Earth Island Institute v. United States Forest Service

FOR PUBLICATION
UNITED STATES COURT OF APPEALS
FOR THE NINTH CIRCUIT
EARTH ISLAND INSTITUTE, a 
California non-profit organization;
CENTER FOR BIOLOGICAL DIVERSITY,
a non-profit organization,
Plaintiffs-Appellants,
v. No. 05-16776
UNITED STATES FOREST SERVICE;
DALE BOSWORTH, Chief of the  D.C. No.
CV-05-01608-MCE
United States Forest Service; JOHN
BERRY, Forest Supervisor for El OPINION
Dorado National Forest,
Defendants-Appellees,
SIERRA PACIFIC INDUSTRIES,
Defendant-intervenor-
Appellee.

Appeal from the United States District Court
for the Eastern District of California
Morrison C. England, District Judge, Presiding
Argued and Submitted
January 9, 2006—San Francisco, California
Filed March 24, 2006
Before: John T. Noonan, A. Wallace Tashima, and
William A. Fletcher, Circuit Judges.
Opinion by Judge William A. Fletcher;
Concurrence by Judge Noonan
3191
3196 EARTH ISLAND INSTITUTE v. USFS
COUNSEL
Rachel Marie Fazio, John Muir Project, Cedar Ridge, Califor-
nia, for the appellants.
Ronald M. Spritzer and Jeffrey S. Dillen, United States
Department of Justice, Washington, D.C., for the appellees.
C. Athena Roussos, Jay Allen Eisen Law Corp., Sacramento,
California; David E. Martinek, Dun & Martinek, Eureka, Cal-
ifornia, for the intervenor-appellee.
OPINION
W. FLETCHER, Circuit Judge:
Plaintiffs Earth Island Institute and the Center for Biologi-
cal Diversity (collectively, “Earth Island”) appeal the district
court’s denial of their motion for a preliminary injunction
enjoining the implementation of two United States Forest Ser-
vice (“USFS”) post-fire restoration projects in the El Dorado
National Forest. Sierra Pacific Industries (“SPI”) has joined
defendants USFS, Dale Bosworth, and John Berry as an inter-
venor.
Earth Island contends that the Final Environmental Impact
Statements (“FEISs”) for both projects fail to meet the
requirements set forth in the National Environmental Policy
Act (“NEPA”), 42 U.S.C. § 4321 et seq., because the USFS
used faulty scientific methodology in developing its tree mor-
EARTH ISLAND INSTITUTE v. USFS 3197
tality guidelines, and because the FEISs failed to consider
adequately the adverse impacts of the projects on the Califor-
nia spotted owl. Earth Island also contends that the FEISs fail
to comply with the National Forest Management Act
(“NFMA”), 16 U.S.C. § 1600 et seq., because the USFS did
not compile sufficient population data for certain bird Man-
agement Indicator Species (“MIS”).
The district court denied Earth Island’s request, finding (1)
that the methodology employed in the FEISs with respect to
the tree mortality guidelines was not arbitrary and capricious;
(2) that the FEISs took a “hard look” at the adverse impacts
of the projects on the California spotted owl, and (3) that the
USFS had gathered sufficient population monitoring data for
certain bird species that have been categorized as MIS. For
the reasons set forth below, we hold that Earth Island has
shown a “strong likelihood of success on the merits” of its
NEPA and NFMA claims, and that it has otherwise satisfied
the requirements for a preliminary injunction. We reverse and
remand to the district court for further proceedings consistent
with this opinion.
I. Statutory and Factual Background
In October 2004, two substantial fires burned portions of
the El Dorado National Forest. The first, known as the Power
Fire, consumed 16,993 acres of National Forest as well as
additional acres of private land. The Power Fire burned at
varying levels of intensity. According to the FEIS prepared
for the Power Fire Restoration Project, approximately 38% of
the forest area burned at low intensity; approximately 13%
burned at moderate intensity, killing 25% to 75% of the trees;
and approximately 48% burned at high intensity, killing 75%
to 100% of the trees as well as burning the duff and litter pro-
tecting the soil. Several Protected Activity Centers (“PACs”),
Home Range Core Areas (“HRCAs”), and Riparian Conserva-
tion Areas (“RCAs”) for the California Spotted Owl were
located in the Power Fire area. Certain MIS cavity-nesting
3198 EARTH ISLAND INSTITUTE v. USFS
birds, notably the hairy woodpecker, black-backed wood-
pecker, and Williamson’s sapsucker, were present in Power
Fire area.
The second fire, known as the Freds Fire, burned 7,700
total acres, 4,600 of which were in the National Forest. Like
the Power Fire, the Freds Fire burned at varying levels of
intensity. According to the FEIS, approximately 12% burned
at low intensity; approximately 11% burned at moderate
intensity, killing 33% to 66% of the trees; and approximately
61% burned at high intensity, killing 66% to 100% of the
trees as well as burning the duff and litter protecting the soil.
An additional 16% of young plantations also burned at high
intensity. California spotted owl PACs, HRCAs, and RCAs
were also located in the Freds fire area. The hairy wood-
pecker, black-backed woodpecker, and Williamson’s sap-
sucker were also present in the Freds Fire area.
In response to the two fires, the USFS undertook the Power
Fire Restoration Project and the Freds Fire Restoration Proj-
ect. Both projects must comply with federal statutes as well
as the relevant regional forest plans.
Under NEPA, federal agencies must prepare detailed envi-
ronmental impact statements on every proposed action that
“significantly affects the quality of the human environment.”
42 U.S.C. § 4332(C). These statements must include a
description and analysis of the environmental impact of the
proposed action, any adverse environmental effects that can-
not be avoided if the action is implemented, alternatives to the
proposed action, the relationship between short-term uses and
long-term productivity, and any irreversible or irretrievable
commitment of resources that would be involved if the action
were to be implemented. Id. In short, NEPA requires that a
federal agency “consider every significant aspect of the envi-
ronmental impact of a proposed action” and “inform the pub-
lic that it has indeed considered environmental concerns in its
decisionmaking process.” Kern v. U.S. Bureau of Land
EARTH ISLAND INSTITUTE v. USFS 3199
Mgmt., 284 F.3d 1062, 1066 (9th Cir. 2002) (internal quota-
tions omitted). NEPA does not contain substantive environ-
mental standards but instead establishes procedural
requirements designed to ensure that agencies take a “hard
look” at the environmental consequences of their actions. Id.
Under the NFMA, the USFS must develop land and
resource management plans for each unit of the National For-
est System. 16 U.S.C. § 1604(a). In developing such plans, “a
systematic interdisciplinary approach to achieve integrated
consideration of physical, biological, economic, and other sci-
ences” must be used. Id. § 1604(b). The NFMA and regula-
tions promulgated thereunder impose substantive
environmental requirements. See 36 C.F.R. § 219.12. Each
forest plan must also comply with NEPA. 16 U.S.C.
§ 1604(g)(1).
The Sierra Nevada Framework Plan Amendment (the 2001
Framework) is the relevant NFMA forest plan for the El
Dorado National Forest. The 2001 Framework is a compre-
hensive forest plan that establishes a “comprehensive conser-
vation strategy” for national forests in the Sierra Nevada
Mountains, including the establishment of PACs for the Cali-
fornia spotted owl, defined as the best available 300 acres of
owl habitat surrounding a known or suspected nesting site.
Earth Island Inst. v. U.S. Forest Serv., 351 F.3d 1291, 1296
(9th Cir. 2003). Logging within PACs under the 2001 Frame-
work is severely restricted, “generally to the reduction of sur-
face and ladder fuels.” Id. In addition, the 2001 Framework
requires that HRCAs, defined as 1000-acre foraging grounds
for the California spotted owl, be maintained surrounding
each PAC. Id.
In 2004, the 2001 Framework was supplemented by a Final
Supplemental Environmental Impact Statement (the 2004
Supplement) that included a revised plan to improve fire pre-
vention and suppression, reduce fuel loads, restore fire-
adapted ecosystems, and promote community assistance. The
3200 EARTH ISLAND INSTITUTE v. USFS
Record of Decision (“ROD”) implementing the 2004 Supple-
ment provides that “[s]ite-specific decisions [regarding timber
sales] will be made on projects in compliance with NEPA,
[the Endangered Species Act], and other environmental laws
following applicable public involvement and administrative
appeal procedures.”
The El Dorado National Forest is also managed under the
El Dorado National Forest Land and Resource Management
Plan (the “LRMP”). The LRMP identifies certain bird Man-
agement Indicator Species (“MIS”) that aid the USFS in
establishing objectives for improving habitat and for evaluat-
ing the quantity and quality of habitat and species population
trends, in accordance with the NFMA.
Both the Power and Freds Project must conform with the
2001 Framework, the 2004 Supplement, and the LRMP, each
of which must in turn comply with NEPA and the NFMA.
The USFS identified four key goals for the two post-fire res-
toration projects:
(1) to reduce long-term fuel loading in order to
reduce future fire severity and resistance to control;
(2) improve roads and establish effective ground
cover in severely burned areas to reduce erosion and
sedimentation to streams in the short term, and to
contribute to long term soil productivity; (3) recover
the economic value of timber killed or severely
injured by the fire, in an expeditious manner, for the
purpose of generating funds to offset the cost of
future restoration activities; and (4) reduce safety
hazards to the public and forest workers.
Freds Fire Restoration, 69 Fed. Reg. 77,175-02 (Dec. 27,
2004).
In furtherance of the third goal, the Power Project was
divided into six timber sales. One of these, the East Panther
EARTH ISLAND INSTITUTE v. USFS 3201
sale, was awarded to SPI, which contracted to remove dead
trees from 1,363 acres of the Power Fire area. The Freds Proj-
ect was divided into two sales. One of these, the Fred Fire
Salvage sale, was also awarded to SPI, which contracted to
remove dead trees from 1,363 acres of the Freds Fire area.
A Notice of Intent to prepare an Environmental Impact
Statement for the Power Fire was published in the Federal
Register on December 22, 2004, and for the Freds Fire on
December 27, 2004. Power Fire Restoration, 69 Fed. Reg.
76,686-01 (Dec. 22, 2004); Freds Fire Restoration, 69 Fed.
Reg. 77,175-01 (Dec. 27, 2004). After a comment period,
Draft Environmental Impact Statements were published in the
Federal Register on March 25, 2005. Environmental Impact
Statements; Notices of Availability, 70 Fed. Reg. 15,315-01
(Mar. 25, 2005).
On June 16, 2005, the Forest Supervisor for the El Dorado
National Forest, John Berry, requested an Emergency Situa-
tion Determination from the Regional Forester pursuant to 36
C.F.R. § 215.10. According to the request, “substantial loss of
economic value to the Federal Government will occur if
implementation of the [Record of Decision] were delayed.
Loss of economic value will in turn jeopardize the implemen-
tation of the project[s] resulting in long term consequences to
the environment.” The request noted that with an Emergency
Determination, RODs for the two projects could be issued as
early as August 1, 2005, but without such a determination, the
RODs might not be issued until the end of October. A delay
would have the consequence of postponing large portions of
the logging operations until summer 2006.
Such a delay would cause deterioration of the timber to be
salvaged, which would thereby reduce USFS’s revenues from
those sales. The Forest Supervisor anticipated that the loss
from delay would be $11.3 million for the Power Project and
$800,000 for the Freds Project. On July 1, 2005, the Regional
3202 EARTH ISLAND INSTITUTE v. USFS
Forester granted the requests for Emergency Situation Deter-
minations for both Projects.
On July 1, 2005, the USFS issued FEISs for both Projects.
Both FEISs used mortality guidelines to predict which trees
will eventually die from their fire-related injuries, and which
trees should therefore be logged. The guidelines are based on
a draft study by Sharon M. Hood, Sheri L. Smith, and Daniel
R. Cluck (the Hood Study). The Hood Study estimated the
probability of mortality for different tree species based,
according to the study, “on an analysis of the largest dataset
available in terms of numbers of trees and species from wild-
fires in California.”
On August 1, 2005, the final RODs for both projects were
issued. The Power ROD chose Alternative Four; the Freds
ROD chose Alternative One. Both RODs stated that the risk
of cutting trees that would otherwise survive is mitigated by
relying upon the data contained in the Hood Study, which
provides models “that allow managers to select the desired
level of predicted mortality based on land management objec-
tives.” Among the stated reasons put forth for having chosen
Alternatives Four and One was the fact that each Alternative
would generate the greatest revenue for the USFS:
$19,056,425 for the Power Project, and $3,345,872 for the
Freds Project.
Earth Island brought suit in federal district court seeking a
preliminary and permanent injunction against the implementa-
tion of the Power and Freds Projects. On August 18, 2005, the
district court granted Earth Island’s motion for a Temporary
Restraining Order (“TRO”) pending that court’s determina-
tion of whether a preliminary injunction should issue. On
August 25, 2005, the district court vacated its August 18 TRO
and denied Earth Island’s motion for a preliminary injunction.
In this expedited appeal, Earth Island argues that the district
court (1) applied an erroneous legal standard for preliminary
EARTH ISLAND INSTITUTE v. USFS 3203
injunctions by requiring a showing of significant irreparable
harm; (2) applied an erroneous legal standard to Earth
Island’s claim that the USFS failed to ensure scientific integ-
rity in its tree mortality guidelines; (3) erred as a matter of law
by concluding that both FEISs met NEPA’s requirement that
they take a “hard look” at the adverse effects of the projects
on the California spotted owl; (4) erred as a matter of law by
concluding that the USFS did not violate the NFMA by fail-
ing to conduct population surveys of certain MIS bird species;
(5) erred as a matter of law by concluding that Earth Island
had not shown the possibility of irreparable harm to the Cali-
fornia spotted owl and certain bird species; and (6) erred as
a matter of law by concluding that the balance of hardships
did not tip in Earth Island’s favor. On September 15, 2005,
Earth Island filed an emergency motion for an injunction
pending appeal with this court, which was denied by a
motions panel on September 22, 2005. After oral argument,
we sua sponte reconsidered our September 22 denial and, in
an order filed January 11, 2006, issued an injunction pending
the issuance of this opinion.
We have jurisdiction pursuant to 28 U.S.C. § 1291(a)(1).
We now reverse the decision of the district court and hold that
Earth Island has met the requirements for a preliminary
injunction. We remand to the district court for further pro-
ceedings consistent with this opinion.
II. Standard of Review
A district court’s decision granting or denying preliminary
injunctive relief may be reversed only if the court abused its
discretion. See Harris v. Bd. of Supervisors, 366 F.3d 754,
760 (9th Cir. 2004); Earth Island, 351 F.3d at 1298. A court
abuses its discretion if it based its decision on an erroneous
legal standard or clearly erroneous findings of fact. Earth
Island, 351 F.3d at 1298. We review findings of fact for clear
error and conclusions of law de novo. See Hawkins v.
Comparet-Cassani, 251 F.3d 1230, 1239 (9th Cir. 2001);
3204 EARTH ISLAND INSTITUTE v. USFS
Brown v. Cal. Dep’t of Transp., 321 F.3d 1217, 1221 (9th Cir.
2003). We “typically will not reach the merits of a case when
reviewing a preliminary injunction . . . . By this we mean we
will not second guess whether the court correctly applied the
law to the facts of the case, which may be largely undevel-
oped at the early stages of litigation. As long as the district
court got the law right, it will not be reversed simply because
the appellate court would have arrived at a different result if
it had applied the law to the facts of the case.” Earth Island
Inst., 351 F.3d at 1298 (quoting Rucker v. Davis, 237 F.3d
1113, 1118 (9th Cir. 2001) (en banc)).
We view Earth Island’s challenges through the lens of the
Administrative Procedures Act (“APA”). Under the APA,
agency decisions may be set aside only if “arbitrary, capri-
cious, an abuse of discretion, or otherwise not in accordance
with law.” 5 U.S.C. § 706(2)(A). Review under this standard
is narrow, and the reviewing court may not substitute its judg-
ment for that of the agency. See U.S. Postal Serv. v. Gregory,
534 U.S. 1, 6-7 (2001). The agency, however, must articulate
a rational connection between the facts found and the conclu-
sions reached. Midwater Trawlers Co-op v. Envtl. Def. Ctr.,
282 F.3d 710, 716 (9th Cir. 2002). We reverse under the arbi-
trary and capricious standard only if the agency has relied on
factors that Congress has not intended it to consider, has
entirely failed to consider an important aspect of the problem,
or has offered an explanation for that decision that runs
counter to the evidence before the agency or is so implausible
that it could not be ascribed to a difference in view or the
product of agency expertise. See Sierra Club v. U.S. Envtl.
Protection Agency, 346 F.3d 955, 961 (9th Cir. 2003),
amended by 352 F.3d 1187 (9th Cir. 2003).
III. Mootness
In light of the extensive logging that has already taken
place pursuant to the Power and Freds RODs, we asked the
parties to advise us whether this appeal may have become
EARTH ISLAND INSTITUTE v. USFS 3205
moot. The USFS concedes that this appeal is not moot with
respect to the Power Project. For the Power Project, the USFS
reported that 100% of the timber harvest has been completed
for the East Panther and Camp Creek sales, but that only 77%
of harvesting has been completed on the Cole Creek sale, only
73% has been completed for the Ellis sale, only 29% has been
completed for the Bear River sale, and only 15% has been
completed for the Rocky Knob sale. The USFS indicated that
due to winter weather conditions, logging operations on the
uncompleted projects has now ceased but are expected to
resume during the spring of 2006.
According to the USFS, as of January 5, 2006, the timber
harvesting that had been authorized for the Freds Project was
completed, and it was expected that all harvested timber
would be removed from the land and transported to mills by
January 9, 2006. However, Earth Island argues that even if
current logging operations have been completed for the Freds
Project, the logging sales contracts were issued for a term of
two years, meaning that SPI has a continuing right to cut trees
that meet the tree mortality guidelines contained in the FEIS.
[1] “A case becomes moot whenever it loses its character
as a present, live controversy of the kind that must exist if we
are to avoid advisory opinions on abstract propositions of law
. . . [T]he question is not whether the precise relief sought at
the time of the application for an injunction was filed is still
available. The question is whether there can be any effective
relief.” Cantrell v. City of Long Beach, 241 F.3d 674, 678 (9th
Cir. 2001) (internal quotations and alterations omitted). In
Neighbors of Cuddy Mountain v. Alexander, 303 F.3d 1059
(9th Cir. 2002), we concluded that challenges brought under
NEPA and the NFMA were not moot simply because logging
operations had already been completed. There, we noted that
remedial measures continued to be available. For example, the
USFS could still study and mitigate the impact of the sales on
species viability, adjust future timber sales to compensate for
the allegedly unlawful sale, and directly monitor bird popula-
3206 EARTH ISLAND INSTITUTE v. USFS
tion trends. Id. at 1066. In Cantrell, we evaluated mootness
under NEPA and concluded that although historic buildings
with important bird habitats had already been torn down, the
defendants in that case could still consider alternatives and
develop ways to mitigate the damage to the birds’ habitat.
Cantrell, 241 F.3d at 678-79.
[2] Similarly, we conclude in this case that the completion
of current logging operations in the Freds Fire area does not
render the controversy about the Freds Project moot. Not only
is it possible that SPI might cut more trees. In addition, as in
Neighbors of Cuddy Mountain and Cantrell, there are a vari-
ety of measures that could provide some effective relief,
including revising the tree mortality guidelines, monitoring of
the California spotted owl, and obtaining more accurate popu-
lation surveys of MIS bird species.
IV. Preliminary Injunction Standard
Earth Island seeks to enjoin logging operations in the
Power and Freds Project areas until its NFMA and NEPA
claims are adjudicated in federal court. It argues that the dis-
trict court applied an erroneous legal standard by requiring
something beyond the “possibility of irreparable harm” when
it denied Earth Island’s request for a preliminary injunction.
Specifically, Earth Island points out that in analyzing the
USFS’s obligation to monitor certain woodpecker species, the
district court found that “immediate and irreparable” injury
had not been shown. With respect to the California spotted
owl, the district court did not discuss harm at all; instead, it
simply concluded that Earth Island had shown no probability
of success on the merits.
[3] A district court “necessarily abuses its discretion when
it bases its decision on an erroneous legal standard or on
clearly erroneous findings of fact.” Rucker v. Davis, 237 F.3d
1113, 1118 (9th Cir. 2001) (en banc), rev’d on other grounds,
Dep’t of Hous. & Urban Dev. v. Rucker, 535 U.S. 125 (2002).
EARTH ISLAND INSTITUTE v. USFS 3207
In this case, we conclude that the district court applied an
improper legal standard when assessing whether a sufficient
level of injury had been shown.
We recently had occasion to pass upon the proper legal
standard governing preliminary injunctive relief in a case
involving the same parties, the same district court judge, and
a very similar set of facts. Earth Island, 351 F.3d 1291. In that
case, Earth Island sought a preliminary injunction against log-
ging operations provided for in the USFS’s Star Fire Restora-
tion Project after the 2001 Star Fire in the Sierra Nevadas. Id.
at 1295. We identified two sets of criteria for preliminary
injunctive relief. Under the “traditional” criteria, a court may
grant a preliminary injunction if a plaintiff shows “(1) a
strong likelihood of success on the merits, (2) the possibility
of irreparable injury to plaintiff if preliminary relief is not
granted, (3) a balance of hardships favoring the plaintiff, and
(4) advancement of the public interest (in certain cases).” Id.
at 1297 (internal quotations omitted). Alternatively, a court
may grant a preliminary injunction if a plaintiff “demonstrates
either a combination of probable success on the merits and the
possibility of irreparable harm or that serious questions are
raised and the balance of hardships tips sharply in his favor.”
Id. at 1298 (internal quotation marks omitted).
[4] In Earth Island, we held that the district court had
applied an improper legal standard by requiring that Earth
Island demonstrate “actual harm . . . as opposed to speculation
that some such harm could possibly occur.” Id. In that case,
the district judge noted that Earth Island had “failed to show
that measures already in place . . . will not afford sufficient
protection” and that Earth Island had “failed to identify any
concrete probability of irreparable harm in any other respect.”
Id. We concluded that “[e]ach of these statements places a
higher burden of proof on the plaintiff than is warranted.” Id.
We emphasized that a preliminary injunction “only requires
plaintiffs to show probable success on the merits and the pos-
sibility of irreparable harm.” Id.
3208 EARTH ISLAND INSTITUTE v. USFS
In its August 25, 2005 order in this case, the district judge
stated, correctly, that in order to prevail on a motion for a pre-
liminary injunction, a party must demonstrate either “(1) a
combination of probable success on the merits and the possi-
bility of irreparable harm; or (2) that serious questions are
raised and the balance of hardships tips sharply in favor of
granting the requested injunction.” However, the district
judge continued, stating that under either standard, a party
must show a “significant threat of irreparable injury.” During
the hearing, the district judge stated that “even if there is
shown to be a probability of success on the merits by the
plaintiffs, the plaintiffs have not shown at this time that there
is a significant threat of irreparable injury by clear and con-
vincing evidence, which is the standard.”
[5] The USFS argues that the district court applied the
appropriate legal standard as set forth in our decision in Oak-
land Tribune, Inc. v. Chronicle Publ’g Co., 762 F.2d 1374
(9th Cir. 1985), because the words “significant threat of irrep-
arable injury” are not the equivalent of the “concrete probabil-
ity of irreparable harm” standard we held to be erroneous in
Earth Island. While it is true that “significant threat” and
“concrete probability” are different words, what matters is
that both standards impose a higher burden of proof on Earth
Island by going beyond the “mere possibility of irreparable
harm” standard. In Oakland Tribune, we first determined that
plaintiff had shown a very low likelihood on the success of
the merits of its claim, thereby justifying the higher standard
of harm. Here, the district court applied the higher standard
from the outset without first determining the probability of
Earth Island’s success on the merits.
[6] We conclude that the district court applied an erroneous
legal standard. As we discuss below, we conclude that Earth
Island has shown a “strong likelihood of success on the mer-
its,” and has also satisfied the other criteria of degree of
injury, balance of hardships, and advancement of the public
interest.
EARTH ISLAND INSTITUTE v. USFS 3209
V. Discussion
A. Likelihood of Success on the Merits
of Earth Island’s NEPA Challenges
[7] NEPA’s procedural requirements require agencies to
take a “hard look” at the environmental consequences of their
actions. A hard look includes “considering all foreseeable
direct and indirect impacts.” Idaho Sporting Cong. v. Ritten-
house, 305 F.3d 957, 973 (9th Cir. 2002). In addition to direct
and indirect impacts, NEPA also requires that agencies assess
the cumulative impacts of their actions, defined as the “incre-
mental impact of the action when added to past, present, and
reasonably foreseeable future actions.” 40 C.F.R. § 1508.7.
[8] A hard look should involve a discussion of adverse
impacts that does not improperly minimize negative side
effects. Native Ecosytems Council v. U.S. Forest Serv., 428
F.3d 1233, 1241 (9th Cir. 2005). Thus, the USFS must “un-
dertake a thorough environmental analysis before concluding
that no significant environmental impact exists.” Id. at 1239
(internal quotations omitted). We review whether the USFS
has taken a hard look under the arbitrary and capricious stan-
dard. When reviewing the adequacy of an FEIS’s hard look,
we follow a “rule of reason” approach, which requires “a
pragmatic judgment whether the [FEIS’s] form, content and
preparation foster both informed decision-making and
informed public participation.” Native Ecosystem Council v.
U.S. Forest Serv., 418 F.3d 953, 960 (9th Cir. 2005); see also
Dep’t of Transp. v. Public Citizen, 541 U.S. 752, 767 (2004).
Under NEPA, “[a]gencies shall insure the professional
integrity, including scientific integrity, of the discussions and
analyses in environmental impact statements. They shall iden-
tify any methodologies used and shall make explicit reference
by footnote to the scientific and other sources relied upon for
conclusions in the statement.” 40 C.F.R. § 1502.24. Agencies
have wide discretion in assessing scientific evidence, but they
3210 EARTH ISLAND INSTITUTE v. USFS
must “take a hard look at the issues and respond[ ] to reason-
able opposing viewpoints.” Earth Island, 351 F.3d at 1301.
“Because analysis of scientific data requires a high level of
technical expertise, courts must defer to the informed discre-
tion of the responsible federal agencies.” Id. “When special-
ists express conflicting views, an agency must have discretion
to rely on the reasonable opinions of its own experts, even if
a court may find contrary views more persuasive. At the same
time, courts must independently review the record in order to
satisfy themselves that the agency has made a reasoned deci-
sion based on its evaluation of the evidence.” Id. (quoting
Marsh v. Or. Natural Res. Council, 490 U.S. 360, 378
(1989)). If an agency has failed to make a reasoned decision
based on an evaluation of the evidence, we may properly con-
clude that an agency has acted arbitrarily and capriciously. Id.
at 1301.
[9] The primary purpose of an FEIS is to allow for
informed public participation and informed decision making.
See Native Ecosystems, 418 F.3d at 965. In furtherance of this
purpose, 40 C.F.R. § 1502.8 requires that FEISs “be written
in plain language and may use appropriate graphics so that
decisionmakers and the public can readily understand them.”
As we have interpreted this regulation, an FEIS “must be
organized and written so as to be readily understandable by
governmental decisionmakers and by interested non-
professional laypersons likely to be affected by actions taken
under the [FEIS].” Or. Envtl. Council v. Kunzman, 817 F.2d
484, 494 (9th Cir. 1987).
1. NEPA Challenges to the FEISs
Analysis of Logging Activities
Earth Island challenges various aspects of the analyses of
logging contained in the two FEISs. Most significantly, it
challenges the analysis of tree mortality and the use of the
proposed mortality guidelines contained in the Power and
EARTH ISLAND INSTITUTE v. USFS 3211
Freds FEISs. It also challenges the FEISs’ conclusions that
there is a lack of adequate soil cover in the project areas, and
that the retention of large snags (dead trees) would result in
a hazardous level of surface fuels.
a. Tree Mortality Guidelines
Earth Island argues that the FEISs’ guidelines for cutting
burned and scorched trees substantially overpredict tree mor-
tality, with the result that many more trees will be cut than are
necessary to meet the legitimate objectives of the Power and
Freds Projects. Earth Island argues that by cutting many trees
that will not die, the USFS will unnecessarily destroy valuable
habitat for the California spotted owl and certain MIS bird
species.
The tree mortality guidelines contained in the FEISs are
based upon the 2005 draft Hood Study. That study predicts
post-fire mortality for trees burned or scorched in forest fires.
The guidelines for both projects are keyed to the percentage
of the tree’s green crown that has been scorched by the fires.
Based on data contained in the Hood Study, the USFS
chose two different marking guidelines, one for areas to be
logged by tractor and one for areas to be logged by helicopter
or skyline. Table 3-5 of both of the FEISs is entitled “Mortal-
ity Guidelines for marking as applied to [various alterna-
tives].” For simplicity of explanation, we will describe the
guidelines only as they apply to yellow pine and white fir, and
only as to helicopter and skyline logging. Yellow pine with
65% or more of the length of their crown scorched are consid-
ered dead, and are to be marked for cutting. The same crown
length scorch percentages apply to white fir over 20” in diam-
eter.
3212 EARTH ISLAND INSTITUTE v. USFS
For the convenience of the reader, we reproduce Table 3-5:
PP/JP/SP White Fir White Fir Incense Red Fir Douglas
[yellow pine] <20" >20" Cedar Fir
Tractor Logging System
Crown Scorch* 75% 95% 80% 100% ? ?
Crown Scorch ?
+ RTB or
Ambrosia Beetle
Helicopter and Skyline Logging Systems
Crown Scorch* 65% 85% 65% 95% ? ?
Crown Scorch
+ RTB or
Ambrosia Beetle ?
*As a percentage of the original live crown. Percentage of green foliage
would not be used for ponderosa or Jeffrey pine until the 2005 needle
flush is visible. Any marking prior to needle flush would be limited to
pines with 100% black crown.
Table 3-6 of the FEISs is entitled “Probability of Tree Mor-
tality.” Both FEISs state, “Table 3-6 indicates the probability
of tree mortality to individual trees meeting the Power [and
Freds] Fire marking guidelines (Correctly Predicted Mortal-
ity) and the predicted survival of trees with less fire damage
than the minimum requirements of the marking guidelines
(Correctly Predicted Survival).” According to Table 3-6, a
yellow pine cut in accordance with the 65% crown length
scorch guideline of Table 3-5 has a 90% “Correctly Predicted
Mortality.” A white fir over 20” in diameter cut in accordance
with the guidelines has a 87% “Correctly Predicted Mortali-
ty.”
For the convenience of the reader, we reproduce Table 3-6.
We have italicized the numbers 90 and 87 to indicate the
“correctly predicted mortality” percentages for yellow pine
and white fir cut in accordance with the 65% crown length
scorch guideline:
EARTH ISLAND INSTITUTE v. USFS 3213
PP/JP/SP White Fir Incense Red Fir Douglas
[yellow pine] Cedar Fir
Tractor Logging System
Correctly Predicted
Mortality (%) 96 95 100 100 —
Correctly Predicted
Survival (%) 51 63 88 — —
Helicopter and Skyline Logging Systems
Correctly Predicted
Mortality (%) 90 87 85 100 —
Correctly Predicted
Survival (%) 65 74 89 — —
Earth Island contends that the findings contained in the
draft Hood Study, as well as in other studies, have substan-
tially different percentage estimates of tree mortality from the
percentages contained in Table 3-6. Earth Island relies upon
the declaration of Dr. Edwin B. Royce in support of its con-
tention. Royce has a Ph.D. in Botany with a specialization in
Forest Plant Ecology from the University of California at
Davis, and a Ph.D. in Applied Physics from Harvard Univer-
sity. Royce has had twelve years of experience in the charac-
terization of forest vegetation.
The USFS challenges the admissibility of Royce’s declara-
tion, as well as other expert declarations offered by Earth
Island, because they were not before the agency during the
administrative review process. We allow extra-record materi-
als if necessary to “determine whether the agency has consid-
ered all relevant factors and has explained its decision.” Sw.
Ctr. for Biological Diversity v. U.S. Forest Serv., 100 F.3d
1443, 1450 (9th Cir. 1996). Because Earth Island’s expert
declarations are offered for this purpose, they were properly
before the district court and are properly before this court on
review.
According to Royce, yellow pine and white fir are both
“thick bark trees” whose bark provides significant protection
3214 EARTH ISLAND INSTITUTE v. USFS
of the living tissue beneath the bark. According to Royce,
“[c]rown kill is commonly the dominant source of fire-
induced mortality in large trees having thick bark, such as
white fir or yellow pine.” However, “[f]all fires, such as the
Power and Fred’s fires, are least damaging to conifers. In part
this is because reserves of stored products of photosynthesis
are high and because new growth throughout the trees is less
vulnerable to damage then, as compared to that same growth
during the summer growing season.” “A low-intensity surface
fire commonly produces only partial crown kill and only min-
imal mortality in larger trees. Trees tolerate partial crown kill
in a surface fire in part because only the lower part of the
crown is normally killed. The lower part of the crown is less
photosynthetically productive than is the upper part.”
For yellow pine, Royce states, “[t]he mortality guidelines
for yellow pine (ponderosa and Jeffrey pine) will permit the
harvest of trees with a substantial probability of surviving if
they were not harvested.” The Royce declaration contains
tables showing predicted tree mortality in the Hood Study, as
well as in three other studies. Those other studies were by
Ryan and Reinhardt, by Stephens and Finney, and by
McHugh and Kolb. Royce states as to yellow pine, “[f]or heli-
copter and skyline logging, mortalities from the [Hood Study]
models are 60% and 70%, as compared to a 90% mortality
claimed in the FEIS and ROD.” Mortalities predicted by the
three other studies are even more at variance from the mortal-
ity predicted in the FEISs. Royce states,
[T]he guidelines for helicopter and skyline logging
will allow the cutting of small trees that have proba-
bilities of mortality between 12% and 57% and large
trees with mortality probabilities between 11% and
32% (compared with 90% mortality claimed by the
FEIS’s). Even if one accepts only the largest of the
probabilities of mortality — that given by the Ryan
and Reinhardt paper — this still translates into a
worst case probability of survival of up to 68% for
EARTH ISLAND INSTITUTE v. USFS 3215
trees that could be cut (meaning 68% of the trees
logged would otherwise survive). (Emphasis in origi-
nal.)
In addition, Royce states that he personally evaluated 39
yellow pine that were part of the data base for the Hood
Study, and “found evidence of measurement errors that would
lead to the development of models that over predict mortali-
ty.” According to Royce,
I measured the average amount of crown kill on
these trees to be 88%, whereas, the [Hood Study]
field workers evaluated the same damage to average
64%. A small part of this difference may be due to
the fact that my measurements were crown volume
measurements, whereas, the [Hood Study] measure-
ments were linear measurements. However, as dis-
cussed [earlier in my declaration], this should
produce only a few percent difference between the
two measurements. I suggest that at least a substan-
tial part of the difference between the two measure-
ments is a result of the [Hood Study] field workers
underestimating crown kill. This will result in a
model that over predicts mortality when correct
crown kill values are used. A corrected [Hood
Study] model yielding lower mortality predictions
would make the proposed guidelines even less
defensible.
For white fir, Royce states, “[t]he mortality guidelines for
large white fir will permit the harvest of trees with a substan-
tial probability of surviving if they were not harvested.” As he
does for yellow pine, Royce presents tables showing predicted
tree mortality in the Hood Study, as well as in the three other
studies. He states as to white fir, “[f]or the helicopter and sky-
line logging of large trees, mortalities from the [Hood Study]
models are between 25% and 45%, as compared to an 87%
mortality claimed in the FEIS and ROD.” Just as with yellow
3216 EARTH ISLAND INSTITUTE v. USFS
pine, mortalities predicted by the three other studies are even
more at variance from the mortalities predicted in the FEISs.
Royce states,
the guidelines for helicopter or skyline logging will
allow the cutting of large trees with mortality proba-
bilities between 7% and 50%, as compared to the
87% mortality claimed in the FEIS’s. Even if one
accepts only the largest of these predictions of mor-
tality — that given by the Ryan and Reinhardt paper
— this still translates into a worst case probability
of survival (in helicopter units) of up to 50% for
trees that could be cut. (emphasis in original).
In response, the USFS relies on a declaration of Sheri L.
Smith. Smith is one of the three co-authors of the draft Hood
Study. She has B.S. and M.S. degrees in Biology and Ento-
mology from Utah State University. She has been an USFS
employee for 15 years. She has been “involved in evaluating
fire-injured trees in California since 1991.”
Smith makes three points in response to the Royce declara-
tion. First, she contends that Royce misunderstood Table 3-6.
As noted above, that table is entitled “Probability of Tree
Mortality.” Royce understood the percentages contained in
that table as indicating the probability that a tree will die. For
example, Table 3-6 indicates that a yellow pine with a mini-
mum of 75% scorched crown (the guideline for tractor log-
ging) has a “correctly predicted mortality” of 96%, and a
“correctly predicted survival” of 51%. Royce understood the
table to mean that 96% of yellow pine with a minimum of
75% scorched crown will die.
Smith responds,
The marking guidelines used for the Power and
Freds fire are based on models that use percent
crown kill . . . . I attach a true and correct copy of
EARTH ISLAND INSTITUTE v. USFS 3217
an excerpt of the models which are the basis for the
Power and Freds marking guidelines (Attachment
A).
Mr. Royce states that Table 3-6 represents that
96% of trees with 75% crown kill will die. This is
simply incorrect. The 96% in Table 3-6 (classifica-
tion table) shows that the marking guideline model
used by the Forest Service is correct in predicting
mortality 96% of the time when our individual study
trees are run through the model. In other words, the
96% is an estimate of the accuracy of the model, not
of the percentage of mortality as Mr. Royce states.
Both FEISs accurately interpret the findings of our
paper in light of the specific marking guideline
model selected by the Forest Service.
The percentages shown in Table 3-6 (classifica-
tion table) of both the Power and Freds EIS are
derived from the entire [Hood Study] dataset, which,
for yellow pine, is based on measurements of 1,969
trees. The models in our paper take these percent-
ages to arrive at a statistical probability of mortality
for an individual tree. The probability of mortality
(Pm) ranges from 0.0 to 1.0, with 1.0 being certain
mortality. The Forest Service selected a Pm of 0.90
for all ground-based harvest units. For yellow pine,
a Pm of 0.90 corresponds to a crown length kill of
75%. Using this criterion, the marking guideline
model correctly predicted mortality for yellow pine
trees in the [Hood Study] dataset 96% of the time
and correctly predicted survival 50% of the time. In
other words, using a Pm of 0.90 (75% crown kill),
only 4% of the trees the model predicts will die,
actually survived, but 50% of the trees predicted to
survive eventually died. . . .
This means that there is a high likelihood of cor-
rectly predicting mortality, but a lower likelihood of
3218 EARTH ISLAND INSTITUTE v. USFS
correctly predicting survival using the 75% crown
scorch model, which means that it is much more
likely that the Forest Service is leaving trees behind
that will later die, than it is taking trees that would
have survived. This is directly contrary to Mr.
Royce’s assertions[.] . . . The same misinterpretation
of the data is true for Mr. Royce’s discussion of yel-
low pine on helicopter and skyline harvest units and
all white fir, all of which misinterpret the percent-
ages in Table 3-6 to be something other than what
they are — a verification of the accuracy of the
marking guideline model using the [Hood Study]
dataset.
(emphasis in original).
Second, Smith responds that Royce estimated the probabil-
ity that a tree will die based on the percentage volume of
crown kill. By contrast, the Hood Study estimated the proba-
bility based on the percentage length of crown kill. Therefore,
according to Smith, Royce’s comparisons are “problematic.”
Third, Smith responds that two of the three other studies
upon which Royce relies for his comparison — the studies of
Stephens and Finney, and of McHugh and Kolb — are based
on “prescribed fires.” According to Smith, “[m]ost often, the
objective of a prescribed fire is to limit the mortality of the
overstory while reducing fuel loadings and ingrowth of smal-
ler trees. Our model is more appropriate to the type of fire
represented by the Freds and Power fires, which killed not
only the ingrowth of smaller trees, but also resulted in high
levels of crown kill of larger trees.”
We analyze Smith’s three points in turn. First, Smith con-
tends Royce has misunderstood Table 3-6. According to
Smith, the percentage figures given for “correctly predicted
mortality” — such as 96% for yellow pine with a minimum
of 75% crown length scorch — do not predict the percentage
EARTH ISLAND INSTITUTE v. USFS 3219
of trees that will die. Rather, these figures predict the accu-
racy of the prediction of the percentage of these trees that will
die. Similarly, according to Smith, the percentage figures for
“correctly predicted survival” — such as 51% for yellow pine
with a minimum of 75% crown length scorch — do not pre-
dict the percentage of trees that will live. Rather, they predict
the accuracy of the prediction of the percentage of these trees
that will live. Smith may well be correct. Indeed, it appears
from the face of Table 3-6 that she may be, for if the table
predicted the percentage of trees that will die or live, as dis-
tinct from predicting the accuracy of the predictions of these
percentages, the percentages given in the table should add up
to 100%. Instead, in the example given for yellow pine with
75% crown length scorch, they add up to 147% (96% plus
51%).
[10] But even if Smith is correct, this does not solve the
problem. Table 3-6 is, to say the least, misleading. Its title is
“Probability of Tree Mortality,” rather than “Probability that
Predictions of Probability of Tree Mortality and Survival are
Correct.” Second, there is no other table in the FEISs provid-
ing the probability of tree mortality. The absence of such a
table is significant. The single most important aspect of the
FEISs is their estimate of the likelihood that trees with certain
amounts of fire damage will die. This is so for the obvious
reason that the justification for cutting burned or scorched
trees is the likelihood that they will die. Any reader of the
FEISs will therefore look for a table providing probability of
tree mortality. The only table in the FEISs that appears to pro-
vide that information is Table 3-6. It is not unforeseeable that
a reader — even an expert reader such as Royce — would
misunderstand the table. Further, the explanation for Table 3-
6 provided by Smith’s declaration in the district court is
nowhere provided in the FEISs. For example, Attachment A
to Smith’s declaration is not provided, or even referred to, in
the FEISs. The absence of such an explanation in the FEISs
obviously increases the chance that the table will be misun-
derstood.
3220 EARTH ISLAND INSTITUTE v. USFS
Further, the Royce declaration provides tree mortality per-
centages given in the draft Hood Study, and he compares
those percentages to those given in Table 3-6. The Hood
Study percentages provided by Royce are not contested by
Smith. (This is not surprising, given that Smith is a co-author
of the Hood Study.) The Hood Study mortality percentages
are substantially lower than the percentages given in Table 3-
6. Yet those percentages are not provided in the text of the
FEISs. If the USFS had been truly interested in educating the
reader as to the actual percentages of trees likely to die from
scorch damage, it would have provided those percentages in
the FEISs. Instead, it provided Table 3-6 (entitled “Probability
of Tree Mortality”) giving percentages of “correctly predicted
mortality.” Even if the USFS could properly rely on the draft
Hood Study instead of other tree mortality studies — a ques-
tion we do not here decide — it should have provided the
mortality percentage figures in the Hood Study rather than, or
in addition to, the figures in Table 3-6.
Second, Smith responds that the comparisons used by
Royce to estimate tree mortality are based on percentage vol-
ume of crown kill rather than percentage length of crown kill.
We note initially that Smith’s response has no application to
the comparison between the percentage numbers given in
Table 3-6 and those given in the Hood Study. Both Table 3-
6 and the Hood Study were based on the same methodology
— correlating percentage length of crown kill with likelihood
of tree death. Further, we note that with respect to studies that
relied on percentage volume rather than percentage length of
crown kill, Royce had explained in his declaration why the
FEISs’ emphasis on the difference in the two type of crown
kill measurements is misleading. He wrote:
Tables 3-4 in both the Power and Fred’s fire
FEIS’s show a relationship between these two mea-
sures of crown kill in which the volume measure is
much larger than the linear measure for the same
amount of damage. However, these tables are valid
EARTH ISLAND INSTITUTE v. USFS 3221
only for young trees with a tapered shape, similar to
the classic conical Christmas tree shape shown in
figure F-1 of the Powers FEIS or figure A-1 of the
Fred’s FEIS. Large, mature trees growing in a forest
surrounded by other trees have a shape more like a
cylinder, tapered only in the topmost branches. . . .
It is to mature trees that the guidelines [in the
FEIS’s] are to be applied to determine if the trees are
dying and hence subject to salvage harvest. In the
context of the salvage of mature trees, the inclusion
of these tables and figures in the FEIS’s is com-
pletely misleading.
My experience measuring mature trees on the
Eldorado-Star fire site . . . was that, within the uncer-
tainty with which these determinations can be made
in the field, the two measures give the same numeri-
cal value for crown kill.
Third, Smith responds that two of the three other studies
used by Royce for comparison were based on “prescribed
fires.” According to Smith, reliance on these two studies —
the Stephens and Finney study and the McHugh and Kolb
study — was inappropriate because the Power and Freds Fires
“resulted in high levels of crown kill of larger trees” com-
pared with the levels of crown kill in the prescribed fires. It
may well be that, as Smith says, the prescribed fires resulted
in lower levels of crown kill in larger trees. But Smith does
not explain why that makes a difference. As those two studies
are described by Royce, they predict tree mortality based on
the actual percentage of crown kill in particular trees. The fact
that fewer trees have that degree of crown kill in a prescribed
fire should not make any difference in the predicted mortality
of those trees that actually have a high level of crown kill.
Further, Smith does not mention the third study upon which
Royce relies — the Ryan and Reinhardt study. Of the three
studies, this one gives the highest probability of tree mortal-
3222 EARTH ISLAND INSTITUTE v. USFS
ity. Royce explicitly discusses and relies on the Ryan and
Reinhardt study in his conclusion that the FEISs overpredict
tree mortality. For example, as indicated above, Royce writes
with respect to yellow pine, “[e]ven if one accepts only the
largest of these probabilities of mortality — that given by the
Ryan and Reinhardt paper — this still translates into a worst
case probability of survival of up to 68% for trees that could
be cut[.]”
We recognize that the FEISs discount two of the studies —
those by Stephens and Finney and by Ryan and Reinhardt.
The Power and Freds FEISs both discount the Stephens and
Finney study on the ground that its purpose was not to pro-
vide salvage guidelines, and that some of its data were
obtained pre- rather than post-fire. The Powers FEIS dis-
counts the Ryan and Reinhardt study as involving only one of
the species (Douglas fir) that occurs in the Power Fire area.
The Freds FEIS discounts that study in slightly different
terms, stating, “[r]esults from the Ryan and Reinhardt 1988
study were obtained mostly for tree species not found in, and
geographic regions not related to, the Sierra Nevada.” But
even if we must discount both of these studies — a question
we do not decide — the McHugh and Kolb study remains.
That study was not discussed, or discounted, in either of the
FEISs.
[11] In the end, we conclude that the USFS abused its dis-
cretion in its estimates of likely tree mortality in both the
Power and Freds FEISs. We will assume, for purposes of our
analysis, that the Smith declaration correctly states what the
percentage numbers in Table 3-6 really mean. But even if the
Smith declaration is correct, Table 3-6 is, for the reasons
given above, extremely misleading. A casual, or even a care-
ful, reader of the FEISs and of Table 3-6 could easily con-
clude that 96% of yellow pine with a minimum of 75% crown
length scorch will die, or that 90% with a minimum of 65%
crown length scorch will die. If those were, in fact, the per-
centages of yellow pine with that degree of fire damage that
EARTH ISLAND INSTITUTE v. USFS 3223
will die, it would be easy to conclude that the USFS is justi-
fied in cutting all yellow pine that satisfy those criteria. But
those are not the percentages of trees that will die.
The Hood Study itself estimates a substantially lower tree
mortality than the percentage numbers provided in Table 3-6.
The other three studies estimate even lower tree mortalities.
It is possible that those who prepared the FEISs, and the For-
est Supervisor who signed the RODs based on the FEISs,
understood Table 3-6 in the way Royce understood it. If this
is so, the USFS abused its discretion, for it failed to take the
requisite “hard look” at the data underlying their analysis and
decision. Kern, 284 F.3d at 1066. It is also possible that those
who prepared the FEISs, and the Forest Supervisor, under-
stood Table 3-6 in precisely the way Smith described it in her
declaration. If this is so, the USFS also abused it discretion,
for it failed to reveal the actual percentages upon which it
relied and it drafted highly misleading FEISs. Native
Ecosystem Council, 418 F.3d at 965. Under the first alterna-
tive, the USFS misunderstood the data; under the second, it
understood but concealed and misrepresented the data. Under
either alternative, it abused its discretion.
b. Soil Cover
Both the Power and the Freds FEISs state that a goal of the
restoration projects is to provide effective ground cover as a
means of reducing erosion and stream sedimentation. Both
FEISs specify that the average soil cover should be between
50% to 60%. The FEISs propose to achieve this amount of
soil coverage by using debris from trees that are logged
according to the mortality guidelines discussed above — in
their words, by using “tops and limbs of dead and dying
trees.”
According to the Power FEIS, there was an average of 11%
to 30% ground cover in areas of moderate-severity burns after
the fire. There was an average of 0% to 10% ground cover in
3224 EARTH ISLAND INSTITUTE v. USFS
areas of high-severity burns. The average projected ground
cover after needles scorched by the fire had fallen to the
ground was estimated to be 51% to 70% in moderate-severity
burn areas, and 20% to 30% in high-severity burn areas.
According to the Freds FEIS, there was an average of 17%
ground cover in moderate-severity burn areas, and an average
of 9% in high-severity burn areas. Following needlecast, the
projected ground cover was 46% in moderate-severity burn
areas, and 19% in high-severity burn areas.
Earth Island argues that these figures for soil coverage are
based on information obtained in site visits immediately after
the fire, and that more recent visits show that effective soil
cover already exceeds 50% to 60% in severely burned areas.
In support of this argument, Earth Island offers the declaration
of Jonathan J. Rhodes. Rhodes has a B.S. in Hydrology and
Water Resources from the University of Arizona, and an M.S.
in Hydrology and Hydrogeology from the University of
Nevada-Reno. He has also finished all required academic
work toward a Ph.D. in forest hydrology at the University of
Washington. He has had more than twenty-two years of expe-
rience as a hydrologist.
Rhodes visited what he described as six high-severity burn
areas within the Power Fire area on August 11, 2005. He
found that needles, twigs, and branches that had fallen since
the fire, as well as new vegetative growth, had increased soil
cover substantially. He stated in his declaration:
I measured soil cover in areas where my visual esti-
mates indicated that soil cover was lowest. Thus, my
measurements include those in areas with soil cover
that is well below the average within the proposed
Power project area as a whole. My measurements
and evaluation of soil cover conditions within the
areas of the proposed Power logging project
unequivocally demonstrate that as of August 11,
2005, soil cover is well-distributed and greater than
EARTH ISLAND INSTITUTE v. USFS 3225
60% in the overwhelming majority of areas burned
at high severity. In the areas that I measured soil
cover, the lowest level of soil cover that I measured
was 59%; soil cover in the other five areas measured
ranged from 66-91%. Even this lowest level of mea-
sured soil cover exceeds the 50% target that Power
and Freds FEIS state should be exceeded.
Rhodes continued, “[b]ased on conditions within the area, it
is unquestionable that soil cover from [conifer regeneration,
and needles, branches, and logs from burned trees] will con-
tinue to steadily increase over the next several years.” He con-
cluded, “there is not a pressing need to try to increase soil
cover via logging in order to reduce soil erosion. This is espe-
cially true because logging always causes increased soil dam-
age and elevated erosion.” (emphasis in original).
The USFS responded with a declaration by Jeffrey TenPas.
TenPas has a M.S. in Soil Science from the University of Cal-
ifornia, Davis. He is an employee of the USFS. He has had
more than fifteen years of experience in soil science. TenPas
assessed soil cover immediately after the Power Fire. He then
revisited the Power Fire area on August 19, 2005, after read-
ing Rhodes declaration. TenPas wrote, “[c]ontrary to Mr.
Rhodes findings, portions of the high severity burned areas in
the Power Project area have existing ground cover much less
than 50%.” He suggested that Rhodes might have been exam-
ining a “more moderately burned area that received signifi-
cant needle cast,” rather than a high-severity burn area.
TenPas stated:
The pattern of recovery [in the Power Project area]
was consistent with expectations. Needle cast had
provided adequate ground cover in low and moder-
ate severity burned areas. Bear clover, where it was
present, covered an estimated 0 to 70% cover at the
scale of 200′ transects. In high severity burned areas,
bear clover was the predominant component in vege-
3226 EARTH ISLAND INSTITUTE v. USFS
tative recovery in areas I visited. Various forbs pro-
vided a trace to 5% cover. In high severity burned
areas without bear clover, ground cover was as low
as 1%, that from a trace of needles.
He concluded, “[i]n sum, my observations indicate that there
remain portions of high severity burned areas where soil
cover is deficient. . . . Timber harvest can contribute addi-
tional cover in these areas.”
[12] As in Earth Island, we conclude that “[a]t this stage,
the record does not allow us to conclude that the Forest Ser-
vice acted arbitrarily and capriciously in relying on its own
data and discounting the alternative evidence offered by the
Plaintiffs.” Earth Island, 351 F.3d at 1302. This is especially
so in light of the deference given to the “reasonable opinions
of [an agency’s] own experts.” Id. at 1301.
c. Fuel Loading
Both the Power and Freds FEISs state that a goal of the res-
toration projects is to reduce fuel loading on the ground so as
to reduce the risk of future catastrophic fires, and to avoid
safety hazards to workers and recreation visitors. Specifically,
the Power FEIS rejected an alternative that would have
retained four large snags per acre outside PACs for the Cali-
fornia spotted owl. The Freds FEIS rejected an alternative that
would have retained four to eight large snags per acre outside
PACs. Earth Island challenges the FEISs’ rejection of these
alternatives, arguing that retaining this number of large snags
would not impede the goals of reducing long-term fuel load-
ing and safety hazards.
According to the Power FEIS, in areas where four large
snags per acre would be retained, fuel levels 25 years after the
fire would be approximately 38 to 40 tons per acre. According
to the Freds FEIS, in areas where four to eight snags per acre
would be retained, fuel loads would be 16 to 50 tons per acre
EARTH ISLAND INSTITUTE v. USFS 3227
by year 25. The USFS contends that these levels of fuel load-
ing would provide significant fuel, contribute to severe soil
heating, and impede effective fire suppression efforts.
[13] As with soil cover, above, we cannot confidently dis-
cern from the present record whether Earth Island’s conten-
tions about fuel-loading have validity. We therefore cannot
say, at this point, that the USFS has acted arbitrarily and
capriciously in rejecting the alternatives that would have
retained four snags per acre in the Power Fire area and four
to eight snags per acre in the Freds Fire area.
2. NEPA Challenge to Analysis of Effects
of Logging on California Spotted Owls
Earth Island argues that the USFS has failed to take a “hard
look” at the effects of the Power and Freds Projects on the
California spotted owl. The California spotted owl was identi-
fied as a “species at risk” in the 2001 Framework. “Species
at risk are those with a high level of concern whose ranges are
not peripheral to the Sierra Nevada and that occur in old for-
est ecosystems.” As reported in the 2004 Supplement, the
Fish and Wildlife Service (“FWS”) declared in February 2003
that the California spotted owl would not be listed as an
endangered species under the Endangered Species Act
because there was “no definite evidence that the population is
decreasing across its range, and various analytical results of
the individual study areas are not wholly supportive of con-
clusions regarding declines in any given study area.”
However, because changes in the Sierra Nevada Frame-
work could affect the California spotted owl, the FWS has
stated that it will continue to monitor the owl. Within the last
year, the FWS issued a “90-day finding” under the Endan-
gered Species Act, 16 U.S.C. § 1533(b)(2)(A), as a prelude to
a possible determination that the California spotted owl
should be listed under the Act. 90-Day on a Petition to List
the California Spotted Owl as Threatened or Endangered, 70
3228 EARTH ISLAND INSTITUTE v. USFS
Fed. Reg. 35,607 (June 21, 2005). The FWS stated that “the
petition presents substantial scientific or commercial evidence
that listing the species may be warranted.” Id. The FWS is
currently engaged in a 12-month review of the California
spotted owl as required by 16 U.S.C. § 4(b)(3)(B). Id. Among
the factors prompting the 12 month review was the 2004 Sup-
plement, as well as new evidence concerning the effects of
fires on the owls. Id. at 35,612.
Under the 2004 Supplement, PACs for California spotted
owls must be maintained regardless of actual occupancy by
owls. After a stand-replacing event such as a major fire, habi-
tat conditions within a 1.5-mile radius around the PAC must
be evaluated in order to identify opportunities for re-mapping
a destroyed or diminished PAC. Only if there is insufficient
suitable habitat for designating a PAC within this 1.5-mile
radius can a PAC be delisted.
Earth Island argues that the FEISs allowed for excessive
cutting of trees that would otherwise survive in areas used by
California spotted owls, did not adequately take into account
studies showing the owls’ use of already-burned areas,
allowed the creation of isolated islands of habitat, and failed
to retain sufficient numbers of large snags for use by owls.
Earth Island relies upon the declaration of Monica Bond, who
has a B.A. in Biology from Duke University and an M.S. in
Wildlife Science from Oregon State University. She is the
lead author of two peer reviewed studies of the California
spotted owl published in 2002 and 2004. Bond contends that
the Power and Fred Projects
will have significant negative effects on the Califor-
nia spotted owl by substantially reducing the amount
of potential foraging habitat within the project sites,
by a) utilizing inaccurate mortality guidelines (see
Declaration of Edwin Royce) which incorrectly cate-
gorize some areas withing the Power and Fred’s
project areas as high-severity burn (and thus unsuit-
EARTH ISLAND INSTITUTE v. USFS 3229
able for owls) when they are in fact live tree areas
available to the owl for foraging purposes; b) ignor-
ing significant new scientific information which
indicates that spotted owls actually utilize forests
burned at even high severities; c) by allowing pock-
ets of forest unburned or burned at low- to moderate
severity within a severe burn area to become isolated
islands of habitat and reducing their value to the
spotted owl; and d) failing to maintain large snags
throughout the areas designated to be logged, for
recruitment as legacy trees when the forest regener-
ates in the future. In addition, these projects call for
extensive clearcut logging of habitat elements impor-
tant to spotted owls (large trees and snags) within
currently occupied spotted owl Protected Activity
Centers (“PACs”) and Home Range Core Areas
(“HRCAs”). All of these activities will result in the
loss of potentially critical foraging grounds that are
necessary to maintain the population of California
spotted owls in the Eldorado National Forest.
The USFS responds with a declaration by Chuck Loffland,
a wildlife biologist employed by the USFS. Loffland does not
describe his academic background. He states that he has been
conducting surveys of the California spotted owl and perform-
ing analyses of effects on the owl since 1989. He concedes
that “it appears from the few studies cited by Ms. Bond that
owls may use burned habitat to some degree,” but that “the
scope and duration of that use are not well studied or well
understood.” He notes that Bond herself states that more sci-
entific study is needed. He states further:
Ms. Bond claims that most or all of the PACs will
be clearcut. . . . This is simply untrue. First, there
will be no salvage activity within suitable habitat.
For the non-core (unsuitable) portion of PACs within
the Power project area, salvage will only remove
hazard trees and trees that are dead with 100%
3230 EARTH ISLAND INSTITUTE v. USFS
crown kill in excess of the 4 largest size class snags
per acre retained. This will leave behind 4 large
snags/acre and all trees that have any green needles
and even a remote chance of surviving. . . . Although
Ms. Bond characterizes project activity in the PACs
as wholly detrimental to the owl, the Forest Service
decided to remove dead and hazard trees from non-
core PAC areas in order to provide protection against
stand-destroying fire. (emphasis in original).
Loffland also disputes some of Bond’s characterizations of
the logging that will take place within specified PACs.
According to the Power FEIS, there were approximately
5,880 acres of suitable owl habitat before the Power Fire; this
habitat was reduced to 2,750 acres as a result of the fire. Ten
PACs were in the Power Fire area before the fire. One was
not affected by the fire and no treatment was proposed; one
was burned at high and moderate intensity, and the USFS
determined that not enough suitable habitat remained to sup-
port a PAC; and the boundaries of the remaining eight were
redrawn to encompass the best remaining habitat within a 1.5
mile radius of the center of the area. Within seven of these
redrawn PACs, the USFS determined that certain areas were
unsuitable for spotted owl habitat (deemed “non-core” areas)
and that salvage logging in these non-core areas could pro-
ceed, but leaving standing the four largest snags (dead trees)
per acre. No logging is permitted within the “core,” or suit-
able habitat, areas of any PAC.
An analysis of the direct effects of the Power Project was
limited to the area that currently remains suitable for nesting
or foraging. Because no logging is allowed in core areas, the
Power FEIS notes that any direct effects would be limited to
hazard zone areas, meaning roads, power line corridors, and
the fire flume. Thus, the potential direct effects were limited
to the removal of roadside hazard trees on 295 acres of key
habitat within PACs. In addition, the Power Project is esti-
EARTH ISLAND INSTITUTE v. USFS 3231
mated to affect 660 acres of the 10,560 total suitable HRCA
acres that are within or adjacent to the Power Project area.
In discussing the indirect effects, the Power FEIS states that
in high- and moderate-intensity burn areas in HRCAs, the
largest 5.8 snags per acre will be retained, but that salvage
harvesting will proceed according to the tractor-harvesting
mortality guidelines. The cumulative effects analysis in the
Power FEIS focused on logging activities proposed for 540
acres of private land within the Power Fire area and 873 acres
of private land outside the area; on a project to thin trees in
order to reduce fuel loads; and on the planned replanting of
700 acres of burned plantation areas.
The Power FEIS states that the effect of the Power Project
may be to reduce the quality of owl habitat, but that the proj-
ect would not reduce the overall amount of owl habitat. The
Power FEIS predicts that to the extent the Power Project
increases future fire resiliency, it may have the effect of
increasing the amount of available habitat in the long run. It
concludes that there would be no trend resulting in the federal
listing of the California spotted owl as an endangered species.
In the Freds Fire area, all or portions of three spotted owl
PACs were present before the fire. One PAC burned at high
and moderate intensity, and the USFS determined that insuffi-
cient habitat remained to support this PAC. The boundaries of
the remaining two PACs were redrawn to encompass the best
1.5 miles of habit surrounding the center of the area. The
Freds FEIS proposed logging in non-core areas of the redrawn
PACs.
According to the Freds FEIS, approximately 3,255 acres of
suitable owl habitat existed on both national forest and private
land before the Freds Fire. After the fire, only 285 suitable
acres remained in areas of low-intensity burn, while approxi-
mately 1,848 acres remained in areas of moderate- and high-
3232 EARTH ISLAND INSTITUTE v. USFS
intensity burn. In addition, some 332 acres of green trees are
scattered across the moderate- and high-intensity burn areas.
According to the direct and indirect effects analysis in the
Freds FEIS, no logging would take place in core areas within
the PACs; no logging would be proposed in areas of low-
intensity burn; and hardwoods and green trees would not be
harvested in areas of moderate- and high-intensity burn. The
FEIS also notes that while the 332 acres of remaining scat-
tered green trees provide some post-fire habitat for the owls,
suitability in those 332 acres would “likely be reduced over
time as more fire-weakened trees die. Dead trees lose foliage
and therefore would reduce canopy closure to levels below
what is preferred by owls.”
The USFS argues that it adopted a conservative approach
to logging in both FEISs in order to ensure that the California
spotted owl is protected. It points to the fact that it assumed
the presence of owls in PACs pending the completion of sur-
veys. It also argues that it considered the information concern-
ing the owl’s use of post-fire habitat and determined that the
findings were too inconclusive to affect its impact analysis.
We have elsewhere interpreted the “hard look” requirement
as entailing both a complete discussion of relevant issues as
well as meaningful statements regarding the actual impact of
proposed projects. In Native Ecosystems Council, we held that
where an EIS used a calculation for determining the summer
range of elk herd that was inconsistent with a specific require-
ment contained in the regional forest plan, the USFS had not
taken a hard look because it did not “provide a full and fair
discussion of the potential effects of the project . . . and did
not inform decisionmakers and the public of the reasonable
alternatives which would avoid or minimize adverse impacts
[to the elk at issue].” 418 F.3d at 965 (internal quotations and
alterations omitted). Likewise, in Klamath-Siskiyou Wildlands
Center v. Bureau of Land Management, 387 F.3d 989, 994-95
(9th Cir. 2004), we held that an environmental assessment
EARTH ISLAND INSTITUTE v. USFS 3233
(“EA”) did not take a “hard look” at the cumulative impacts
of a proposed action where the EA had sections devoted to
discussing direct, cumulative, and foreseeable actions, but
gave no objective or qualified assessment of the combined
environmental impacts of the information presented. Most
recently, we concluded that where an EIS stated without
meaningful explanation that a post-fire salvage project would
have a negative impact on black-backed woodpeckers but
would not result in a trend toward federal listing, this did not
constitute a “hard look.” Ecology Ctr., Inc. v. Austin, 430 F.3d
1057, 1067 (9th Cir. 2005).
In Earth Island, Earth Island claimed that the FEIS for the
Star Fire Restoration Project did not adequately analyze the
cumulative impact of the project’s destruction of an HRCA on
the California spotted owl. Earth Island, 351 F.3d at 1306-07.
While the FEIS in Earth Island had acknowledged that spot-
ted owls exhibit “high site fidelity” and that a pair of owls had
returned to the project area, it “never assessed the potential
role of the remaining suitable habitat within the former
HRCA for a maintained [PAC] despite the acknowledged
presence of owls in the area.” Id. at 1307. For this reason, we
held that the “omission amounts to an insufficient consider-
ation of cumulative impact under NEPA.” Id.
We conclude that the Power and Freds FEISs have not
taken the requisite “hard look” at the effects of the two proj-
ects on the California spotted owl. We concluded above that
the FEISs did not adequately analyze probable tree mortality.
The likely consequence of the apparent overprediction of tree
mortality is excessive logging. This likely excessive logging,
in turn, is likely to produce adverse effects on the California
spotted owl that are not adequately analyzed in the FEISs.
Further, it is likely that the projects will substantially
reduce potential foraging habitat because the FEISs’ designa-
tion of non-core areas, where logging will occur, is based
upon the USFS’s determination that because these areas were
3234 EARTH ISLAND INSTITUTE v. USFS
heavily burned they are not likely to be suitable owl habitat.
According to Bond — both in her declaration and in her pub-
lished work — the California spotted owl uses burned areas
for foraging in the short-term, and these areas may also pro-
vide important benefits in the long-term.
[14] The FEISs cannot assume that simply because the owl
habitat studies are preliminary, the adverse impacts discussed
therein will not occur. Rather, the FEISs must respond explic-
itly and directly to conflicting views in order to satisfy
NEPA’s procedural requirements. By removing trees that
might survive in areas it assumes to be unsuitable for Califor-
nia spotted owl habitat, the FEISs allow logging in what could
well be suitable habitat. The FEISs do not explain in any
detail how their determinations that habitat was “unsuitable”
were made, and do not investigate or analyze how redrawing
the boundaries of the PACs and HRCAs might negatively
impact the owls. We therefore conclude that the FEISs do not
satisfy the requirement under NEPA that the agency take a
“hard look” and that there be a “full and fair discussion”
allowing informed public participation and informed decision-
making.
B. Likelihood of Success on the Merits
of Earth Island’s Challenge under the NFMA
[15] Earth Island argues that the FEISs’ reliance on the
Breeding Bird Survey (“BBS”) does not satisfy its obligations
under the NFMA to conduct population surveys for certain
Management Indicator Species (“MIS”) bird species. The
NFMA requires that a forest plan “comply with substantive
requirements of the [NFMA] designed to ensure continued
diversity of plant and animal communities and the continued
viability of wildlife in the forest. . . .” Austin, 430 F.3d at
1063; see also 16 U.S.C. § 1604(g)(3)(B). The 2001 Frame-
work identifies certain birds as MIS species, for which
increased population monitoring is required. The 2004 Sup-
EARTH ISLAND INSTITUTE v. USFS 3235
plement incorporates the 2001 Framework’s population-
monitoring requirements.
The El Dorado National Forest Land and Resource Man-
agement Plan (“LMRP”), as well as both the Freds and Power
FEISs, list cavity-nesting birds, including the black-backed
woodpecker, hairy woodpecker, and Williamson’s sapsucker,
as MIS species. An MIS species is a bellwether, or class rep-
resentative, “for other species that have the same special habi-
tat needs of population characteristics.” Inland Empire Pub.
Lands Council v. U.S. Forest Serv., 88 F.3d 754, 762 n.11
(9th Cir. 1996). The 2001 Framework states, “[p]opulation
and/or habitat monitoring will be conducted for all MIS and
species at risk. Varying levels of monitoring will be con-
ducted depending on the level of concern associated with each
species; as the level of concern about a species increases, the
investment in monitoring increases.” The 2001 Framework
allows for a very limited degree of habitat monitoring in lieu
of actual population monitoring, stating that “coarse habitat
relationships constitute a relatively insensitive index to the
status of populations and would only be appropriate for spe-
cies with a lower level of concern or for which the status of
the population were also being monitored.”
According to the 2001 Framework, the hairy woodpecker
and Williamson’s sapsucker are low-vulnerability MIS spe-
cies. Low-vulnerability species are monitored to determine
changes in their distribution. Distribution data consist of
“changes in the presence of species across a number of sam-
ple locations” and is a “spatially explicit version of frequency
of occurrence data.” In addition, the 2001 Framework notes
that in an area as large as the Sierra Nevadas, “changes in the
distribution of species represent ecologically significant infor-
mation on the status and change of populations.” Appendix E
of the 2001 Framework makes explicit that population data
must be collected for the hairy woodpecker and Williamson’s
sapsucker. The black-backed woodpecker is not specifically
3236 EARTH ISLAND INSTITUTE v. USFS
mentioned in the 2001 Framework, but it is designated as an
MIS species in the LMRP and in the two FEISs.
As a preliminary matter, we conclude that NFMA regula-
tions promulgated in 1982 apply to the 2001 Framework and
2004 Supplement. These regulations require population moni-
toring. See 36 C.F.R. § 219. Because the 2001 Framework
and 2004 Supplement were developed based on regulations in
effect before November 9, 2000, transitional rules, now con-
tained at 36 C.F.R. § 219.14, govern this case. The applicable
regulation provides:
For units with plans developed, amended, or revised
using the provisions of the planning rule in effect
prior to November 9, 2000, the Responsible Official
may comply with any obligations relating to man-
agement indicator species by considering data and
analysis relating to habitat unless the plan specifi-
cally requires population monitoring or population
surveys for the species. Site-specific monitoring or
surveying of a proposed project or activity area is
not required, but may be conducted at the discretion
of the Responsible Official.
36 C.F.R. § 219.14(f) (emphasis added).
The district court found that in light of the hairy wood-
pecker and Williamson’s sapsucker’s classification as “low
vulnerability” species, a lower level of monitoring is envi-
sioned for these birds, and that the use of BBS data satisfies
this lower level of monitoring. Earth Island argues that the
district court erred because the BBS data are insufficient and
inaccurate.
The BBS is a cooperative program sponsored by the gov-
ernments of the United States and Canada to monitor several
North American bird species. Under the BBS, the William-
son’s sapsucker and black-backed woodpecker are listed in
EARTH ISLAND INSTITUTE v. USFS 3237
the “red” category, meaning that the results are “very impre-
cise” and the data suffer from low regional abundance and
small sample sizes. The hairy woodpecker is listed in the
“blue” category, which reflects data of “moderate precision”
and of “moderate abundance,” but which still “may not pro-
vide valid results.” The BBS notes other potential problems
in its overall data collection, specifically that the trends do not
take into account activity outside of the range of the survey;
that the surveys are only conducted by roadside, and “may not
be representative of regional habitat changes”; and that within
the range of the survey, many habitats are not well covered
and that species within those habitats are consequently
“poorly sampled.”
Although not controlling on this court, Sierra Club v.
Eubanks, 335 F. Supp. 2d 1070 (E.D. Ca. 2004), is instruc-
tive. In Eubanks, the same district judge as in this case
granted a preliminary injunction against the logging and pro-
posed timber sale provided in the Red Star Restoration Proj-
ect. Id. at 1073. The Tahoe Forest Plan expressly provided for
annual population monitoring of MIS species, and it listed
each species “ ‘for which population trend data is expected to
be obtained.’ ” Id. at 1081 (quoting the administrative record).
The USFS argued that population survey information was not
required as long as MIS habitats were adequately analyzed,
and that because the Red Star Restoration Project would not
diminish the habitat for MIS species, actual monitoring was
not required. Id. The district court agreed with this premise,
stating that “[h]abitat analysis is an acceptable substitute for
population trend data if there is enough underlying data to
support such an analysis. . . .” Id. at 1082. Nevertheless, the
district court concluded that the Red Star Restoration Project
failed to comply with the El Dorado LRMP, the Sierra
Nevada Framework, and the NFMA because underlying data
for MIS species that could provide the necessary information
for a habitat analysis did not exist. Id.
In its order denying Earth Island’s request for a preliminary
injunction, the district court distinguished Eubanks by con-
3238 EARTH ISLAND INSTITUTE v. USFS
cluding that in this case, such underlying data, in the form of
the BBS, do exist. In support of this conclusion, the district
court relied upon Forest Conservation Council v. Jacobs, 374
F. Supp. 2d 1187, 1207 (N.D. Ga. 2005). In Jacobs, the dis-
trict court found that the BBS provided sufficient data. How-
ever in Jacobs, the USFS relied on other data in addition to
the BBS. Id. at 1205.
We do allow the USFS to conduct habitat analyses in place
of population monitoring under certain circumstances. In
Inland Empire Public Lands Council, various environmental
groups claimed that an EIS did not perform a proper popula-
tion analysis under the NFMA for several sensitive species
living in a project area, including the black-backed wood-
pecker. 88 F.3d at 759. We upheld the USFS’s use of a habitat
management analysis where the USFS had (1) consulted field
studies showing how many acres of territory an individual
species needed; (2) assumed that the amount of acreage
remained constant no matter the actual size of the individual
species’ territory; and then (3) examined the proposed alterna-
tives to see how many acres of necessary habitat remained
after the timber was harvested. Id. We also upheld the USFS’s
decision to not engage in a detailed analysis of one species
because nesting and feeding habitat requirements were not
available, stating that “an analysis that uses all the scientific
data currently available is a sound one.” Id. at 762.
[16] The USFS relies upon Inland Empire to argue that by
analyzing the amount of habitat affected, rather than direct
population counts of the birds, it has satisfied the NFMA.
This argument fails for two reasons. First, a plain reading of
the regulations does not support the USFS’s argument. The
transitional rules state that the USFS may use habitat analysis
in lieu of population monitoring only when a forest plan does
not specifically require population monitoring. The 2001
Framework and 2004 Supplement provide that the USFS may
use “population monitoring and/or habitat analyses.” How-
ever, in discussing the hairy woodpecker and Williamson’s
EARTH ISLAND INSTITUTE v. USFS 3239
sapsucker, the Framework expressly requires “population
monitoring,” specifically in the form of “distribution data.” It
is difficult to see how distribution data could effectively be
gathered in the absence of actual population monitoring, and
we reject the USFS’s argument that it is under no obligation
to determine population trends for the hairy woodpecker or
Williamson’s sapsucker.
Second, although we agree that the USFS could have relied
on habitat monitoring for the black-backed woodpecker, the
USFS has not conducted a habitat analysis on the level of that
found satisfactory in Inland Empire. See Idaho Sporting
Cong. v. Thomas, 137 F.3d 1146 (9th Cir. 1998) (allowing
habitat analysis under the NFMA in certain circumstances).
The FEISs discuss various studies of black-backed woodpeck-
ers that confirm their preference for burned forest habitat.
Table 3-52 of the Power FEIS groups cavity-nesting birds into
three different species groups, lists the areas included in the
project areas that are “assumed to provide high and moderate
capability habitat,” and then lists the number of available
acres of high and moderate capability habitat. There is no
indication that the USFS consulted current or accurate field
studies to arrive at these numbers, and there is no identifica-
tion of the methodology used in determining what constitutes
suitable habitat.
[17] As we stated in Native Ecosystems Council, “[o]ur
case law permits the Forest Service to meet the wildlife spe-
cies viability requirements by preserving habitat, but only
where both the Forest Service’s knowledge of what quality
and quantity of habitat is necessary to support the species and
the Forest Service’s method for measuring the existing
amount of that habitat are reasonably reliable and accurate.”
428 F.3d at 1250. In Austin, we noted that the black-backed
woodpecker is “particularly dependent upon post-fire land-
scapes.” 430 F.3d at 1065. There, we concluded that the
USFS had failed to provide a factual basis sufficient to satisfy
the NFMA because the EIS did not indicate how the USFS
3240 EARTH ISLAND INSTITUTE v. USFS
determined that habitat levels were not critically low or how
it planned to generate additional habitat to counteract losses
from logging. Id. at 1068. We likewise conclude here that
because the USFS’s habitat analysis of the black-backed
woodpecker has not provided a factual basis for determining
the quantity or quality of suitable habitat, its analysis does not
satisfy the NFMA.
We also conclude that the USFS’s reliance on BBS data to
meet its population monitoring obligation was arbitrary and
capricious. Table 3-53 in the Power FEIS (Table 3-54 in the
Freds FEIS) is entitled “Population trend of cavity-nesting
birds.” Immediately beneath the table, the FEIS states,
“NOTE: based upon Breeding Bird Survey routes in the
Sierra Nevada physiographic province, from 1996-2003.”
Table 3-53 then proceeds to list the population status of the
hairy woodpecker as “Decreasing” and the population status
of the Williamson’s sapsucker as “Unknown.” The population
status of the black-backed woodpecker is also listed as “Un-
known.” A footnote to the hairy woodpecker figure states,
“Data is not statistically significant; tendency is estimated
using the population trend classification system described in
Siegel and deSante (1999).”
[18] The 2004 Supplement specifically mentions the need
for annual monitoring of MIS species, stating that only after
a period of annual monitoring will there be “sufficient under-
standing of important habitat characteristics that we can confi-
dently monitor habitat without annual monitoring of species’
distribution and abundance.” We therefore disagree with the
USFS that annual monitoring is not required. The USFS has
not complied with 36 C.F.R. § 219.19 because it has not suffi-
ciently monitored the population of the hairy woodpecker and
Williamson’s sapsucker. We hold that the BBS alone cannot
satisfy the population monitoring requirement, and the USFS
has acted arbitrarily and capriciously under the NFMA in
relying upon it. Cf. Colo. Wild v. U.S. Forest Serv., 299 F.
Supp. 2d 1184, 1189 (D. Colo. 2004) (holding that where an
EARTH ISLAND INSTITUTE v. USFS 3241
FEIS stated that it “lacks qualitative data” for one species, had
“no specific data” for another, and was “currently in the pro-
cess of establishing a monitoring program” for a third, the
USFS had not met its monitoring obligations under the 1982
regulations); Forest Guardians v. U.S. Forest Serv., 180 F.
Supp. 2d 1273, 1282 (D. N.M. 2001) (holding that under the
1982 regulations, the BBS did not satisfy the monitoring
requirement where the district court could not tell whether
population surveys had been conducted for the area at issue).
With respect to the black-backed woodpecker, we also hold
that the USFS has acted arbitrarily and capriciously by relying
on inadequate habitat monitoring data. See Lands Council v.
Powell, 395 F.3d 1019, 1036 (9th Cir. 2005) (holding that the
USFS violated the NFMA where the data for a habitat analy-
sis were outdated and featured inaccurate estimates).
[19] The USFS’s approval and implementation of both
Projects without appropriate or sufficient population and habi-
tat data is contrary to the NFMA and governing provisions of
the forest plan. The district court erred in finding otherwise.
C. Overall Likelihood of Success on the Merits
[20] For the foregoing reasons, we conclude that Earth
Island has shown a “strong likelihood of success on the mer-
its” of both its NEPA and NFMA challenges to the Power and
Freds FEISs and RODs. Earth Island, 351 F.3d at 1297. We
now evaluate the remaining criteria for granting a preliminary
injunction.
D. Possibility of Irreparable Injury, Balance of Hardships,
and Advancement of the Public Interest
Because Earth Island has shown “a strong likelihood of
success on the merits,” it need only show “the possibility of
irreparable injury” if preliminary relief is not granted, and that
the “balance of hardships” tips in its favor. Id. Further, in this
case it is appropriate that it be required to show “the advance-
3242 EARTH ISLAND INSTITUTE v. USFS
ment of the public interest.” Id.; Kootenai Tribe v. Veneman,
313 F.3d 1094, 1125 (9th Cir. 2002) (“[W]here the purpose
of the challenged action is to benefit the environment, the
public interest must be taken into account in balancing the
hardships”).
[21] The “possibility of irreparable injury” has clearly been
shown. We have stated that “[e]nvironmental injury, by its
nature, can seldom be adequately remedied by money dam-
ages and is often permanent or at least of long duration, i.e.,
irreparable.” Nat’l Parks & Conservation Ass’n v. Babbitt,
241 F.3d 722, 737 (9th Cir. 2001) (internal quotations omit-
ted). In this case, Earth Island contends that the Power and
Freds Projects may result in the unnecessary cutting of trees
that would otherwise survive, in harm to the California spot-
ted owl, and in harm to several MIS bird species. Earth Island
has certainly shown that there is a possibility of such injury,
and that such injury is irreparable.
[22] The balance of hardships also tips in Earth Island’s
favor. The USFS and SPI contend, with some reason, that
they will suffer economic losses if we enjoin the timber sales.
But in Babbitt, we stated that a cruise ship’s “loss of antici-
pated revenues . . . does not outweigh the potential irreparable
damage to the environment.” Id. at 738. Further, in Earth
Island we noted the importance of preserving the public’s
interest in “preserving precious, unreplenishable resources.”
Earth Island, 351 F.3d at 1309 (quoting Kootenai Tribe, 313
F.3d at 1125).
Finally, we believe that a preliminary injunction advances
the public interest. The preservation of our environment, as
required by NEPA and the NFMA, is clearly in the public
interest.
Conclusion
We have noticed a disturbing trend in the USFS’s recent
timber-harvesting and timber-sale activities. See, e.g., Ecol-
EARTH ISLAND INSTITUTE v. USFS 3243
ogy Ctr., Inc. v. Austin, 430 F.3d 1057 (9th Cir. 2005) (hold-
ing that the USFS’s post-fire treatment of old-growth forest
stands in the Lolo National Forest violated both the NFMA
and NEPA, and that the EIS failed to explain adequately the
adverse impacts of the proposed plan on the black-backed
woodpecker); Lands Council v. Powell, 395 F.3d 1019 (9th
Cir. 2005) (reversing the district court’s grant of summary
judgment to the USFS because its EIS did not take a “hard
look” at past timber harvests or current trout habitat condi-
tions); Idaho Sporting Cong. v. Rittenhouse, 305 F.3d 957
(9th Cir. 2002) (remanding to the district court to enjoin two
timber sales approved in violation of the NFMA and NEPA).
See also Utah Envtl. Cong. v. Bosworth, 421 F.3d 1105 (10th
Cir. 2005) (holding that the USFS did not properly monitor
MIS species and did not consider a reasonable range of alter-
natives in a proposed timber-harvesting project); Sierra Club
v. Eubanks, 335 F. Supp. 2d 1070 (E.D. Cal. 2004) (granting
a preliminary injunction against salvage logging provided for
in the USFS’s post-fire Red Star Restoration Project); Sierra
Club v. Bosworth, 199 F. Supp. 2d 971 (N.D. Cal. 2002)
(rejecting the USFS’s argument that post-fire salvage burning
was needed to prevent a future fire and enjoining implementa-
tion of post-fire salvage logging); Colo. Wild v. U.S. Forest
Serv., 299 F. Supp. 2d 1184 (D. Colo. 2004) (granting a pre-
liminary injunction of a timber salvage project because the
USFS failed to gather population data for MIS species); For-
est Guardians v. U.S. Forest Serv., 180 F. Supp. 2d 1273 (D.
N.M. 2001) (reversing authorization of a timber sale in the
Cibola National Forest because of the USFS’s failure to col-
lect adequate MIS population data).
It has not escaped our notice that the USFS has a substan-
tial financial interest in the harvesting of timber in the
National Forest. We regret to say that in this case, like the
others just cited, the USFS appears to have been more inter-
ested in harvesting timber than in complying with our envi-
ronmental laws.
3244 EARTH ISLAND INSTITUTE v. USFS
[23] We reverse the district court’s denial of Earth Island’s
request for a preliminary injunction, and we remand to the
district court for proceedings consistent with this opinion. Our
injunction pending appeal shall remain in effect for 30 days
following the issuance of the mandate in order to allow the
district court sufficient time to fashion such preliminary
injunctive relief as it deems appropriate consistent with this
opinion.
REVERSED AND REMANDED.
NOONAN, Circuit Judge, concurring:
There is no doubt that the district court articulated the
wrong standard as to the possibility of harm that the plaintiffs
must show. Earth Island v. United States Forest Service, 351
F.3d 1291, 1298 (9th Cir. 2003) (Earth Island I). There is also
no doubt in my mind that the financial interest of the Forest
Service requires further investigation and evaluation. See id.
at 1309 (Noonan, J., concurring). That the parties have not
pursued this problem does not give the Forest Service a pass.
If it is indeed a biased adjudicator, its determination is a nul-
lity. Tumey v. Ohio, 273 U.S. 510, 521 (1927).
It is not too difficult for a court of appeals to discern the
correct legal standard for an injunction and to see the problem
created by a financial interest on the part of the adjudicator.
It is more difficult for this court to master the massive record
in an environmental case and to be confident in its discrimina-
tion between expert opinions. Recognizing the mastery of the
available data that distinguishes the majority opinion, I cannot
say that I am sure as to Earth Island’s probable success. I
would remand to the district judge (1) to obtain information
as to the importance of the sales to the Forest Service’s opera-
tion; (2) to apply the correct legal standard; and (3) to make
EARTH ISLAND INSTITUTE v. USFS 3245
its own estimate of the probability of Earth Island’s success
on the merits.