A. WALLACE TASHIMA, United States Circuit Judge.
Now pending before the court are the parties' cross-motions for summary judgment, which have been fully briefed and argued and, on March 17, 2014, taken under submission. For the reasons set forth below, Defendants' and Intervenors-Defendants' motions are
In 2003, Defendants completed a revision of the 1988 Payette National Forest Land and Resource Management Plan. FS005768 [ROD at 1]. The Intermountain Regional Forester received a number of appeals of that revision, asserting that it failed adequately to address the risk of disease transmission between bighorn sheep and domestic sheep, and thus to protect bighorn sheep populations. Id. The Chief of the Forest Service agreed and instructed the Regional Forester to "reanalyze the potential impacts of domestic sheep grazing on bighorn sheep viability." Id.
The 2010 Final Supplemental Environmental Impact Statement (FSEIS) and Record of Decision (ROD) are the product of that remand. In 2008, the Forest Service released a Draft SEIS that considered the effects on bighorn sheep viability of various wildlife management alternatives. FS017430-575. In 2010, following notice and comment, Defendants released the FSEIS and ROD. FS005762-98 (ROD); FS005028-5761 (FSEIS). Those documents formalized Defendants' decision to adopt an alternative ("Alternative 7O modified") that reduces domestic sheep grazing on the Payette National Forest (Payette) by approximately 70%. FS005781 [ROD at 14]. Defendants concluded that it is necessary to limit domestic sheep grazing to protect bighorn sheep against the risk of disease transmission from domestic sheep. FS005777-83 [ROD at 10-16].
Plaintiffs brought suit for declaratory and injunctive relief under the National Environmental Policy Act (NEPA), 42 U.S.C. §§ 4321-4370h, challenging the adequacy of the FSEIS and ROD. Plaintiffs now move for summary judgment, and Defendants and Intervenors-Defendants cross-move.
Summary judgment is proper where there is no genuine issue of material fact
"The Administrative Procedure Act (`APA') provides authority for the court's review of decisions under NEPA...." N. Idaho Cmty. Action Network v. U.S. Dep't of Transp., 545 F.3d 1147, 1152 (9th Cir. 2008). "Under the APA, the district court may only set aside agency actions that are `arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.'" Id. (quoting 5 U.S.C. § 706(2)(A)). A decision is arbitrary and capricious
Id. at 1152-53 (quoting Lands Council v. McNair, 537 F.3d 981, 987 (9th Cir.2008) (en banc)). "When specialists express conflicting views, an agency must have discretion to rely on the reasonable opinions of its own qualified experts." Marsh v. Or. Natural Res. Council, 490 U.S. 360, 378, 109 S.Ct. 1851, 104 L.Ed.2d 377 (1989). Indeed, "[a] court generally must be at its most deferential when reviewing scientific judgments and technical analyses within the agency's expertise under NEPA." Native Ecosystems Council v. Weldon, 697 F.3d 1043, 1051 (9th Cir.2012) (internal quotation marks omitted).
NEPA is a procedural statute; "it does not dictate the substantive results of agency decision making." Id. Rather, "[i]ts purpose is to ensure that federal agencies take a `hard look' at the environmental consequences of their proposed actions before deciding to proceed." Id. "[T]he agency must, at a minimum, support its conclusions with studies that the agency deems reliable." N. Plains Res. Council, Inc. v. Surface Transp. Bd., 668 F.3d 1067, 1075 (9th Cir.2011) (citing Lands Council, 537 F.3d at 994). "[T]he agency must `explain the conclusions it has drawn from its chosen methodology, and the reasons it considered the underlying evidence to be reliable.'" Id. (quoting Lands Council, 537 F.3d at 994). "An agency will have acted arbitrarily and capriciously only when `the record plainly demonstrates that [the agency] made a clear error in judgment in concluding that a project meets the requirements' of NEPA." Weldon, 697 F.3d at 1043 (alteration in original) (quoting Lands Council, 537 F.3d at 994).
Plaintiffs contend that Defendants violated NEPA in three ways: by (A) failing adequately to support their assumption that domestic sheep transmit deadly bacteria to bighorn sheep; (B) failing to take a "hard look" at potential risk factors to bighorn sheep viability other than domestic sheep grazing; and (C) using inadequate data and models. Plaintiffs also contend that Defendants (D) violated this Court's previous orders by relying on the findings and conclusions of a committee formed in violation of the Federal Advisory Committee Act (FACA).
Plaintiffs first argue that Defendants failed adequately to support their assumption that domestic sheep transmit a deadly pathogen to bighorn sheep. Plaintiffs say that the assumption lacks adequate support for two reasons: (1) it failed to account for expert agency comments, in violation of 40 C.F.R. § § 1500.1(b), 1502.24; and (2) it failed properly to address the relevance of unavailable or incomplete scientific information, in violation of 40 C.F.R. § § 1502.22, 1500.2(b).
40 C.F.R. § 1500.1(b) provides that "expert agency comments" are one of a few factors that are "essential to implementing NEPA." 40 C.F.R. § 1500.1(b); see also id. § 1502.24 ("Agencies shall insure the professional integrity, including scientific integrity, of the discussions and analyses in environmental impact statements."). Plaintiffs contend that Defendants violated this requirement by failing to consider input from the Agricultural Research Service (ARS), the in-house research agency of the U.S. Department of Agriculture. Plaintiffs focus on the allegedly overlooked input of two ARS scientists in particular, Drs. Knowles and Highland. Plaintiffs argue that Defendants failed to solicit input from the ARS, Knowles, or Highland; ignored a paper (the Knowles and Rink paper) casting doubt on the link between domestic sheep grazing and bighorn health issues; and ignored Knowles' criticism of a paper (the Lawrence, et al., paper) that found such a link. Defendants respond, first, that the ARS, Knowles, and Highland are not relevant "experts," because they have no expertise in "wildlife management." Second, they argue that they did account for opposing viewpoints, including Knowles'.
NEPA case law demands that courts give deference to agencies' reasonable selection of viewpoints among competing experts. See Marsh, 490 U.S. at 378, 109 S.Ct. 1851; Weldon, 697 F.3d at 1051. The evidence demonstrating that disease transmission occurs between domestic and bighorn sheep is by no means conclusive. See FS005778 [ROD at 11] (acknowledging this uncertainty). But nothing in the record suggests that it was unreasonable for Defendants to have concluded that that extensive (albeit developing) body of evidence "indicate[s] that contact of wild bighorn populations with domestic sheep does pose a risk of disease transmission and die-offs in the free-ranging bighorn populations." FS005092 [FSEIS at 3-11]; see also FS005778 [ROD at 11] (concluding despite the "scientific uncertainty" that "the majority of literature" supports the plausibility of disease transmission). To the extent that Plaintiffs merely express an opposing position in that debate, Defendants did not act arbitrarily and capriciously by rejecting it.
Plaintiffs' better argument is that Defendants simply consulted the wrong kinds of experts, making an informed decision impossible. Indeed, Plaintiffs go so far as to argue that Defendants'"lost the ... deference otherwise accorded [to] agencies making predictions within their special expertise" because Defendants did not have special expertise.
Defendants' initial response to Plaintiffs' argument is too facile, but their second suffices. Defendants respond initially that they did have special expertise (in fact, that Plaintiffs did not) because the relevant decision in this case is a "wildlife management decision." Defendants are correct that "wildlife management" is one component of the decision, or perhaps the broader category of decision, but there is a "disease transmission" decision subsidiary to the "wildlife management" decision (whether domestic sheep in fact transmit disease to bighorn sheep) that is Plaintiffs' focus. Thus, Plaintiffs distinguish the experts in this case with great particularity: veterinarians, infectious animal disease experts, pathologists, wildlife biologists, and wildlife ecologists. Defendants did not consult the first three types of experts, Plaintiffs argue, who are experts in disease transmission, so Defendants did not have relevant expertise to understand the transmission question.
Plaintiffs appear to be correct that Defendants did not retain experts with specialized expertise in disease transmission,
Plaintiffs further argue that the FSEIS and ROD failed to obtain or properly to address unavailable or incomplete information linking domestic sheep to bighorn sheep deaths in natural conditions. Plaintiffs argue, in effect, that the disease transmission science is uncertain. They contend that Defendants did not acknowledge this uncertainty, that the transmission science is "incomplete," and that no evidence shows conclusively that transmission can happen in the wild. While Plaintiffs are correct that the science does not definitively demonstrate that disease transmission occurs, they are again incorrect to say that Defendants did not acknowledge the uncertainty.
Plaintiffs next argue that Defendants failed to take a "hard look" at "the multitude of other risk factors" that contribute to bighorn sheep health, including that: (1) disease transmission is endemic to bighorn sheep populations; (2) wolves transmit disease to bighorn sheep; and (3) bighorn sheep contract disease by grazing on lands off of the Payette.
Some data to which Plaintiffs point show that "once [a particular pathogen] [has] been introduced to bighorn sheep populations, [it] may become endemic and continue cycling for decades." FS005088 [FSEIS at 3-7]; see also FS018503 (raising this point in a 2007 meeting). Contrary to Plaintiffs' suggestion, Defendants considered and addressed this possibility. They responded, in effect, that one risk (endemic disease cycling) does not take away from other risks (transmission from domestic sheep). FS006134. So "[e]ven though other events may lead to die-offs of bighorn sheep, the Forest Service still ha[d] a responsibility to address the risk of disease posed by its management decisions relating to domestic sheep grazing." Id.; see also FS006137 (noting the need to manage the risk of disease from domestic sheep even if some bighorn populations carry disease endemically); FS006225 (same); FS006234 (same). Defendants further concluded that there was insufficient information on endemic disease transmission to model the possibility accurately. FS005447. And they cited research showing that "pathogens evolve as they move within and between species" and that certain domestic sheep carry pathogens that are lethal to bighorn sheep but not domestic sheep — suggesting that there are still pathogen strains capable of transfer, even if domestic sheep have already transferred one lethal strain. FS005095 [FSEIS at 3-14].
Defendants' response is reasoned and reasonable. NEPA does not require that agencies eliminate all risks contributing to a problem before they can address one of those risks. Cf. Lands Council, 537 F.3d at 1001 (explaining that agencies need not even "present every uncertainty in its EIS" because "such a requirement might inadvertently prevent [agencies]
Plaintiffs next argue that Defendants failed to address "how the reintroduction of wolves has affected bighorn sheep movements." See FS006150 (summarizing this concern); FS061817 (raising the issue). The record, however, demonstrates in at least two ways that Defendants adequately considered the effect of wolves.
First, the record shows that Defendants directly addressed the effect of wolves. FS006150 (responding to the concern); FS067823-24 (same). Defendants concluded that wolves did not significantly affect the risk of disease transmission to bighorn sheep based on the facts that: (1) bighorn sheep and wolves had cohabited the Payette for many years and that bighorn sheep had "survived in large quantities"; (2) bighorn sheep are generally located in "terrain too severe for wolves"; and (3) no research has shown that wolves affect bighorn susceptibility to disease.
Second, the telemetry data used in Defendants' Risk of Contact Model was collected from beginning two years after the reintroduction of wolves on the Payette. FS005108. Thus, the outcome of that contact analysis accounted for the presence of wolves, FS006150, and found a continued risk of contact between bighorn and domestic sheep, FS005169 [FSEIS at 3-88]. Defendants thus addressed Plaintiffs' concern, and their response is reasonable.
Plaintiffs lastly argue that Defendants did not adequately consider that bighorn sheep will continue to graze off of the Payette and to contact domestic sheep there. Plaintiffs point out that Defendants' own analysis suggested that off-Payette grazing would result in "substantial" contact and threaten bighorn sheep viability. But Defendants did not account for off-Payette grazing in their Disease Model. See FS005075 [FSEIS 2-18].
The court rejects Plaintiffs' arguments for the same reasons as above. Defendants recognized the increased risk of contact from off-Payette grazing but they cannot control that risk factor. FS005782 [ROD at 15]. Rather, they can control grazing only on the Payette,
In sum, contrary to Plaintiffs' contentions, the three additional risk factors do not "upset [Defendants'] assumption that domestic sheep on the Payette transmit bacteria to wildlife bighorns which then kills them." The factors merely add risk to bighorn sheep viability without "upset[ting] [Defendants'] assumption" about the risk that continues to exist from domestic sheep. Defendants addressed the challenges and responded reasonably in rejecting them. The court therefore rejects Plaintiffs' three "hard look" challenges.
NEPA requires that agencies "insure the professional integrity, including scientific integrity, of the discussions and analyses in environmental impact statements." 40 C.F.R. § 1502.24; see also Lands Council, 537 F.3d at 994 (stating the agency must "support its conclusions... with studies that the agency, in its expertise, deems reliable"). Agencies must also give "up-front disclosures of relevant shortcomings in [their] data or models." Lands Council v. Powell, 395 F.3d 1019, 1032 (9th Cir.2005).
Plaintiffs argue that Defendants' "Source Habitat," "Risk of Contact," and "Disease" models failed to satisfy NEPA's standard of professional and scientific integrity and that Defendants failed to disclose the models' shortcomings.
Plaintiffs contend that Defendants' Source Habitat Model — used to identify habitat suitability and where bighorn sheep are likely to spend time — assumed that bighorn sheep were "more likely [to] foray into habitat containing available forage (considered `suitable habitat')."
This argument is based on a mischaracterization of Defendants' model. Defendants first classified habitats — as either "source habitat," "connectivity area," or "non-habitat" — along a number of metrics, including the type of terrain and horizontal visibility. FS005103-04 [FSEIS at 3-22 to 3-24]; FS005677-78. None of the metrics appears to be defined by available forage. Next, Defendants modeled bighorn sheep preferences for the habitats. FS005119-20 [FSEIS at 3-38 to 3-39]. The record does not indicate that Defendants merely
This modeling did not equate suitable habitat with occupied habitat. Indeed, Defendants consistently acknowledged that "availability of suitable habitat does not infer occupied habitat" and thus that "the inference that suitable habitat is an accurate proxy for occupied habitat is not useful in assessing the persistence of bighorn sheep populations." FS005080-81 [FSEIS at 2-23 to 2-24]; see also FS005733. Although Plaintiffs impute this assumption to Defendants, Defendants' models appear considerably more complex.
Second, where Plaintiffs' criticisms were presented to the agency, the agency responded reasonably, stating that the model incorporated topographic and landscape factors, FS005420 (responding to comment); see also FS005677 (describing the model as incorporating such factors), and that the model was adjusted to account for the fact that bighorn sheep had "higher `preference' for non-habitat and connectivity areas" when on forays than when in their core herd home range, FS005422 (responding to comment).
Third, the model's reliability and integrity is corroborated by the process by which it was developed. The model was initially designed over a decade ago by the Hells Canyon Initiative, FS005677-78, which used the model "to identify promising areas for reintroducing bighorn sheep in and around the canyon." Defendants' experts updated the model to include more accurate data and conditions, and they tested the model with actual telemetry data. FS005678-79. Finally, the model was presented on numerous occasions to Defendants' broader team of experts, including the experts at the CADMS, for review and comment. See, e.g., FS060242-23; FS060262-78; FS064542-43. This process suggests that Defendants conscientiously, "in [their] expertise, deem[ed] [the model] reliable." Lands Council, 537 F.3d at 994. In light of all of the foregoing factors, the court cannot say that Defendants "made a clear error in judgment" in determining that the Source Habitat Model "meets the requirements of [NEPA]." Id.
The Risk of Contact Model identifies the probability that a bighorn sheep will enter into an area where domestic sheep graze (defined as an "allotment"). FS005684. The model has two broad components: one that addresses where bighorn sheep are likely to spend most of their time (the "core herd home range" analysis) and another that addresses when, how frequently, and how far bighorn sheep are likely to foray (the "foray" analysis). Id.; see also FS005108 [FSEIS at 3-27]. Plaintiffs contend that the model is "crippled by shortcomings in the telemetry data used for the model"; in particular, that no data show that "any of the bighorn sheep studied in the 12-year period had actually ever made contact with a domestic sheep on the Payette."
Second, and more importantly, where Plaintiffs' criticisms were addressed to the agency, the agency responded reasonably. Contrary to Plaintiffs' contention, Defendants discussed their model methodology in depth. FS005103-31; FS005672-721. See Citizens to Pres. Overton Park v. Volpe, 401 U.S. 402, 420, 91 S.Ct. 814, 28 L.Ed.2d 136 (1971) (stating that a court's "review is to be based on the full administrative record that was before the Secretary at the time he made his decision"). Further, Defendants reasonably addressed criticisms of the Risk of Contact Model that were addressed to the agency. See, e.g., FS006203-07 (addressing bandwidth criticism of foray analysis); FS006108-09 (discussing risk of contact from straying domestic sheep). Again, Defendants need not affirmatively address every uncertainty but rather "acknowledge and respond to" relevant comments. See Lands Council, 537 F.3d at 1001. Plaintiffs point to no relevant comments that were addressed to the agency to which it did not respond.
Third, the reliability of the Risk of Contact Model is corroborated by the process by which Defendants designed it.
Plaintiffs lastly assail the Disease Model. That model estimates the likelihood that a bighorn sheep will come into contact with a domestic sheep when the bighorn sheep intersects the domestic sheep allotment, that the bighorn sheep will contract disease from the domestic sheep, and that the bighorn sheep will return to its herd and transmit the disease. FS005107 [FSEIS at 3-26]. Plaintiffs argue that the model is "drastically oversimplified" because it inflates the probability that bighorn sheep in a domestic sheep allotment will make "effective contact" with a domestic sheep (contact that transmits disease) and transmit disease back to its herd. Specifically, Plaintiffs contend that Defendants should have considered the probability of physical contact between bighorn and domestic sheep (not only that bighorn sheep will be present in a domestic sheep allotment) and the probability that disease will be transferred, including consideration of "susceptibility, infectious dose being shed, duration of shedding, and degree and duration of physical contact."
The court rejects Plaintiffs' challenge to the Disease Model. Plaintiffs volunteer more variables that they contend Defendants should have included in the model to improve its accuracy. Defendants respond that inclusion of the additional variables would make the model too "complex and unwieldy," and "would have required the Agency to attempt to make more use of the available data than would have been scientifically defensible." Instead, Defendants opted for a simpler model that effectively accounted for Plaintiffs' additional variables by positing a range of probabilities of disease outbreak.
This is not a case in which an agency "entirely failed to consider an important aspect of the problem," N. Idaho Cmty. Action Network, 545 F.3d at 1152, i.e., the existence of variables that affect the likelihood of a disease outbreak.
Finally, the process by which Defendants developed the model again corroborates its reliability. Defendants built the model on published, peer-reviewed models. FS005081 [FSEIS at 2-24]; FS005713. And Defendants consulted widely with their own experts, including CADMS, for review and comment.
The court therefore rejects Plaintiffs' challenges to Defendants' models. Plaintiffs cite Lands Council v. Powell to support their claim that Defendants erred by not more thoroughly discussing the shortcomings of their models. Lands Council, 395 F.3d at 1032 ("Although there are some disclosures of the model's shortcomings..., nowhere do the disclosures cover the limitations of [the model] shown by the Lands Council and now conceded by the Forest Service."). But Defendants here do not concede any undisclosed limitations, and the record reveals none. Plaintiffs' criticisms are not shortcomings but methodological disagreements with Defendants' reasoned choices. Or relevant shortcomings were disclosed. See, e.g., FS005778 [ROD at 11] (acknowledging uncertainty in the models). Defendants did not have "to affirmatively present every uncertainty." Lands Council, 537 F.3d at 1001. Defendants needed only respond to "significant" and "reasonably support[ed]" uncertainties. Id. Defendants did so, and their choices do not appear arbitrary and capricious.
Lastly, Plaintiffs argue that Defendants improperly relied on the findings and conclusions of the RADT committee,
This "principal assumption" is the same assumption that Plaintiffs earlier attack as uninformed and unsupported by expert comments and complete information. See supra Part III.A. As discussed supra, Defendants relied primarily on their extensive review of disease transmission research and literature to substantiate the link. See FS005087-95 [FSEIS at 3-6 to 3-14]. This court's previous orders permit "the Forest Service's use of the underlying science that may exist to support the Committees' recommendations." Idaho Wool Growers Ass'n v. Schafer, No. CV 08-394 S BLW, 2009 WL 3806371, at *3 (D.Idaho Nov. 9, 2009); see also Wool Growers I, 637 F.Supp.2d at 880 ("If, indeed, the Committees represented only a mechanism to collect and summarize all available data relevant to the issue at hand, that same, underlying information would exist to support future agency decisions as well."). Defendants'"principal assumption" satisfies that "underlying science" criterion.
Plaintiffs rightly note that the document cited in the FSEIS as support for the "principal assumption" provides little support. FS005100 [FSEIS at 3-19] (citing the 2003 FEIS). Indeed, the Chief of the Forest Service remanded the 2003 FEIS for the very reason that it failed adequately to address "viability or the potential for disease transmission." FS005768 [ROD at 1]. But the citation appears, at most, misplaced. The citation does not suggest that Defendants were relying on the RADT committee findings (indeed the 2003 FEIS predates the RADT committee), especially when Defendants defend the assumption extensively in the literature review section of the FSEIS. Because Defendants' "principal assumption" relied on "underlying science," not the RADT committee report, it did not violate this court's previous orders.
For the reasons set forth above,
1. Defendants' and Intervenors-Defendants' Motions for Summary Judgment (Docs. 44, 51) are
2. Plaintiffs' Motion for Summary Judgment (Doc. 37) is
3. Judgment shall be entered consistent herewith.
Knowles' 2010 criticism postdates the finalization of the FSEIS and ROD, so Defendants did not address it there. FS070049-50. Rather, they amply and reasonably respond to the criticism in their declarations and in letters responding to Knowles' letter. Plaintiffs are correct t that Defendants' post-hoc defense of the Lawrence paper cannot justify its use in the FSEIS and ROD. But Defendants did not need to justify use of the paper because Knowles' criticisms were not presented to the agency. See Lands Council, 537 F.3d at 1001 (explaining that agencies need not "affirmatively present every uncertainty" but rather "acknowledge and respond to comments by outside parties that raise significant scientific uncertainties and reasonably support that such uncertainties exist").
The important point with respect to both documents (the 2006 outline and 2010 letter) is that Defendants considered opposing viewpoints that were presented to them and reasonably rejected them. See infra note 4 and accompanying text; cf. Alaska Survival v. Surface Transp. Bd., 705 F.3d 1073, 1087 (9th Cir.2013) ("All that NEPA requires is that the lead agency consider these concerns [of other agencies] and explain why it finds them unpersuasive."). Defendants did not also need to credit Knowles' criticism to satisfy NEPA.
This debate simply underscores the extent to which Plaintiffs' criticisms of Defendants' models are technical, methodological disagreements. It is clear from the record that Defendants developed their methodology reasonably and with care. Plaintiffs' disagreements do not render Defendants' reasoned decisions arbitrary and capricious, given Defendants' extensive review of the models, their responses to criticism, and the other indicia of reliability detailed supra. See Weldon, 697 F.3d at 1053 ("The mere fact that [a plaintiff] disagrees with [a defendant's] methodology does not constitute a NEPA violation.... [W]e may not insert our opinions in the place of those of [experts]. Rather, we are required to apply the highest level of deference in our review of the Forest Service's scientific judgments in selecting [a given] methodology." (citation omitted)).
Defendants respond in two relevant ways. First, they contend that their range of probabilities also accounts for the probability of cohabitation. In other words, the cohabitation question is not antecedent to the probability index; it is included within the range. Also, Defendants contend that no data exist to estimate the probability of actual physical contact (as opposed to general intersection with an allotment). Second, Defendants contend that Thurmond's calculation relies on bad math and scant (if any) data. In any case, Defendants note that 5% is not even one order of magnitude greater than 1.28%.
This exchange again illustrates the extent to which Plaintiffs' challenges amount to methodological disagreements. And Defendants' response is reasonable: the probability index could be said to include the cohabitation variable, data on cohabitation is lacking, and Defendants' low probability (5%) is not far removed from Thurmond's proposed figure (1.28%).
