTMENT OF
THE INTERIOR; U.S. FISH & WILDLIFE
SERVICE; RYAN ZINKE, in his official
capacity as Secretary of the Interior,
Defendants-Appellees,

and

HAWAII LONGLINE ASSOCIATION,
Intervenor-Defendant-Appellee.

Appeal from the United States District Court
for the District of Hawaii
Susan Oki Mollway, Chief District Judge, Presiding

Argued and Submitted June 14, 2016
Honolulu, Hawaii
2 TURTLE ISLAND RESTORATION NETWORK V. USDOC

Filed December 27, 2017

Before: Sidney R. Thomas, Chief Judge, and Consuelo M.
Callahan and Mary H. Murguia, Circuit Judges.

Opinion by Judge Murguia;
Dissent by Judge Callahan

SUMMARY *

Environmental Law

The panel affirmed in part, and reversed in part, the
district court’s judgment in favor of federal agencies in an
action brought by plaintiff environmental groups
challenging the decision of the National Marine Fisheries
Service (“NMFS”) to allow a Hawaii-based swordfish
fishery to increase its fishing efforts, which may result in the
unintentional deaths of endangered sea turtles; and
challenging the decision of the U.S. Fish and Wildlife
Service (“FWS”) to issue a “special purpose” permit to the
NMFS, which authorized the fishery to incidentally kill
migratory birds.

The panel held that the FWS’s decision to issue a special
purpose permit to the NFMS on behalf of a commercial
fishery was arbitrary and capricious. The panel held that the
FWS’s interpretation of 50 C.F.R. § 21.27 as authorizing it
to grant an incidental take permit to the NMFS did not
conform to either the Migratory Bird Treaty Act’s

*
This summary constitutes no part of the opinion of the court. It has
been prepared by court staff for the convenience of the reader.
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 3

conservation intent or the plain language of the regulation.
The panel therefore reversed the district court’s grant of
summary judgment affirming the FWS’s decision to issue
the permit.

The panel held that NMFS’s 2012 Biological Opinion’s
“no jeopardy” finding as to the loggerhead sea turtles was
arbitrary and capricious because the scientific data suggested
that the loggerhead population would significantly decline,
and the agency failed to sufficiently explain the discrepancy
in its opinion and record evidence. Specifically, the panel
held that the climate-based model predicted that the
proposed action would exacerbate the loggerheads’ decline,
and the Biological Opinion was structurally flawed to the
extent the NMFS failed to incorporate those findings into its
jeopardy analysis. The panel therefore reversed the district
court’s grant of summary judgment upholding this portion of
the Biological Opinion.

The panel otherwise affirmed the district court’s grant of
summary judgment to defendants, and remanded. The panel
held that the NMFS’s no jeopardy conclusion regarding the
leatherback turtles found support in the scientific record, and
therefore was sufficient to withstand judicial review.
Specifically, the panel held that it could not conclude that the
2012 Biological Opinion violated the Endangered Species
Act or that the NMFS otherwise acted arbitrarily and
capriciously in determining that the fishery would have no
appreciable effect on the leatherback turtle population. The
panel also held that the NMFS’s consideration of climate
change in the Biological Opinion was neither arbitrary,
capricious, nor contrary to the NMFS’s obligation to base its
jeopardy decision on the best scientific data it could obtain.

Judge Callahan dissented in part. Judge Callahan agreed
with the majority that the 2012 Biological Opinion was not
4 TURTLE ISLAND RESTORATION NETWORK V. USDOC

arbitrary and capricious in determining that the Hawaii-
based shallow-set fishery expansion would have no
appreciable effect on the leatherback sea turtle population,
and that the 2012 Biological Opinion adequately considered
the impact of global climate change; and dissented from the
remainder of the majority opinion. Judge Callahan would
uphold the Migratory Bird Treaty Act Permit and the
loggerhead sea turtle Biological Opinion.

COUNSEL

David L. Henkin (argued) and Paul H. Achitoff, Earthjustice,
Honolulu, Hawaii, for Plaintiffs-Appellants.

Brian C. Toth (argued), Ellen J. Durkee, Dean K. Dunsmore,
and Kristen L. Gustafson, Attorneys; Jeffrey H. Wood,
Acting Assistant Attorney General; Environment & Natural
Resources Division, United States Department of Justice,
Washington, D.C.; Philip Kline, Office of the Solicitor,
United States Department of the Interior, Portland, Oregon;
Elena Onaga, Office of General Counsel, National Oceanic
& Atmospheric Administration, United States Department
of Commerce, Honolulu, Hawaii; for Defendants-Appellees.

Ryan P. Steen (argued) and Jeffrey W. Leppo, Stoel Rives
LLP, Seattle, Washington, for Intervenor-Defendant-
Appellee.
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 5

OPINION

MURGUIA, Circuit Judge:

Plaintiffs Turtle Island Restoration Network and the
Center for Biological Diversity challenge the decision of the
National Marine Fisheries Service (“NMFS”) to allow a
Hawaii-based swordfish fishery to increase its fishing
efforts, which may result in the unintentional deaths of
endangered sea turtles. Plaintiffs also challenge the decision
of the U.S. Fish and Wildlife Service (“FWS”) to issue a
“special purpose” permit to the NMFS, which authorizes the
fishery to incidentally kill migratory birds.

Plaintiffs brought suit against the agencies under various
environmental statutes that the NMFS and the FWS are
charged with administering, including the Magnuson-
Stevens Fishery Conservation and Management Act (the
“Magnuson-Stevens Act”), the Endangered Species Act of
1973 (“ESA”), the Migratory Bird Treaty Act (“MBTA”),
and the National Environmental Policy Act (“NEPA”). The
Hawaii Longline Association subsequently intervened to
represent the interests of the swordfish fishery in defense of
the agencies’ actions. We have jurisdiction under 28 U.S.C.
§ 1291, and we affirm in part, and reverse and remand in
part.

BACKGROUND

I. Regulatory Framework

In response to concerns about overfishing, Congress
enacted the Magnuson-Stevens Act to promote the long-term
biological and economic sustainability of marine fisheries in
U.S. federal waters. See 16 U.S.C. § 1801(b). Under this Act,
the NMFS and eight regional councils develop
6 TURTLE ISLAND RESTORATION NETWORK V. USDOC

“management plans” for the nation’s fisheries, which the
Secretary of Commerce may approve, partially approve, or
reject. Id. §§ 1801(b)(4), 1852(h)(1), 1854(a)(3). The
Magnuson-Stevens Act demands that a management plan be
consistent with the national standards set out in the Act and
“any other applicable law,” id. § 1853(a)(1)(C), including
the ESA, id. §§ 1531–43, and the MBTA, id. §§ 703–12.

The ESA provides for the conservation of fish, wildlife,
and plant species that are at risk of extinction by requiring
federal agencies to ensure that actions they authorize, fund,
or carry out are “not likely to jeopardize the continued
existence” of any ESA-listed species. 16 U.S.C.
§ 1536(a)(2). Agencies proposing actions that may affect an
ESA-listed species must consult with either the NMFS or the
FWS—depending on the species involved—which then
reviews the proposed action and prepares a “biological
opinion” (“BiOp”) that evaluates whether and the extent to
which the action may impact the species. Id. § 1536(b);
50 C.F.R. § 402.12. If the NMFS or the FWS finds that the
proposed action would not jeopardize any species’ continued
existence, it issues a statement permitting the “taking” of a
particular number of protected animals “if such taking is
incidental to, and not the purpose of, the carrying out of an
otherwise lawful activity.” 16 U.S.C. § 1539(a)(1)(B).

The FWS also has authority to enforce the MBTA, id.
§§ 703–12; 50 C.F.R. § 10.1, which strictly prohibits the
taking of any migratory bird the Act protects except under
the terms of a valid permit issued by the Secretary of the
Interior, id. § 703(a). The Secretary of the Interior has issued
regulations authorizing various types of exemptions to the
MBTA permitting the taking of migratory birds under
certain circumstances. See 16 U.S.C. § 704(a).
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 7

In addition to the substantive mandates of the ESA and
the MBTA, both the NMFS and the FWS are subject to
NEPA’s procedural requirements. See Robertson v. Methow
Valley Citizens Council, 490 U.S. 332, 348 (1989). NEPA is
concerned with process alone and “merely prohibits
uninformed—rather than unwise—agency action.” Id. at
351. NEPA requires federal agencies to prepare
environmental impact statements (“EIS”) detailing the
effects of any proposed action that stands to have a
significant impact on the environment. See 42 U.S.C.
§ 4332(C); Robertson, 490 U.S. at 350. An agency may also
prepare an environmental assessment (“EA”) to determine
whether an EIS is needed. 40 C.F.R. §§ 1501.4(b),
1508.9(a)(1); Te-Moak Tribe of W. Shoshone of Nev. v. U.S.
Dep’t of Interior, 608 F.3d 592, 599 (9th Cir. 2010). If the
EA shows that the proposed action may significantly affect
the environment, then the agency must prepare a full EIS. W.
Watersheds Project v. Abbey, 719 F.3d 1035, 1050 (9th Cir.
2013). Otherwise, the agency issues a finding of no
significant impact and the proposed action can proceed
without further study. Id.

II. The Hawaii-Based Longline Fishing Industry

“Longline” fishing is a commercial fishing method that
involves reeling out—or “setting”—a single, horizontal
mainline to which shorter “branchlines” are attached at
intervals. Each dangling branchline carries baited hooks. A
typical longline set can use several hundred or thousand
individual hooks, allowing a single fishing vessel to spread
its efforts over a large area. While the mainline is in the
water, the fishing equipment often ensnares birds, sea turtles,
and other marine wildlife in addition to the target fish. This
incidental taking of non-target animals is known as
“bycatch.”
8 TURTLE ISLAND RESTORATION NETWORK V. USDOC

The NMFS collects bycatch statistics by tracking the
number of times a non-target animal is hooked or entangled
by fishing gear. The most commonly observed non-target
animal interactions are with Northern Pacific loggerhead and
leatherback sea turtles, both of which are currently listed
under the ESA as “endangered.” See 50 C.F.R. § 17.11. In
addition, several types of albatross interact often with the
longline fisheries, including the black-footed albatross and
the Laysan albatross.

There are two separately regulated longline fisheries
based out of Hawaii: the deep-set fishery—which targets
tuna—and the shallow-set fishery, which targets swordfish.
The two fisheries are managed by the Fishery Ecosystem
Plan for Pelagic Fisheries of the Western Pacific Region
(“Pelagics FMP”), developed by the Western Pacific Fishery
Management Council (“Council”) in accordance with the
Magnuson-Stevens Act and implemented by the NMFS. In
2001, the shallow-set fishery was closed by court order due
to the NMFS’s failure to prepare an EIS analyzing the impact
of longline fishing on the sea turtle population, which the
court found was a violation of the agency’s NEPA
obligations. See Leatherback Sea Turtle v. Nat’l Marine
Fisheries Serv., No. 99-00152, 1999 WL 33594329 (D.
Haw. Oct. 18, 1999). In response, the NMFS issued an EIS
and a BiOp in which the agency concluded that the shallow-
set fishery was adversely affecting several species of sea
turtles. In 2002, the NMFS issued regulations prohibiting all
Hawaii-based swordfish longlining.

The Council subsequently developed various measures
to minimize turtle bycatch, and in 2004 the NMFS
reauthorized shallow-set longlining subject to new
restrictions designed to reduce the number and severity of
interactions between protected turtles and fishing gear. In
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 9

part, the NMFS strictly limited the number of interactions
the fishery could have with leatherback and loggerhead sea
turtles to a maximum of 16 and 17, respectively, per fishing
season. Further, the NMFS imposed an annual limit of 2,120
shallow sets, which represents fifty percent of the average
number of sets deployed prior to the fishery’s closure in
2001.

In 2008, the NMFS proposed an amendment to the
Pelagics FMP (“Amendment 18”) that would remove the
2,120 annual set limit, allowing gear deployments to
increase to their pre-2001 maximums, and also increase the
number of sea turtle interactions allowed each year. After
consulting internally pursuant to the ESA, the NMFS
produced a BiOp concluding that Amendment 18 would not
jeopardize the sea turtles. The NMFS issued a final rule
implementing Amendment 18 in December 2009. 74 Fed.
Reg. 65,640 (Dec. 10, 2009).

Plaintiffs initiated suit against the NMFS on the grounds
that the 2009 rule violated the ESA and the MBTA. See
Turtle Island Restoration Network v. U.S. Dep’t of
Commerce, 834 F. Supp. 2d 1004, 1007 (D. Haw. 2011).
Plaintiffs’ MBTA claim was based on the fishery’s
incidental take of migratory seabirds without an MBTA
permit. The parties settled the case, and the NMFS entered
into a consent decree that required it to withdraw its no
jeopardy BiOp, reinstate the 2004 annual turtle-interaction
caps, and issue a new BiOp after deciding whether to
reclassify various population segments of sea turtles under
the ESA. Id. at 1023–25. The other remaining provisions of
the 2009 rule remained in effect, including the removal of
annual set limits.

The NMFS later proposed raising the shallow-set
fishery’s annual turtle interaction cap to 26 (with
10 TURTLE ISLAND RESTORATION NETWORK V. USDOC

leatherbacks), and 34 (with loggerheads) and otherwise
continuing to operate the fishery in accordance with the
provisions of Amendment 18 to the Pelagics FMP. In
January 2012, the NMFS issued a new BiOp concluding that
the shallow-set fishery would not jeopardize the continued
existence of either the loggerhead or leatherback turtles if it
operated under higher caps on turtle interactions.

While it was engaged in the re-consultation process, the
NMFS submitted an application to the FWS for a special
purpose permit that would allow the shallow-set fishery to
take migratory seabirds in connection with swordfish
longlining. The FWS issued a final EA in which it
considered denying the permit, granting the permit as
requested, and granting the permit while requiring the
NMFS to conduct new research on additional ways to avoid
seabird interactions. See 77 Fed. Reg. 1501 (Jan. 10, 2012).
The FWS ultimately concluded that none of the alternatives
would have a significant adverse impact on the seabirds’
population levels. Accordingly, the FWS issued a finding of
“no significant impact.” In August 2012, the FWS granted a
three-year special purpose permit authorizing the shallow-
set fishery to kill a maximum of 191 black-footed albatross,
430 Laysan albatross, 30 northern fulmars, 30 sooty
shearwaters, and one short-tailed albatross. Of those birds,
only the short-tailed albatross is listed under the ESA,
50 C.F.R. § 17.11(h).

Plaintiffs subsequently filed this lawsuit under the ESA,
the MBTA, and their implementing regulations, challenging
the NMFS’s final rule approving the continued operation of
the shallow-set fishery and the FWS’s issuance of a
migratory bird permit to the NMFS. After the parties moved
for summary judgment, the district court ruled in the
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 11

agencies’ favor on all of Plaintiffs’ claims. Plaintiffs timely
appealed.

STANDARD OF REVIEW

We review challenges to final agency action decided on
summary judgment de novo and pursuant to Section 706 of
the Administrative Procedure Act (“APA”). Turtle Island
Restoration Network v. Nat’l Marine Fisheries Serv.,
340 F.3d 969, 973 (9th Cir. 2003). Review is based on the
administrative record. Camp v. Pitts, 411 U.S. 138, 142
(1973).

The APA requires courts to “hold unlawful and set aside
agency action, findings, and conclusions found to be . . .
arbitrary, capricious, an abuse of discretion, or otherwise not
in accordance with law,” “in excess of statutory
jurisdiction,” or “without observance of procedure required
by law.” 5 U.S.C. § 706(2)(A), (C)–(D). “The scope of
review under the ‘arbitrary and capricious’ standard is
narrow and a court is not to substitute its judgment for that
of the agency.” Motor Vehicle Mfrs. Ass’n of U.S., Inc. v.
State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983).
Nevertheless, we require the agency to “examine the
relevant data and articulate a satisfactory explanation for its
action,” and we will strike down agency action as “arbitrary
and capricious if the agency has relied on factors which
Congress has not intended it to consider, entirely failed to
consider an important aspect of the problem, offered an
explanation for its decision that runs counter to the evidence
before the agency,” or if the agency’s decision “is so
implausible that it could not be ascribed to a difference in
view or the product of agency expertise.” Id.

Separate from the APA, we also give deference to an
agency’s interpretation of the statutes and regulations that
12 TURTLE ISLAND RESTORATION NETWORK V. USDOC

define the scope of its authority. Chevron, U.S.A., Inc. v.
Natural Resources Defense Council, Inc. compels us to defer
to an agency’s reasonable interpretation of its enabling
legislation. 467 U.S. 837, 843 (1984). Under the Chevron
analysis, we must first exhaust the traditional tools of
statutory construction to determine whether Congress has
“directly spoken to the precise question at issue.” Id. at 842.
If we determine that the statute is silent or ambiguous on the
question at hand, then at Chevron step two we must respect
the agency’s interpretation so long as it “is based on a
permissible construction of the statute.” Id. at 843. A
permissible construction is one that is not “arbitrary,
capricious, or manifestly contrary to the statute.” Id. at 844;
see also Judulang v. Holder, 132 S. Ct. 476, 483 n.7 (2011)
(recognizing that Chevron step two is equivalent to the
APA’s arbitrary and capricious standard).

Chevron deference applies only to agency decisions
rendered through formal procedures. United States v. Mead
Corp., 533 U.S. 218, 226–27 (2001). However, under Auer
v. Robbins, we must also defer to an agency’s interpretation
of its own ambiguous regulations, which controls unless
“plainly erroneous or inconsistent with the regulation,” or
where there are grounds to believe that the interpretation
“does not reflect the agency’s fair and considered judgment
of the matter in question.” Christopher v. SmithKline
Beecham Corp., 132 S. Ct. 2156, 2159 (2012) (quoting Auer
v. Robbins, 519 U.S. 452, 461–62 (1997)). Similarly, “we
must ensure that the interpretation is not inconsistent with a
congressional directive; a court need not accept an agency’s
interpretation of its own regulations if that interpretation is
inconsistent with the statute under which the regulations
were promulgated.” Marsh v. J. Alexander’s LLC, 869 F.3d
1108, 1116–17 (9th Cir. 2017) (internal changes, quotation
marks and citations omitted). Our review of an agency’s
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 13

construction of a statute or regulation that does not qualify
for either Chevron or Auer deference is de novo, although
we may still accord the agency’s opinion some weight.
Satterfield v. Simon & Schuster, Inc., 569 F.3d 946, 952–53
(9th Cir. 2009) (citing Skidmore v. Swift & Co., 323 U.S.
134, 140 (1944)).

DISCUSSION

I. “Special Purpose” Permit

Plaintiffs argue that the FWS acted arbitrarily and
capriciously by issuing a special purpose permit to the
NMFS on behalf of a commercial operation—the shallow-
set fishery—that provides no benefit to migratory birds.
Plaintiffs specifically contend that, in issuing this permit, the
FWS ignored or violated its obligations under the MBTA.

The MBTA is a strict liability criminal statute that
Congress enacted for the “object and purpose . . . to aid in
the restoration of [game and other wild] birds.” 16 U.S.C.
§ 701. The MBTA states in expansive language that, unless
otherwise permitted by the Secretary of the Interior, “it shall
be unlawful at any time, by any means or in any manner, to
pursue, hunt, take, capture, kill, [or] attempt to take, capture,
or kill . . . any migratory bird.” 16 U.S.C. § 703(a). The
MBTA also limits the FWS’s authority to authorize the
killing of migratory birds absent specified regulations
“[s]ubject to the provisions and in order to carry out the
purposes of the conventions” underlying the Act. Id.
§ 704(a). The conventions underlying the MBTA stipulate
that migratory birds may only be killed under “extraordinary
conditions,” where birds have “become seriously injurious
to the agricultural or other interests in any particular
community.” Humane Soc’y of the U.S. v. Glickman,
14 TURTLE ISLAND RESTORATION NETWORK V. USDOC

217 F.3d 882, 885 (D.C. Cir. 2000) (internal quotation marks
omitted).

Pursuant to the MBTA, the FWS has enacted a
permitting program for narrow categories of migratory bird
takings, such as scientific collecting, rehabilitation, hunting,
and depredation control. See 16 U.S.C. §§ 704(a), 712(2)
(empowering the FWS to promulgate implementing
regulations); 50 C.F.R. §§ 21.21–21.61 (authorizing the
issuance of various types of permits). The FWS has also
established a “special purpose” permit that allows a person
to “lawfully take . . . migratory birds . . . for any purpose not
covered by the standard form permits” included elsewhere
in the regulations. 50 C.F.R. § 21.27(a). The FWS may issue
such a permit for “special purpose activities related to
migratory birds,” where the applicant “makes a sufficient
showing” that the activity would be “of benefit to the
migratory bird resource, important research reasons, reasons
of human concern for individual birds, or other compelling
justification.” Id.

Here, the FWS interpreted § 21.27 as authorizing it to
grant a special purpose permit sanctioning the incidental take
of migratory birds to the NMFS, thereby allowing a
commercial activity—longline fishing—that does not
concern bird conservation. In its decision to issue the permit,
the FWS found that the “commercial fishery carries no
intrinsic benefit for migratory bird resources,” “the take that
occurs is neither directed by, nor is the result of, important
research,” and that “the take that occurs does not result from
concern for individual birds.” However, the FWS found that
“compelling justification” existed to permit the continued
operation of the shallow-set fishery, which the FWS
believed “provides a net benefit to the Nation” economically
and “serves as a benchmark internationally for employing
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 15

effective seabird mitigation techniques and serves as an
example of responsible conservation practices by a fishery.”
The FWS also noted that “[c]losure of this fishery would
likely result in replaced effort by foreign longline fleets to
supply swordfish demand, where use of bycatch mitigation
methods would not likely follow international best
practices.”

We conclude that the FWS’s decision to issue a special
purpose permit to the NMFS on behalf of a commercial
fishery was arbitrary and capricious. Although the FWS’s
interpretation of § 21.27 would ordinarily deserve deference,
see Mead, 533 U.S. at 226–27, we cannot conclude that such
deference is appropriate in this case. Deference to the FWS’s
interpretation is not warranted because the plain language of
this regulation is not reasonably susceptible to the FWS’s
new interpretation. The other “standard form permits” the
MBTA regulations authorize govern discrete types of
takings, such as scientific collecting, taxidermy, and
rehabilitation, and although § 21.27 is intended to allow the
FWS to authorize activities not otherwise permitted by the
regulations, it is still a narrow exception to the MBTA’s
general prohibition on killing migratory birds. See Marsh,
869 F.3d at 1116–17 (“[W]e must always ensure that the
interpretation is not inconsistent with a congressional
directive . . . .”); Ctr. for Biological Diversity v. Salazar,
706 F.3d 1085, 1092 (9th Cir. 2013) (“[W]e must interpret
[a] regulation as a whole, in light of the overall statutory and
regulatory scheme . . . .” (internal quotation marks
omitted)). The FWS’s construction of § 21.27’s “special
purpose activit[y]” exception as applying to basic
commercial activities like fishing that have no articulable
“special purpose” is therefore inconsistent with the existing
permitting scheme that the FWS has enacted. The FWS must
read the “special purpose” provision in the context of the
16 TURTLE ISLAND RESTORATION NETWORK V. USDOC

regulation’s other requirements that, taken together, fail to
turn § 21.27 into a general incidental take exception: the
permit must “relate[] to migratory birds” and may issue only
upon a “sufficient showing of . . . [a] compelling
justification.” 50 C.F.R. § 21.27.

The FWS unpersuasively argues that the phrase “related
to migratory birds” is not a restriction on its permitting
authority, but merely a description of what can be permitted.
The FWS specifically maintains that longline fishing is
“related to migratory birds” because it incidentally interacts
with them. Although nothing in the regulation requires that
the permitted activity directly concern migratory birds, it
nevertheless strains reason to say that every activity that
risks killing migratory birds “relate[s] to” those birds. See
50 C.F.R. § 21.27. The FWS’s approach to the regulation
renders the majority of its text superfluous. See Nat’l Ass’n
of Home Builders v. Defs. of Wildlife, 551 U.S. 644, 669
(2007) (cautioning against reading an agency regulation in a
way that renders part of it redundant).

The FWS’s interpretation of § 21.27 as authorizing it to
grant an incidental take permit to the NMFS does not
conform to either the MBTA’s conservation intent or the
plain language of the regulation. We therefore conclude that
the FWS’s grant of a special purpose permit to the NMFS
was arbitrary and capricious. 1

1
Because we conclude that the FWS acted arbitrarily and capriciously
in issuing the incidental take permit to the NMFS under § 21.27, we need
not reach Plaintiffs’ additional argument concerning whether the FWS’s
action also violated NEPA.
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 17

II. 2012 “No Jeopardy” BiOp

Plaintiffs also argue that the NMFS violated the ESA by
failing to properly assess the shallow-set fishery’s impacts
on endangered sea turtles. The ESA permits federal agencies
to authorize actions that will result in the taking of
endangered or threatened species only if the projected take
“is not likely to jeopardize the continued existence of” any
listed species. 16 U.S.C. § 1536(a)(2). “Jeopardize the
continued existence of means to engage in an action that
reasonably would be expected, directly or indirectly, to
reduce appreciably the likelihood of both the survival and
recovery of a listed species in the wild by reducing the
reproduction, numbers, or distribution of that species.”
50 C.F.R. § 402.02 (emphasis added).

Where listed marine species are concerned, the NMFS
prepares a BiOp evaluating the effects of the proposed action
on the survival and recovery of the listed species. 16 U.S.C.
§ 1536(c). The agency specifically considers the proposed
action’s direct, indirect, and cumulative effects on a listed
species in relation to the environmental baseline, and opines
on whether the action is likely to jeopardize the species’
survival. 50 C.F.R. § 402.14(g)(4); see also Nat’l Wildlife
Fed’n v. Nat’l Marine Fisheries Serv., 524 F.3d 917, 924
(9th Cir. 2008). Where a species is already in peril, an
agency may not take an action that will cause an “active
change of status” for the worse. Nat’l Wildlife Fed’n,
524 F.3d at 930.

When formulating a BiOp, the NMFS must base its
conclusions on evidence supported by “the best scientific
and commercial data available.” 16 U.S.C. § 1536(a)(2);
50 C.F.R. § 402.14(g)(8). This requirement “prohibits [an
agency] from disregarding available scientific evidence that
is in some way better than the evidence [it] relies on.” San
18 TURTLE ISLAND RESTORATION NETWORK V. USDOC

Luis & Delta-Mendota Water Auth. v. Jewell, 747 F.3d 581,
602 (9th Cir. 2014) (citation and internal quotation mark
omitted). “The determination of what constitutes the ‘best
scientific data available’ belongs to the agency’s ‘special
expertise’ . . . .” Id. (emphasis in original) (citation omitted).

In 2012, the NMFS issued a BiOp concluding that the
removal of the annual limit of 2,120 shallow-set lines in the
fishery might result in the incidental “take” of Northern
Pacific loggerhead and leatherback sea turtles, but would not
jeopardize the continued existence of either species for the
next 25 years. To establish the environmental baseline, the
NMFS used existing studies on loggerhead and leatherback
interactions with all Pacific longline fisheries (domestic and
international) from 2000 to 2009. The NMFS ultimately
found that the Hawaii-based shallow-set fishery is currently
responsible for killing two to three loggerheads and
leatherbacks (each) per year. The NMFS also determined
that the impacts associated with anthropogenic climate
change were likely beginning to affect both sea turtle
species, but lacked sufficient data to quantify the threat that
climate change posed to the turtles.

The NMFS then attempted to predict the impact that
allowing the fishery to deploy 5,500 longline sets per year—
the approximate maximum annual number of sets before the
fishery was first closed out of concern for the sea turtle
populations—would have on the loggerheads and
leatherbacks. The NMFS ultimately projected that setting
5,500 lines would kill no more than one adult, female
loggerhead turtle and four adult, female leatherback turtles.
The NMFS then employed population viability assessment
models to forecast the risk that killing small numbers of
adult, female sea turtles would lead to the species’
extinction. The NMFS concluded from the results that the
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 19

proposed action could not reasonably be expected to
appreciably reduce the likelihood of survival of either the
loggerhead or the leatherback turtles.

The NMFS’s “no jeopardy” conclusion was not affected
by the agency’s consideration of the cumulative effects of
worsening climates. And, the NMFS’s analysis of
“spillover” trends suggested that the proposed increase in
Hawaii-based swordfishing would benefit sea turtles overall.
Because domestic fisheries operate under more stringent
conservation measures than foreign fleets that compete to
provide swordfish to U.S. consumers, the NMFS predicted
that increasing domestic fishery yields would displace
foreign fishing activities in the same area that the Hawaii-
based shallow-set fishery operates, resulting in a net
decrease in mortalities for the affected turtle species.
However, because the NMFS concluded that the projected
decrease in turtle deaths from this “spillover” effect was not
precise enough to incorporate into its population assessment
models, the NMFS did not incorporate these benefits into its
no jeopardy finding.

A. Population Viability Assessment Models

Plaintiffs argue that the 2012 BiOp’s conclusion that the
proposed action would not appreciably impact loggerhead
and leatherback sea turtles is unsupported by the scientific
methods the FWS relied on. To project the impact of the
shallow-set fishery’s operations on the sea turtle species’
likelihood of survival, the NMFS ran a climate-based
population forecast model and relied primarily on the results
of this model, “along with inputs from multiple experts and
sources, where available.” The climate-based model showed
a significant decline in loggerhead numbers over the next
generation even without the proposed action of removing the
fishery’s set limits: 99.5% of the tests showed the loggerhead
20 TURTLE ISLAND RESTORATION NETWORK V. USDOC

falling below the quasi-extinction threshold within 25 years.
When the model was run incorporating the anticipated
mortality associated with the fishery’s operations without set
limits, the results were similar. The NMFS specifically
found that “[v]irtually all the loggerhead climate model runs
. . . indicat[ed] high extinction risk with high model
confidence.” The additional loss to the loggerhead
population from the proposed action ranged from 4 to 11%.
As for the leatherback turtles, the climate-based model
showed an increase in leatherback population over the next
generation without a change in the fishery’s set limits, and
even with the proposed action the “extinction risk
remain[ed] in the low category,” although the results
predicted a “measurable loss to the population” of 16 to
30%.

Based on the results from the model, the NMFS decided
that it did not “believe that the small effect posed by the
lethal takes in this fishery, when considered together with
the environmental baseline and the cumulative effects, will
be detectable or appreciable” and “that the additional risk to
the [loggerhead turtles] that would result from loss of one
adult female annually is considered negligible.” Similarly,
the NMFS concluded “that the proposed action would have
a negligible impact on the risk to . . . the western Pacific
leatherback population as a whole.” Therefore, the NMFS
opined that increasing the maximum annual number of sets
at the fishery would not jeopardize either species.

1. Loggerhead Turtles

With respect to the loggerhead turtles, the NMFS
violated the APA’s requirement that the agency articulate a
rational connection between the population viability model
upon which the NMFS relied and its no jeopardy conclusion.
The BiOp acknowledged that the climate-based model
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 21

predicted a decline in loggerhead populations to a level that
“represents a heightened risk of extinction,” but still upheld
a finding of “no jeopardy” on the grounds that there was
“little to no difference in the extinction risk when the annual
removal of one adult female loggerhead resulting from the
proposed action is considered in the model.” We rejected
similar logic in National Wildlife Federation, holding that
“where baseline conditions already jeopardize a species, an
agency may not take action that deepens the jeopardy by
causing additional harm.” 524 F.3d at 930 (noting that listed
species’ “slow slide into oblivion is one of the very ills the
ESA seeks to prevent”). In National Wildlife Federation, the
NMFS had prepared a BiOp in which it determined that
hydropower dam operations would not jeopardize threatened
and endangered salmon populations. Id. at 925. NMFS,
however, had already determined that baseline
environmental conditions posed a risk of jeopardy to the
species. Id. Therefore, to reach a conclusion of “no
jeopardy,” the agency completely excluded from the
environmental baseline all impacts from “nondiscretionary”
federal activities such as operations relating to irrigation,
flood control, and power generation. We held that this
exclusion was improper and that baseline conditions must be
factored into the jeopardy analysis, cumulatively with the
entirety of agency actions. The relevant inquiry is therefore
whether the “action effects, when added to the underlying
baseline conditions,” are such that they would cause
jeopardy. Id. at 929.

Here, the NMFS improperly minimized the risk of
bycatch to the loggerheads’ survival by only comparing the
effects of the fishery against the baseline conditions that
have already contributed to the turtles’ decline. The BiOp’s
no jeopardy opinion is premised on the proportionally low
risk that the shallow-set fishery poses to the loggerheads
22 TURTLE ISLAND RESTORATION NETWORK V. USDOC

relative to other threats, such as international fishing and
climate change: the NMFS specifically found that although
“any level of take and mortality can have an adverse effect
on the overlying population . . . the expected level of take
from the action, including a small number of mortalities, is
extremely small when considered together with all impacts
considered in the Status of the Species, Baseline and
Cumulative Effects sections, including other federally
authorized fisheries and foreign fisheries.” As in National
Wildlife Federation, the agency reached an arbitrary
conclusion by only comparing the prospective harm to the
loggerheads that is attributable to the proposed action—the
death of a single adult, female loggerhead per year—to the
much greater harm resulting from factors beyond the fishery.
Based on this impermissible comparison, the agency
concluded that the proposed action’s adverse impacts would
not appreciably reduce the loggerheads’ likelihood of
survival. See Nat’l Wildlife Fed’n, 524 F.3d at 930.

The NMFS relies heavily on the conservative nature of
its calculations to support the difference between its
conclusion and the climate-based model’s results. The
NMFS asserts that it rounded up its calculation of maximum
adult female mortality, modeled the viability of turtle
populations using the maximum potential number of annual
interactions opposed to the average number of interactions
reported in previous years, and estimated the number of sea
turtle deaths based on the assumption that the shallow-set
fishery would immediately operate at 5,500 sets each year.
In reality, the increase in sets is expected to be gradual over
many years. The ESA, however, requires agencies to
rigorously ensure their actions will not “tip [the loggerhead]
species from a state of precarious survival into a state of
likely extinction.” See Nat’l Wildlife Fed’n, 524 F.3d at 930.
The agency may not reject the “best scientific data” in favor
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 23

of its belief that “incidental take . . . would be reduced to the
best extent possible” and “the vast majority of the
loggerhead sea turtle takes from the proposed action are
expected to be non-lethal.”

The NMFS also notes that the climate-based model used
an assumed fraction of the current turtle population size
(50%) as a proxy for extinction, and explains that
“population decline below that” number “does not
necessarily mean that” the species is “unrecoverable” or
“functionally extinct.” But, given the agency’s endorsement
of the climate-based model and its expert’s decision to use a
“quasi-extinction threshold” to reflect a decline in the turtle
population to numbers insufficient to ensure the population’s
viability, this logic does not support the NMFS’s
determination that the projected population declines would
not appreciably threaten the loggerheads’ survival.

Another rationale presented in the BiOp is that “spillover
effect is reasonably certain to contribute to a reduction in
loggerhead mortalities . . . due to reduced effort in foreign
fisheries.” Shortly thereafter, however, the NMFS noted that
data on foreign fishery bycatch are “likely incomplete or
inaccurate.” The NMFS went on to state that “mortality
reduction data associated with spillover effects are not as
robust as those analyzed for the direct effects of the proposed
action.” For those reasons, the NMFS did not incorporate the
estimated sea turtle mortalities that would be avoided due to
a potential spillover effect into its population assessment
models.

The NMFS’s model showed the loggerhead species are
on a path toward extinction, which accords with the fact that
the NMFS recently raised the Pacific loggerhead’s ESA
listing from “threatened” to “endangered.” The NMFS also
found that “effects” to the loggerhead “are likely to occur as
24 TURTLE ISLAND RESTORATION NETWORK V. USDOC

a result of worsening climate change,” which the NMFS
“expect[s] to continue and therefore may impact sea turtles
and their habitats in the future.” Rising levels of marine
debris “could also increase entanglements.” Even though the
NMFS was unable to quantify the risks of climate change
and its associated impacts, the agency recognized that they
would be detrimental to the loggerheads.

The climate-based model predicted that the proposed
action would exacerbate the loggerheads’ decline, and the
BiOp is structurally flawed to the extent the NMFS failed to
incorporate those findings into its jeopardy analysis. Nat’l
Wildlife Fed’n, 524 F.3d at 927. Because the NMFS has not
articulated a rational connection between the best available
science and its conclusion that the loggerhead sea turtles
would not be affected by the increased fishing efforts, the
agency’s determination that the loggerhead “population will
remain large enough to retain the potential for recovery” is
arbitrary and capricious.

2. Leatherback Turtles

Plaintiffs also argue that the 2012 BiOp improperly
concluded that the fishery would have no appreciable impact
on the leatherback turtle population. Unlike its conclusion
concerning the loggerheads, however, the NMFS’s no
jeopardy conclusion regarding the leatherback turtles finds
support in the scientific record and, therefore, is sufficient to
withstand judicial review.

Plaintiffs specifically argue that the NMFS erred in
limiting the “temporal scale” of its analysis to 25 years,
despite the fact that the fishery’s operations have no related
limitation and the NMFS determined that impacts on the sea
turtles due to increasing temperatures “are expected to occur
slowly over the next century.” However, the NMFS was
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 25

entitled to rely on the climate-based population assessment
model, even though that model could only predict changes
in the turtle population for 25 years. See San Luis & Delta-
Mendota Water Auth. v. Locke, 776 F.3d 971, 997 (9th Cir.
2014) (“[T]he agency has substantial discretion to choose
between available scientific models, provided that it explains
its choice.”); The Lands Council v. McNair, 537 F.3d 981,
988 (9th Cir. 2008) (explaining that the court may not “act
as a panel of scientists that instructs the [agency] how to . . .
choose[] among scientific studies”), overruled on other
grounds by Winter v. Nat. Res. Def. Council, Inc., 555 U.S.
7 (2008). The constraints in the available data supply a
reasonable justification for the NMFS to limit its analysis.
Accordingly, we cannot conclude that the 2012 BiOp
violated the ESA or that the NMFS otherwise acted
arbitrarily and capriciously in determining that the fishery
would have no appreciable effect on the leatherback turtle
population.

B. Consideration of the Effects of Climate Change

Lastly, Plaintiffs argue that the 2012 BiOp failed to
evaluate the impacts of global climate change. Plaintiffs
specifically maintain that the NMFS acted arbitrarily by
dismissing the effects of global warming on sea turtles as
uncertain without further study.

In the 2012 BiOp, the NMFS explained that the effects
from climate change on listed turtle species include rising
sand temperatures and sea levels, beach erosion, increased
storm activity, and changes in ocean temperature and
chemistry. The BiOp also summarized studies anticipating
that climate change will impact, among other traits and
behaviors, turtle gender ratios, nesting habitat, and
reproductive capacity. However, the NMFS determined that
there was no available data from which it could credibly
26 TURTLE ISLAND RESTORATION NETWORK V. USDOC

project the impacts that climate change would have on the
loggerhead or leatherback turtle survival rates. With respect
to the loggerhead turtles, the NMFS explained that “current
scientific methods are not able to reliably predict the future
magnitude of climate change and associated impacts or the
adaptive capacity of this species.” The NMFS also stated that
“leatherbacks are probably already beginning to be affected
by impacts associated with anthropogenic climate change in
several ways,” but noted that it did “not have information to
predict what the population would do” or “what impact other
climate-related changes may have such as increasing sand
temperatures, sea level rise, and increased storm events.” As
the NMFS observed elsewhere in the BiOp, the effects of
climate change will not be globally uniform, and the
uncertainty of the rate, magnitude, and distribution of such
effects on different temporal and spatial scales—not to
mention the turtles’ ability to adapt to these effects—have
not been comprehensively studied. Consequently, the NMFS
decided that climate change effects could not be “reliably
quantified” nor “qualitatively described or predicted” by the
agency at the time.

Here, we cannot conclude from the NMFS’s lack of
precision that it failed to adequately consider the effects of
climate change on the sea turtles. On the whole, the BiOp
demonstrated that the NMFS considered a variety of ways in
which climate change may affect the sea turtles, but simply
concluded that the data available was too indeterminate for
the agency to evaluate the potential sea-turtle impacts with
any certainty. Cf. Greenpeace Action v. Franklin, 14 F.3d
1324, 1326–27, 1336 (9th Cir. 1993) (holding that the
agency’s no jeopardy conclusion was not arbitrary because
the BiOp at issue demonstrated that the agency had based its
no jeopardy decision on the best available scientific data,
even though the data was “uncertain”); Stop H-3 Ass’n v.
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 27

Dole, 740 F.2d 1442, 1460 (9th Cir. 1984) (sustaining a
BiOp that stated “we have very little data for providing an
opinion, but feel it would be unreasonable to request [an
additional] study which would be unlikely to provide
definitive results. . . . Based on the available information,
which we grant is weak, it is our opinion the proposed
project is not likely to jeopardize the continued existence of
the Oahu Creeper”). Plaintiffs have failed to sufficiently
refute the NMFS’s stated inability to offer more specific
predictions on the effects of climate change, and they have
not alleged that less speculative scientific information is
available that the agency overlooked. San Luis & Delta-
Mendota, 747 F.3d at 602 (“[W]here [superior] information
is not readily available, we cannot insist on perfection: [T]he
‘best scientific . . . data available,’ does not mean the best
scientific data possible.” (citation and internal quotation
marks omitted)). Accordingly, the NMFS’s consideration of
climate change in the BiOp was neither arbitrary, capricious,
nor contrary to the NMFS’s obligation to base its jeopardy
decision on the best scientific data it could obtain. See
16 U.S.C. § 1536(a)(2).

CONCLUSION

We conclude that the FWS’s grant of an incidental take
permit to the NMFS in reliance on the “special purpose
permit” provision in 50 C.F.R. § 21.27 was arbitrary and
capricious because the FWS’s interpretation of § 21.27 does
not conform to either the MBTA’s conservation intent or the
plain language of the regulation. We therefore reverse the
district court’s grant of summary judgment affirming the
FWS’s decision to issue the permit.

We also conclude that NMFS’s 2012 BiOp’s no jeopardy
finding as to the loggerhead sea turtles was arbitrary and
capricious because the scientific data suggested that the
28 TURTLE ISLAND RESTORATION NETWORK V. USDOC

loggerhead population would significantly decline, and the
agency failed to sufficiently explain the discrepancy in its
opinion and the record evidence. We therefore reverse the
district court’s grant of summary judgment upholding this
portion of the BiOp. We otherwise affirm the district court’s
grant of summary judgment to Defendants.

AFFIRMED in part; REVERSED in part; and
REMANDED. Each party shall bear its own costs on
appeal.

CALLAHAN, Circuit Judge, dissenting in part:

I agree with the majority that the 2012 Biological
Opinion (“BiOp”) is not arbitrary and capricious in
determining that the Hawaii-based shallow-set fishery
expansion would have no appreciable effect on the
leatherback sea turtle population, and that the 2012 BiOp
adequately considers the impact of global climate change.
However, I dissent from the remainder of the majority
opinion.

First, the majority errs in rejecting the U.S. Fish and
Wildlife Service’s (“FWS”) issuance of a special purpose
permit (the “Permit”) under the Migratory Bird Treaty Act
(“MBTA”) to the National Marine Fisheries Service
(“NMFS”) for the incidental take of migratory birds. The
majority determines that issuing the Permit runs afoul of the
pertinent regulation’s plain language and the MBTA’s
conservation-oriented purpose. That conclusion, however,
reflects a misapplication of our deferential standard of
review under Auer v. Robbins, 519 U.S. 452 (1997), because
both the regulation—50 C.F.R. § 21.27—and the MBTA
itself accommodate FWS’s view. See Auer, 519 U.S. at 461;
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 29

Marsh v. J. Alexander’s LLC, 869 F.3d 1108, 1116–17 (9th
Cir. 2017). Moreover, the Permit accords with FWS’s past
practice, and thereby reflects its considered judgment—
another basis for granting deference under Auer.
Christopher v. SmithKline, 132 S. Ct. 2156, 2166 (2012).

Second, the majority errs in rejecting the 2012 BiOp’s
assessment of the proposed shallow-set fishery expansion’s
effects on the endangered loggerhead sea turtle. NMFS’s
BiOp concludes that the proposed action would not
jeopardize the continued survival and recovery of the
loggerhead, as is required to green-light the project under the
Endangered Species Act (“ESA”). The majority dismisses
the BiOp as arbitrary and capricious because, among other
things, it concludes that the scientific evidence does not
support NMFS’s no-jeopardy conclusion, and it perceives a
conflict with our case law. I disagree. While the record data
shows that the loggerhead is in decline, NMFS reasonably
concluded that the fishery expansion would not appreciably
reduce the likelihood of the loggerhead’s survival and
recovery. Nor did NMFS misapply our decision in National
Wildlife Federation v. National Marine Fisheries Service,
524 F.3d 917 (9th Cir. 2008) (“NWF”): it considered the
incremental impact of the proposed action along with
degraded baseline conditions. That is precisely what NWF
requires.

The majority’s contrary conclusion is a classic example
of the judiciary exceeding its authority by substituting an
agency’s judgments with its own. This complex case relies
on the technical and scientific findings of experts tasked with
the responsibility of protecting our Nation’s species-in-peril.
It is in this context that our respect for a coordinate branch
of government is at its zenith. Indeed, we are “‘at our most
deferential’ when reviewing scientific judgments and
30 TURTLE ISLAND RESTORATION NETWORK V. USDOC

technical analyses within the agency’s expertise,” Lands
Council v. McNair, 629 F.3d 1070, 1074 (9th Cir. 2010)
(quoting Balt. Gas & Elec. Co. v. Natural Res. Def. Council,
462 U.S. 87, 103 (1983)) (adjustment omitted) (“Lands
Council II”), and should only reject an agency’s action if it
is plainly arbitrary and capricious, see Motor Vehicle Mfrs.
Ass’n of U.S., Inc. v. State Farm Mut. Auto. Ins. Co.,
463 U.S. 29, 42–43 (1983). Yet instead of anchoring its
analysis in well-established principles of agency deference,
the majority sets sail on a voyage of discovery, leaving in its
wake our precedent and the doctrinal moorings of Auer v.
Robbins. I dissent, respectfully.

I.

A.

Under Auer v. Robbins, we must defer to an agency’s
reasonable interpretation of its own regulation. See
Christopher, 132 S. Ct. at 2166. Deference is not warranted,
however, “when the agency’s interpretation is plainly
erroneous or inconsistent with the regulation,” or when it
does not reflect the agency’s “considered judgment.” Id.
(internal quotation marks omitted). A lack of “considered
judgment” may be evidenced by (i) an “agency[]
interpretation [that] conflicts with a prior interpretation,”
(ii) “when it appears that the interpretation is nothing more
than a convenient litigating position,” or (iii) when the
interpretation amounts to a “post hoc rationalization
advanced by an agency seeking to defend past agency action
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 31

against attack.” Id. (internal quotation marks and adjustment
omitted). 1

At issue is FWS’s issuance of a special purpose permit
allowing NMFS to authorize incidental take of migratory
birds that are protected under the MBTA. 50 C.F.R. § 21.27
authorizes FWS to issue permits for the take of migratory
birds protected under the MBTA in certain circumstances.
In full, the regulation provides that

[p]ermits may be issued for special purpose
activities related to migratory birds, their
parts, nests, or eggs, which are otherwise
outside the scope of the standard form
permits of this part. A special purpose permit
for migratory bird related activities not
otherwise provided for in this part may be

1
Auer’s continued vitality is a matter of considerable debate. Justice
Antonin Scalia, the progenitor of the doctrine named after the 1997 case,
Auer v. Robbins, which he authored, called for its abolition eighteen years
later in Perez v. Mortgage Bankers Association, 135 S. Ct. 1199, 1213
(2015) (Scalia, J., concurring). He appears to have shared this view with
at least two other justices, Justices Samuel Alito and Clarence Thomas.
See id. at 1210 (Alito, J., concurring); id. at 1213 (Thomas, J., concurring).
See also John C. Eastman, The President’s Pen and the Bureaucrat’s
Fiefdom, 40 HARV. J.L. & PUB. POL’Y 639, 641 (2017). Also, Justice Neil
Gorsuch has openly criticized Chevron deference, see Gutierrez-Brizuela
v. Lynch, 834 F.3d 1142, 1156 (10th Cir. 2016) (Gorsuch, J., concurring)
(citing Marbury v. Madison, 5 U.S. 137 (1803))—a less controversial
deference doctrine because it provides for a check-and-balance between
two branches of government (Congress and the Executive), whereas Auer
involves the Executive’s interpretations of its own actions. At any rate,
my conclusion that the Permit is a lawful exercise of FWS’s authority does
not rely on the continued validity of Auer. Applying traditional tools of
statutory construction, the Permit is lawful agency action because it is
consistent with (i) the regulatory text of § 21.27, (ii) § 21.27’s greater
context, and (iii) the purposes of both § 21.27 and the MBTA itself.
32 TURTLE ISLAND RESTORATION NETWORK V. USDOC

issued to an applicant who submits a written
application containing the general
information and certification required by part
13 and makes a sufficient showing of benefit
to the migratory bird resource, important
research reasons, reasons of human concern
for individual birds, or other compelling
justification.

50 C.F.R. § 21.27. The majority declines to defer to FWS’s
issuance of the Permit because it finds that FWS’s action is
plainly contrary to § 21.27 and the MBTA and is therefore
ultra vires. Because I conclude that issuing the Permit does
not depart from FWS’s past practice, is not inconsistent with
§ 21.27’s text, and comports with the MBTA’s conservation-
oriented purpose, I would defer to FWS’s determination.

1.

Appellants Center for Biological Diversity, et al.
(“CBD”) argue that FWS’s Permit should not be accorded
Auer deference because, CBD asserts, it does not align with
FWS’s past practice.

To determine whether an agency has departed from past
practice, the first step is—manifestly—to define the practice.
Christopher, 132 S. Ct. at 2167–68. A practice is a policy or
mode of operating that is defined by articulable parameters;
simply showing that a current action differs from a prior one
in some way does not establish a departure from past
practice. Cf. FCC v. Fox Television Stations, Inc., 556 U.S.
502, 538 (2009) (agency departed from past practice by
deeming broadcasts of non-literal uses of expletives as
actionable only upon repetition); Dillmon v. Nat’l Transp.
Safety Bd., 588 F.3d 1085, 1090–91 (D.C. Cir. 2009)
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 33

(agency departed from past practice of deferring to an ALJ’s
credibility determinations).

CBD argues that by issuing the Permit, FWS has
changed course from its prior position that it lacks authority
to grant permits to allow unintentional bird taking—i.e.,
incidental taking—for an activity that is not directed at
migratory birds. The majority does not base its decision on
this rationale and for good reason: FWS has long-issued
incidental take permits for all manner of activities whose
only relationship to migratory birds is that they affect the
birds. For example, since at least 1996, FWS has authorized
incidental take of migratory birds for commercial activities
through Endangered Species Act (“ESA”) Habitat
Conservation Plans (“HCPs”). 2 A benefit of entering into an
HCP is that it comes with an ESA § 10 incidental take
permit. See 16 U.S.C. § 1539(a)(1)(B), (a)(2). That permit
“double[s]” as a § 21.27 special purpose permit under the
MBTA. See Dep’t of Commerce, Habitat Conservation
Plan Assurances (“No Surprises”) Rule, 63 Fed. Reg. 8859,
8862–63 (Feb. 23, 1998). Critically, the take that occurs
results from activities that are unrelated to migratory birds—
e.g., natural gas drilling, homebuilding, and myriad other
types of land development—except that they result in
incidental bird deaths—the very ill that CBD insists infects
the Permit at issue here.

2
See Fish and Wildlife Service and National Marine Fisheries
Service, Habitat Conservation Planning and Incidental Take Permit
Processing Handbook App’x 5 (Nov. 4, 1996) (“1996 HCP Handbook”);
see also Fish and Wildlife Service and National Marine Fisheries Service,
Habitat Conservation Planning and Incidental Take Permit Processing
Handbook 16-9 (Dec. 21, 2016) (“2016 HCP Handbook”) (“FWS
routinely issues consolidated ESA and [MBTA] permits for ESA-listed
bird species.”).
34 TURTLE ISLAND RESTORATION NETWORK V. USDOC

FWS has also issued incidental take permits for bald and
golden eagles—which are migratory birds—for activities
that, too, are not directed at migratory birds. See 50 C.F.R.
§§ 22.11; 22.26. And in 1998, FWS issued a special purpose
permit allowing the incidental take of migratory raptors by a
wind farm due to collisions and electrocutions. See FWS
Region 6, Fed. Fish & Wildlife Permit No. PRT-808690
(1998). In short, CBD’s categorical assertion that “FWS has
always understood [§] 21.27 does not authorize incidental
take as the Permit allows” is plainly wrong.

Identifying one error in CBD’s consistency-with-past-
practice argument reveals another. CBD asserts that, “until
[FWS] issued to NMFS the permit at issue exempting
commercial longline fishing from the MBTA’s take
prohibition, the only Special Use Permits FWS had ever
issued authorizing incidental take of non-endangered
migratory birds were specifically intended to promote
migratory bird conservation . . . .” If CBD means to say that
past permits were always associated with activities that had
as their purpose bird conservation, then the preceding
paragraph refutes this contention. But if CBD means
something more capacious—i.e., that such activities must
incorporate bird conservation strategies—then the Permit
addresses this concern. NMFS regulates the Hawaii-based
shallow-set longline fishery under a program that is
expressly geared at reducing seabird bycatch. See 50 C.F.R.
§ 665.815(a)(1), (2), (4). Indeed, since the program took
effect in 2004, incidental take of seabirds by the fishery has
plunged nearly 90 percent. Thus, whatever CBD means by
activities that “promote migratory bird conservation,”
FWS’s issuance of the Permit is consistent with the agency’s
historical practice of tying incidental take permits to
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 35

conservation measures. If there is a past practice to be
discerned, that is it. 3

2.

CBD insists that FWS’s past statements belie the
agency’s assertion that the Permit accords with historical
practice. CBD points to a 2009 regulation governing take
under the Bald and Golden Eagle Protection Act
(“BGEPA”). See Dep’t of the Interior, Eagle Permits; Take
Necessary to Protect Interests in Particular Localities,
74 Fed. Reg. 46,836, 46,862 (Sept. 11, 2009). BGEPA
allows for the take of bald and golden eagles—which species
also fall under the purview of the MBTA—pursuant to an
MBTA permit. 50 C.F.R. § 22.11; see 50 C.F.R. § 22.26. In
response to a public comment, the regulation’s preamble
notes that “[n]o permit is currently available to authorize
incidental take under the MBTA.” 74 Fed. Reg. at 46,862.
CBD seizes on this language as evidence that the Permit is
unlawful.

CBD’s argument proves too much. If the cited statement
means that FWS does not issue incidental take permits for
migratory birds as a categorical rule, then all other instances

3
To be sure, what I articulate as FWS’s past practice does not
precisely align with FWS’s own description of its policy for issuing special
purpose permits, which broadly encompasses “incidental take of migratory
birds” pursuant to agency “activities.” Courts are not permitted to make
sense of an agency action by supplying a rationale not offered by the
agency itself. Bowman Transp., Inc. v. Arkansas-Best Freight Sys., Inc.,
419 U.S. 281, 285–86 (1974) (citing SEC v. Chenery Corp., 332 U.S. 194,
196 (1947)). But my description of the agency’s past practice does not
supply a rationale for an otherwise arbitrary and capricious agency action.
My observation that FWS’s issuance of the Permit is consistent with
FWS’s historical policy simply demonstrates that CBD has not met its
burden of showing that FWS has departed from past practice.
36 TURTLE ISLAND RESTORATION NETWORK V. USDOC

of such permits would be unlawful. Yet CBD spills pages of
ink distinguishing the Permit here from other take permits
granted under the aegis of § 21.27, without suggesting that
those permits are similarly unlawful. Moreover, under
CBD’s interpretation, the cited statement is irreconcilable
with FWS’s other pronouncements permitting take for, e.g.,
migratory birds that are also ESA-listed species. See
2016 HCP Handbook at 16-9. Cf. Boise Cascade Corp. v.
EPA, 942 F.2d 1427, 1432 (9th Cir. 1991) (where possible,
courts avoid statutory interpretations that result in
inconsistencies).

A more natural reading of FWS’s statements—and one
that comports with FWS’s past practice—is that the agency
recognizes that the MBTA lacks a programmatic framework
for issuing incidental take permits. To be sure, a
comprehensive regulation governing incidental take would
be preferable. It could set forth uniform criteria for issuing
permits, thereby offering predictability for the regulated and
environmental communities. 4 But the fact that there exists a
better way to authorize incidental take does not mean that it
is the only lawful way of doing so. Neither the majority nor
CBD provides a persuasive explanation for why § 21.27
does not support case-by-case issuance of permits
authorizing incidental take. 5

4
FWS is in the process of drafting a regulation that would do just that,
though it appears the process has stalled. See Dep’t of the Interior,
Migratory Bird Permits; Programmatic Environmental Impact Statement,
Notice of Intent, 80 Fed. Reg. 30,032 (May 26, 2015).
5
CBD also references statements from a 1996 version of FWS’s
Habitat Conservation Handbook. The Handbook describes the process
governing HCPs under the ESA. Because the Handbook is, at most, a
guidance document, it lacks the force and effect of law. See Perez v.
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 37

Undeterred, CBD takes aim at yet another non-MBTA
regulation. This one—the so-called “No Surprises Rule”—
implements the HCP provision of the ESA. See 63 Fed. Reg.
at 8862–63. The rule explains that an ESA § 10 incidental
take permit, issued in conjunction with an HCP, may
“double” as a special purpose permit under the MBTA for
ESA-listed species. FWS explains that issuing an ESA § 10
permit in lieu of an MBTA § 21.27 special purpose permit is
appropriate because the ESA is more species-protective than
the MBTA. Id. For example, HCPs require an “operating
conservation program designed to conserve the species and
minimize and mitigate the impacts of take of the listed
species of migratory birds to the maximum extent
practicable.” Id. at 8863. CBD extracts from this statement
the conclusion that special purpose permits may not be used
to cover incidental take of non-ESA-listed species because
such species will not enjoy the superior protections of the
ESA.

CBD’s reasoning founders on a logical fallacy. The No
Surprises Rule provides that, because an ESA take permit
comes with greater protections than an MBTA permit, a

Mortg. Bankers Ass’n, 135 S. Ct. 1199, 1203–04 (2015); see generally
1996 HCP Handbook. And to the extent it is probative of FWS’s “past
practice,” it is of little value because the current Handbook is internally
contradictory. One chapter states that “[n]on ESA-listed, migratory birds
can be covered or otherwise addressed in the HCP and incidental take
permit.” 2016 Handbook at 3-28. But another chapter states that “if an
MBTA protected species is not ESA-listed, the FWS does not have a way
to authorize incidental take.” 2016 Handbook at 7-7. An internal
contradiction is archetypal evidence of a lack of “considered [agency]
judgment,” and so the Handbook’s description of FWS’s MBTA
permitting authority is neither persuasive nor deserving of deference. See
Christopher, 132 S. Ct. at 2166.
38 TURTLE ISLAND RESTORATION NETWORK V. USDOC

party need not also apply for an MBTA permit: the latter is
subsumed under the former. See id. at 8862–63. But that
does not mean that ESA-level protections are necessary to
authorize take under the MBTA. Put another way, the No
Surprises Rule says nothing about whether it is appropriate
to issue a special purpose permit for incidental take under
the MBTA for non-ESA-listed species. 6

By analogy, consider a hypothetical state’s labeling
requirements for perishable foodstuffs. The default
regulation for all perishable foods requires the use-by date to
be no more than thirty days from the sell-by date. But certain
perishable foods are on a “highly perishable” list, and are
subject to stricter regulations requiring the affixed use-by
date to be no more than a week from the sell-by date. Now
consider a particular perishable food that is not subject to the
stricter regulations because it is not on the applicable list.
Does this mean it is not governed by the laxer default rule?
Not at all. Yet that is CBD’s logic here: that because the
ESA’s heightened protections apply to some migratory
birds, other non-ESA birds are not subject to the MBTA’s
take provision. In fact, nothing about FWS’s incidental take
policy toward ESA-listed migratory birds forecloses the

6
CBD offers no reason why the rationale for issuing ESA § 10 permits
in lieu of an MBTA § 21.27 permit—that the ESA affords species greater
protections—is not equally applicable to standalone § 21.27 permits for
non-ESA-listed species. FWS, in its discretion, may require a § 21.27
permittee to implement the same types of conservation measures that are
codified under the ESA. FWS effectively did just that with the shallow-
set fishery here. Because the fishery incorporates conservation measures
that have dramatically reduced seabird bycatch, FWS’s issuance of the
Permit is consistent with its rationale for covering migratory birds under
ESA § 10.
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 39

agency from issuing incidental take permits for non-ESA-
listed migratory birds.

B.

While FWS’s issuance of the shallow-set fishery
incidental take permit reflects its considered judgment and is
consistent with its past practice, we may still be compelled
to withhold deference if its interpretation of § 21.27 is
“plainly erroneous or inconsistent with the regulation.”
Christopher, 132 S. Ct. at 2166 (internal quotation marks
omitted). The majority relies on this rationale in concluding
that we should not afford FWS’s action Auer deference, but
its reasoning is based on flawed logic and a misinterpretation
of the MBTA.

1.

The majority claims that the “special purpose activit[y]”
exception to the general ban on permitting take does not
apply here because fishing lacks an “articulable special
purpose.” What qualifies a purpose as “special”? The
majority never quite answers this question, except to
obliquely note that “special purpose” must be read “in the
context of the regulation’s other requirements . . . .” Those
requirements are, according to the majority, that the activity
authorized by the permit “relate[] to migratory birds,” be
paired with a “compelling justification,” and have a
conservation purpose. 7 But the majority never explains what

7
See Klem v. City of Santa Clara, 208 F.3d 1085, 1092 (9th Cir. 2000)
(“the question . . . is whether the Secretary’s interpretation is justified
when considered together with the text of [the regulation], taken in
context”); cf. FDA v. Brown & Williamson Tobacco Corp., 529 U.S. 120,
133 (2000) (noting the “fundamental canon of statutory construction that
40 TURTLE ISLAND RESTORATION NETWORK V. USDOC

it means to “relate[] to migratory birds,” except to posit a
proposition in the negative—namely, that not all activities
that risk killing migratory birds “relate[] to those birds.”
Landowners, environmental practitioners, and FWS will be
hard-pressed to decipher this delphic explanation. Do some
activities that do not have as their purpose the conservation
of migratory birds “relate to those birds”? Which ones? And
how do we know?

The Auer inquiry is more straightforward. We consider
the agency’s interpretation relative to the regulation and the
governing statute. Marsh, 869 F.3d at 1116–17. We must
assure ourselves that the agency has fairly construed its own
regulation, while also keeping one eye trained on Congress’
intent. Id. To that end, “‘[we] need not accept an agency’s
interpretation of its own regulations if that interpretation is
. . . inconsistent with the statute under which the regulations
were promulgated.’” Id. at 1117 (quoting Mines v. Sullivan,
981 F.2d 1068, 1070 (9th Cir. 1992)).

My analysis proceeds as follows: I disaggregate § 21.27
into its relevant textual parts, consider each part against the
regulation’s broader structure and context, and then assess
FWS’s interpretation against the MBTA.

• “Permits may be issued for special purpose activities
. . . which are otherwise outside the scope of the
standard form permits of this part.” The regulation
does not define “special purpose activit[y].” It is also a
regulatory term of art that is not susceptible to
interpretation by reference to dictionary definitions.

the words of a statute must be read in their context and with a view to their
place in the overall statutory scheme” (internal quotation marks omitted)).
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 41

Deploying a wider net, we expand our analysis to the
regulation’s structure and context. The latter part of the
sentence is instructive. It indicates that a “special
purpose activit[y]” is one that is not covered by an
expressly identified permitting scheme. Contrary to
CBD’s assertion, nothing in the context of the regulation
indicates that to be “special” an activity’s purpose must
be directed at migratory birds. 8 See Klem, 208 F.3d at
1092.

• Special purpose permits must be “related to migratory
birds . . . .” The term “relate” has several dictionary
definitions (an inauspicious start for the majority),
including, as is pertinent here: “[t]o refer to,” “[t]o have
reference to,” “[t]o have some connection with; to stand
in relation to,” or “[t]o connect, to link; to establish a
relation between.” Oxford English Dictionary (3d ed.
2009) (goo.gl/grzBqC) (last accessed Dec. 8, 2017).
Whether the first two definitions could flex to embrace
an activity whose purpose is not directed at migratory
birds is debatable. But we need not parse those
definitions because the last two plainly do: an activity
like commercial fishing indisputably has “some
connection with” migratory birds.

• An applicant for a special purpose permit must
“make[] a sufficient showing of benefit to the migratory
bird resource, important research reasons, reasons of

8
CBD asserts that an “ongoing fishing business . . . has no ‘special
purpose’ beyond catching fish.” But this observation only begs the
question: what is a “special purpose”? CBD offers no explanation, except
to march out a parade of horribles, warning that if the Permit is allowed to
stand then the court will have ushered in a brave new world in which
“every activity that happens to somehow harm birds” will qualify for an
incidental take permit.
42 TURTLE ISLAND RESTORATION NETWORK V. USDOC

human concern for individual birds, or other
compelling justification.” FWS invoked the “other
compelling justification” category as the regulatory hook
for issuing the Permit. FWS discerned a “compelling
justification” in its determination that the Permit would
“provide a[n economic] net benefit to the Nation” and
would “serve[] as a benchmark internationally for
employing effective seabird mitigation techniques and
serves as an example of responsible conservation
practices by a fishery.”

The majority concludes that FWS’s rationale is
inadequate, observing that FWS fails to “read the ‘special
purpose’ provision in the context of the regulation’s other
requirements that, taken together, fail to turn § 21.27 into a
general incidental take exception.” 9 The problem for CBD
and the majority, however, is that nothing in § 21.27
suggests—let alone requires—that all special purpose
activities must have as their objective migratory bird
conservation to satisfy the “compelling justification”
standard. In fact, § 21.27's text reveals just the opposite.
The first eligible category is for activities that provide a
“benefit to the migratory bird resource.” Thus, one type of
permit is for an activity that is directed at bird conservation.
But another listed category—“important research
reasons”—includes not even a gloss of conservation intent.
Nor does anything in § 21.27 indicate that a characteristic of
the first stand-alone category—“benefit to the migratory bird
resource”—modifies all those that follow. Rather, the most
natural reading is that special purpose permits are

9
The majority correctly adheres to the doctrine that “all the words
used in a list should be read together and given related meaning when
construing a statute or regulation.” Aguayo v. U.S. Bank, 653 F.3d 912,
927 (9th Cir. 2011).
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 43

appropriate for activities that are either directed at bird
conservation or at other activities that may or may not have
a conservation purpose—e.g., scientific research.

Lest there be any doubt, the immediately following
subsection makes clear that permits may be issued for non-
conservation-related purposes. Section 21.27(a) describes
the criteria for issuing a special purpose permit. See
50 C.F.R. § 21.27(a). It explains that such a permit “is
required before any person may sell, purchase, or barter
captive-bred, migratory game birds . . . .” Id. Nothing in
this subsection suggests that selling, purchasing, or bartering
birds serves the purpose of conserving those birds. Nor do
those terms have an inherent conservation-oriented
meaning—quite the opposite. 10

In sum, the catch-all category “other compelling
justification” is not limited to activities whose purpose is
conserving migratory birds. And the majority provides no
other limiting condition, except to warn against transforming
§ 21.27 into a “general incidental take exception.” But no
party argues that § 21.27 grants FWS a roving license to
permit incidental take whenever it chooses. The question is,
instead: where the agency’s interpretation is not
irreconcilable with the regulation’s text and reflects the
agency’s “considered judgment” (i.e., it is consistent with
past practice), who gets to decide, the courts or the agency?
Auer provides the answer: we defer to the agency in which
Congress has vested regulatory authority to craft policy
based on its expert judgment. See Christopher, 132 S. Ct. at
2166–67 (internal quotation marks omitted). Accordingly, I

10
To be sure, the quoted phrase applies only to captive-bred birds.
But the point is that the regulation expressly contemplates issuing special
purpose permits for something other than conserving migratory birds.
44 TURTLE ISLAND RESTORATION NETWORK V. USDOC

conclude that FWS’s interpretation of “other compelling
justification” as including economic benefits and the benefit
of teaching other nations good conservation techniques is not
“plainly erroneous or inconsistent with the regulation.” Id.
at 2166 (internal quotation marks omitted).

2.

The Permit also comports with the MBTA’s
conservation purpose. The majority is correct that in passing
the MBTA Congress sought to promote migratory bird
conservation. 11 But the statute also expressly provides for
non-conservation-related take of migratory birds. As is
relevant here, the MBTA allows FWS to consider economic
factors in determining whether to permit, among other
things, the taking, killing, possessing, or sale of migratory
birds or their parts. 16 U.S.C. § 704(a). Stated in full,
§ 704(a) provides that:

Subject to the provisions and in order to carry
out the purposes of the [migratory bird treaty]
conventions . . . the [FWS] is authorized and
directed, from time to time, having due
regard to the zones of temperature and to the
distribution, abundance, economic value,
breeding habits, and times and lines of
migratory flight of such birds, to determine

11
See Humane Soc’y of U.S. v. Watt, 551 F. Supp. 1310, 1319 (D.D.C.
1982), aff’d, 713 F.2d 865 (D.C. Cir. 1983) (“‘The United States . . . [and]
Great Britain . . . , being desirous of saving from indiscriminate slaughter
and insuring the preservation of such migratory birds as are either useful
to men or are harmless, have resolved to adopt some uniform system of
protection which shall effectively accomplish such objects . . . .’”)
(quoting 39 Stat. 1702 (Convention on the Protection of Migratory Birds)
incorporated by reference into the MBTA at 16 U.S.C. § 703(a)).
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 45

when, to what extent, if at all, and by what
means, it is compatible with the terms of the
conventions to allow hunting, taking,
capture, killing, possession, sale, purchase
shipment, transportation, carriage, or export
of any such bird, or any part, nest, or egg
thereof, and to adopt suitable regulations
permitting and governing the same . . . .

Id. (emphasis added).

But how—the reader may ask—can we reconcile the
statute’s conservation-oriented focus with its provisions
allowing for the killing of migratory birds? One way is to
interpret § 704(a) as permitting bird deaths—by way of
hunting, incidental take, or other means—to the extent that
doing so does not threaten the overall conservation of
migratory birds. Indeed, we would not be the first court to
adopt this interpretation. See Humane Soc’y v. Watt, 551 F.
Supp. 1310, 1319 (D.D.C. 1982), aff’d, 713 F.2d 865 (D.C.
Cir. 1983) (“It does not necessarily follow from the MBTA’s
evident purposes of conservation that the statute creates a
presumption against hunting . . . .”).

The Permit is consistent with this accommodation of
competing statutory directives: it allows for the take of
migratory birds when paired with measures designed to
minimize such take. Neither CBD nor the majority contends
that, if such measures are followed, the MBTA’s broad goal
of conserving migratory birds is threatened.

3.

The majority has one lure left in its tackle box, but I
decline to take the bait. The majority suggests that because
the MBTA generally prohibits take, a presumption attaches
46 TURTLE ISLAND RESTORATION NETWORK V. USDOC

against reading § 21.27 as authorizing incidental take. The
majority reasons that “although § 21.27 is intended to allow
the FWS to authorize activities not otherwise permitted by
the regulations, it is still a narrow exception to the MBTA’s
general prohibition on killing migratory birds.”

While it is true that the MBTA generally prohibits taking
migratory birds, the majority’s observation is a red herring
because the statute and regulations provide for numerous
exceptions to the general rule. 12 The pertinent question turns
on the scope of the exception to the prohibition, not the
existence of the general prohibition in the first place. As
discussed, § 21.27 is ambiguous and accommodates FWS’s
view that the Permit supports a “special purpose activit[y]”
that is anchored in a “compelling justification.”

* * *

Because issuing the Permit follows FWS’s past practice,
is not plainly erroneous or inconsistent with § 21.27, and
comports with the MBTA’s conservation-oriented purpose,
I would hold it to be a lawful exercise of FWS’s authority.

12
See 16 U.S.C. § 703(a) (“except as permitted by regulations . . . it
shall be unlawful . . . to . . . take . . . any migratory bird . . . .” (emphasis
added)); 50 C.F.R. §§ 21.13 (taking certain mallard ducks); 21.15
(incidental take for military readiness activities); 21.23 (taking for
scientific research); 21.24 (taking for taxidermy); 21.25 (“dispos[ing]” of
migratory waterfowl); 21.26 (killing Canada geese); 21.27 (“special
purpose activities” not covered by other permits); 21.29 (taking for
raptors).
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 47

II.

The majority also errs in rejecting NMFS’s loggerhead
turtle BiOp as arbitrary and capricious. The majority’s
analysis rests on a misapprehension of both binding case law
and the administrative record in this case.

Section 7(a)(2) of the ESA requires all federal agencies
to ensure that any discretionary “action” they authorize,
fund, or implement does not “jeopardize the continued
existence” of an ESA-listed species. 16 U.S.C. § 1536(a)(2);
50 C.F.R. § 402.03. 13 To “jeopardize” means “to engage in
an action that reasonably would be expected, directly or
indirectly, to reduce appreciably the likelihood of both the
survival and recovery of a listed species in the wild by
reducing the reproduction, numbers, or distribution of that
species.” 50 C.F.R. § 402.02. Put another way, “[t]o
‘jeopardize’—the action ESA prohibits—means to ‘expose
to loss or injury’ or to ‘imperil.’” NWF, 524 F.3d at 930. As
we have previously explained,

[e]ither of these [terms] implies causation,
and thus some new risk of harm. Likewise,
the suffix “-ize” in “jeopardize” indicates
some active change of status: an agency may
not “cause a species to be or to become” in a
state of jeopardy or “subject a species to”
jeopardy . . . .

[A]n agency may not take action that will
tip a species from a state of precarious
survival into a state of likely extinction.

13
“Section 7 . . . appl[ies] to all actions in which there is discretionary
Federal involvement or control.” 50 C.F.R. § 402.03.
48 TURTLE ISLAND RESTORATION NETWORK V. USDOC

Likewise, even where baseline conditions
already jeopardize a species, an agency may
not take action that deepens the jeopardy by
causing additional harm.

Id. (emphasis added).

Our discussion of “jeopardy” in NWF must be read in the
context of the regulatory standard. To “deepen[] the
jeopardy” of a species is to “reduce appreciably” a species’
chance at continued survival and recovery. See 50 C.F.R.
§ 402.02. It cannot—as CBD and the majority suggest—
simply mean exacerbating a species’ already “imperiled”
existence, no matter how de minimis the impact. An
“endangered species” like the loggerhead is, by definition, a
“species which is in danger of extinction throughout all of a
significant portion of its range.” 16 U.S.C. § 1532(6)
(defining “endangered”). If the ESA prohibited any action
that worsened—no matter how marginally—a species’
current plight, then it is difficult to conceive of an action that
could survive § 7 consultation. That is not the standard: the
question is not whether the agency action will negatively
affect the species, but whether in doing so it will appreciably
reduce its likelihood of survival and recovery. NWF,
524 F.3d at 930 (the operative inquiry is whether the action
will “cause[] some new jeopardy”—i.e., whether it will “tip
a species from a state of precarious survival into a state of
likely extinction” (emphasis added)).

In NWF, we rejected a BiOp that excluded certain
discretionary agency actions from the jeopardy analysis, and
which also failed to consider degraded baseline conditions.
Id. at 933. The BiOp assessed the effects of dam operations
on the Chinook salmon, an ESA-listed species. Id. at 925–
26. We faulted NMFS for departing from its past practice
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 49

and taking a novel approach in evaluating dam operation
impacts. First, NMFS labeled several operations as
nondiscretionary, thereby “excluding them from the
requisite ESA jeopardy analysis.” Id. at 928–29. Second,
NMFS considered only the marginal impact of certain
discretionary dam operations in its jeopardy analysis. Id. at
929–30. As concerns the second error, NMFS considered
only whether those actions were “‘appreciably’ worse than
baseline conditions.” Id. at 930. Only if they were would
NMFS then conduct a jeopardy analysis. Id.

We held that NMFS’s methodology collided with the
plain text of the regulations. Section 402.02 explains that an
agency action “jeopardizes” a species if it “reduce[s]
appreciably the likelihood of” the species’ “survival and
recovery,” when considering the action’s direct, indirect, and
cumulative impacts measured against the environmental
baseline. 50 C.F.R. §§ 402.02; 402.14(g)(4). NMFS
executed a different procedure. Instead of weighing the
proposed action in the context of the species’ continued
existence, it assessed the action against then-current baseline
conditions. See NWF, 524 F.3d at 930.

By way of example, consider a hypothetical scenario in
which a residential subdivision is planned for an area
inhabited by the endangered arroyo toad. See Rancho Viejo,
LLC v. Norton, 334 F.3d 1158, 1160 (D.C. Cir. 2003)
(Roberts, J., dissenting from denial of rehearing en banc).
The development requires a federal permit, thereby
triggering ESA § 7 consultation. Sierra Club v. Bureau of
Land Mgmt., 786 F.3d 1219, 1224 (9th Cir. 2015)
(consultation required where a private project is “funded,
authorized, or constructed by any federal agency”). The toad
is already threatened by the combined effects of climate
50 TURTLE ISLAND RESTORATION NETWORK V. USDOC

change and habitat fragmentation. 14 Existing developments
have substantially reduced the toad’s habitat, and it teeters
on the precipice between survival and extinction. The
proposed development would reduce the toad’s habitat by an
additional 10 percent, which, in the agency’s estimation,
does not amount to an “appreciable” negative impact when
compared to the habitat destruction that has already taken
place. Thus, under the methodology rejected by this court in
NWF, the agency would not have engaged in a jeopardy
analysis.

The pertinent question under NWF, however, is whether
the proposed development would have an appreciable
impact on the toad’s survival and recovery. Comparing only
the marginal impact against already degraded baseline
conditions conceals this inquiry. Only by considering the
impact of the proposed development “‘within the context of
other existing human activities that impact the listed
species’”—i.e., in the context of climate change effects and
an already diminished natural habitat—can the agency
determine whether the proposed action will consign the toad
to a fate of oblivion. See NWF, 524 F.3d at 930. Similarly,
the flaw NWF identified in that case was NMFS’s failure to
account for the “existing human activity” of dam operations,
which impacted the salmon’s survival. See id. at 930–31.
The court held that NMFS should have considered the
proposed agency action—continued dam operations—
together with degraded baseline conditions, instead of
against those conditions. See id. at 931.

Turning to the matter before us, NMFS undertook the
analysis required by NWF. NMFS considered, among other

14
See U.S. Fish and Wildlife Service, Arroyo Toad 5-Year Review:
Summary and Evaluation 10, 16 (Aug. 2009).
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 51

things, the (i) the current status of the loggerhead sea turtle,
(ii) the direct effects of the proposed action on the
loggerhead based on climate-based and classical modeling,
(iii) the impact of climate change and other cumulative
effects, and (iv) whether the proposed action would result in
an appreciable reduction in the likelihood of the
loggerhead’s survival and recovery. The majority arrives at
a contrary conclusion by fixating on the BiOp’s statement
that the incremental harm of the proposed action is “the
death of a single adult, female loggerhead per year,” which
is an “‘extremely small . . . level of take from the action.’”
The majority insists that NMFS ran afoul of NWF by
comparing the marginal impact of the fishery “to the much
greater harm resulting from factors beyond the fishery.” But
NMFS’s consideration of the marginal impact of the fishery
did not drive its jeopardy analysis à la NWF. Instead, NMFS
considered the “adverse effect on the overlying population
. . . when considered together with all impacts considered in
the Status of the Species, Baseline and Cumulative Effects
sections, including other federally authorized fisheries and
foreign fisheries.” NMFS explained that,

[d]espite the projected population decline
over one generation, we expect the overall
population to remain large enough to
maintain genetic heterogeneity, broad
demographic representation, and successful
reproduction. The proposed action will have
a small effect on the overall size of the
population, and we do not expect it to affect
the loggerheads’ ability meet their lifecycle
requirements and to retain the potential for
recovery.
52 TURTLE ISLAND RESTORATION NETWORK V. USDOC

Thus, unlike in NWF, where NMFS failed to consider direct,
indirect, and cumulative effects, here, NMFS incorporated
the marginal impact of the fishery in assessing whether the
action—combined with baseline conditions—would “tip
[the loggerhead] from a state of precarious survival into a
state of likely extinction.” See id. at 930. It concluded it
would not, and we owe that determination deference. 15 See
Lands Council II, 629 F.3d at 1074 (“Review under the
arbitrary and capricious standard is narrow and we do not
substitute our judgment for that of the agency.”) (internal
quotation marks omitted)).

The majority also criticizes NMFS for relying on “the
conservative nature of its calculations to support the
difference between its conclusion and the climate-based
model’s results.” As a first matter, the majority does not
explain where the model results diverge from NMFS’s
finding of no-jeopardy. Nor could it plausibly do so: an

15
NMFS included in its analysis an assessment of “spillover”
effects—i.e., the impact of the expanded domestic shallow-set fishery on
foreign fisheries. NMFS found that without the expansion, foreign
fisheries would move in and occupy the area. And because the implicated
foreign nations generally have weaker environmental laws than does the
United States, NMFS concluded “with reasonable certainty, that [under the
agency action] there will be a reduction of [loggerhead and leatherback sea
turtle] mortalities as a result of the spillover effect.” NMFS estimated the
reduction to be “11 fewer interactions in the central and north Pacific . . .
or four fewer [loggerhead and leatherback sea turtle] mortalities.”

This data amply supports NMFS’s no-jeopardy conclusion. However,
NMFS did not incorporate its findings into the jeopardy analysis because
it concluded that “data on foreign fisheries is likely incomplete or
inaccurate.” Thus, while the “spillover” effects data is compelling, I—like
the agency—do not rely on it in assessing the reasonableness of NMFS’s
ultimate determination.
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 53

analysis of the record data in the BiOp supports NMFS’s
conclusion. The climate-based model showed that, in 99.5
percent of the tests, the loggerhead would fall below the
quasi-extinction threshold (“QET”) in 25 years without the
proposed action. NMFS similarly found that “[w]hen the
same model is run with the proposed action, the mortality of
1 adult female, the results are similar with 99.5% to 100%
of the runs falling below the QET.” 16 Indeed, the model
showed that while the proposed action would have a
“detectable influence on the loggerhead population, there is
no significant difference in the risk of extinction between the
default, climate-based trends and the forecast considering
the direct effects of the proposed action.” In other words,
the risk of extinction is virtually the same whether or not the
shallow-set fishery is expanded. Accordingly, NMFS
reasonably concluded that the proposed action would not
“reduce appreciably the likelihood” of the loggerheads’
“survival and recovery.” See 50 C.F.R. § 402.02.

At any rate, the majority is simply wrong that NMFS
relied on its conservative estimates to arrive at its no-
jeopardy conclusion. In fact, NMFS relied on (i) the results
of the climate change model showing no statistically
significant difference in the risk of extinction to the
loggerhead with or without the proposed agency action; and

16
The additional loss of one adult female per annum from the
proposed action results in a projected reduction in the overall population
of 4 to 11 percent, due to a loss of that single turtle’s “reproductive
potential” over the course of generations. But, contrary to the majority’s
assessment, NMFS did not credit this numerical loss because it had low
confidence in the data. NMFS noted that the estimated loss does “not
account for the high mortality rate expected of these hatchlings from other
sources, including climate-based threats.” In other words, the reduction
due to a loss of reproductive potential is significantly overstated.
54 TURTLE ISLAND RESTORATION NETWORK V. USDOC

(ii) a “qualitative analysis” reflecting that the loss of one
additional female loggerhead per year would still allow the
loggerhead population to “remain large enough to maintain
genetic heterogeneity, broad demographic representation,
and successful reproduction.” 17

Accordingly, because NMFS’s path “may reasonably be
discerned” and “a reasonable basis exists for its decision,” I
would affirm NMFS’s loggerhead BiOp. Pac. Coast Fed’n
of Fishermen’s Ass’ns v. Blank, 693 F.3d 1084, 1091 (9th
Cir. 2012) (internal quotation marks and citation omitted);
Bowman Transp., Inc. v. Arkansas-Best Freight System, Inc.,
419 U.S. 281 (1974) (“[W]e will uphold a decision of less
than ideal clarity if the agency’s path may reasonably be
discerned.”).

17
NMFS’s use of conservative data inputs is relevant not because it is
the sole basis for its no-jeopardy conclusion (as discussed, it isn’t), but
because it reflects the reasonableness of its findings. For example, NMFS
considered the lost “reproductive potential” of all “unborn hatchlings,”
even though hatchlings have a “high mortality rate.” It also assumed that
the shallow-set fishery would immediately operate at 5,500 sets every
year, even though the increase is likely to be gradual over time. And its
climate model did not incorporate the results of anticipated indirect
effects—namely, beneficial “spillover” effects—of the domestic fishery’s
displacement of international fisheries.

As discussed, NMFS’s no-jeopardy conclusion is not unreasonable
even without considering the conservative nature of its inputs.
Recognizing that those inputs are more conservative than actual conditions
warrant therefore only weakens the majority’s erroneous conclusion that
NMFS’s action is arbitrary and capricious. See George v. Bay Area Rapid
Transit, 577 F.3d 1005, 1011 (9th Cir. 2009) (“The party challenging an
agency’s action as arbitrary and capricious bears the burden of proof
. . . .”).
 TURTLE ISLAND RESTORATION NETWORK V. USDOC 55

CONCLUSION

FWS acted within its authority when it issued a special
purpose permit to NMFS under the MBTA. Its decision
aligns with past practice, is not “plainly erroneous or
inconsistent with [50 C.F.R. § 21.27],” and comports with
the MBTA’s conservation-oriented purpose. The majority
errs in holding otherwise. Similarly, NMFS’s no-jeopardy
finding for the loggerhead sea turtle is rationally related to
the evidence in the record, satisfies its statutory obligation to
consider direct, indirect, and cumulative impacts, and is
faithful to our decision in NWF. Because we should uphold
the MBTA Permit and the loggerhead BiOp, I must
respectfully dissent.