[HCC.gif] [homeB.gif] [sitemapH.gif] [casespacer.gif] [lefttopcorner.gif] FOR PUBLICATION UNITED STATES COURT OF APPEALS FOR THE NINTH CIRCUIT NATIONAL PARKS & CONSERVATION ASSOCIATION, Plaintiff-Appellant- Cross-Appellee, v. Nos. 99-36065 BRUCE BABBITT, Secretary, United 99-36094 States Department of the Interior; ROBERT STANTON, Director, D.C. No. CV-97-00456-JKS National Park Service, Defendants-Appellees, OPINION and HOLLAND AMERICA LINE-WESTOURS, INC., Defendant-Intervenor- Appellee-Cross-Appellant. Appeal from the United States District Court for the District of Alaska James K. Singleton, District Judge, Presiding Argued and Submitted July 31, 2000--Anchorage, Alaska Filed February 23, 2001 Before: Dorothy W. Nelson, Stephen Reinhardt, and Sidney R. Thomas, Circuit Judges. Opinion by Judge Reinhardt _________________________________________________________________ 2413 COUNSEL Catherine E. Stetson, Hogan & Hartson L.L.P., Washington, D.C., for the plaintiff-appellant. Sean H. Donahue, United States Department of Justice, Envi- ronment and Natural Resources Division, Washington, D.C., for the defendants-appellees. 2418 Cynthia Pickering Christianson, Anchorage, Alaska, for the defendant-intervenor-appellee. _________________________________________________________________ OPINION REINHARDT, Circuit Judge: Glacier Bay National Park and Preserve is a place of "unri- valed scenic and geological values associated with natural landscapes" and "wildlife species of inestimable value to the citizens." The Bay was proclaimed a national monument in 1925 and a national park in 1980. UNESCO designated Gla- cier Bay an international biosphere reserve in 1986 and a world heritage site in 1992. Not surprisingly, many people wish to visit the park. As there are no roads to Glacier Bay, most tourists arrive by boat. To be more specific, most -- approximately 80% of the park's visitors -- arrive on large, thousand-passenger cruise ships. In 1996 the National Park Service (Parks Service) com- menced implementation of a plan that increased the number of times cruise ships could enter Glacier Bay each summer season immediately by 30% and overall by 72% if certain conditions were met. In its environmental assessments, the Parks Service acknowledged that this plan would expose the park's wildlife to increased multiple vessel encounters, noise pollution, air pollution, and an increased risk of vessel colli- sions and oil spills. The Parks Service also acknowledged that it did not know how serious these dangers to the environment were, or whether other dangers existed at all. Nevertheless, declaring that its plan would have "no significant impact" on the environment, the Parks Service put it into effect without preparing an environmental impact statement (EIS). The plaintiff National Park and Conservation Association (NPCA), a nonprofit citizen organization, alleges that the 2419 Parks Service's failure to prepare an EIS violated the National Environmental Policy Act (NEPA), 42 U.S.C. S 4321 et seq. It seeks an order requiring the Parks Service to prepare an EIS and enjoining implementation of the plan pending its comple- tion. The district court ruled that an EIS was not required because the Parks Service had made its findings after ade- quately "canvassing the existing knowledge base. " We reverse the district court's ruling and remand with instructions to enjoin the plan's increases in vessel traffic, including any portion already put into effect, until the Parks Service has completed an EIS. FACTUAL AND PROCEDURAL HISTORY There may be no place on Earth more spectacular than the Glacier Bay. Located in the Alaskan panhandle, surrounded by snow-capped mountain ranges, Glacier Bay extends sixty miles inland and encompasses ten deep fjords, four of which contain actively calving tidewater glaciers, and approximately 940 square miles of "pristine" marine waters. The air quality, though fragile, is still unspoiled and permits those fortunate enough to be visitors a crisp, clear view of the Bay with its glacier faces as well as the opportunity to breathe the fresh and invigorating air. The park is the habitat for an extraordi- nary array of wildlife. On the land, pioneer plant communities grow in areas recently exposed by receding glaciers. Moose, wolves, and black and brown bears roam the park's spruce and hemlock rain forest. Bald eagles, kittiwakes, murrelets, and other seabirds nest along the shore; sea otters, harbor seals, Steller sea lions, harbor and Dall's porpoises, minke, killer, and humpback whales reside in the bay. The Steller sea lion and the humpback whale, two of the marine mammal species that inhabit Glacier Bay, are imper- iled. The Steller sea lion was listed as a threatened species under the Endangered Species Act (ESA), 16 U.S.C. 1531 et seq., in 1990. The worldwide population of the species 2420 declined by as much as 48% in the thirty years prior to 1992.1 Glacier Bay has several "haul-out" sites where hundreds of Steller sea lions gather. The humpback whale, "the most gamesome and lighthearted of all the whales," Herman Mel- ville, Moby Dick, 123 (Harrison Hayford & Hershel Parker, eds., W.W. Norton & Co. 1967) (1851), has been listed as an endangered species since the enactment of the ESA in 1973. Until a moratorium was instituted in 1965, commercial whal- ing decimated the worldwide population of humpback whales. Today only 10,000 to 12,000 remain.2 A subpopulation of humpbacks spends the summer feeding season in southeast Alaska, including the waters of Glacier Bay; other humpbacks remain there throughout the year. Watercraft -- cruise ships, tour boats, charter boats, and private boats -- provide primary access to Glacier Bay's attractions. Approximately 80% of the park's visitors are cruise ship passengers. According to the Parks Service's envi- ronmental assessment, the "key attraction of the visit to Gla- cier Bay . . . [is] [t]he glaciers at the head of the West Arm [of the Bay.] [They] are larger, more active, and considered by the [cruise-ship] companies to offer a more spectacular expe- rience."3 The ships linger at the glaciers from between fifteen _________________________________________________________________ 1 In 1997 the Steller sea lion was reclassified, for the worse, as an endan- gered species. Greenpeace v. Nat'l Marine Fisheries Serv., 80 F. Supp. 2d 1137, 1139 (W.D. Wash. 2000). 2 In 1991 the U.S. National Marine Fisheries Service (Fisheries Service) established a Final Recovery Plan for the humpback. The plan sets a long- term goal of restoring 60% of the species's pre-whaling population (about 125,000) and a more immediate goal of maintaining and enhancing "cur- rent or historical habitats used by humpback whales by reducing distur- bance from human-produced underwater noise in important habitats when humpback whales are present and [encouraging] government entities at all levels to correct existing impacts on habitats of humpback whales." 3 To get there, cruise ships cross the entrance to the Bay and traverse its western bank, both areas particularly frequented by humpback whales. See Christine M. Gabriele, Population Characteristics of Humpback Whales in Glacier Bay and Adjacent Waters 9 (1994). 2421 minutes to an hour and provide a large, high viewing platform from which to witness the crack and crash of the great ice masses as they cast off huge shards of floating ice. Although the ships' height permits an unobstructed view of the park's geologic features, it limits close views of the wildlife and veg- etation that form such a significant feature of the park. Between 1968 and 1978, vessel traffic in Glacier Bay increased dramatically. In 1978 the U.S. National Marine Fisheries Service (Fisheries Service) produced a "biological opinion" based on its studies of the humpback whale popula- tion in Glacier Bay. The biological opinion expressed concern over the "uncontrolled increase of vessel traffic " in the whales' departure from the Bay during 1978 and 1979, and cautioned that a continued increase in the amount of vessel traffic "would likely jeopardize the continued existence of the humpback whale population frequenting southeast Alaska." The Fisheries Service recommended that the Parks Service regulate the number of vessels entering Glacier Bay; restrict vessels from approaching and pursuing whales; and conduct studies on whale feeding behavior, the effect of vessels on whale behavior, and the acoustic environment.4 The Parks Service soon thereafter promulgated regulations governing the entry and activity of cruise ships and other ves- sels in Glacier Bay. The regulations provided that only two cruise ships could enter the bay each day, with a maximum of 89 cruise ship entries between June 1 and August 31. Smaller boats, designated "private/pleasure craft," were limited to twenty-one entries per day with a seasonal maximum of 538 _________________________________________________________________ 4 The acoustic environment appears to be very important to humpback whales. As many a schoolchild is aware, humpbacks produce a variety of sounds, including moans, grunts, screams, and long complex "songs." According to the Parks Service, "[p]ostulated functions for whale vocal- izations include maintenance of distance among individuals, species and individual recognition, maintenance of social organization, localization of underwater topography, and contextual information about feeding, court- ship, or alarm." 2422 entries. Vessels were prohibited from intentionally position- ing themselves within a quarter of a nautical mile of a whale or attempting to pursue a whale. Within "designated whale waters," vessels had to operate at a constant speed of ten knots or less and follow a mid-channel course. In 1983 the Fisheries Service issued a second biological opinion, which concluded in part: [I]f the existing restrictions on the operation of ves- sels within the Bay were removed, the associated disturbance would be likely to jeopardize the contin- ued existence of the Southeast Alaska humpback whale stock. . . . [A]ny increase in vessel traffic in Glacier Bay probably will add to the level of traffic encountered by humpback whales in southeast Alaska, and thereby add to cumulative impacts to the humpback whale. Nevertheless, the Fisheries Service opinion stated that a slight increase in vessel traffic was tolerable, provided that the num- ber of individual whales entering the bay did not fall below the 1982 level and that appropriate corrective measures were taken. Accordingly, in 1984 the Parks Service promulgated a Vessel Management Plan (VMP) and regulations that pro- vided for a 20% increase, in increments, in the previously authorized vessel entry quotas. This overall increase -- allow- ing for a total of 107 cruise ship entries per season -- was fully realized in 1988. In September 1992 the Parks Service completed an internal draft of a new VMP that proposed to increase the then current level of cruise ship entries in Glacier Bay by an additional 72%. On February 19, 1993, the Fisheries Service issued a third biological opinion expressing its concern "about the decline in humpback whale use of Glacier Bay," and stating that there were "no studies to show that this decline is not due to avoidance of vessel traffic." Although the Fisheries Service 2423 did not oppose the draft VMP, it urged the Parks Service "to take a conservative approach in all management actions that may affect humpback whales" and to implement particular research and monitoring programs.5 The Fisheries Service did not find that the Parks Service's proposed action was "likely to jeopardize the continued existence and recovery " of Steller sea lions. As mandated by NEPA, the Parks Service investigated whether a substantial cruise-ship increase would significantly affect the environment in Glacier Bay. See 42 U.S.C. S 4332(2)(C); 40 C.F.R. S 1508.27. In May 1995 the Parks Service issued a combined proposed VMP and environmental assessment (EA). An EA is a document that, under NEPA, (1) provides "sufficient evidence and analysis for determining whether to prepare an environmental impact statement or a finding of no significant impact;" (2) aids an agency's com- pliance with NEPA when no EIS is necessary; and (3) facili- tates preparation of an EIS when one is necessary. 40 C.F.R. S 1508.9(a). An EA is a "less formal and less rigorous" docu- ment than an EIS. Conner v. Burford, 848 F.2d 1441, 1446 (9th Cir. 1998). The combined VMP/EA reported the existence of environ- mental questions that went far beyond the potential impact on the humpback whales. It also described and assessed six alter- _________________________________________________________________ 5 "Because there has not been systematic monitoring of humpback whale prey density, distribution and type, and noise produced by vessels, it is impossible to ascribe whale distribution shifts to one cause or another. However, it is [the Fisheries Service's] opinion that for the next three sea- sons (1993, 1994, and 1995), the levels of vessel use combined with vessel operation requirements as described in the September 25, 1992, Vessel Management Plan and environmental assessment are not likely to jeopar- dize the continued existence of the North Pacific population of humpback whales." This opinion did not consider the localized effects of the pro- posed action on the portion of the North Pacific whale population that actually uses Glacier Bay, nor did it address possible effects that were sig- nificant but less than likely to jeopardize the population's continued exis- tence. 2424 native approaches for managing vessels in Glacier Bay, rang- ing from Alternative Four's reduction in vessel traffic by between 14% and 22%, to Alternative One's maintenance of the status quo, to Alternative Five's increase of cruise ship entries by 72%. Notwithstanding the environmental problems it recognized, the Parks Service expressed its preference for Alternative Five. This alternative maintained the limit of two cruise ship entries per day, but increased the total number of seasonal entries from 107 to 184.6 It did not increase seasonal entries for other vessels. The Parks Service conducted six public hearings on the VMP. The Parks Service received approximately 450 com- ments, approximately 85% of which opposed Alternative Five and favored Alternative Four. The Sierra Club, the Alaska Wildlife Alliance, and the plaintiff NPCA spoke out against Alternative Five at the hearings, and submitted expert opinion and evidence in opposition to the Parks Service's findings. On March 20, 1996, the Parks Service announced its decision to implement a modified version of Alternative Five as its new VMP. Under this modified plan, the seasonal entry quota for cruise ships would increase by 30% for 1996 and 1997, and by as much as 72% thereafter if certain conditions were met. Also, the entry quotas for charter boats and private/pleasure craft would increase by 8% and 15%, respectively. An accom- panying revised EA, titled "Impacts of the Modified Alterna- tive," discussed the effects of the new VMP on threatened and endangered marine mammals, other marine mammals, birds, and the human environment, including air quality. The revised EA included the following observations: - Steller sea lions using open water "would be sub- _________________________________________________________________ 6 The EA identifies the "season " as lasting from June 1 through August 31. The proposed increase permits vessel entries every day of the season. Accordingly, under Alternative Five, there will be thirty-eight more days of vessel traffic each season. 2425 ject to increased vessel traffic and its related distur- bance. Little is known about the effects of the disturbance." - The increased vessel traffic would expose the har- bor seal, harbor porpoise, Dall's porpoise, humpback whale, killer whale, and minke whale to "increased levels of disturbance," causing the animals to expend energy reserves and possibly compromising "the sur- vival and reproduction of individual animals." In addition, "the potential for daily and seasonal expo- sure of humpback whales to underwater noise would increase." "The effect of increased levels of distur- bance" on these cetacean populations, it concluded, was "unknown." - Marine mammals "using open-water habitats would be subject to increased vessel traffic and its related disturbance. However, little is known about the effects of the disturbance." The risk of vessels colliding with marine mammals would increase, although "the degree of increase is unknown." Simi- larly, there would be an increased risk of ship colli- sions, other accidents, and associated fuel spills. "The rate of actual spills could increase, but the degree of increase in unknown." - "The degree to which disturbance and displace- ment would affect the humpback whale populations in Glacier Bay is unknown. Several mitigation mea- sures implemented under this alternative would reduce the risk of whale/vessel interactions and the level of potential effects on individual whales. The implementation of oil-spill response plans by the cruise ship industry could reduce oil spill risks to individual whales." 2426 - It was "unknown" whether populations of Marbled Murrelets and Kittlitz Murrelets would change under the VMP. - "The overall effect on bald eagle populations is unknown." - "It is unknown if waterfowl populations would change under this alternative." Finally, the revised EA acknowledged that the increase in cruise ship entries would "result in more violations of state air quality standards," but stated that the "biological effects of these air pollutants from stack emissions are unknown." At the same time it released its revised VMP and EA, 7 the Parks Service also released a proposed Finding of No Signifi- cant Impact (FONSI). As its title suggests, a FONSI states the reasons why an agency's proposed action will not have a sig- nificant effect on the environment and, therefore, it believes that the preparation of an EIS is unnecessary under NEPA. See 40 C.F.R. S 1508.13. The Parks Service's Glacier Bay FONSI stated, in relevant part: The [Parks Service] has determined that the modi- fied alternative [Five] . . . can be implemented with no significant adverse effect to natural and cultural resources as documented by the environmental assessment. Key environmental issues associated with the modified alternative include effects on marine mammals and birds from vessel disturbance and air quality degradation from cruise ship stack emissions. Some disturbance to these resources would be expected. However, the mitigation strate- gies included in this action would significantly _________________________________________________________________ 7 The revised VMP and EA will hereinafter be referred to simply as the VMP and EA. 2427 reduce environmental effects resulting from vessel entries. The NPCA submitted objections to the VMP/EA and the pro- posed FONSI on April 19, 1996. The Parks Service adopted the VMP/EA and issued its FONSI and final regulations on the plan on May 30. See 61 Fed. Reg. 27,008, codified at 16 C.F.R. 13.65(b). On May 2, 1997, the NPCA filed suit against Secretary Babbitt and Dennis J. Galvin, Acting Director of the Parks Service. The NPCA requested declaratory and injunctive relief requiring the Parks Service "to rescind the new VMP and prohibiting any activities to be conducted pursuant to these rules until such time as [the Parks Service has] complied with NEPA by preparing an adequate EIS for use in evaluat- ing the new VMP before reimplementing it." Holland Amer- ica Line Westours, Inc. ("Westours"), one of the two major cruise ship operators in Glacier Bay, intervened as a defen- dant pursuant to Federal Rule of Civil Procedure 24(a). The NPCA, the Parks Service, and Westours each filed motions for summary judgment. On August 24, 1999, the dis- trict court issued an order denying NPCA's motion, granting the Parks Service's motion, and denying Westours's motion as moot. The court observed that the EA made it "fairly clear that interactions between whales and vessels might be seri- ously disruptive to wildlife residing in the Park. " It also acknowledged "that the effects cruise ship operations have on Glacier Bay National Park and the animals that live there are unknown, either before ship operations increased or since," and that the "Modified Alternative Description[EA] attached to the FONSI contains a long list of uncertainties about the potential effects of increased vessel traffic." While it pro- posed that further studies be conducted, it determined that the existence of the numerous uncertainties was not sufficient to require an EIS. "The EA in this case," the district court wrote, "thoroughly canvasses all existing information and recognizes 2428 that theoretical harms might occur, but concludes that there is no evidence suggesting that a significant risk exists that the harms will occur." It concluded: [A] modest increase in the number of visitors may be allowed while the studies are commissioned and the existing data base increased with care taken through ameliorization [sic] to recognize and eliminate prob- lems as they arise. The district court granted the Parks Service's cross-motion for summary judgment and dismissed the case. The NPCA appealed, and Westours cross-appealed. ANALYSIS I. THE PARKS SERVICE VIOLATED NEPA A. Standard of Review We review a district court's decision to grant or deny a motion for summary judgment de novo. Metcalf v. Daley, 214 F.3d 1135, 1141 (9th Cir. 2000) (citation omitted). In review- ing an agency's decision not to prepare an EIS under NEPA, we employ an arbitrary and capricious standard, Blue Moun- tains Biodiversity Project v. Blackwood, 161 F.3d 1208, 1211 (9th Cir. 1998), cert. denied, 527 U.S. 1003 (1999), that requires us to determine whether the agency has taken a "hard look" at the consequences of its actions, "based [its decision] on a consideration of the relevant factors," id., and provided a "convincing statement of reasons to explain why a project's impacts are insignificant." Metcalf, 214 F.3d at 1142. B. Environmental Assessment [1] NEPA requires that an Environmental Impact Statement (EIS) be prepared for all "major Federal actions significantly affecting the quality of the human environment." 42 U.S.C.A. 2429 S 4332(2)(C). However, if, as here, an agency's regulations do not categorically require the preparation of an EIS, then the agency must first prepare an Environmental Assessment (EA) to determine whether the action will have a significant effect on the environment. See 40 C.F.R.S 1501.4; Salmon River Concerned Citizens v. Robertson, 32 F.3d 1346, 1356 (9th Cir. 1994). If the EA establishes that the agency's action "may have a significant effect upon the .. . environment, an EIS must be prepared." Foundation for N. Am. Wild Sheep v. United States Dep't of Agric., 681 F.2d 1172, 1178 (9th Cir. 1982) (emphasis added); see also Blue Mountains , 161 F.3d at 1212. If not, the agency must issue a Finding of No Signifi- cant Impact (FONSI), see Blue Mountains, 161 F.3d at 1212; 40 C.F.R. SS 1501.4, 1508.9, accompanied by" `a convincing statement of reasons' to explain why a project's impacts are insignificant." Blue Mountains, 161 F.3d at 1212 (quoting Save the Yaak Comm. v. Block, 840 F.2d 714, 717 (9th Cir. 1988). [2] Whether there may be a significant effect on the envi- ronment requires consideration of two broad factors:"context and intensity." See 40 C.F.R. S 1508.27; 42 U.S.C. S 4332(2)(C); see also Sierra Club v. United States Forest Serv., 843 F.2d 1190, 1193 (9th Cir. 1988). 8 Context simply _________________________________________________________________ 8 "Significantly" as used in NEPA requires considerations of both con- text and intensity: (a) Context. This means that the significance of an action must be analyzed in several contexts such as society as a whole (human, national), the affected region, the affected interests, and the locality. Significance varies with the setting of the proposed action. For instance, in the case of a site-specific action, signifi- cance would usually depend upon the effects in the locale rather than in the world as a whole. Both short- and long-term effects are relevant. (b) Intensity. This refers to the severity of impact. Responsible officials must bear in mind that more than one agency may make decisions about partial aspects of a major action. The following should be considered in evaluating intensity: 2430 delimits the scope of the agency's action, including the inter- ests affected. Intensity relates to the degree to which the _________________________________________________________________ (1) Impacts that may be both beneficial and adverse. A signifi- cant 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 prox- imity to historic or cultural resources, park lands, prime farm- lands, 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 individ- ually insignificant but cumulatively significant impacts. Signifi- cance exists if it is reasonable to anticipate a cumulatively significant impact on the environment. Significance cannot be avoided by terming an action temporary or by breaking it down into small component parts. (8) The degree to which the action may adversely affect dis- tricts, sites, highways, structures, or objects listed in or eligible for listing in the National Register of Historic Places or may cause loss or destruction of significant scientific, cultural, or his- torical 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 or requirements imposed for the protection of the environment. 40 C.F.R. S 1508.27. 2431 agency action affects the locale and interests identified in the context part of the inquiry. Here, the context is Glacier Bay National Park, its natural setting, its variegated non-human inhabitants, and its pure but fragile air quality; intensity must be established in this case by using three of the standards enu- merated in S 1508.27: (1) the unique characteristics of the geographic area; (2) the degree to which VMP Alternative Five's possible effects on the human environment are highly uncertain; and (3) the degree of controversy surrounding those possible effects. The unique characteristics of Glacier Bay are undisputed and of overwhelming importance. Accordingly, we next consider the agency's determination in light of the degree of uncertainty manifested, and the degree of contro- versy generated. Either of these factors may be sufficient to require preparation of an EIS in appropriate circumstances. Sierra Club, 843 F.2d at 1193, 1194; Blue Mountains, 161 F.3d at 1212-14. In the end, we conclude that the Parks Ser- vice clearly erred and that the high degree of uncertainty and the substantial controversy regarding the effects on the quality of the environment each necessitates preparation of an EIS. C. Uncertainty [3] An agency must generally prepare an EIS if the envi- ronmental effects of a proposed agency action are highly uncertain. See Blue Mountains, 161 F.3d at 1213 ("significant environmental impact" mandating preparation of an EIS where "effects are `highly uncertain or involve unique or unknown risks' ").9 Preparation of an EIS is mandated where uncertainty may be resolved by further collection of data, see id. at 1213-14 (lack of supporting data and cursory treatment of environmental effects in EA does not support refusal to _________________________________________________________________ 9 Greenpeace Action v. Franklin , 14 F.3d 1324 (9th Cir. 1993), is not to the contrary. There the court simply noted that the cases cited did "not stand for the proposition that the existence of uncertainty mandates the preparation of an impact statement." 14 F.3d at 1334 n.11. Blue Mountains has since filled that gap. 161 F.3d at 1213-14. 2432 produce EIS), or where the collection of such data may pre- vent "speculation on potential . . . effects. The purpose of an EIS is to obviate the need for speculation by insuring that available data are gathered and analyzed prior to the imple- mentation of the proposed action." Sierra Club, 843 F.2d at 1195. [4] Here, scientific evidence presented in the Parks Ser- vice's own studies revealed very definite environmental effects. The uncertainty was over the intensity of those effects. The FONSI reported increased daily and seasonal exposure of humpback whales and other denizens of the Bay to underwater noise (and predicted a range of adverse behav- ioral responses), "traffic effects" (including increased risk of collision, affecting whales, harbor seals, sea otters, murrelets, and molting waterfowl), and increased risk of oil pollution for all animal life in the Park. An increase in cruise ships would also "result in more violations of state air quality standards for cruise ship stack emissions."10 Among the specific effects set forth in the VMP/EA upon which the FONSI was based were that increased vessel entry into the Bay would: subject stellar sea lions to additional disturbance; increase the escape pat- terns of various types of whales; potentially increase mortality rates and change the social patterns of the harbor seal; pre- clude sea otters from colonizing the upper Bay; and increase disturbance of feeding murrelets, other seabird nesting colo- nies, and bald eagles. [5] The EA describes the intensity or practical conse- quences of these effects, individually and collectively, as "un- known." See pages 2425-2427, supra . The uncertainty _________________________________________________________________ 10 The initial EA reported that"[t]he airsheds of the steep-walled fjords of Glacier Bay are susceptible to visible stack emissions becoming trapped by temperature inversions and light winds. . . .[T]he tolerance for air pol- lution in Glacier Bay is extremely low." In a study conducted in 1986 and 1987, 25 of 77 cruise ships's emissions were over twice the permitted duration. 2433 manifested through the EA stems from two sources: an absence of information about the practical effect of increased traffic on the Bay and its inhabitants; and a failure to present adequate proposals to offset environmental damage through mitigation measures. The lack of data regarding the practical effect of increased traffic, like the failure to investigate envi- ronmental impacts in Blue Mountains, 161 F.3d at 1213, undermines "[t]he [Parks Service's] EA . . . [which] is where the [agency's] defense of its position must be found." Id. That document states that "[l]ittle is known about the effects of the [cruise ship] disturbance" on steller sea lions; "[t]he effect of increased levels of disturbance" on Glacier Bay's cetacean populations is "unknown"; and "the degree of increase [in oil spills as a result of increased traffic] is unknown." It also states that the effect of noise and air pollution on murrelets, bald eagles, and waterfowls remains "unknown" because unstudied. Moreover, the extent to which air pollution will diminish the beauty and quality of the natural environment is also unknown.11 The Parks Service's EA does, however, establish both that such information may be obtainable and that it would be of substantial assistance in the evaluation of the environmental impact of the planned vessel increase. The EA proposes a park research and monitoring program to "fill information needs, and understand the effects of vessel traffic on air quality, marine mammals [and] birds . .. to assist in the prediction, assessment, and management of potential effects on the human, marine, and coastal environments of Glacier Bay resulting from human use of the environment with partic- ular emphasis on traffic." That is precisely the information and understanding that is required before a decision that may have a significant adverse impact on the environment is made, and precisely why an EIS must be prepared in this case. _________________________________________________________________ 11 The initial EA acknowledged, however, that the proposed increase in vessel traffic "could significantly impair the visual and photographic scene at the glacier faces in Tarr Inlet, the high point of a Glacier Bay cruise, and elsewhere in the park." 2434 The Parks Service proposes to increase the risk of harm to the environment and then perform its studies. It has in fact already implemented the first part of its VMP. This approach has the process exactly backwards. See Sierra Club, 843 F.2d at 1195. Before one brings about a potentially significant and irreversible change to the environment, an EIS must be pre- pared that sufficiently explores the intensity of the environ- mental effects it acknowledges. A part of the preparation process here could well be to conduct the studies that the Park Service recognizes are needed. That might be done here by performing the studies of the current vessel traffic and extrap- olating or projecting the effects of the proposed increase.12 Ultimately, the Park Service may develop other means for obtaining the information it currently lacks. The point is, however, that the "hard look" must be taken before, not after, the environmentally-threatening actions are put into effect. [6] The Parks Service's lack of knowledge does not excuse the preparation of an EIS; rather it requires the Parks Service to do the necessary work to obtain it. In Blue Mountains, we found that general statements about possible environmental effects failed the "hard look" test required under NEPA. 161 F.3d at 1213. Here, the Parks Service's repeated generic state- ment that the effects are unknown does not constitute the req- _________________________________________________________________ 12 With respect to the cruise ships, the addition of thirty-eight extra days of cruising may simply require the service to determine the current effects of vessel traffic, and extrapolate or project from that data the effects of increased traffic, taking into account all available relevant information, including technological change affecting the cruise ship-industry. It appears that the impact of the other increases in vessel traffic may also be determined by studying the effects of the current traffic rates. We do not decide here, however, how the EIS should be conducted. That is for the Park Service to determine initially in the context of the applicable statutes and regulations. We intend only to observe that in this case, unlike some others, an actual study can probably be conducted on the basis of existing conditions and that it is not necessary to consider the intensity of the effect of vessel traffic entirely in the abstract. There are practical consequences resulting from the current level of traffic that may be studied and consid- ered in the final report. 2435 uisite "hard look" mandated by the statute if preparation of an EIS is to be avoided. See id. (" `general statements about "possible" effects and "some risk" do not constitute a "hard look" absent a justification regarding why more definitive information could not be provided' ") (citing Neighbors of Cuddy Mountain v. United States Forest Serv., 137 F.3d 1372, 1380 (9th Cir. 1998)). The Park Service's statement of rea- sons does not provide a convincing explanation as to why the requisite information could not be obtained prior to placing the VMP into effect. Id. In short, the information currently provided by the Parks Service in its EA leaves us with the firm impression that, absent successful mitigation measures, there is a substantial possibility that the VMP will signifi- cantly affect Glacier Bay Park, including the air, the water, and the various species that inhabit the Park. [7] The second source of uncertainty is the Parks Service's ability to offset the environmental impact of the increase in vessel traffic through its proposed mitigation measures. An agency's decision to forego issuing an EIS may be justified in some circumstances by the adoption of such measures. Wet- lands Action Network v. United States Army Corps of Eng'rs, 222 F.3d 1105, 1121 (9th Cir. 2000); Friends of Payette v. Horseshoe Bend Hydroelectric Co., 988 F.2d 989, 993 (9th Cir. 1993). "If significant measures are taken to`mitigate the project's effects, they need not completely compensate for adverse environmental impacts.' " Wetlands Action Network, 222 F.3d at 1121 (quoting Friends of Payette, 988 F.2d at 993). While the agency is not required to develop a complete mitigation plan detailing the "precise nature of the mitigation measures," the proposed mitigation measures must be "devel- oped to a reasonable degree." Id.13 A " `perfunctory descrip- _________________________________________________________________ 13 In Laguna Greenbelt, Inc. v. United States Dep't of Transp., 42 F.3d 517, 528 n.11 (9th Cir. 1994), the court held that "scientific uncertainties in the mitigation measures" need not be discussed during the EIS discus- sion period. Such uncertainties must be discussed, however, during the EIS preparation period. 40 C.F.R. S 1502.22 2436 tion,' " Okanogan Highlands Alliance v. Williams, 236 F.3d 468, 473 (9th Cir. 2000) (quoting Neighbors of Cuddy Mountain, 137 F.3d at 1380), or " `mere listing' of mitigation measures, without supporting analytical data," is insufficient to support a finding of no significant impact. Id. (quoting Idaho Sporting Congress v. Thomas, 137 F.3d 1146, 1151 (9th Cir. 1998)). In evaluating the sufficiency of mitigation measures, we consider whether they constitute an adequate buffer against the negative impacts that may result from the authorized activity. Specifically, we examine whether the mit- igation measures will render such impacts so minor as to not warrant an EIS. See Greenpeace Action, 14 F.3d at 1332. [8] There is a paucity of analytic data to support the Parks Service's conclusion that the mitigation measures would be adequate in light of the potential environmental harms. By contrast, in Okanogan the Forest Service conducted computer modeling to predict the quality and quantity of environmental effects, discussed the monitoring measures to be put in place, ranked the probable efficacy of the different measures, detailed steps to achieve compliance should the measures fail, and identified the environmental standards by which mitiga- tion success could be measured. Id. at 473-75. Because the Forest Service "considered extensively the potential effects and mitigation processes," the court found that discussion of the mitigation measures was adequate to constitute the con- vincing statement of reasons to permit preparation of a FONSI. Id. at 477 (emphasis omitted). In this case, however, the Parks Service did not conduct a study of the anticipated effects of the mitigation measures nor did it provide criteria for an ongoing examination of them or for taking any needed corrective action (except for the plan to conduct "studies"). As with the rest of its proposal, it planned to act first and study later. The Parks Service first described its proposed mitigation measures in the initial EA. That document reflects the uncer- tainty that exists as to whether the mitigation measures would 2437 work: moreover, it is unclear from that document whether the measures are sufficiently related to the effects they are designed to cure. The Parks Service simply noted, for exam- ple, that mitigation measures "could mitigate some potential effects to humpbacks in concentrated whale-use areas"; "could reduce whale/vessel collisions and reduce the noise emanating from the ships"; "[s]pecial-use-area closures and restrictions implemented under . . . alternative[five] may off- set some of the expected disturbance." Air pollution measures "would be expected to contribute to a reduction in cruise ship stack emissions over time." Further, the service stated that it: intends to institute a comprehensive research and monitoring program to fill informational needs and quantity the effects of vessel traffic on air quality, marine mammals, birds and visitor-use enjoyment. The monitoring program, developed within one year of the record of decision, will stipulate research and protection actions [Parks Service] will undertake to ensure that environmental effects do not exceed acceptable levels. . . The final EA was similarly uncertain with respect to the proposed measures' effects. It recognized that a 10-knot speed restriction to offset the increased vessel traffic might disturb the creatures in the park, but that "very little is known about the effects of the disturbance." The EA also stated that the increase in seasonal entries "could reduce whale/vessel colli- sions and reduce the noise level emanating from the ships . . . , [f]ollow-up research and monitoring will be essential to define humpback whale use patterns in Glacier Bay resulting from this alternative"; and that "requiring cruise ships to implement oil-spill response plans could mitigate the effects of oil spills." As for air pollution, "the magnitude of increased violations would presumably be reduced over time. " There is no indication, however, as to how long any such reduction might take or how great a reduction might ultimately be accomplished. In short, there is no evidence that the mitiga- 2438 tion measures would significantly combat the mostly "un- known" or inadequately known effects of the increase in vessel traffic. The EA's speculative and conclusory state- ments are insufficient to demonstrate that the mitigation mea- sures would render the environmental impact so minor as to not warrant an EIS. See Greenpeace Action, 14 F.3d at 1332. It is instructive to contrast this case with Wetlands Action Network. In that case, the court made clear that, though the mitigation measures were underdeveloped, the imposition of special conditions, enforced through a permit, and reviewed by various other agencies ensured that the measures would be enforced in a manner that properly reduced negative environ- mental impact. See Wetlands Action Network, 222 F.3d at 1121. Here, there were no such special conditions applied to the Parks Service or the tour boat operators in connection with the measures designed to mitigate the effects of the proposed increase in cruise ships. In fact, whether the mitigation mea- sures are fully enforced or not, the EA reflects significant uncertainty as to whether they could provide an adequate buffer against the harmful effects of the plan. This is under- standable in light of the fact that the extent of those harmful effects is itself unknown, and that in such circumstance it is particularly difficult to estimate the effects of mitigation mea- sures. That fact, however, does not help make the measures adequate for purposes of avoiding the preparation of an EIS -- quite the contrary. As with the question of the extent of the unremediated injury that might otherwise occur, the question of the impact of the proposed mitigation measures must be studied as part of the preparation of an EIS rather than after the injury has transpired. The fact that the agency plans to test the effect of its mitigation measures does not relieve it of the obligation to prepare an EIS prior to the time of the threatened environmen- tal damage. Rather, the Parks Service's testing proposal shows that the information necessary to determine the impact of any mitigation measures, like the information relating to 2439 the extent of the injurious effects, may well be obtainable before any environmental injury occurs. The proposed mitiga- tion studies thus argue in favor of preparing an EIS, not against it. The district court found that the agency's decision not to prepare an EIS was justified because the uncertainty reflected the existing state of knowledge. The passage of cruise ships through Glacier Bay is not a new development, however. Both the vessels' sailings and the dispute over the potential environmental damage have existed for a substantial period of time. No new scientific developments are required in order to obtain the requisite information. The Parks Service itself pro- poses to conduct studies which it anticipates may provide the answers. We simply hold that, under these circumstances, where significant environmental damage may occur to a trea- sured natural resource, the studies must be conducted first, not afterwards. D. Controversy The district court also found that NPCA had not made a sufficient showing of controversy to require preparation of an EIS. Again, we disagree. [9] Agencies must prepare environmental impact state- ments whenever a federal action is "controversial," that is, when "substantial questions are raised as to whether a project . . . may cause significant degradation of some human envi- ronmental factor," Northwest Envtl. Def. Ctr. v. Bonneville Power Admin., 117 F.3d 1520, 1539 (9th Cir. 1997) (quoting LaFlamme v. FERC, 852 F.2d 389, 397 (9th Cir. 1988)) (Reinhardt, J., concurring in part and dissenting in part), or there is "a substantial dispute [about] the size, nature, or effect of the major Federal action." Blue Mountains , 161 F.3d at 1212 (citing Greenpeace Action, 14 F.3d at 1335; Sierra Club, 843 F.2d at 1190). A substantial dispute exists when evidence, raised prior to the preparation of an EIS or FONSI, 2440 see Greenpeace Action, 14 F.3d at 1334 (holding that party may not establish controversy post hoc, when at the time of the agency's action no controversy existed), casts serious doubt upon the reasonableness of an agency's conclusions.14 See Idaho Sporting Congress, 137 F.3d at 1150; Blue Mountains, 161 F.3d at 1212. NEPA then places the burden on the agency to come forward with a "well-reasoned expla- nation" demonstrating why those responses disputing the EA's conclusions "do not suffice to create a public contro- versy based on potential environmental consequences." LaFlamme, 852 F.2d at 401 (citing Jones v. Gordon, 792 F.2d 821, 829 (1986)).15 The term "well reasoned explanation" is simply a less direct way of saying that the explanation must be "convincing." See Metcalf, 214 F.3d at 1142. [10] After publication of the initial EA, and before publica- tion of the EA and FONSI, the Parks Service received 450 comments on the VMP, approximately 85% of which opposed Alternative Five and favored Alternative Four. This volume of negative comment is more than sufficient to meet the "out- pouring of public protest" discussed in Greenpeace Action.16 _________________________________________________________________ 14 Although a court should not take sides in a battle of the experts, Greenpeace Action, 14 F.3d at 1333; see also Wetlands Action Network, 222 F.3d at 1120-21, it must decide whether the agency considered con- flicting expert testimony in preparing its FONSI, and whether the agency's methodology indicates that it took a hard look at the proposed action by reasonably and fully informing itself of the appropriate facts. See Idaho Sporting Congress, 137 F.3d at 1150 (precluding the agency from relying on expert opinion in the absence of hard data). 15 Consensus among the parties over the proposed measures is ordinarily sufficient to satisfy the agency's burden. See Greenpeace Action, 14 F.3d 1334 (holding that consensus among objecting parties before agency declined to prepare EIS sufficient to rebut showing of controversy); Bon- neville Power, 117 F.3d at 1536 (holding that where agency cooperated with objecting parties, and alleviated most of those parties concerns, agency need not prepare EIS). No such consensus existed here. 16 That case distinguished other showings of public controversy made in cases such as Wild Sheep, 681 F.2d at 1175 and n.10, by requiring that any "outpouring of public protest" be contemporaneous with the period for commentating on the EA, and holding that a post hoc dispute was insuffi- cient to establish the necessary type of controversy. Greenpeace Action, 14 F.3d at 1334. That is not the case here. 2441 See 14 F.3d at 1334. More important, to the extent the com- ments urged that the EA's analysis was incomplete, and the mitigation uncertain, they cast substantial doubt on the ade- quacy of the Parks Service's methodology and data. The dis- pute between NPCA and the Parks Service thus goes beyond a disagreement of qualified experts over the "reasoned con- clusions" as to what the data reveal. See Greenpeace Action, 14 F.3d at 1335.17 Here, the agency's conclusions were not reached by reasoned extrapolation from the data, rather the data were simply insufficient. The Parks Service acknowl- edged that an increase in vessel traffic would have an environ- mental impact. The data, however, did not establish the intensity of that impact, nor the efficacy of the mitigation measures designed to offset that unquantified impact. NPCA asserted that the effects on the environment would likely be substantial. The Parks Service responded that the extent of the effects was unknown. Therein lay the controversy. In its response, the agency proposed to determine the extent of the effects by implementing the VMP and studying its conse- quences. As we have stated in Section I.C, the absence of cur- rently available information does not excuse the Parks Service from preparing an EIS when there is a reasonable possibility that such information can be obtained in connection with the preparatory process. The agency's response is therefore not sufficient to resolve the controversy. Accordingly, the case before us is controlled by Blue Mountains, Idaho Sporting Congress, and Sierra Club; and Greenpeace Action is not to the contrary. Here, preparation of an EIS is mandated by the "controversy," as well as by the "uncertainty," factor of the intensity provision. 40 C.F.R. S 1508.27(b). II. INJUNCTIVE RELIEF [11] NPCA asks this court to "reverse the District Court _________________________________________________________________ 17 Where there is conflict in the data, or the evidence supports several conflicting opinions, the agency may rely upon the opinion of its expert. Wetlands Action Network, 222 F.3d at 1121. 2442 and remand the case with instructions to enjoin further imple- mentation of the 1996 Vessel Management Plan until a full Environmental Impact Statement is prepared." To determine whether injunctive relief is appropriate, "even in the context of environmental litigation," we apply "the traditional balance of harms analysis." Forest Conservation Council v. United States Forest Serv., 66 F.3d 1489, 1496 (9th Cir. 1995) (cita- tions omitted); see also Amoco Prod. Co. v. Village of Gam- bell, 480 U.S. 531, 541, 542 (1987) (holding that unless Congress directed otherwise, "a court must balance the com- peting claims of injury" in determining whether injunctive relief is appropriate). "Environmental injury, by its nature, can seldom be adequately remedied by money damages and is often permanent or at least of long duration, i.e., irrepara- ble," Amoco Prod. Co., 480 U.S. at 545. When the "proposed project may significantly degrade some human environmental factor," injunctive relief is appropriate. Alaska Wilderness Recreation & Tourism Assoc. v. Morrison, 67 F.3d 723, 732. See also Sierra Club, 843 F.2d at 1195; Save the Yaak, 840 F.2d at 722. [12] NPCA has made the requisite showing for injunctive relief. As we concluded in Section I, an EIS is required. We so held because of the significant adverse impact on the envi- ronment that might result from the implementation of the VMP. See Blue Mountains, 161 F.3d at 1216 (an EIS is required of an agency in order that it explore, more thor- oughly than an EA, the environmental consequences of a pro- posed action whenever "substantial questions are raised as to whether a project may cause significant [environmental] deg- radation"). Where an EIS is required, allowing a potentially environmentally damaging project to proceed prior to its prep- aration runs contrary to the very purpose of the statutory require- ment.18 Here, the Parks Service has already undertaken a 30% _________________________________________________________________ 18 We have recognized, nevertheless, that in "unusual circumstances" an injunction may be withheld, or, more likely, limited in scope. See Forest 2443 increase in cruise-ship traffic preliminary to a seventy-two percent increase. The potential effects of its action extend beyond the endangered marine mammal population to the rest of the wildlife at Glacier Bay, as well as the Park's air quality. Kittiwakes, murrelet, eagles, sea otters, seals, sea lions, por- poises, and killer and minke whales, as well as the better known humpbacks, are affected. Until an EIS is prepared and the effects of increased vessel traffic on the inhabitants and air quality of Glacier Bay are properly examined, there is a suffi- cient possibility of environmental harm that the VMP may not be implemented. Westours argues that the damage to its business should be considered when addressing injunctive relief, that its financial losses outweigh the potential damage to the environment, and that NPCA "is not entitled to an injunction against the 32 sea- sonal cruise ship entries" that the Parks Service authorized for the 2000-2004 seasons. As a general rule, only the federal government may be a defendant in a NEPA action. Wetlands Action Network, 222 F.3d at 1114. An exception may be made in the remedial phase of a case where the contractual rights of the applicant are affected by the proposed remedy. See Forest Conservation Council, 66 F.3d at 1495. Here, Wes- _________________________________________________________________ Conservation Council, 66 F.3d at 1496. Amoco Production Co. is not to the contrary. There, the Supreme Court rejected a presumption of irrepara- ble injury where an agency failed adequately to investigate the conse- quences of its proposed action, see Amoco Prod. Co., 480 U.S. at 544-45: it required courts to undertake the traditional "balance of harms" analysis. Id. at 545. We have fully weighed the competing interests using our tradi- tional equitable jurisdiction, and conclude that injunctive relief is appro- priate. Finally, in Sierra Club v. Marsh, 872 F.2d 497 (1st Cir. 1989), then-Circuit Judge Breyer held that, because NEPA is a purely procedural statute, the requisite harm is the failure to follow the appropriate proce- dures. See id. at 500 (because NEPA can do no more than require the agency to produce and consider a proper EIS, the harm that NEPA intends to prevent is imposed when a decision to which NEPA obligations attach is made without the informed environmental consideration that NEPA requires). Marsh also justifies injunctive relief in this case. 2444 tours was permitted to intervene, and appears before us as a party-defendant. Westours has asserted financial damages premised upon its contracts of carriage. Its loss of anticipated revenues, however, does not outweigh the potential irrepara- ble damage to the environment. Moreover, neither Westours nor those of its passengers who may be unable to view Gla- cier Bay at the time they originally planned have cause to claim surprise as a result of any injunction. The plaintiffs filed their objections to the plan approximately five years ago and just one year later sought an injunction. If the passengers who booked cruises on any "excess" tours were not warned by Westours of the pending litigation, their interests were not well served by that company. Thus, while Westours has standing to object to our grant of injunctive relief, its evidence fails to tilt the balance of harms in its favor. For the purposes of injunctive relief, we may admit evi- dence not before the district court to show that an agency has rectified a NEPA violation after the onset of legal proceed- ings. See Friends of the Clearwater v. Dombeck , 222 F.3d 552, 560 (9th Cir. 2000). Here, Westours asks us to consider evidence addressing one aspect of NPCA's complaint, the humpback whale population. We will assume that such evi- dence is admissible. Westour's evidence shows a short term increase in the humpback whale population, but completely fails to address the other environmental effects at issue here. Nor does it address the deficiencies in the Parks Service's EA and FONSI. Accordingly, Westours's evidence is insufficient to dissuade us from granting injunctive relief. Finally, we note that where the question of injunctive relief "raise[s] intensely factual issues," the scope of the injunction should be determined in the first instance by the district court. See Alaska Wilderness, 67 F.3d at 732. Here, however, there are no such intensely factual issues and the scope of the injunction to which NPCA is entitled is quite plain. It is appropriate, therefore, for us to decide the injunction question on this appeal. 2445 We direct the district court to enjoin the further increases in vessel traffic, and to return traffic levels to their pre-1996 levels. We agree, however, with the Parks Service's conten- tion that "the current vessel regulations are in important respects more environmentally protective than the 1984 regu- lations they replaced." The NPCA does not contend (and the record would not support a finding) that the 1996 VMP's establishment of protected "whale waters" in parts of Glacier Bay or implementation of oil-spill response plans pose an actual or potential threat to the environment. Accordingly, there is no basis for enjoining that part of Alternative Five. Our order is limited to the thirty to seventy-two percent increase in the seasonal entry quota for cruise ships, and the eight percent increase for charter boats and fifteen percent increase for "private/pleasure" craft. All entry quotas shall be returned to the levels preceding the introduction of Alterna- tive Five. The other measures adopted by the Parks Service in its VMP shall remain in effect, and will not be subject to the injunction. Notwithstanding the above, we recognize that the issuance of the mandate in this case may be delayed for some period of time as a result of the en banc procedures that this court follows, and that such delay may occur even if we ultimately decide not to grant a rehearing en banc. We do not believe, however, that a sufficient emergency exists to warrant our ordering that the injunction become effective prior to the completion of our customary process. Accordingly, we cannot know in advance how close to the opening of this year's cruising season the mandate will issue, or even whether issu- ance will be delayed until after the season has already begun. Nor can we determine how great may be the practical disrup- tion that would occur were last minute cancellation of a sig- nificant number of voyages required. Accordingly, we leave it to the district court to decide upon the effective date of the injunction it is ordered to issue, and, specifically, to decide in its informed discretion, and on the basis of any evidence that may be presented, whether the injunction should take effect 2446 prior to the completion of this year's cruising season, some- time this September. CONCLUSION Much of the briefing and argument in this appeal has focused on the impact of the VMP on the imperilled hump- back whale population. However, a variety of other non- human inhabitants of the Park -- bald eagles, kittiwakes, murrelets, sea otters, harbor seals, Steller sea lions, harbor and Dall's porpoises, minke, and killer whales -- are, as the EA reflects, affected, and the already fragile air quality is as well. The existence of adverse effects is not uncertain. What is uncertain is the extent of the likely environmental injury, and the impact of the proposed mitigation measures. The Parks Service's own experts, whose integrity the government com- mended at oral argument, admitted both the likelihood of cer- tain harms to the environment of Glacier Bay and their uncertainty about the likelihood of other harms. In giving insufficient respect to their experts' evaluation of harm, declaring that no significant environmental effects were likely, and implementing the vessel traffic increase without complying with the requirements of NEPA, the Parks Ser- vice's decision-makers made a "clear error of judgment." Marsh v. Or. Natural Res. Council, 490 U.S. 360, 378 (1989). Glacier Bay Park is too precious an ecosystem for the Parks Service to ignore significant risks to its diverse inhabitants and its fragile atmosphere. We reverse the decision below and remand with instructions that the district court issue an injunction enjoining the granting of permits to vessels pursu- ant to the 1996 increase in vessel entry quotas pending the Parks Service's completion of an EIS. Permits shall be limited in number to those authorized prior to the issuance of the EA, the VMP, and the FONSI. The district court shall provide in its order for whatever specific actions it deems necessary to ensure that the number of cruise ships and other vessels authorized to enter Glacier Bay (pending completion of an 2447 EIS) shall not exceed the number authorized prior to the 1996 increase. It shall have the discretion, however, to determine the effective date of its injunction, including whether the order shall be made effective prior to the completion of this year's cruising season. REVERSED AND REMANDED WITH INSTRUCTIONS. 2448 [leftbottomcorner.gif] [topofpage.gif]