Opinion by Judge BERZON for sections I and II; Opinion by Judge GOULD for sections III and IV.
OPINION
BERZON, Circuit Judge.“[Wjhile in life the great whale’s body may have been a real terror to his foes, in his death his ghost[became] a powerless panic to [the] world.” Herman Melville, Moby Dick 262 (W.W. Norton & Co. 1967) (1851). This modern day struggle over whale hunting began when the United States granted support and approval to the Makah Tribe’s (“the Tribe’s”) plan to resume whaling.
The Tribe, a traditional Northwest Indian whale hunting tribe, had given up the hunt in the 1920s. In recent years, the Tribe leadership came to regret the cultural impact on the Tribe of the lapse of its whale hunting tradition. As part of a general effort at cultural revival, the Tribe developed plans to resume pursuing gray whales off the coast of Washington State and in the Strait of Juan de Fuca. The worldwide hunt for whales in the years the real-life Captain Ahabs roamed the high seas, however, seriously depleted the worldwide stock of the cetaceans. As a result of the near extinction of some species of whales, what had been a free realm for ancient and not-so-ancient mariners became an activity closely regulated under both federal and international law. This case is the second in which we have considered whether the federal government’s approval of the Tribe’s plans to pursue once again the Leviathan of the deep runs afoul of that regulation. See Metcalf v. Daley, 214 F.3d 1135 (9th Cir.2000).
The plaintiffs, citizens and animal conservation groups,1 challenge, as did the plaintiffs in Metcalf, the government’s failure to prepare an environmental impact statement (“EIS”) pursuant to the National Environmental Policy Act of 1969 (“NEPA”), 42 U.S.C. § 4321 et seq. They also contend that the Tribe’s whaling plan cannot be implemented because the Tribe has not complied with the Marine Mammal Protection Act of 1972 (“MMPA”), 16 U.S.C. § 1361 et seq. Having reviewed the environmental assessment (“EA”) prepared by the government agencies and the administrative record, we conclude that there are substantial questions remaining as to whether the Tribe’s whaling plans will have a significant effect on the environment. The government therefore violated NEPA by failing to prepare an EIS before approving a whaling quota for the Tribe. We also conclude that the MMPA applies to the Tribe’s proposed whale hunt.
*1010I. Background
A. The Whales
The record discloses that there are two genetically distinct North Pacific gray whale populations — an eastern stock, also known as the California gray whale, and a western stock, confined to East Asian waters. See Steven L. Swartz et al., Review of Studies on Stock Identity in the Gray Whale (Eschrichtius robustus) (hereinafter “Review of Studies ”) (written by scientists employed by National Marine and Fisheries Service (“NMFS”) and National Oceanic and Atmospheric Administration (“NOAA”), at 1 (2000)). The California gray whales migrate annually between the North Pacific and the West Coast of Mexico. These whales were at one time nearing extinction and were therefore listed on the Endangered Species Act list. See Endangered Species Act of 1973, 16 U.S.C. § 1531 et seq.; Metcalf, 214 F.3d at 1138. Protected by the endangered species designation and by other conservation measures, the California gray whale stock revived, so that by 1994 the whale was removed from the endangered species list. See 59 Fed.Reg. 31,094 (Jun. 16, 1994). The NMFS has determined that the eastern North Pacific gray whale stock has now recovered to between 17,-000 and 26,000 whales, a number near its carrying capacity.2 See 63 Fed.Reg. 16,-701, 16,704 (Apr. 6, 1998); 58 Fed.Reg. 3121, 3122 (Jan. 7, 1993); John Calambok-idis et al., Final Report, Range and Movements of Seasonal Resident Gray Whales from California to Southeast Alaska, at 11 (Dec.2000); Review of Studies at 11, 14.3 Most of the migrating whales pass through the Olympic Coast National Marine Sanctuary [“Marine Sanctuary”],4 adjacent to the Makah Tribe’s home territory on the coast of Washington State, on their way to the Bering and Chukchi Seas, and again when heading south for the winter.
Not all of the gray whales, however, make the entire journey to the Far North each summer. On this much the parties agree, although they disagree about the habits of the nonmigrating whales as they pertain to this case.
The plaintiffs contend that a separate group of gray whales remains in and around the Marine Sanctuary waters and within the Strait of Juan de Fuca (south of Vancouver Island and east of the Pacific Ocean) during the summer and early fall, rather than migrating to the Bering and Chukchi Seas with the other eastern stock North Pacific gray whales. See Appendix (map depicting area). This resident group, plaintiffs maintain, arrives in the late spring with the northward migration and remains in the area for the summer, leaving only when the larger contingent of behemoths migrate south for the winter.
The government, in contrast, posits that the whales in the Marine Sanctuary area and the Strait of Juan de Fuca are not a distinct group but rather a rotating one changing from year to year, albeit with some repeat visitors. See Environmental Assessment on Issuing a Quota to the Makah Indian Tribe for a Subsistence Hunt on Gray Whales for the Years 2001 and 2002, at 22-29 (July 12, 2001) [“Final EA”]. The government points to several studies that suggest that if there is any *1011identifiable whale subgroup, it is a much larger one than the plaintiffs suppose. This larger subgroup is denominated by the government the Pacific Coast Feeding Aggregation (“PCFA”). The PCFA, the government maintains, does not migrate all the way north for the summer but ranges over a long stretch of the Pacific Coast from California to Southern Alaska. See id. According to this analysis, although some whales in the PCFA show a tendency to return to the same area along the Pacific Coast, most of them move around among different areas along the West Coast rather than staying in a particular area. Some frequent different locations throughout the summer, and others visit different places each year.
Despite this disagreement among the parties about the habits of the nonmigrat-ing whales, there are some concepts that are not disputed. Scientists, including those relied upon by the government agencies, generally support the assessment that there is a fairly small number of whales who spend some or all of the summer in the general area of the planned Tribe hunt, and that some of these whales return to the area for more than one summer, albeit not necessarily in successive years. See, e.g., John Calambokidis et al., Final Report, Gray Whale Photographic Identification in 1999: Collaborative Research by Cascadia Research, the National Marine Mammal Laboratory, and Humboldt State University, at 8-10 (Dec.2000) (prepared for the National Marine Mammal Laboratory [“NMML”]); Calambokidis et al., Range and Movements, supra, at 3-4, 6-9, 11 (funded by NMML with participation of Dr. James Darling, the plaintiffs’ main expert); Darling Decl. ¶¶ 2-3;5 Re-vieio of Studies at 10, 14-15 (NMFS and NMML study); Jennifer Leigh Quan, Univ. of Wash. Sch. of Marine Affairs, Thesis, Summer Resident Gray Whales of Washington State: Policy, Biological and Management Implications of Makah Whaling, at 1, 7-11 (2000).
Further, while the parties disagree in their assessment of the scientific literature as it pertains to many details regarding the behavior of these returning whales, they agree — and our review of the administrative record confirms — -that overall, the best current scientific evidence indicates that each summer about sixty percent of the whales in the area around Neah Bay and the Strait of Juan de Fuca are returning whales. See Final EA at 24, 27; Review of Studies at 1, 13-15, 20-21 (finding that there are “identifiable gray whales, termed ‘summer residents’ [that] have been observed to return each summer to the same areas at various locations along the Pacific Northwest Coast” for at least part of the season) (citing studies done within the Tribe’s hunting grounds off the Washington Coast and near Vancouver Island); Quan, supra, at 4, 9-10 (supporting the finding that approximately 61% of the whales found in the Tribe’s whaling area were repeat visitors). See also, Darling Decl. ¶¶ 2-3, 7 (approximately sixty percent of the whales identified in a separate area, off central Vancouver Island in any one summer, are seen repeatedly over multiple years, with some whales having returned to the region each summer for more than twenty years).
The total number of whales frequenting the area of the planned Makah Tribe hunt each summer is not known. It is common ground, however, that the whales in the Tribe’s proposed whaling area are a relatively small subgroup of the larger number of nonmigrating whales that forego the complete trip to the North. See Final EA *1012at 26 (seventy whales sighted in the area in 2000); Calambokidis et al., Range and Movements, swpra, at 6 (between forty and forty-five whales were spotted south of Vancouver Island in two summer months during 1998); Darling Decl. ¶¶ 2-3 (the number of whales south of Vancouver Island in the Strait of Juan de Fuca is likely to be similar to the thirty-five to fifty whales observed as residents off the central coast of Vancouver Island); Quan, supra, at 10 (from 1993-1998 ten to thirty-five individual whales were identified in the outer coast near Neah Bay and the Strait of Juan de Fuca).
B. The Makah Tribe and Its Efforts to Resume Whaling
The Tribe is composed of Native Americans whose traditional territory is in Washington State, on the northwestern Olympic Peninsula. In 1855, the United States entered into a treaty with the Tribe, the Treaty of Neah Bay, providing that the Tribe would give up most of its land on the Olympic Peninsula. See 12 Stat. 939, 940 (1855). In exchange, the Tribe was given, inter alia, the “right of taking fish and of whaling or sealing at usual and accustomed grounds and stations....” Id. That the Treaty of Neah Bay is the only treaty between the United States and a Native American tribe that specifically protects the right to hunt whales suggests the historic importance of whaling to the Makah Tribe.
Despite the central place of whaling in their lives, the Tribe ended their whaling expeditions in the late 1920s. Explanations regarding the reasons for the abandonment of this custom include: the federal government’s discouragement and lack of assistance; a decline in demand for whale oil; social and economic dislocation within the Tribe; and the drastic decline of the gray whale population.
Then came, in the early 1990s, both a renewed interest within the Tribe in reviving its traditional whaling customs and the removal of the California gray whale from the Endangered Species Act list. The Tribe therefore determined to resume its traditional whale hunting. In the seventy years since the last hunt, however, whaling had become an activity tightly regulated internationally, through the International Whaling Commission, and domestically, through the Whaling Convention Act,6 and the MMPA, as well as through more general federal environmental legislation. Pursuant to the ICRW, aboriginal subsistence whaling is permitted,7 but such whal*1013ing must conform to quotas for various whale stocks issued by the IWC.
In 1996 the NOAA entered into a written agreement with the Tribe to obtain an aboriginal subsistence quota from the IWC. The United States presented a proposal for such a quota at the annual IWC meeting in June 1996. The proposal proved controversial, however, and some members of the IWC blocked its passage. The House of Representatives Committee on Resources also passed a unanimous bipartisan resolution opposing the Tribe’s hunting proposal. In the face of this opposition the United States withdrew its request.
Before the United States began its next attempt to gain IWC approval, some animal conservation organizations, whale watching groups, and other citizens wrote a letter to the NOAA expressing concern about the prospect of a renewed whale hunt in the waters off the continental United States. The letter charged that the government had violated NEPA by agreeing to help the Tribe obtain hunting rights without conducting an EA. The NOAA quickly produced for public comment a Draft EA, concluding that the Tribe’s hunt would have no significant environmental impact.
A few months later, the NOAA and the Tribe entered into a new agreement similar to the prior one except that the new version required that the Tribe’s management plan provide time and area restrictions “including ... confining hunting activities to the open waters of the Pacific Ocean outside the Tatoosh-Bonilla Line.” Agreement Between the NOAA and the Makah Tribal Council, at 5 (1997). This provision sought to reduce the likelihood that the Tribe would take nonmigrating whales. Four days after this agreement was reached, the NMFS issued a final EA and a finding of no significant impact (“FONSI”) concerning the proposed hunt.
The United States thereupon presented a joint proposal with the Russian Federation to the IWC’s 1997 annual meeting. The joint proposal combined the desired Makah Tribe quota with the Russian request for a whaling quota for its Siberian aboriginal people, the Chukotka, into a single request for permission to take 620 whales over a five-year period. See IWC Chairman’s Report of the 49th Annual Meeting, at 19 (Oct.1997).
Delegates at the IWC meeting again disagreed about whether the Tribe qualified under the aboriginal subsistence exception. Rather than resolving the disagreement, the delegates papered it over with ambiguous language: The new Schedule approved by a majority of IWC members limited use of the California gray whale quota to aboriginal groups “whose traditional aboriginal subsistence needs have been recognised,” but did not say who was to recognize those needs, or how. See id. at 20. So it remained unclear whether a majority of the members considered the Tribe entitled to the aboriginal subsistence exception, or whether instead such recognition was to be conferred by the country issuing the quota. In March 1998, the NMFS announced a quota permitting the Tribe to take8 five gray whales in a one-year period and allowing no more than thirty-three strikes9 over a five year *1014period. See 63 Fed.Reg. 16,701 (Apr. 6, 1998).
Meanwhile, on the day the 1997 FONSI was released and before the IWC and NMFS quotas were issued, a group of concerned citizens and animal conservation organizations filed a complaint in federal court against the federal defendants for violations of NEPA, the WCA, and the Administrative Procedure Act. The primary allegation was that the EA was a deficient effort, put together simply to justify the prior agreement allowing the Tribe to hunt whales. After the district court granted summary judgment for the defendants, the Tribe began whaling and in 1999 killed one whale.
The whale’s demise did not bring this prolonged dispute to an end, for this court reversed the district court in Metcalf. We held that the EA was invalid because it was not produced until after the agreement with the Tribe had been consummated. Id. at 1143-46. A new EA must be drafted, we ordered, “under circumstances that ensure an objective evaluation free of the previous taint.” Id. at 1146. Because we viewed the government defendants’ actions as having been undertaken improperly, we stated that when the new EA was completed and returned to the courts for evaluation, it should be subject to “additional scrutiny [and] the burden shall be on the Federal Defendants to demonstrate ... that they have complied with [the] requirement” to evaluate the environmental impact of the proposal objectively and in good faith. Id.
After the decision in Metcalf, the federal defendants dissolved the agreement with the Makah Tribe (over the Tribe’s protest) and began the EA process anew. The NMFS and the NOAA published a new Draft EA in January 2001. The Draft EA, like the 1997 EA, presented as the most desirable option a whale quota targeted at migrating whales. The restriction was to be accomplished by limiting the hunt to the area west of the Tatoosh-Bonilla line and to months when northward or southward migration was underway. Draft EA at 7. Similarly, the proposed Makah Management Plan only allowed whaling in the “open waters of the Pacific Ocean which are outside the Tatoosh-Bonilla Line.” Management Plan for Makah Treaty Gray Whale Hunting for the Years 1998-2002 (pre-amendment), at 6.
Before the Final EA issued but after the comments period on the Draft EA had closed, the Tribe amended the Management Plan. The amended plan, in contrast to the earlier ones, does not contain any general geographic limitations on the whale hunt. Instead, the new plan allows for the taking of five whales in any one calendar year, with the aggregate number taken from 1998 to 2002 not to exceed twenty whales. See Makah Management Plan at 3. No more than thirty-three whales can be struck between 1998 and 2002, and the number of gray whales struck between 2001 and 2002 cannot exceed fourteen. Id. The amended plan does limit the number of strikes- — but not the number of takes — likely to affect non-migrating whales: For 2001 and 2002, the plan limits to five the number of strikes (1) during the months of the migration, between June 1 and November 30; and (2) at all times in the Strait of Juan de Fuca. Id.
On July 12, 2001, the NOAA and NMFS published a Final EA, based on the amended Management Plan and once again found no significant environmental *1015impact. The Draft EA did not evaluate the amended Management Plan, so there has been no opportunity for public comment on the important amendments. Nor did any of the scientific studies relied on in the EA specifically evaluate the impact of the revised Management Plan. Rather, to the extent those studies and comments discuss the proposed hunt at all, they assume a hunt limited to areas west of the Tatoosh-Bonilla line.
The final step in the administrative saga took place when the NOAA and the NMFS issued a Federal Register notice on December 13, 2001 announcing a quota for the "land{ingj" of five gray whales in 2001 and 2002 and approving the latest Makah Management Plan. 66 Fed.Reg. 64,378 (Dec. 13, 2001).
C. The Current Litigation
The plaintiffs filed this action in January 2002, alleging violations of both NEPA and the MMPA. The Tribe intervened. In April, the plaintiffs moved for a preliminary injunction to prevent an anticipated whale hunt, but the district court denied the motion. Concluding that the federal agencies had taken the requisite "hard look" at the risks associated with the whale hunt and that the court was required to defer to their decision, the district court determined that the plaintiffs did not have a probability of success on the merits. The district court also held that the Treaty of Neah Bay's preservation of the Tribe's whaling rights takes precedence over the MMPA's requirements; the plaintiffs therefore were unlikely to prevail on their MMPA claim as well. The plaintiffs appealed these rulings.
While the preliminary injunction decision was on appeal, the district court granted summary judgment to the defendants. The plaintiffs now appeal the summary judgment order. We consolidated the two appeals and dismissed the preliminary injunction appeal as moot. Now before us is the appeal from the summary judgment order.10
II. NEPA Analysis
A. Standard of Review
The Administrative Procedure Act governs judicial review of agency decisions under NEPA. If an agency decides not to prepare an EIS, the decision not to do so may be overturned only if it is "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law." Native Ecosystems Council v. Dombeck, 304 F.3d 886, 891 (9th Cir.2002) (quoting 5 U.s.c. § 706(2)(A)); Tillamook County v. U.S. Army Corps of Eng'rs, 288 F.3d 1140, 1143 (9th Cir.2002).
More specifically, this court must determine whether the agencies that prepared the EA took a "`hard look' at the environmental consequences" of the proposed action. Blue Mountains Biodiversity Project v. Blackwood, 161 F.3d 1208, 1211 (9th Cir.1998) (citing Oregon Natural Res. Council v. Lowe, 109 F.3d 521, 526 (9th Cir.1997)). The court must defer to an agency conclusion that is "fully informed and well-considered," but need not rubber stamp a "clear error of judgment." Blue Mountains, 161 F.3d at 1211 (quoting Save the Yaak Comm. v. Block, 840 F.2d 714, 717 (9th Cir.1988) and Marsh v. Ore*1016gon Natural Res. Council, 490 U.S. 360, 378, 109 S.Ct. 1851, 104 L.Ed.2d 377 (1989)).11
B. NEPA Standards
NEPA is a statute that aims to promote environmentally sensitive governmental decision-making, without prescribing any substantive standards. Robertson v. Methow Valley Citizens Council, 490 U.S. 332, 353, 109 S.Ct. 1835, 104 L.Ed.2d 351 (1989); Tillamook County, 288 F.3d at 1143. Toward that end, the statute requires, with some exceptions, that all federal agencies consider the environmental impact of their actions. If a federal action “significantly [affects] the quality of the human environment,” then the implementing agency or agencies must prepare an EIS providing a detailed and comprehensive analysis of the potential environmental impacts of the proposed action. See 42 U.S.C. § 4332(C); 40 C.F.R. § 1502 et seq. The EIS must also suggest and -analyze the environmental impact of alternatives to the proposed action. 42 U.S.C. §§ 4332(C) & (E).
Regulations promulgated by the Council on Environmental Quality provide factors that agencies must consider in deciding whether to prepare an EIS and emphasize the importance of involving the public in NEPA evaluations. 40 C.F.R. §§ 1500.2, 1502.4(b). The public must be given an opportunity to comment on draft EAs and EISs, and public hearings are encouraged to facilitate input on the evaluation of proposed actions. See. 40 C.F.R §§ 1503.1, 1506.6.
The CEQ regulations define the term “significantly” for purposes of NEPA as requiring analysis of both the “context” and the “intensity” of the action. Of great importance for purposes of this case, the context of the action includes “society as a whole (human, national), the affected region, the affected interests, and the locality.” 40 C.F.R. § 1508.27(a) (emphasis added).
“Intensity” refers to the severity of the impact, and includes the following considerations:
(1) Impacts that may be both beneficial and adverse. A significant effect may exist even if the Federal agency believes that on balance the effect will be beneficial.
(2) The degree to which the proposed action affects public health or safety.
(3) Unique characteristics of the geographic area such as proximity to historic or cultural resources, park lands, prime farmlands, wetlands, wild and scenic rivers, or ecologically critical areas.
(4) The degree to which the effects on the quality of the human environment are likely to be highly controversial.
(5) The degree to which the possible effects on the human environment are highly uncertain or involve unique or unknown risks.
(6) The degree to which the action may establish a precedent for future actions with significant effects or represents a decision in principle about a future consideration.
(7) Whether the action is related to other actions with individually insignificant but cumulatively significant impacts....
(8) The degree to which the action may adversely affect districts, sites, highways, structures, or objects listed in or *1017eligible for listing in the National Register of Historic Places or may cause loss or destruction of significant scientific, cultural, or historical resources.
(9) The degree to which the action may adversely affect an endangered or threatened species or its habitat that has been determined to be critical under the Endangered Species Act of 1973.
(10) Whether the action threatens a violation of Federal, State, or local law....
40 C.F.R. § 1508.27.
Before deciding whether to complete an EIS, government agencies may prepare a less formal EA which “briefly provides sufficient evidence and analysis for determining whether to prepare an environmental impact statement or a finding of no significant impact.” Tillamook County, 288 F.3d at 1144 (citation omitted). See also Nat’l Parks & Conservation Ass’n v. Babbitt, 241 F.3d 722, 728, 730 (9th Cir.2001); 40 C.F.R. §§ 1501.3, 1501.4, 1508.9, 1508.27. If the EA results in a finding of no significant impact — a FONSI, in NEPA lingo — then no EIS need be completed. See 40 C.F.R. §§ 1500.3, 1500.4(q), 1500.5(1), 1508.13.
Critically for this case, to prevail on the claim that the federal agencies were required to prepare an EIS, the plaintiffs need not demonstrate that significant effects will occur. A showing that there are “ ‘substantial questions whether a project may have a significant effect’ on the environment” is sufficient. Blue Mountains, 161 F.3d at 1212 (quoting Idaho Sporting Cong. v. Thomas, 137 F.3d 1146, 1150 (9th Cir.1998)) (emphasis added). The plaintiffs in this case point to a number of CEQ significance factors as pertinent to raising substantial questions concerning a possible significant effect on the environment but concentrate on three, to which we now turn.
C. Impact on Public Safety
The plaintiffs first focus on the possible impact of the Tribe’s whaling proposal on public safety. The proposal provides that whales will be hunted by a combination of traditional and contemporary methods: Whales must first be struck with harpoons but, once struck, are to be killed, in an effort to make the killing as humane as possible, with high-powered rifles. The plaintiffs maintain that the long range of the rifles and the possibility that injured whales will lash out at nearby people and boats present serious human safety issues.
The government EA analyzes these risks in some detail and finds them insignificant. See Final EA at 63-65. In so concluding, the EA relies in large part on a study by Kirk H. Beattie, the safety expert hired by the Tribe. Beattie made specific safety recommendations, largely adopted in the Makah Management Plan, including: shooting only if the Tribe’s boat is thirty feet or closer to the targeted whale; pointing the rifle away from the shoreline if within 500 yards of it; and having a safety officer on the chaser boat to ensure a clear line of fire for the rifleman. See Kirk H. Beattie, Report, Minimizing the Potential Injury or Death from Rifle Fire to Non Participants in Makah Gray Whale Hunts; Final EA at 64; Ma-kah Management Plan at 6. Beattie went on to conclude that the proposed whale hunt will be far less dangerous than deer hunting, in which there is a risk of injury from ricocheted bullets of approximately one in four million. See Beattie, supra, at 5.
The plaintiffs argue that the government cannot rely on Beattie because he is not an independent expert. Their contention is wrong. The government may rely on experts hired by other parties so long as the agency objectively evaluates the qualifications and analysis of the ex*1018pert. See Friends of Earth v. Hintz, 800 F.2d 822, 834-35 (9th Cir.1986). The government has done just that: Beattie is an undisputed expert, and the EA evaluates his findings.
Furthermore, the EA specifically discounts the opinion of the plaintiffs’ expert, Roy Kline, that firing away from the shoreline is not a solution because the bullets could ricochet 1,700 meters off the line of fire. Because Kline did not consider the specifics of the Tribe’s hunt, including the exact kind of weapons to be used and the mitigating safety measures, the agencies concluded that his concerns were not warranted. See Final EA at 63-65.
The EA points, in addition, to several more factors that reduce the risk to the public: There will be only a few whale hunts and such hunts will not take place near populated areas. Final EA at 64, 69. The Coast Guard also has established a 500 yard exclusionary zone to keep the public from danger during a Tribe whale hunt. Final EA at 64, 69; see also Regulated Navigation Area, 63 Fed.Reg. 52,603 (Oct. 1, 1998). The Coast Guard regulations further require the Tribe to broadcast warnings over the radio and display a warning flag marking the hunting vessel. See id.
We must defer to the expertise of the agency in evaluating scientific evidence. “[W]hen the record reveals that an agency based a finding of no significant impact upon relevant and substantial data, the fact that the record also contains evidence supporting a different scientific opinion does not render the agency’s decision arbitrary and capricious.” See Wetlands Action Network v. U.S. Army Corps of Eng’rs, 222 F.3d 1105, 1120-21 (9th Cir.2000) (citations omitted). The government does not need to show that there is no risk of injury, but only that the risk is not significant.
The agencies’ finding that public safety is not endangered is neither arbitrary nor capricious. Were there no substantial questions raised as to other aspects of the environmental impact of the Tribe’s hunt, no EIS would be required with respect to the public safety concerns alone.
D. Controversy and Uncertainty
Under the CEQ regulations, we must consider whether the effects of the Tribe’s whaling on the human environment are “likely to be highly controversial,” 40 C.F.R. § 1508.27(b)(4), and also whether the “possible effects ... are highly uncertain or involve unique or unknown risks.” 40 C.F.R. § 1508.27(b)(5). A proposal is highly controversial when there is “a substantial dispute [about] the size, nature, or effect of the major Federal action rather than the existence of opposition to a use.” Blue Mountains, 161 F.3d at 1212 (quoting Greenpeace Action v. Franklin, 14 F.3d 1324, 1335 (9th Cir.1993)). Put another way, a proposal can be considered controversial if “substantial questions are raised as to whether a project ... may cause significant degradation of some human environmental factor.” Nat’l Parks, 241 F.3d at 736 (quoting Northwest Envtl. Def. Ctr. v. Bonneville Power Admin., 117 F.3d 1520, 1539 (9th Cir.1997) (Reinhardt, J., concurring in part and dissenting in part)).
There is no disagreement in this case concerning the EA’s conclusion that the impact of the Makah Tribe’s hunt on the overall California gray whale population will not be significant. What is in hot dispute is the possible impact on the whale population in the local area where the Tribe wants to hunt. In our view, the answer to this question'—of greatly increased importance with the revision of the Makah Management Plan so as expressly to allow hunting of local nonmigrating ani*1019mate — is sufficiently uncertain and controversial to require the full EIS protocol.
Our reasoning in this regard is as follows: The government agrees that a relatively small group of whales comes into the area of the Tribe’s hunt each summer, and that about sixty percent of'them are returning whales (although, again, not necessarily whales returning annually). Even if the eastern Pacific gray whales overall or the smaller PCFA group of whales are not significantly impacted by the Makah Tribe’s whaling, the summer whale population in the local Washington área may be significantly affected. Such local effects are a basis for a finding that there will be a significant impact from the Tribe’s hunts. See 40 C.F.R. § 1508.27(a). Thus, if there are substantial questions about the impact on the number of whales who frequent the Strait of Juan de Fuca and the northern Washington Coast, an EIS must be prepared.
The crucial question, therefore, is whether the hunting, striking, and taking of whales from this smaller group could significantly affect the environment in the local area. The answer to this question is, we are convinced, both uncertain and controversial within the meaning of NEPA. No one, including the government’s retained scientists, has a firm idea what will happen to the local whale population if the Tribe is allowed to hunt and kill whales pursuant to the approved quota and Ma-kah Management Plan. There is at least a substantial question whether killing five whales from this group either annually or every two years, which the quota would allow, could have a significant impact on the environment.
The government estimates that a conservative allowable take from a group of 222 to 269 whales is 2.5 whales per year,12 while a less conservative approach would allow killing up to six whales per year from the PCFA. Final EA at 57. Thus, with a smaller group, it would appear that a take of less than 2.5 whales per year could exceed the allowable Potential Biological Removal level or “PBR” established under the MMPA’s standards.
Some of the scientists relied upon by the government worry that takes from the local resident whale population may deplete the number of local whales in the area off the coast of Washington State and in and around the Strait of Juan de Fuca. See Review of Studies at 15 (“[The whales’] fidelity to specific locations could subject them to differential harvests and potential depletions if there are unregulated local takes.”) (emphasis added); Quan, supra, at 13 (finding that there could be an adverse impact on the local whale population in the area of the Tribe’s hunt if the whales’ site fidelity is based on social or familial recruitment); see also Darling Decl. ¶ 7 (“[I]t remains a reasonable possibility that removals of resident whales would deplete their presence in specific areas from which they would require an extended time peri*1020od to recover.”)- These concerns, it should be noted, were expressed at a time when it was expected that the Tribe’s hunt would be structured so as to avoid targeting the nonmigrating whales in the area, a restriction that has in large part been lifted.
The government tries in two ways to minimize the importance of the possible local impact. First, the government maintains that the PCFA — or summer resident whale group, if one exists — is not genetically distinct from the other California gray whales.13 For purposes of applying the CEQ regulations, this consideration is irrelevant. If California gray whales disappear from the area of the Strait of Juan de Fuca, the Marine Sanctuary, or both, that would be a significant environmental impact even if the PCFA whales populating the rest of the Pacific Coast in the summer are genetically identical to the local whales, and even if the PCFA whales are genetically identical to the migrating whales.
Second, the government implies that any whales taken from the local resident group will be replaced in the local area by other whales from the PCFA, so the number of whales locally will not decline. The EA describes the PCFA as composed of whales that move from one feeding area to another rather than staying in one locale for all the summer months. That some of the whales who return, whether annually or intermittently, to the area of the proposed hunt also visit other areas of the coast cannot, however, eliminate concern about the local impact. The fact remains that a majority of the fairly small number of whales identified in the Makah Tribe’s hunting area have been there in previous years, wherever else they have also journeyed. Whether there will be fewer or no whales in the pertinent local area if the hunt is permitted depends not on whether the whales who frequent that area also travel elsewhere, but upon the opposite inquiry: whether whales who heretofore have not visited the area will do so, thereby replenishing the summer whale population in the area, if some of the returning whales are killed.
It is on this latter question that the scientific uncertainty is at its apogee. Almost all of the scientific experts relied upon in the EA state that the effect of taking whales who demonstrate some site fidelity within the Tribe’s hunting area is uncertain. Quan, for example, suggests that much depends on how whales are recruited to the area, an open question requiring further study. See Quan, supra, at 11-13. If the local whales are recruited randomly, removing four whales annually from the Tribe’s hunt area should not have any long-term impact. If the whales are recruited familially, however, “the annual removal of four gray whales could directly [affect the number of whales] observed and utilizing the area.” Quan, supra, at 13.
Similarly, Darling states that “the recruitment mechanism that influences or maintains the resident group of gray whales found in Washington is not known. As a result, it is difficult to predict at this time how the harvesting of resident whales could affect the resident population.” Darling Deck ¶ 10. See also Calambokidis et al., Range and Movements, supra, at 4 (“It is unclear how loyal these [seasonal resident] animals are to the feeding grounds, how they adopt this alternate feeding strategy, and their range of movements.”); Review of Studies, at 20 (“Relatively little is known about how individuals *1021choose feeding grounds throughout their lives.... It is plausible that females may learn their migration route and preferred feeding areas from their mothers. . . . A summer hunt that is localized and very coastal has the potential to adversely affect such localized feeding groups and could lead to distributional changes and local extirpation.")14
The EA's only substantive attempt to address the impact of the Tribe's whaling on the number of whales in the area of the Marine Sanctuary and the Strait of Juan de Fuca is as follows: "With the extreme movements of whales in the [FCFA] both within and between seasons a limit of five strikes over two years should also alleviate any potential local depletion issues." Final EA at 58.15 The EA's conclusion simply does not follow from its premise: That PCFA whales do not spend all summer or every summer in the area of the Tribe's hunt does not eliminate the possibility that the killing of returning whales present in any given year may lead to a depletion of whales in the local area. Obviously, with the demise of some returning whales, fewer whales with the habit of returning to that area in the summer will survive. As the underlying studies establish, the local impact of the Tribe's whaling therefore turns on whether different PCFA whales will fill in for the killed, struck, or frightened whales no longer in the area. This critical question is never analyzed, numerically or otherwise, in the EA.
In short, the record establishes that there are "substantial questions" as to the significance of the effect on the local area. Despite the commendable care with which the EA addresses other questions, the EA simply does not adequately address the highly uncertain impact of the Tribe's whaling on the local whale population and the local ecosystem. This major analytical lapse is, we conclude, a sufficient basis for holding that the agencies' finding of no significant impact cannot survive the level of scrutiny applicable in this case.16
And because the BA simply does not adequately address the local impact of the Tribe's hunt, an EIS is required. See Blue Mountains, 161 F.3d at 1213 (ordering the Forest Service to prepare an MS where the EA's treatment of one important environmental factor was "cursory and inconsistent"); Nat'l Parks, 241 F.3d at 735-36 (requiring preparation of an EIS when the EA admitted that it was not known how serious the dangers of the proposed action were and the EA failed adequately to address opposing expert studies).
E. Precedential Effect
There is a second consideration that buttresses the conclusion that an EIS must be *1022prepared. If approval of a single action will establish a precedent for other actions which may cumulatively have a negative impact on the environment, an EIS may be required. See 40 C.F.R. § 1508.27(b)(6). The plaintiffs argue that the approval of the Tribe’s hunting quota could have such a significant precedential impact on future IWC quotas. Approval of a whaling quota for one group for a limited time period is not binding, however, on future IWC or WCA decisions regarding other groups, or even regarding the same group in the future. This factor is therefore insufficient on its own to demonstrate a significant environmental impact.
There is nonetheless sufficient merit to plaintiffs’ concerns to lend support to the conclusion that there are substantial questions concerning whether the Makah Tribe’s hunt will adversely affect the environment. As noted, it appears that the IWC quota language concerning the aboriginal subsistence exception was left purposely vague. The quota issued jointly to Russia and the United States was limited to whaling by aboriginal groups “whose traditional aboriginal subsistence needs have been recognised.” Conspicuously absent from this phrase is any delineation of who must do the recognizing or how.
Prior to adoption of this language, the understanding among IWC members was that only the IWC could decide which groups met the subsistence exception. The 1997 IWC gray whale quota, as implemented domestically by the United States, could be used as a precedent for other countries to declare the subsistence need of their own aboriginal groups, thereby making it easier for such groups to gain approval for whaling.17 If such an increase in whaling occurs, there will obviously be a significant impact on the environment.
The EA does not specifically address the impact of the quota on any IWC country besides the United States. Instead, the EA only analyzes the possible precedent with regard to other Native American tribes in this country and Canada.
We agree that because Canada is not a member of the IWC, it will not be heavily swayed one way or the other by approval of the Makah Tribe’s whaling quota. But we cannot agree with the agencies’ assessment that because the Makah Tribe is the only tribe that has an explicit treaty-based whaling right, the approval of their whaling is unlikely to lead to an increase in whaling by other domestic groups. And the agencies’ failure to consider the prece-dential impact of our government’s support for the Makah Tribe’s whaling in future IWC deliberations remains a troubling vacuum. We conclude that the possible impact on the heretofore narrow aboriginal subsistence exception supports our conclusion that an EIS is necessary.
* * ^ *
In sum, given the substantial uncertainty and controversy over the local impact of the Makah Tribe’s whaling and its possible precedential effect, an EIS should have been prepared. Of course scientific inquiry rarely yields certainty. But here the *1023agencies' inquiry itself was deficient. Thus, an EIS is required.
There is no doubt that the government put much effort into preparing the lengthy environmental assessment now before us. While a notable attribute of the creatures we discuss in this opinion, girth is not a measure of the analytical soundness of an environmental assessment. No matter how thorough, an EA can never substitute for preparation of an EIS, if the proposed action could significantly affect the environment. See Sierra Club v. Marsh, 769 F.2d 868, 874-76 (1st Cir.1985).
We stress in this regard that an EIS serves different purposes from an EA. An BA simply assesses whether there will be a significant impact on the environment. An EIS weighs any significant negative impacts of the proposed action against the positive objectives of the project. Preparation of an EIS thus ensures that decision-makers know that there is a risk of significant environmental impact and take that impact into consideration. As such, an EIS is more likely to attract the time and attention of both policymakers and the public.
In addition, there is generally a longer time period for the public to comment on an EIS as opposed to an EA, and public hearings are often held. See id. at 875-76. Furthermore, preparation of an EIS could allow additional study of a key scientific issue, the local recruitment scheme of the whales in the Makah Tribe's hunting area. See, e.g., Review of Studies at 21 ("A better understanding of site fidelity and potential stock structure will be gained through continuation and expansion of photographic identification and satellite tagging research on the feeding grounds.
Because the agencies have not complied with NEPA, we set aside the FONSI, suspend implementation of the Agreement with the Makah Tribe, and vacate the approved whaling quota for the Tribe.
III. MMPA Analysis
. Will Anderson, Fund for Animals, Humane Society of the United States, Australians for Animals, Cetacean Society International, West Coast Anti-Whaling Society, Sandra Abels, Cindy Hansen, Patricia Ness, Robert Ness, Lisa Lamb, Margaret Owens, Charles Owens, Peninsula Citizens for the Protection of Whales, Dan Spomer, Sue Miller, and Steph Dutton.
. Carrying capacity is the largest number of a species that a given ecosystem can sustain.
. The studies and other documents cited are in the Administrative Record for the 2001 EA.
. The Marine Sanctuary was designated in 1994 under the National Marine Sanctuaries Act. The Act provides for the identification and protection of marine environment areas of special national significance. See 16 U.S.C. § 1431 et seq. The Marine Sanctuary covers more than 3300 square miles of ocean off the Olympic Peninsula of Washington State.
. This Declaration was submitted to the NOAA and is part of the administrative record.
. The International Convention for the Regulation of Whaling ("ICRW”) was established in 1946 to restrict and regulate whaling. 62 Stat. 1716, 161 U.N.T.S. 72 (Dec. 2, 1946). The ICRW created the International Whaling Commission ("IWC”), comprised of one member from each of the ratifying countries. The IWC is empowered to set international whaling regulations and annual whaling quotas. Id. at arts. Ill, V § 1. The United States signed the Convention, 62 Stat. 1716 (1946), and implemented it domestically in the Whaling Convention Act of 1949 ("WCA”), 16 U.S.C. § 916 et seq. See also 50 C.F.R. § 230.1 (WCA implementing regulations).
. This exception originated in the first quota (termed a "Schedule”) approved under the ICRW, which stated that "[it] is forbidden to take or kill gray whales ... except when the meat and products of such whales are to be used exclusively for local consumption by the aborigines." 62 Stat. at 1723 (emphasis added). The articulation of the aboriginal subsistence exception has varied in ICRW Schedules over time. The precise reach of the exception has remained unclear. See, e.g., Brian Trevor Hodges, The Cracking Facade of the International Whaling Commission as an Institution of International Law: Norwegian Small Type Whaling and the Aboriginal Subsistence Exemption, 15 J. Envtl. L. & Litig. 295, 304-05 (2000); Nancy C. Doubleday, Aboriginal Subsistence Whaling: The Right of Inuit to Hunt Whales and Implications for International Environmental Law, 17 Denv. J. Int'I L. & Poly 373, 384-94 (1989).
. To "take” a whale under the Makah Management Plan means "to flag, buoy or make fast to a whale catcher, including a canoe, chase boat or support boat.” See Makah Management Plan for Makah Treaty Gray Whale Hunting for the Years 1998-2002, as amended by Council Resolution No. 57-01 on May 30, 2001 ["Makah Management Plan”], at 2.
. A "strike” is defined in the Makah Management Plan as "any blow or blows delivered to a whale by a harpoon, lance, rifle, explosive device or other weapon. When used as a verb, 'strike' means the act of delivering such *1014a blow or blows to a whale. A harpoon blow is a strike only if the harpoon is embedded in the whale. Any rifle shot which hits a whale is a strike.” Makah Management Plan at 2 (emphasis added). For purposes of the quota, multiple blows to one whale are counted as a single strike. Id.
. The parties dispute whether the 1997 administrative record should have been considered part of the administrative record in this case. They agree, however, that we may use the material contained in the Excerpts of Record prepared for the preliminary injunction appeal, and we have done so. Further, we have relied only on studies cited in the 2001 EA and material in the 2001 administrative record. We therefore have no need to resolve the dispute regarding the scope of the administrative record in the summary judgment appeal and do not do so.
. It might be argued that our review may be less deferential than usual, given the specific directive in Metcalf that we give any later EA "additional scrutiny” and place the burden of demonstrating objective evaluation on the government. See Metcalf, 214 F.3d at 1146. We find it unnecessary to apply this directive from Metcalf, because we conclude in this section that NEPA was violated based on the traditional standard of review set forth above.
. The government's calculation of the acceptable Potential Biological Removal level ("PBR”) number for the PCFA is not without controversy. The PBR is calculated based on an MMPA formula which strives to prevent any marine mammal from being reduced below its optimum sustainable population level. The EA relies on an estimate that there are 222 to 269 whales in the PCFA. Final EA at 28, 57. Studies, however, suggest that these figures are not representative and overestimate the actual number of whales in the group. See Calambokidis et al., Range and Movements, supra, at 9; Calambokidis et al., Gray Whale Photographic Identification in 1999, supra, at 4, 12. Based on the higher range of 222 to 269, the EA finds that a conservative estimate allows for the taking of 2.5 whales per year without jeopardizing the PCFA population. See Final EA at 57. The Makah Management Plan, however, is set at five per year — the higher end of the range of acceptable removal levels (2.5 to 6) for this possibly exaggerated number of whales.
. The studies upon which the government relies are not definitive on this issue. See Final EA at 28; Review of Studies at 14 (stating that it is not known whether the "summer residents are genetically distinct”); Quan, supra, at 4 ("The biological significance of the seasonal residents is unclear.”).
. Swartz and his colleagues go on to recommend that "[s]mall-localized feeding groups should be closely monitored and management adapted to detect and avoid adverse population changes resulting from harvests." Id. at 21.
. The EA also quotes from the IWC's Scientific Committee: "[T]he Committee agreed that there is a need for better understanding of site fidelity and potential stock substructure in eastern gray whales to improve advice on management." Final EA at 29. Far from negating scientific uncertainty, this conclusion by an international group of experts supports the conclusion that there are unresolved issues critical to assessing the possible local environmental impact of the Tribe's hunt.
. Although we reach our result under ordinary NEPA analysis, we note in addition that the amendments to the Makah Management Plan between the Draft and Final EAs, followed by an apparent change in the agencies' position regarding the importance of targeting only nonmigrating whales, provides some basis for doubting that the government has met the special burden, imposed by Metcalf, of establishing its objectivity.
. There have been disputes in the IWC in recent years over efforts by other countries to gain approval for quotas for whaling communities viewed domestically, but not internationally, as meeting the aboriginal exception. See, e.g., Leesteffy Jenkins & Cara Romanzo, Makah Whaling: Aboriginal Subsistence or a Stepping Stone to Undermining the Commercial Whaling Moratorium, 9 Colo. J. Int'l Envtl. L. & Pol’y 71, 88-99 (1998); see also James Brooke, Japan Cuts Whaling Rights for Native Peoples of Arctic, N.Y. Times, May 25, 2002, at A4 (describing Japan’s efforts to block whaling quotas for Alaskan and Siberian aboriginal people until quota was approved for its own whaling communities).