JOHN O. MARSH, JR., SECRETARY OF THE ARMY, ET AL., PETITIONERS V. OREGON NATURAL RESOURCES COUNCIL, ET AL. No. 87-1704 In the Supreme Court of the United States October Term, 1988 On Writ of Certiorari to the United States Court of Appeals for the Ninth Circuit Brief for the Petitioners PARTIES TO THE PROCEEDING The petitioners are John O. Marsh, Jr., in his official capacity as Secretary of the Army, and Elvin R. Heiberg, III, in his official capacity as Chief of Engineers of the United States Army. The respondents are Oregon Natural Resources Council, Oregon Guides and Packers Association, Inc., Rogue Fly-fishers, Inc., and Rogue River Guides Association. TABLE OF CONTENTS Questions presented Parties to the proceedings Opinions below Jurisdiction Statutes and regulations involved Statement Summary of argument Argument: I. NEPA does not require federal agencies to include in their environmental impact statements a complete plan to mitigate environmental damage or a "worst case" analysis of potential environmental harm A. NEPA's requirement that federal agencies prepare a detailed statement on the environmental consequences of major federal actions does not impose a substantive obligation to mitigate adverse environmental effects or a procedural obligation to develop a "complete mitigation plan" B. NEPA does not require federal agencies to prepare a "worst case" analysis of potential environmental harm if relevant information concerning significant environmental effects is unavailable or too costly to obtain II. The Army Corps of Engineers gave adequate consideration to information concerning environmental effects that was presented after the completion of the Corps' supplemental environmental impact statement Conclusion OPINIONS BELOW The amended opinion of the court of appeals (Pet. App. 1a-29a) is reported at 832 F.2d 1489. The original opinion of the court of appeals is reported at 820 F.2d 1051. The opinion of the district court (Pet. App. 30a-51a) is reported at 628 F. Supp. 1557. JURISDICTION The court of appeals entered its judgment (Pet. App. 68a) on November 24, 1987. The court of appeals denied the government's petition for rehearing on December 10, 1987 (Pet. App. 69a). On February 18, 1988, Justice O'Connor extended the time for filing a petition for a writ of certiorari to and including March 30, 1988. On March 23, 1988, Justice O'Connor further extended the time for filing the petition to and including April 14, 1988, and the petition was filed on that date. This Court granted the petition for a writ of certiorari on June 27, 1988, consolidating this case with Robertson v. Methow Valley Citizens Council, No. 87-1703. /1/ The jurisdiction of this Court rests on 28 U.S.C. 1254(1). STATUTES AND REGULATIONS INVOLVED Section 102 of the National Environmental Policy Act of 1969, 42 U.S.C. 4332, provides, in relevant part: The Congress authorizes and directs that, to the fullest extent possible: * * * (2) all agencies of the Federal Government shall -- * * * * * (C) include in every recommendation or report on proposals for legislation and other major Federal actions significantly affecting the quality of the human environment, a detailed statement by the responsible official on -- (i) the environmental impact of the proposed action, (ii) any adverse environmental effects which cannot be avoided should the proposal be implemented, (iii) alternatives to the proposed action, (iv) the relationship between local short-term uses of man's environment and the maintenance and enhancement of long-term productivity, and (v) any irreversible and irretrievable commitments of resources which would be involved in the proposed action should it be implemented. Pertinent excerpts from the federal regulations involved in this case are reproduced in the Addenda. QUESTIONS PRESENTED 1. Whether the National Environmental Policy Act requires federal agencies to include in each environmental impact statement: (a) a fully developed plan to mitigate environmental harm; and (b) a "worst case" analysis of potential environmental harm if relevant information is unavailable or too costly to obtain. 2. Whether the Army Corps of Engineers gave adequate consideration to new information concerning environmental effects that was presented after the completion of the environmental impact statement. STATEMENT Petitioners seek reversal of a court of appeals decision overturning a district court's determination that the Army Corps of Engineers had fully complied with the National Environmental Policy Act of 1969 (NEPA), 42 U.S.C. 4321 et seq., in authorizing the completion of the Elk Creek Dam. 1. In 1962, after thirty years of consideration and planning, Congress authorized the Rogue River Basin Project, a three-part flood control project on the Rogue River. Flood Control Act of 1962, Pub. L. No. 87-874, Section 203, 76 Stat. 1192-1193. The Project, which includes three multiple purpose dams to be built and administered by the Army Corps of Engineers, was designed primarily for flood control, but it was also envisioned to serve other purposes, including the enhancement of fish and wildlife, municipal and industrial water supplies, irrigation and recreation. See Elk Creek Lake Environmental Impact Statement Supp. No. 1 (Dec. 1980) (hereinafter Elk Creek Supplemental Statement) (Summary). /2/ Two of the Project's three dams -- the Lost Creek Dam on the Rogue River and the Applegate Dam on the Applegate River -- have been completed and are in operation. This case concerns the third structure -- the Elk Creek Dam -- located 1.7 miles above Elk Creek's confluence with the Rogue River. The dam is to be a 238-foot-high roller compacted concrete structure. When completed, the dam will control runoff from 132 square miles of the 135-square-mile Elk Creek watershed. At full pool, the lake created by the dam will be 6.2 miles long, have 18 miles of shoreline, and cover 1,290 acres of land. It will hold 101,000 acre feet of water, a particularly important feature for downstream water users during the late summer and early fall low river flow periods. See Elk Creek Supplemental Statement 3. The Elk Creek Dam is designed to be operated in conjunction with the Lost Creek Dam. Water releases will be controlled from a "multiport" structure, which will discharge waters from five reservoir levels ranging from 1,525 to 1,705 feet above sea level. This multiport design will enable reservoir managers to control the temperature, turbidity, and volume of released water, thereby enhancing water quality during periods of low water flow and minimizing adverse effects downstream. Elk Creek Supplemental Statement 7-10. 2. In 1971, the Corps completed and filed with the Council on Environmental Quality (CEQ) the Final Environmental Impact Statement for the Elk Creek portion of the Rogue River Basin Project. Pre-construction activities began in 1972 with the relocation of a county road along the southwestern portion of the Elk Creek Valley and the acquisition of more than 2,600 acres of land for project management purposes and recreation. Project activity stopped, however, when a controversy arose over the dam's effect on the downstream water quality in the Rogue River. The Corps suspended the project pending resolution of the continuing dispute. See Elk Creek Supplemental Statement Preface 1, 3. In December 1980, the Corps released the Elk Creek Supplemental Statement, a supplemental environmental impact statement prepared pursuant to Section 102 of NEPA, 42 U.S.C. 4332. The Elk Creek Supplemental Statement specifically discussed concerns that Elk Creek Dam could adversely affect water quality and aquatic ecosystems and, in particular, could increase the turbidity of the Rogue River. Turbidity received special attention because high turbidity levels impair flyfishing, a major source of recreation on the Rogue, which has been designated, pursuant to the Wild and Scenic Rivers Act of 1968, 16 U.S.C. (& Supp. IV) 1271 et seq., as a part of the national wild and scenic river system (16 U.S.C. 1274(a)(5)). The Statement also discussed changes in river temperature, which can affect aquatic species. See Elk Creek Supplemental Statement 33-34, 35-37. With respect to turbidity, the Corps summarized its conclusions as follows (Elk Creek Supplemental Statement (Summary)): The effects of the proposed Elk Creek project on turbidity levels in the Rogue River downstream of Elk Creek (have) been a major item of public concern. Moderate to high levels of turbidity during periods of high runoff (are) characteristic of Elk Creek under existing conditions. The concern has been that the reservoir would accumulate turbid water and prolong the duration of turbid flows downstream of the dam. Studies comparing actual turbidity levels and durations in Elk Creek with anticipated conditions with the project in operation indicate that although the duration of turbid flow downstream would be increased, the maximum level of turbidity in the stream would be reduced. The Corps based that conclusion, in part, on the results of a water sampling program, initiated in 1972, to determine the natural levels of turbidity in the project area (Elk Creek Supplemental Statement 21-22). The results indicated that Elk Creek, in its natural state, experiences episodes of moderate to high turbidity during periods of high runoff (id. at 22). The Elk Creek Dam would be expected to modulate this condition by prolonging the duration of downstream turbid flow at reduced levels of turbidity. The Corps also relied on the results of its 1974 report, entitled Rogue River Basin, Water Temperature and Turbidity (A.R. Doc. No. 4) (hereinafter 1974 Water Quality Report), which analyzed data on the Rogue River from several years, representing low, high, average, and worst-known flow conditions. See Elk Creek Supplemental Statement 33. Using computer simulations, the report projected temperature, flow, and turbidity in the Rogue River Basin for the Lost Creek and Elk Creek dams operating in combination during low, high, average, and worst flow conditions (ibid.). The report also analyzed the projects under several different operational conditions -- e.g., for flood control only, for irrigation uses, and for other purposes -- and compared the results to the system without the dams in place. It concluded that during the summer the entire project would increase turbidity "'by 2 to 3 JTU in low and average runoff years and by up to 6 to 9 JTU during high runoff years'" (ibid. (quoting the 1974 Water Quality Report)). /3/ By itself, the Elk Creek Dam was projected to contribute from "'1 to 3 JTU's'" (ibid.). The report further concluded that the Elk Creek Dam would increase temperature from "'one to two degrees F(ahrenheit)'" (ibid.). The Corps also relied upon a further study, prepared in 1979, entitled the Elk Creek Lake Water Quality Update Study (contained in A.R. Doc. No. 6, App. C) (hereinafter 1979 Water Quality Report). See Elk Creek Supplemental Statement 33-34. That report reevaluated the turbidity potential of the Elk Creek watershed and compared the 1974 Water Quality Report's computer-based projections for turbidity at Lost Creek with actual turbidity measurements taken after completion of that dam in 1977 (Elk Creek Supplemental Statement 33-34). The results "increased technical confidence in the mathematical model predictions of the (1979 Water Quality Study) and reinforced the conclusions of the 1974 Water Temperature and Turbidity Report" (Elk Creek Supplemental Statement 34). Based on the foregoing studies, the Elk Creek Supplemental Statement concluded that the increase in turbidity caused by the Elk Creek Dam would have only minor effects on fish production but that the combined effect of Lost Creek and Elk Creek dams on the turbidity of the Rogue River would impair angling to some degree (Elk Creek Supplemental Statement 36-37). The Statement also discussed measures to be taken to mitigate the dam's effects on aquatic life (id. at 35-37). The most significant effect on fish -- the loss of spawning grounds for coho salmon and steelhead trout -- had already been mitigated through the construction of the Cole M. Rivers Fish Hatchery, with the capacity to release nearly 700,000 juvenile salmon and trout each year (id. at 35-36). In addition to the fish hatchery, sediment settling within the reservoir and the dam's multiport intake design would minimize turbidity effects on fish production (id. at 37). The Statement concluded that "(a)verage annual downstream turbidity will be the same with or without the project" (ibid.). The Elk Creek Supplemental Statement also discussed the project's effects on vegetation and wildlife and set forth an inventory of the major species affected by the project, including endangered and threatened species (Elk Creek Supplemental Statement 23-27, 38, App. B). No endangered species were found to exist in the project area, but up to 100 black tail deer were believed to reside there and 17 elk had been sighted at the location (id. at 26). Construction of the Elk Creek reservoir, which would inundate 1,290 acres of wildlife habitat, would adversely affect the deer and elk, as well as game bird populations (id. at 38). The Elk Creek Statement suggested that mitigation could be accomplished by managing selected lands to improve the quality of habitat, thereby augmenting their wildlife carrying capacity (Elk Creek Supplemental Statement 6, 38). The Statement specifically endorsed the use of habitat manipulative techniques, including development of palatable browse plants for deer and vegetation interfaces -- or "edge" -- for quail, variation of foliage height, and placement of snags in the reservoir (ibid.). It explained that the specific mitigation measures would be developed based upon the results of a wildlife compensation plan currently underway at Applegate Dam and upon the further recommendations of federal and state agencies (id. at 6). The Statement fully recognized, however, that even with mitigation, some species would be adversely affected by the project (id. at 38, 46). 3. The Corps' Division Engineer reviewed the Elk Creek Supplemental Statement and, on February 19, 1982, concluded that the project should proceed, subject to congressional approval of funding (Pet. App. 52a-58a). He noted that the project would have environmental impacts, but "(a)ll practical means have been incorporated into the project plans to avoid or minimize environmental harm and to mitigate for loss of fish and wildlife habitat" (id. at 57a-58a). Furthermore, "(t)he benefits to be realized through construction of Elk Creek Lake outweigh the economic and environmental costs, and completion of this project represents the course of action, which, on balance, serves the overall public interest" (id. at 58a). The Corps proceeded to conduct a water quality study pursuant to Section 404 of the Clean Water Act (33 U.S.C. (& Supp. IV) 1344), which was completed in 1983 (see Pet. App. 33a). It also redesigned certain portions of the dam (ibid.). The Corps, which has promulgated regulations setting forth procedures for agency compliance with NEPA (see 33 C.F.R. 230.1 et seq.), prepared an environmental assessment, in accordance with 33 C.F.R. 230.9, to determine whether another supplemental environmental impact statement was necessary to discuss those changes. The Corps determined that they resulted in no significant impacts. See 33 C.F.R. 230.10. It accordingly prepared a supplemental information report pursuant to 33 C.F.R. 230.11(d), which provides for issuance of such reports "to provide supplemental information to a point of concern" previously discussed in an environmental impact statement. See Pet. App. 33a. In August 1985, Congress appropriated the funds necessary to complete the Elk Creek Dam and directed the Corps to finish the project without further delay (Pet. App. 33a-34a & n.5). See Act of Aug. 15, 1985, Pub. L. No. 99-88, Tit. I, 99 Stat. 314; see also S. Rep. 99-82, 99th Cong., 1st Sess. 97 (1985). The dam is now approximately one-third built, with intake and outlet structures complete. More than six miles of road and utilities have been relocated and the river has now been rechannelled through the dam. 4. In October 1985, respondents Oregon Natural Resources Council, et al., brought this action in district court to enjoin construction of the Elk Creek Dam. The district court held an evidentiary hearing, rejected each of respondents' six NEPA-related claims, and entered judgment against respondents (Pet. App. 30a-51a). First, the court concluded that the Elk Creek Supplemental Statement properly considered the Elk Creek Dam's cumulative environmental impact by analyzing its effects in terms of the entire Rogue River Basin Project (id. at 37a-38a). The court next held that the Statement adequately described the affected area (id. at 38a-39a) and the environmental effects of the proposed project, including the opportunities to mitigate environmental harm (id. at 40a-42a). The court rejected respondents' argument that the Corps'discount factor failed to provide a reasoned basis for choosing among alternatives (id. at 42a-45a). The court also rejected respondents' argument that the Corps' scientific methodology was unreliable and that a "worst case" analysis was therefore required (id. at 45a-46a). Finally the court held that the Corps was not required to prepare another supplemental environmental impact statement to address new information, which was not available to the Corps at the time it prepared the Elk Creek Supplemental Statement, concerning the project's potential environmental effects (Pet. App. 46a-49a). Respondents cited as new information a report prepared by the Oregon Department of Fish and Wildlife (ODFW) entitled Lost Creek Dam Fisheries Evaluation Phase I Completion Report (A.R. Doc. No. 92) (hereinafter ODFW Report) which described the effects of constructing the Lost Creek Dam on fisheries resources. /4/ Respondents also cited as new information a Soil Conservation Service survey of Elk Creek soil characteristics (hereinafter SCS Report). /5/ The district court agreed with the Corps that neither the ODFW Report nor the SCS Report contained significant new information requiring supplementation of the existing Elk Creek studies. 5. Respondents appealed, and a divided court of appeals affirmed in part and reversed in part (Pet. App. 1a-2a). The court first held that the Elk Creek Supplemental Statement failed to satisfy NEPA because it lacked a detailed plan to mitigate harm to wildlife (id. 4a-7a). The court stated (id. at 7a): "The importance of a mitigation plan cannot be overestimated. It is a determinative factor in evaluating the adequacy of an environmental impact statement." The court next held that the Corps must prepare a new supplemental environmental impact statement to discuss the information contained in the ODFW and SCS reports (id. at 7a-13a). The court rejected the Corps' scientific evaluation of the new information. It concluded, contrary to the Corps and the district court, that the new information is significant (id. at 11a). The court further concluded that "the Corps' assessment of turbidity due to Elk Creek Dam is subject to uncertainty" (Pet. App. 14a) and that the Corps must therefore either conduct additional research or prepare a "worst case" assessment of environmental harm (id. at 13a-15a). Finally, the court held that the Corps failed to take "a hard look at the cumulative environmental impacts" (id. at 16a) and must therefore supplement that portion of the Corps study (id. at 17a). The court affirmed the remainder of the district court's determinations (id. at 17a-20a) and remanded the case for "entry of appropriate injunctive relief" (id. at 21a). Judge Wallace dissented in part (Pet. App. 21a-29a). He principally concluded, after a careful canvassing of the scientific evidence, that the Corps acted reasonably in declining to prepare a new supplemental environmental impact statement in response to the ODFW and SCS Reports (id. at 21a-28a). SUMMARY OF ARGUMENT The Army Corps of Engineers has fully complied with NEPA in authorizing the completion of the Elk Creek Dam. The court of appeals erred in holding that the Corps' Elk Creek Supplemental Statement was deficient for failing to include a "complete mitigation plan" for wildlife impacts and for failing to include a "worst case analysis" of turbidity effects. The court also erred in holding that the Corps is obligated to prepare a second supplemental environmental impact statement to discuss "new information." 1. NEPA directs federal agencies proposing actions that significantly affect the quality of the human environment to prepare a detailed statement on the environmental consequences of the proposed action. NEPA Section 102(2), 42 U.S.C. 4332(2). See Baltimore Gas & Elec. Co. v. Natural Resources Defense Council, Inc., 462 U.S. 87, 97 (1983). NEPA's mandate to the agencies is "essentially procedural" (Vermont Yankee Nuclear Power Corp. v. Natural Resources Defense Council, Inc., 435 U.S. 519, 553 (1978)), and the only role for the courts is to insure that the agency has taken a "hard look" at environmental consequences. Kleppe v. Sierra Club, 427 U.S. 390, 410 n.21 (1976). The court of appeals exceeded its proper role here. The court interpreted NEPA to create new substantive and procedural obligations, found nowhere in the statutory language or in the Council on Environmental Quality's (CEQ's) regulations implementing the Act. a. The court of appeals erred, first, in suggesting that the agency must undertake "significant measures * * * to mitigate the project's effects" (Pet. App. 5a). As we explain in somewhat greater detail in the companion case, Robertson v. Methow Valley Citizens Council, No. 87-1703, NEPA does not dictate what actions the agency should take in response to expected environmental consequences. NEPA requires the agency to consider and inform the public of potential environmental impacts; it does not require the agency to "'elevate environmental concerns over other appropriate considerations.'" Baltimore Gas & Elec. Co., 462 U.S. at 97-98 (citation omitted). The court of appeals' suggestion that NEPA creates a substantive mitigation duty is accordingly incorrect. The court of appeals also misconstrued NEPA's procedural requirements. The court mistakenly held that federal agencies must include in every environmental impact statement a "complete mitigation plan" for dealing with environmental harm. No one disputes that a federal agency can and should consider mitigation opportunities when assessing the environmental consequences of a proposed action. Indeed, the CEQ's implementing regulations direct that mitigation opportunities be considered and discussed during the preparation of an environmental impact statement. But the court's inflexible requirement that an agency must prepare a "complete mitigation plan" is foreign to NEPA and would impede the goal of sound agency decisionmaking. As we explain here and in Robertson, a federal agency has broad discretion to determine the form and scope of the mitigation discussion in light of the particular circumstances, provided only that the environmental review as a whole satisfies NEPA's specific requirements and that the resulting decision is not arbitrary or capricious. See Baltimore Gas & Elec. Co., 462 U.S. at 97-98, 100, 105-106. In this case, the Elk Creek Supplemental Statement's discussion of mitigation measures was sufficiently thorough for the decision at hand, namely, whether to proceed with completion of the Elk Creek Dam. The Statement identified mitigation measures but left the details to future development. The impacts on wildlife were limited, and the discussion of mitigation measures could be correspondingly limited as well. The agency, overall, took a hard look at environmental consequences and measures to ameliorate those consequences. That is all that NEPA requires. Baltimore Gas & Elec. Co., 462 U.S. at 97. b. The court of appeals also erred in requiring the Corps to prepare a "worst case analysis" with respect to "scientific uncertainty" concerning the Elk Creek Dam's effect on downstream turbidity. NEPA itself prescribes no particular method for dealing with uncertainty; the CEQ introduced worst case analysis as its prescribed method for addressing that problem. The CEQ's experience with worst case analysis revealed, however, that the method did not fulfill NEPA's objectives, and the agency therefore has replaced it with a more manageable methodology that identifies "reasonably foreseeable" environmental impacts. See 40 C.F.R. 1502.22. The court of appeals' conclusion that "NEPA case law" (Pet. App. 14a n.8) requires that agencies continue to employ worst case analysis is based upon a misreading of those cases, which merely interpreted the worst case regulation. The court of appeals' requirement that the Corps conduct a worst case analysis is particularly unreasonable here. The court based its finding of scientific "uncertainty" on a single state agency comment that the Corps addressed in the Elk Creek Supplemental Statement. That comment did not give rise to "incomplete information" that is "essential to a reasoned choice among alternatives" (40 C.F.R. 1502.22). Thus, neither the old nor the new CEQ regulation is applicable here. 2. The court of appeals also erred in requiring the Corps to prepare a second supplemental environmental impact statement to address the ODFW Report's information concerning fish mortality and the SCS Report's information concerning soil characteristics. An agency's decision whether to prepare or supplement an environmental impact statement is subject to judicial review under the Administrative Procedure Act, which states that the agency's decision must be upheld unless "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law" (5 U.S.C. 706). In this instance, the Corps considered the information and determined that it was either unreliable or lacked sufficient bearing on the proposed action to warrant preparation of another supplemental statement. The Corps' judgment on this matter was supported by sound scientific reasoning, was clearly reasonable, and was entitled to deference. The court of appeals' action was particularly inappropriate here, where the district court had held a hearing to consider additional scientific evidence and affirmed the agency's conclusion. ARGUMENT I. NEPA DOES NOT REQUIRE FEDERAL AGENCIES TO INCLUDE IN THEIR ENVIRONMENTAL IMPACT STATEMENTS A COMPLETE PLAN TO MITIGATE ENVIRONMENTAL DAMAGE OR A "WORST CASE" ANALYSIS OF POTENTIAL ENVIRONMENTAL HARM The National Environmental Policy Act of 1969 directs federal agencies proposing actions that significantly affect the quality of the human environment to prepare a detailed statement on the environmental consequences of the proposed action. NEPA Section 102(2), 42 U.S.C. 4332(2). This procedural requirement serves two purposes: First, it obligates the agency to consider the environmental effects of the proposal; and second, it assures the public that the agency has considered environmental concerns in the decisionmaking process. Baltimore Gas & Elec. Co. v. Natural Resources Defense Council, Inc., 462 U.S. 87, 97 (1983). The federal agencies are well versed in NEPA's requirements, they commonly utilize environmental impact statements as a central element of their planning processes, and they have prepared thousands of such statements over the past 18 years. The court of appeals' decision in this case departs from precedent and established practice to put forward a novel interpretation of NEPA's requirements that would profoundly alter the federal agencies' obligations. The court's decision would require all environmental impact statements to contain detailed and demonstrably effective plans for mitigating environmental harm and, whenever the agency lacks information or adequate scientific methodology to determine precisely a potential environmental effect, to include a "worst case" analysis. Neither NEPA, the Council on Environmental Quality's implementing regulations, nor any of this Court's precedents provides a basis for imposing those requirements. As we demonstrate below and in our brief in the companion case, Robertson v. Methow Valley Citizens Council, No. 87-1703, the court's rigid directives are not only contrary to federal law, they would actually impede the goal of informed agency decisionmaking by preventing federal agencies from carrying out their resource management responsibilities through systematic project development. /6/ A. NEPA's Requirement That Federal Agencies Prepare A Detailed Statement on the Environmental Consequences of Major Federal Actions Does Not Impose A Substantive Obligation To Mitigate Adverse Environmental Effects Or A Procedural Obligation To Develop A "Complete Mitigation Plan" 1. This Court has clearly articulated the respective roles that NEPA assigns to the federal agencies and the courts. Section 102(2)(C) of NEPA (42 U.S.C. 4332(2)(C)) requires federal agencies to prepare an environmental impact statement that describes "'every significant aspect of the environmental impact of a proposed action'" and that "inform(s) the public that it has indeed considered environmental concerns in its decisionmaking process." Baltimore Gas & Elec. Co., 462 U.S. at 97 (quoting Vermont Yankee Nuclear Power Corp. v. Natural Resources Defense Council, Inc., 435 U.S. 519, 553 (1978)). While "NEPA does set forth significant substantive goals for the Nation, * * * its mandate to the agencies is essentially procedural." Vermont Yankee Nuclear Power Corp., 435 U.S. at 558. Furthermore, "the only procedural requirements imposed by NEPA are those stated in the plain language of the Act" (id. at 548). "The procedural duty imposed upon agencies (by Section 102(2)(C)) is quite precise, and the role of the courts in enforcing that duty is similarly precise." Kleppe v. Sierra Club, 427 U.S. 390, 406 (1976). "The only role for a court is to insure that the agency has taken a 'hard look' at environmental consequences; it cannot 'interject itself within the area of discretion of the executive as to the choice of the action to be taken'" (id. at 410 n.21 (citation omitted)). Furthermore, the courts should defer to the judgment of the Council on Environmental Quality (CEQ) -- the agency responsible for assuring proper implementation of NEPA (42 U.S.C. 4344) -- which has promulgated regulations for preparing environmental impact statements. See 40 C.F.R. 1501.1 et seq. "CEQ's interpretation of NEPA is entitled to substantial deference." Andrus v. Sierra Club, 442 U.S. 347, 358 (1979). The court of appeals departed from these principles, both here and in its later decision in Methow Valley Citizens Council v. Regional Forester, 833 F.2d 810 (1987), cert. granted, No. 87-1703 (June 27, 1988). The court concluded that NEPA requires federal agencies to assume significant new substantive and procedural obligations with respect to mitigation of environmental harm. It stated, in particular, that environmental impact statements must include a discussion of measures to mitigate adverse environmental consequences, adding that "(a)s long as significant measures are undertaken to mitigate the project's effects, the measures need not compensate completely for adverse environmental impacts" (Pet. App. 5a). The court's suggestion that NEPA imposes a substantive duty to mitigate environmental harm is obviously incorrect. /7/ Section 102 of NEPA, by its express terms, requires federal agencies to prepare a "detailed statement" of a proposed project's environmental consequences, including "adverse environmental effects which cannot be avoided," "alternatives to the proposed action," the relationship between "local short-term uses" and "long-term productivity," and "any irreversible and irretrievable commitments of resources" (42 U.S.C. 4332). Section 102 thus assures that all federal agencies will take environmental concerns into account in their decisionmaking processes. See Weinberger v. Catholic Action, 454 U.S. 139, 143 (1981); Strycker's Bay Neighborhood Council, Inc. v. Karlen, 444 U.S. 223, 227 (1980); Andrus v. Sierra Club, 442 U.S. 347, 350-351 (1979). See also 40 C.F.R. 1501.1. But Section 102 does not dictate what actions the agency should take in response to expected environmental consequences; indeed, "the Act mandates no particular substantive outcomes" (City of New York v. United States Dep't of Transportation, 715 F.2d 732, 748 (2d Cir. 1983), cert. denied, 465 U.S. 1055 (1984)). Instead, NEPA leaves substantive decisions to the agency. As this Court has repeatedly emphasized, NEPA requires the agency to consider and inform the public of potential environmental impacts; it does not require the agency to "'elevate environmental concerns over other appropriate considerations.'" Baltimore Gas & Elec. Co., 462 U.S. at 97-98 (quoting Strycker's Bay Neighborhood Council, Inc., 444 U.S. at 227). Accord Vermont Yankee Nuclear Power Corp., 435 U.S. at 558; Kleppe v. Sierra Club, 427 U.S. at 410 n.21. The court of appeals' suggestion that NEPA imposes a substantive duty to mitigate adverse environmental effects is thus patently incorrect. /8/ 2. Here, as in Methow Valley Citizens Council, the court of appeals' misperception of NEPA's function has led the court to misconstrue the statute's formal requirements. The court's erroneous premise that NEPA requires federal agencies to mitigate adverse environmental effects prompted it to construct a rigid and unwarranted procedural regime for developing mitigation measures. The court concluded that whenever a federal agency prepares an environmental impact statement, it must include a thorough discussion of measures to mitigate the adverse environmental impacts of a proposed action and "must analyze the mitigation measures in detail and explain the effectiveness of the measures" (Pet. App. 5a). The court specifically stated (id. at 7a): The importance of the mitigation plan cannot be overestimated. It is a determinative factor in evaluating the adequacy of an environmental impact statement. Without a complete mitigation plan, the decisionmaker is unable to make an informed judgment as to the environmental impact of the project -- one of the main purposes of an environmental impact statement. The court of appeals' inflexible new procedural requirements are foreign to NEPA and would ultimately impede the goal of sound agency decisionmaking. No one questions that a federal agency can and should consider mitigation opportunities when assessing the environmental consequences of a proposed action. The crucial issue here is the role that mitigation considerations play in the agency's environmental review. The court of appeals concluded that NEPA requires mitigation and that the government must therefore develop a "complete mitigation plan" as an essential element of an environmental impact statement (Pet. App. 7a). The government urges a fundamentally different proposition: NEPA requires environmental analysis and disclosure, and the government must therefore discuss mitigation in a manner that aids reasoned decisionmaking. See 40 C.F.R. 1502.1. "As Vermont Yankee made clear, NEPA does not require agencies to adopt any particular internal decisionmaking structure." Baltimore Gas & Elec. Co., 462 U.S. at 100. The agency has broad discretion to determine the form and scope of the mitigation discussion in light of the particular circumstances, subject only to the requirements that the environmental review as a whole satisfies NEPA's specific requirements and that the resulting decision is not arbitrary or capricious. See id. at 97-98, 105-106. The need for this flexibility becomes obvious when one considers the broad variety of contexts in which mitigation considerations may become relevant. The federal agencies must prepare an environmental impact statement (or explain why a statement is not necessary) for the full gamut of government activities, ranging from proposals for legislation, to physical construction and improvements, to the issuance of permits. In addition, federal agencies frequently find it necessary or desirable to provide a series of such statements at various stages of a single project's development. /9/ The relevance of and opportunities for mitigation vary considerably depending on the nature and complexity of the activity and the stage at which the environmental impact statement is being prepared. It would not be feasible, and often would be counterproductive, to develop a "complete mitigation plan" in every circumstance. /10/ Obviously, the agency's obligation to discuss mitigation measures will depend on the nature of the proposed agency action. The CEQ's implementing regulations, which are binding on all federal agencies (Exec. Order No. 11,991, 3 C.F.R. 123, 124 (1978)), therefore, provide a flexible approach to the consideration and discussion of such measures. The regulations, which broadly define the term "mitigation" to include virtually any technique that avoids, eliminates, ameliorates, or compensates for environmental harm (40 C.F.R. 1508.2), /11/ require a federal agency to consider mitigation opportunities at three stages of the NEPA review process. First, the CEQ regulations direct that when a federal agency takes the initial step of defining the "scope" of an environmental impact statement, it should consider the adoption of mitigation measures, not already included in the proposed action, as one type of alternative to the proposed action. See 40 C.F.R. 1508.25. /12/ This provision thus assures that when an agency determines the relevant range of alternatives, it will consider, in addition to the "(n)o action alternative" and "(o)ther reasonable courses of actions" (40 C.F.R. 1508.25(b)), the possibility of modifying the proposal to reduce or eliminate environmental harm. Second, the CEQ regulations direct that once the agency has defined the scope of its environmental review, it should address mitigation in the body of the environmental impact statement when discussing proposal alternatives and environmental consequences. See 40 C.F.R. 1502.14, 1502.16. These provisions, which instruct the agency to consider "appropriate mitigation measures not already included in the proposed action or alternatives" (40 C.F.R. 1502.14), thus encourage the agency to consider mitigation opportunities that were not identified in the scoping process. Third, the CEQ regulations state that if, upon completion of the environmental analysis, the agency chooses to go forward with the proposal, it must indicate in its record of decision "whether all practicable means to avoid or minimize environmental harm from the alternative selected have been adopted, and if not, why they were not." 40 C.F.R. 1505.2(c). The record of decision must also summarize a monitoring and enforcement program "where applicable for any mitigation" (40 C.F.R. 1505.2(c)), and the agency must confirm that the mitigation measures are ultimately implemented (40 C.F.R. 1505.3). These provisions ensure that the agency will consider mitigation opportunities in making its decision and will institute any measures that are chosen. The CEQ's regulations are "designed to make the environmental impact statement process more useful to decisionmakers and the public; and to reduce paperwork and the accumulation of extraneous background data, in order to emphasize the need to focus on real environmental issues and alternatives" (Exec. Order No. 11,991, 3 C.F.R. 123 (1978)). They permit a federal agency to consider mitigation opportunities in a systematic fashion and in accordance with NEPA's statutory framework. The regulations neither broaden nor restrict the scope of the inquiry that NEPA mandates; instead, they provide a blueprint for compliance. /13/ Furthermore, they preserve the agency's basic discretion under NEPA to determine in any particular case how far-ranging, thorough, or detailed the mitigation discussion should be. /14/ The CEQ regulations, which have the characteristic NEPA hallmark of adaptability, can be employed throughout the entire spectrum of agency decisionmaking. Furthermore, they ensure that an agency can target its discussion of mitigation to the particular decision at hand. Given NEPA's "quite precise" procedural instructions (Kleppe, 427 U.S. at 406) and the CEQ's carefully crafted implementing regulations, there is no need or warrant for courts to impose additional procedures. The court of appeals' rigid new requirement that all environmental impact statements must contain a "complete mitigation plan" not only exceeds the court's proper role (see, e.g., id. at 410 n.21), it is both unworkable and unnecessary. As this case demonstrates, that judicially coined requirement seriously impedes the NEPA review process with no corresponding benefits. 3. The Corps conscientiously complied with NEPA and the CEQ regulations in preparing the Elk Creek Supplemental Statement and, specifically, in discussing mitigation measures. The Statement discussed opportunities to mitigate harm to both fish and non-aquatic wildlife. The court of appeals did not question the adequacy of the mitigation measures for fish. /15/ The court found, however, that the discussion of wildlife mitigation measures was inadequate because it was not sufficiently detailed (Pet. App. 5a-7a). As we have explained (pages 6-7, supra), the Elk Creek Supplemental Statement determined that the project would have relatively limited impacts on wildlife. Elk Creek Supplemental Statement 23-27, 38. Construction of the Elk Creek Dam and the accompanying reservoir would inundate 1,290 acres of wildlife habitat, which would adversely affect black tail deer and elk, as well as certain game bird populations. No endangered species were found to exist in the project area, and the project was expected to affect no more than 100 deer and a lesser number of elk, reducing the size of the deer herd and displacing (but not reducing) the elk herd (ibid.). /16/ The Elk Creek Supplemental Statement suggested that these effects could be mitigated by managing selected lands nearby to improve the quality of the habitat and thereby increasing their wildlife carrying capacity. Specifically, the Statement suggested certain habitat manipulative techniques -- development of palatable browse plants for deer and vegetation interfaces (or "edge") for quail, variation of foliage height, and placement of snags in the reservoir -- to improve nearby habitat (id. at 6, 38). The Statement was clear, however, that the specific wildlife mitigation measures would be developed based upon the results of a wildlife compensation plan currently underway at the Applegate Dam and upon the further recommendations of federal and state wildlife agencies (id. at 6). The Statement also indicated that, even with adoption of mitigation techniques, some species would be adversely affected by the project (id. at 38, 46). This discussion of wildlife mitigation measures fully complied with NEPA and the CEQ regulations, and it was sufficiently thorough for the decision at hand; namely, whether to proceed with completion of the Elk Creek Dam (Pet. App. 53a). The level of detail that an agency employs in its environmental discussions must, of course, be determined in light of "feasibility" and "(c)ommon sense" (Vermont Yankee Nuclear Power Corp., 435 U.S. at 551). Furthermore, the lower courts have specifically held that where mitigation details are contingent on presently unknown factors, it is appropriate for an agency to restrict its discussion accordingly. See, e.g., Concerned Citizens on I-190 v. Secretary of Transportation, 641 F.2d 1, 5-6 (1st Cir. 1981). /17/ Here, where the impacts on wildlife were limited, the discussion of mitigation measures could be correspondingly limited as well. The Corps and the district court both agreed that the Elk Creek Supplemental Statement's mitigation discussion satisfied NEPA's requirements (Pet. App. 41a-42a, 56a-57a, 57a-58a, 61a). The agency, overall, took a "hard look" at environmental consequences and measures to ameliorate those consequences; this is all that NEPA requires. See Baltimore Gas & Elec. Co., 462 U.S. at 97. /18/ The court of appeals, therefore, erred in concluding that the Elk Creek Supplemental Statement failed to address mitigation adequately. Moreover, the court's blanket requirement that agencies prepare a "complete mitigation plan" is both unworkable and unwise. The court's requirement pervasively conflicts with CEQ regulations /19/ and would expunge from NEPA the flexibility that allows vastly different federal agencies to perform meaningful environmental analyses tailored to their functions and needs. As we explain in our brief in Robertson (at pages 25-38), agencies would no longer be able to plan and implement their activities through staged project development (see note 9, supra). As a result, the court's decision would reduce, rather than enhance, the agency's ability to conduct informed and well-reasoned decisionmaking. And the courts themselves would be forced to expend considerable energy adjudicating the ultimately discretion-laden question whether a "complete mitigation plan" is sufficiently detailed to qualify as "complete." This Court has repeatedly emphasized that the courts have a limited role to play under NEPA. E.g., Baltimore Gas & Elec. Co., 462 U.S. at 97-98. In essentially rewriting NEPA to suit its own vision of environmental policy, the court of appeals here has clearly overstepped its proper bounds. B. NEPA Does Not Require Federal Agencies To Prepare A "Worst Case" Analysis Of Potential Environmental Harm If Relevant Information Concerning Significant Environmental Effects is Unavailable Or Too Costly To Obtain 1. The court of appeals also erred in concluding (Pet. App. 13a) that NEPA requires federal agencies to conduct a "worst case" analysis whenever relevant scientific information on environmental effects is unavailable or too costly to obtain. NEPA itself prescribes no particular method for dealing with the problem of scientific uncertainty; instead, the CEQ introduced worst case analysis as a method for addressing that problem. The CEQ's experience with worst case analysis ultimately revealed that the method did not effectively fulfill NEPA's objectives, and the agency therefore has replaced it with a more manageable methodology that identifies "reasonably foreseeable" environmental impacts. The court of appeals nevertheless insisted that "NEPA case law" (id. at 14a n.8) requires that agencies continue to employ worst case analysis. The court's decision is fatally infirm. Worst case analysis is a product of the CEQ regulations only, and the case law on that subject merely interprets those regulations. The CEQ originated the concept of worst case analysis in its 1978 implementing regulations. See 43 Fed. Reg. 55,978, 55,997. Those regulations, which were promulgated pursuant to presidential order, represented the CEQ's first attempt to provide an authoritative blueprint for NEPA compliance that would be binding on every federal agency. See id. at 55,978. Section 1502.22, which came to be known as the worst case requirement, described the government's obligation when information relevant to a proposed action's significant adverse environmental effects was either unavailable or incomplete. See 40 C.F.R. 1502.22 (1985). /20/ Under that section, the agency was required, if possible, to obtain the information and to include it in the environmental impact statement (40 C.F.R. 1502.22(a) (1985)). If the information were truly unavailable or too costly to obtain, the agency was required to "weigh the need for the action against the risk and severity of possible adverse impacts" and to provide a "worst case analysis and an indication of the probability or improbability of its occurrence" (40 C.F.R. 1502.22(b) (1985)). The CEQ gave no explanation why it chose worst case analysis over other methods for dealing with uncertainty (see 43 Fed. Reg. 55,984 (1978)). /21/ Moreover, the CEQ's regulation gave no explanation of how an agency was to go about conducting a worst case analysis. As a result, federal agencies encountered substantial difficulties in implementing the worst case requirement. The CEQ published some interpretive information (46 Fed. Reg. 18,026 (1981)) but continued to receive requests for clarification from government agencies and the private sector. Many of these inquiries criticized the regulation for failing to identify clearly the circumstances in which the worst case analysis requirement was triggered. As a result, the Council drafted guidance on the regulation and proposed to incorporate an initial threshold of probability before a worst case analysis would be required. See 48 Fed. Reg. 36,486 (1983). After receiving conflicting public comments, however, CEQ withdrew the proposed guidance to examine the issue further before publishing a new proposal. 49 Fed. Reg. 4803 (1984). Meanwhile, the courts' interpretation of the worst case regulation, accompanied by judicial second-guessing of the agencies' scientific determinations, rendered the method largely unworkable. The courts repeatedly overturned agency worst case analyses and instructed the agency to evaluate rigorously a broad spectrum of highly remote possibilities, declaring "the specific worst case to be examined by the agency on remand, including the particular scientific models and studies to be applied and the research to be performed." Note, Federal Agency Treatment of Uncertainty in Environmental Impact Statements Under the CEQ's Amended NEPA Regulation Section 1502.22: Worst Case Analysis or Risk Threshold?, 86 Mich. L. Rev. 777, 808 (1988). See, e.g., Save Our Ecosystems v. Clark, 747 F.2d 1240, 1244-1246 (9th Cir. 1984); Southern Oregon Citizens Against Toxic Sprays, Inc. v. Clark (SOCATS), 720 F.2d 1475, 1479 (9th Cir. 1983), cert. denied, 469 U.S. 1028 (1984) (White and O'Connor, JJ., dissenting); Sierra Club v. Sigler, 695 F.2d 957 (5th Cir. (1983). See generally Note, supra, 86 Mich. L. Rev. at 798-804. In 1984, following this Court's denial of the government's petition for a writ of certiorari in SOCATS, the CEQ published an advance notice of proposed rulemaking to solicit comments on possible amendment of the worst case regulation. See 49 Fed. Reg. 50,744 (1984). The following year, the CEQ published a proposed amendment that suggested eliminating the worst case analysis requirement and replacing it with a requirement that the agency consider reasonably foreseeable impacts based upon credible scientific support. See 50 Fed. Reg. 32,234, 32,238 (1985). The CEQ explained that, based on an intensive review, it had concluded that worst case analysis "is a unsatisfactory approach to the analysis of potential consequences in the face of missing information" (id. at 32,236). It noted that worst case analysis is unworkable, in part, because of the "limitless nature of the inquiry established by this requirement; that is, one can always conjure up a worse 'worst case' by adding an additional variable to a hypothetical scenario" (ibid.). /22/ The CEQ further observed (ibid. (citation omitted)): Experts in the field of risk analysis and perception stated that the "worst case analysis" lacks defensible rationale or procedures, and that the current regulatory language stands "without any discernible link to the disciplines that have devoted so much thought and effort toward developing rational ways to cope with problems of uncertainty. It is, therefore, not surprising that no one knows how to do a worst case analysis . . . . " The CEQ ultimately decided to replace the worst case regulation with a new regulation that, in that agency's judgment, provides "a wiser and more manageable approach to the evaluation of reasonably foreseeable significant adverse impacts in the face of incomplete or unavailable information in an EIS." 51 Fed. Reg. 15,620 (1986). /23/ The new regulation states that if "incomplete information relevant to reasonably foreseeable significant adverse impacts is essential to a reasoned choice among alternatives and the overall costs of obtaining it are not exorbitant, the agency shall include the information in the environmental impact statement" (40 C.F.R. 1502.22(a)). If information relevant to reasonably foreseeable significant adverse impacts cannot be obtained, the agency must include four items in the environmental impact statement: (1) a statement that the information is incomplete or unavailable; (2) a statement of the relevance of the incomplete or unavailable information to evaluating impacts; (3) a summary of existing credible scientific evidence relevant to evaluating impacts; and (4) the agencies' evaluation of the impacts based upon theoretical approaches or research methods generally accepted in the scientific community (40 C.F.R. 1502.22(b)). /24/ The revised regulation accordingly requires "federal agencies to evaluate the reasonably foreseeable significant adverse impacts of an action, even in the face of unavailable or incomplete information" (51 Fed. Reg. 15,621 (1986)). It "specifies that the evaluation must be carefully conducted, based upon credible scientific evidence, and must consider those reasonably foreseeable significant adverse impacts which are based on scientific evidence" (ibid.). In short, the amended regulation "retains the duty to describe the consequences of a remote, but potentially severe impact, but grounds the duty in evaluation of scientific opinion rather than in the framework of a conjectural 'worst case analysis.'" 50 Fed. Reg. 32,237 (1985). 2. The court of appeals refused to give effect to the CEQ's carefully deliberated regulatory revision, holding here -- and later in Methow Valley Citizens Council -- that NEPA requires worst case analysis irrespective of the CEQ regulations. See Pet. App. 13a-15a & n.8. The court incorrectly reasoned that the CEQ's worst case regulation is a codification of prior NEPA case law" and that "the rules embodied in the regulation remain in effect even though the regulation was rescinded" (id. at 14a n.8). In support of that assertion, the court of appeals summarily cited its previous decision in Save Our Ecosystems. But Save Our Ecosystems simply cites SOCATS for that proposition. See 747 F.2d at 1244. SOCATS, in turn, simply cites Sierra Club v. Sigler, 695 F.2d at 971. There the trail ends. Neither Sigler nor any previous case holds that NEPA requires the federal agencies to conduct a worst case analysis. The question in Sigler was whether the Army Corps of Engineers was required to comply with the CEQ's worst case analysis regulation under the particular facts of that case. The Fifth Circuit first observed that while there "is some language in NEPA which may be said to endorse generally the concept of a worst case analysis" (695 F.2d at 969). The court next noted that NEPA does contemplate that agencies would engage in "'(r)easonable forecasting and speculation'" concerning future environmental effects, including the "probabilities of the occurrence" and the "'cost of uncertainty -- i.e., the costs of proceeding without more and better information'" (id. at 970). /25/ The court then stated that the "CEQ's worst case analysis merely codifies these judicially created principles" (id. at 971). Thus, the Fifth Circuit concluded that the worst case regulation was the CEQ's mechanism for implementing certain judicially recognized NEPA objectives. The court did not hold that the worst case methodology was the only allowable mechanism for effecting those general goals. Accord Note, supra, 86 Mich. L. Rev. at 797-798; see also id. at 798 n.105, 801-802 & nn. 120-121, 812-814, 819-820. Indeed, the court of appeals' conclusion here that NEPA requires a worst case analysis is inconsistent with that court's own prior case law, which recognizes that "(a) reasonably thorough discussion of the significant aspects of the probable environmental consequences is all that is required by an EIS." Trout Unlimited v. Morton, 509 F.2d 1276, 1283 (9th Cir. 1974). /26/ The CEQ has adhered to the objectives identified in Sigler, but has determined that "the 'worst case analysis' requirement is an unproductive and ineffective method of achieving those goals" (51 Fed. Reg. 15,618, 15,620 (1986)). The CEQ has therefore adopted "a wiser and more manageable approach to the evaluation of reasonably foreseeable significant adverse impacts in the face of incomplete or unavailable information" (id. at 15,620). The court of appeals' decision in this case would effectively rescind the CEQ's new regulation and, contrary to that court's own decision in Trout Unlimited, would install worst case analysis as a permanent feature of NEPA compliance. This aspect of the court of appeals's decision is obviously contrary to this Court's holding that the CEQ's judgment is entitled to substantial deference, even when the agency alters its prescribed procedures for NEPA compliance. Andrus v. Sierra Club, 442 U.S. 347, 358 (1979). See also, e.g., American Trucking Ass'ns v. Atchison, T. & S.F. Ry., 387 U.S. 397, 416 (1967); Sierra Club v. Sigler, 695 F.2d 957, 967, 972 (5th Cir. 1983) (recognizing that NEPA requires adherence to CEQ regulations, even when CEQ changes its interpretation, and that CEQ regulations are entitled to substantial deference). /27/ And, as is true of its novel mitigation requirements, the court of appeals' holding here would have severe consequences. As the CEQ has explained, requiring a worst case analysis forces federal agencies to conduct wasteful and time-consuming inquiries into speculative impacts that lack "a firm connection between credible science and the hypothetical consequences of an agency's proposed action" (50 Fed. Reg. 32,236 (1985)). This sort of "pure conjecture * * * lacking a credible scientific basis is not useful to either the decisionmaker or the public" (ibid. (emphasis in original)). The CEQ, after careful deliberation, properly elected to eliminate its worst case analysis requirement. The court of appeals improperly relied on a "contrived 'common law' rationale for overruling the CEQ" (Note, supra, 86 Mich. L. Rev. at 820). 3. The court of appeals' imposition of the worst case requirement here was particularly far-reaching because there were no meaningful "gaps in relevant information or scientific uncertainty" (40 C.F.R. 1502.22 (1985)) in this case. The court found that "the Corps' assessment of turbidity due to Elk Creek Dam is subject to uncertainty" based on a single comment from the ODFW addressing the Elk Creek Supplemental Statement's observation that Elk Creek provides a small portion of the Rogue River's total flow (Pet. App. 14a). According to the court of appeals, "(t)he state agency specifically disagreed with the Corps' finding that Elk Creek would make only a small contribution to total flow in the Rogue River" (ibid.). In fact, the Corps had considered and addressed conditions in which Elk Creek would contribute higher flows to the Rogue. /28/ Whatever the merits of the ODFW's comment, it does not, we submit, give rise to a case of incomplete information within the meaning of either the old or the new version of 40 C.F.R. 1502.22. Section 1502.22's requirements do not come into play whenever a commenting party disagrees with the agency; rather, they apply only when information "essential to a reasoned choice among alternatives" is lacking. /29/ The Corps, an expert agency charged with making such determinations, found no such lack of essential information here. Its determination, which is entitled to deference, is both reasonable and manifestly correct. Accordingly, neither the old nor the new CEQ regulation concerning incomplete information is applicable to this situation. II. THE ARMY CORPS OF ENGINEERS GAVE ADEQUATE CONSIDERATION TO INFORMATION CONCERNING ENVIRONMENTAL EFFECTS THAT WAS PRESENTED AFTER THE COMPLETION OF THE CORPS' SUPPLEMENTAL ENVIRONMENTAL IMPACT STATEMENT The court of appeals also erred in requiring (Pet. App. 7a) the Corps to prepare a second supplemental environmental impact statement to address the ODFW Report's information concerning fish mortality and the SCS Report's information concerning soil characteristics. The Corps had considered the information in both instances and determined that it was either unreliable or lacked sufficient bearing on the proposed action to warrant preparation of another supplemental statement. The district court, which heard and considered the testimony of witnesses on these technical subjects, agreed that the Corps' decision to forgo preparation of another supplemental environmental impact statement was reasonable (Pet. App. 46a-49a). The court of appeals nevertheless reversed that judgment. As Judge Wallace observed in dissent (Pet. App. 21a-28a), the court of appeals departed from its proper appellate role. Rather than examining the district court's decision under the appropriate standard of review, the court assumed the role of a de novo factfinder charged with responsibility for making scientific determinations (id. at 28a). 1. Federal agencies frequently develop or receive new and updated information, prior to a project's completion, arguably relevant to the project's environmental effects. When the agency receives such information after the publication of its environmental impact statement, it must determine whether the information is sufficiently important to require reinvocation of NEPA's formal environmental evaluation procedures. /30/ That determination requires a pragmatic assessment of the information's value in the decisionmaking process. /31/ Furthermore, the determination, like other agency judgments concerning the extent of environmental analysis required, typically "requires a high level of technical expertise and is properly left to the informed discretion of the responsible federal agencies" (Kleppe v. Sierra Club, 427 U.S. at 412). The scope of judicial inquiry is correspondingly limited. See ibid.; Baltimore Gas & Elec. Co., 462 U.S. at 97-98; Vermont Yankee Nuclear Power Corp., 435 U.S. at 554-555. The Administrative Procedure Act (APA) identifies the appropriate standard of judicial review. A reviewing court may not set aside an agency's decision to forgo preparation or supplementation of an environmental impact statement unless the agency action is "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law" (5 U.S.C. 706(2)(A)). See Wisconsin v. Weinberger, 745 F.2d 412, 418 (7th Cir. 1984); Webb v. Gorsuch, 669 F.2d 157, 159 (4th Cir. 1983); Cabinet Mountains Wilderness/Scotchman's Peak Grizzly Bears v. Peterson, 685 F.2d 678, 681, 685 (D.C. Cir. 1982); Hanly v. Kleindienst, 471 F.2d 823, 828-830 (2d Cir. 1972), cert. denied, 412 U.S. 908 (1973). /32/ While the agency employs a "reasonableness standard" in determining whether a supplemental environmental impact statement is necessary (see, e.g., Friends of the River v. FERC, 720 F.2d 93, 109 (D.C. Cir. 1983)), the courts must apply the APA standard in reviewing the agency's choice. There is no warrant for creating a less deferential standard of judicial review. /33/ In applying the "arbitrary or capricious" standard, the court of appeals and the district court "must consider whether the decision was based on a consideration of the relevant factors and whether there has been a clear error of judgment" (Citizens to Preserve Overton Park v. Volpe, 401 U.S. 402, 416 (1971)). The courts must, however, give appropriate deference to the agency, especially with respect to scientific judgments within an agency's expertise. See Baltimore Gas & Elec. Co., 462 U.S. at 103. Here, as with other NEPA inquiries, "(a)bsent a showing of arbitrary action, (courts) must assume that the agencies have exercised this discretion appropriately" (Kleppe, 427 U.S. at 412). See Baltimore Gas & Elec. Co., 462 U.S. at 97-98; Vermont Yankee Nuclear Power Corp., 435 U.S. at 554-555. As the Seventh Circuit has observed (Wisconsin, 745 F.2d at 418 (emphasis in original)): an agency cannot have acted arbitrarily or capriciously in deciding not to file a (supplemental environmental impact statement) unless the new information provides a seriously different picture of the environmental landscape such that another hard look is necessary. 2. The court of appeals paid little apparent heed to these principles or, indeed, to the decision of the district court. Rather than inquiring whether the agency made a "clear error in judgment," the court essentially assumed the role of a de novo factfinder. The court gave no deference to the Corps' scientific determinations; instead, it engaged in a cursory -- and flawed -- evaluation of the scientific evidence. The court ultimately overturned the Corps' decision based on an estimation that "some of the proffered information is probably accurate" (Pet. App. 11a) and a mistaken impression that other information "was never considered" (Pet. App. 12a). As Judge Wallace observed in dissent (Pet. App. 21a-28a), the court not only employed the incorrect legal standard, its ensuing conclusions with respect to the ODFW and SCS Reports are inconsistent with those of the district court, the expert agency, and the agency's independent consultants. a. The ODFW Report, which was part of an extensive study of anadromous fishery resources financed by the Corps, describes the effects of the Lost Creek Dam on downstream fisheries resources. Respondents based their argument that the ODFW Report contained significant new information on an internal ODFW memorandum (the Cramer Memorandum, see note 4 supra) from two ODFW staff biologists to their superior, extrapolating the draft ODFW Report results to Elk Creek. /34/ The Cramer Memorandum suggested, based on the results of the draft ODFW Report, that the release of water from the Elk Creek reservoir could increase water temperatures in the Rogue River and thereby decrease survival of newly hatched chinook salmon (Addendum, infra, 3a). /35/ The Memorandum further suggested, based again on information in the draft ODFW Report, that impoundment and release of that water could lead to an increase in epizootic (population-wide) fish disease (id. at 4a-5a). /36/ Finally, the Memorandum suggested, based on complaints from fishermen, that reservoir water releases from September to October could affect angling on the Rogue River (id. at 3a-4a). See also Pet. App. 8a-10a, 47a-48a, 62a-66a. The Corps evaluated the Cramer Memorandum and the ODFW Report, using both its own staff and outside experts, and determined that those documents did not present sufficiently reliable new information to justify the preparation of a second supplemental environmental impact statement. The Corps prepared a supplemental information report (the Elk Creek SIR, see note 4, supra) to respond to the concerns raised. First, the Corps questioned the Cramer Memorandum's reliance on the ODFW Report's mathematical model relating temperature increases to fish mortality. The Corps found that model to be of questionable worth since it had not been validated and did not take into account the Elk Creek Dam's beneficial effects during egg incubation and fry rearing. See Addendum, infra, 9a-10a. /37/ Furthermore, the Corps' modeling of downstream temperatures, using 1978 through 1979 as validation years, indicated that the Elk Creek Dam would have little effect on water temperatures (id. at 10a). /38/ The Corps also noted that the Elk Creek Dam's multilevel intake tower, which allows water discharges from different temperature zones within the reservoir, would minimize adverse effects on the fisheries resources (id. at 10a, 13a). See also Pet. App. 22a-26a, 46a-49a. The Corps also found the Cramer Memorandum's predictions of an increase in fish disease to be unreliable. The incidents of increased Lost Creek fish mortality documented in the ODFW Report had never been traced to dam construction nor has there been any recurrence of epizootic disease at the location since 1981. See Addendum, infra, 10a-11a). See also Cramer Dep. 34-41 (Dec. 4, 1985); A.R. Doc. No. 12 (Critique of Lost Creek Fisheries Evaluation) 2-3. Corps-funded ODFW studies had failed to detect disease-causing organisms in the Lost Creek outflow; furthermore, there were no similar epizootic outbreaks following closure of the Applegate Dam (id. at 11a). Thus, the Corps concluded that there was no firm evidence to implicate the Elk Creek reservoir as a potential harborer of disease organisms (ibid.). See also Pet. App. 22a-26a, 46a-49a. /39/ In reaching these conclusions, the Corps relied not only on the judgment of its own experts, but also on that of two independent experts hired specifically to consider and evaluate the significance of the ODFW Report. Both of these consultants criticized the methodology and conclusions of the ODFW Report. See Pet. App. 22a-23a. /40/ Furthermore, the ODFW biologists admitted, in depositions and at the hearing in district court, that the report's conclusions were open to question. See Pet. App. 23a-24a, 63a-65a. Thus, the Corps had a sound basis for deciding that the Cramer Memorandum and the ODFW Report did not present significant new information requiring a supplemental environmental impact statement. The district court correctly concluded that the Corps' decision was reasonable (Pet. App. 47a-49a). b. The Corps was also justified in concluding that the SCS Report did not contain significant new information concerning turbidity. The SCS Report -- which consisted exclusively of a collection of soil maps (see note 5, supra) -- was used by respondents in district court to urge that soils in the Elk Creek area have higher sediment-producing capacity than the Corps had estimated and might make a greater contribution to reservoir turbidity owing to the wave action of the impounded water on the reservoir banks. The Corps, however, had already studied turbidity questions in a far more comprehensive and meaningful manner. As we explained (pages 4-6, supra), the Corps completed an exhaustive study on the projected turbidity from Lost Creek and Elk Creek dams -- the 1974 Water Quality Report -- and concluded that the Rogue River Project, as a whole, would increase turbidity in specific amounts -- from 2 to 3 JTU in average runoff years and from 6 to 9 JTU in high runoff years. The results of the 1974 study were confirmed by a verification study -- the 1979 Water Quality Report -- conducted at Lost Creek after that dam was built. In addition, the Corps had considered and addressed the relationship between soil characteristics and turbidity in the Elk Creek Supplemental Statement. The Corps explained that soil characteristics do not play a significant role in predicting turbidity; rather, "soil disturbances are the chief contributing factor to turbidity in runoff" (Elk Creek Supplemental Statement 115). Thus, the SCS Report had limited relevance to actual turbidity levels. /41/ The Corps provided further elaboration on that point in the hearing on respondents' motion for a preliminary injunction. The Corps' engineering geologist explained that the SCS used a "different frame of reference" than the Corps in conducting its study. Tr. 71 (Jan. 16, 1986) (Sturgill testimony). When the SCS conducted its soil survey, it described the soil type and texture. That description provided guidance on the soil's potential, in the abstract, to produce suspended sediment. The Corps, by contrast, approached the question from "an engineering standpoint" and examined whether the soil would actually produce turbidity under the conditions experienced at the reservoir (id. at 70-72). /42/ The Corp's approach also took into account "management" techniques to reduce turbidity (id. at 70). For example, the Corps could reduce the effects of wave action on erosive soils by coating the shore with gravel (see Tr. 57 (Jan. 16, 1986) (Borine testimony)). See Pet. App. 26a-27a; 48a-49a. When the Corps received the soil maps constituting the SCS Report, it reviewed them and concluded that the information -- though correct as far as it went -- would not change the Corps' turbidity projections. /43/ The Corps accordingly concluded that the SCS Report did not provide significant new information. The district court correctly held that the Corps' conclusion was reasonable (Pet. App. 48a-49a). 3. The court of appeals' action in this case cannot be squared with this Court's decisions in Baltimore Gas & Elec. Co., Vermont Yankee Nuclear Power Corp., and Kleppe, or with the dictates of limited review contained in the APA. As this Court recognized in Vermont Yankee Nuclear Power Corp., a party dissatisfied with the results of an administrative proceeding can always raise a claim that "new information" has been discovered after the administrative record has been closed (435 U.S. at 554-555). If the reviewing court refuses to give substantial deference to the agency's expert judgment on the significance of the new information, there will be few cases in which no plausible basis for requiring additional studies can be asserted in litigation. A Court is not entitled to "substitute its judgment for that of the agency as to the environmental consequences of its actions" (Kleppe, 427 U.S. at 410 n.21); it is particularly inappropriate for a court of appeals to do so where, as here, a district court has reviewed the administrative record, heard additional scientific evidence on the environmental significance and probable accuracy of the new information at issue, and has affirmed the expert agency's scientific determination. /44/ The court of appeals thus has exceeded its reviewing function under this Court's precedents and, in the process, misjudged the significance of the ODFW and SCS Reports. This error -- which would substantially and needlessly burden the Corps' completion of the Elk Creek Dam -- should be corrected by this Court, as should the other errors we have discussed. CONCLUSION The judgment of the court of appeals should be reversed with instructions to remand the case to the agency for further proceedings. /45/ Respectfully submitted. CHARLES FRIED Solicitor General ROGER J. MARZULLA Assistant Attorney General LAWRENCE G. WALLACE Deputy Solicitor General JEFFREY P. MINEAR Assistant to the Solicitor General PETER R. STEENLAND, JR. VICKI L. PLAUT Attorneys SEPTEMBER 1988 /1/ We have simultaneously submitted a separate brief in Robertson. /2/ For the Court's convenience, we have lodged copies of the Elk Creek Supplemental Statement with the Clerk of the Court. /3/ A "JTU" -- or Jackson Turbidity Unit -- is a unit of measurement for water clarity. As a general rule, the turbidity limit for drinking water is 5 JTU, turbidity of 10 JTU impairs flyfishing, turbidity of 20 JTU impairs other fishing methods, and a long-term turbidity of 50 JTU alters fish behavior. Elk Creek Supplemental Statement 21. /4/ As the district court noted (Pet. App. 47a), respondents actually relied on an internal memorandum by two ODFW staff biologists (hereinafter Cramer Memorandum) that expressed concern, based on the draft ODFW Report, that the Elk Creek Dam could adversely affect fishery resources. We have reproduced the Cramer Memorandum, which is not a part of the administrative record, as Addendum B, infra 3a-6a. See also Plaintiffs' Brief in Support of Plaintiffs' Motions for a Temporary Restraining Order Etc. Exh. 14 (Nov. 1, 1985). After obtaining the opinions of two independent fisheries experts, the Corps issued a supplemental information report (hereinafter Elk Creek SIR) explaining, among other matters, why the Cramer Memorandum and the ODFW Report do not contain significant new information (Pet. App. 59a-67a). For convenience, we have reproduced the Elk Creek SIR as Addendum C, infra, 7a-13a. /5/ The SCS Report, which consists of a series of maps, appears in the record as hearing exhibits 1 through 8. See Pet. App. 47a n.15. /6/ The court of appeals also erred in holding that the Elk Creek Supplemental Statement's discussion of cumulative impacts of the entire Rogue River Project was inadequate (Pet. App. 16a-17a). The 1974 Water Quality Report, which was incorporated into the Supplemental Statement by reference, had adequately addressed the cumulative impacts; thus, there was no need for the Supplemental Statement to revisit that issue. The significance of the court's factbound ruling, however, is limited to this case. We therefore have not sought correction by this Court of that error. /7/ The court derived that principle from Friends of Endangered Species, Inc. v. Jantzen, 760 F.2d 976, 987 (9th Cir. 1985). As we explain (note 10, infra), the Jantzen passage does not support the court of appeals' suggestion that environmental impact statements must always provide for some mitigation; read in context, the passage states that if an agency substantially mitigates the environmental effects of its actions, it may not need to prepare any environmental impact statement at all. See 760 F.2d at 987. The court of appeals' imposition of a substantive duty to mitigate was even more explicit in Methow Valley Citizens Council v. Regional Forester, where the court stated that Section 102 of NEPA "requires that 'action be taken to mitigate the adverse effects of major federal actions'" (833 F.2d at 819, citing Stop H-3 Ass'n v. Brinegar, 389 F. Supp. 1102 (D. Haw. 1974)). See 87-1703 Pet. App. 17a. As we explain in our brief to this Court in that case, the court's reliance on Stop H-3 Ass'n (which was reversed on other grounds, 533 F.2d 434 (9th Cir.), cert. denied, 429 U.S. 999 (1976)) is misplaced. /8/ When Congress has wished to impose substantive environmental protection obligations on federal agencies, it has done so explicitly. For example, Congress specified in the Endangered Species Act of 1973 (ESA), Pub. L. No. 93-205, 87 Stat. 884, that federal agencies shall "insure that any action authorized, funded, or carried out * * * is not likely to jeopardize the continued existence of any endangered species" (16 U.S.C. 1536(a)(2)). See Roosevelt Campobello Int'l Park v. EPA, 684 F.2d 1041, 1049 (1st Cir. 1982). Similarly, Congress specified in Section 4(f) of the Department of Transportation Act of 1966, Pub. L. No. 89-670, 80 Stat. 934, that the Secretary of Transportation may approve a transportation project requiring use of a public park, recreation area, wildlife refuge, or historic site only if "there is no prudent and feasible alternative to using that land; and * * * the program or project includes all possible planning to minimize harm to the (land)" (49 U.S.C. 303). These statutes, which were enacted in the same general time frame as NEPA, demonstrate that when Congress wishes to impose substantive mitigation obligations it does so expressly. The court of appeals' effort to write such a duty into NEPA would usurp Congress's deliberate contrary choice. /9/ The CEQ has encouraged agencies to utilize the practice of "tiering" environmental impact statements. See 40 C.F.R. 1502.20, 1508.28. Tiering is simply a method of implementing the familiar and sensible concept of staged project development. See, e.g., D. Dannenbring & M. Starr, Management Science: An Introduction 548 (1981) ("A project is itself a combination of many varied and complex tasks or activities. These tasks are interdependent in that most cannot begin until some other task has been completed."). Tiering "is appropriate when it helps the lead agency to focus on the issues which are ripe for decision and exclude from consideration issues already decided or not yet ripe" (40 C.F.R. 1508.28). /10/ Prior to the Ninth Circuit's decisions here, legal challenges involving an agency's discussion of mitigation measures typically arose out of the agency's determination that mitigation measures obviated the need for an environmental impact statement. In response, the courts consistently recognized that a "Finding of No Significant Impact" (40 C.F.R. 1508.13) is appropriate if otherwise significant environmental effects will be remedied by mitigation measures built into the project. See Friends of Endangered Species, Inc. v. Jantzen, 760 F.2d at 987; Louisiana v. Lee, 758 F.2d 1081, 1083 (5th Cir. 1985), cert. denied, 475 U.S. 1044 (1986); Cabinet Mountains Wilderness/Scotchman's Peak Grizzly Bears v. Peterson, 685 F.2d 678, 681-684 (D.C. Cir. 1982); Preservation Coalition, Inc. v. Pierce, 667 F.2d 851, 860-861 (9th Cir. 1982; City & County of San Francisco v. United States, 615 F.2d 498, 501 (9th Cir. 1980); Maryland-National Capital Park & Planning Comm'n v. U.S. Postal Service, 487 F.2d 1029, 1039 (D.C. Cir. 1973). See also 46 Fed. Reg. 18,026 (1981). The Ninth Circuit specifically stated in this context that "so long as significant measures are undertaken to 'mitigate the project's effects,' they need not completely compensate for adverse environmental impacts." Friends of Endangered Species, Inc. v. Jantzen, 760 F.2d at 987 (emphasis in original; citation omitted). The court of appeals here mistakenly cited Jantzen for the radically different proposition that an environmental impact statement must always provide for some mitigation of environmental harm. See Pet. App. 5a. See note 7, supra. /11/ The CEQ regulations state that mitigation includes (40 C.F.R. 1508.20): (a) Avoiding the impact altogether by not taking a certain action or parts of an action. (b) Minimizing impacts by limiting the degree or magnitude of the action and its implementation. (c) Rectifying the impact by repairing, rehabilitating, or restoring the affected environment. (d) Reducing or eliminating the impact over time by preservation and maintenance operations during the life of the action. (e) Compensating for the impact by replacing or providing substitute resources or environments. /12/ Once an agency decides that it will prepare an environmental impact statement, it commences "an early and open process for determining the scope of issues to be addressed and for identifying the significant issues related to a proposed action." 40 C.F.R. 1501.7. The "scope" of an environmental impact statement "consists of the range of actions, alternatives, and impacts to be considered" in the document. 40 C.F.R. 1508.25. /13/ Thus, the CEQ regulations require the agency to consider mitigation measures with reference to NEPA's specific criteria, including the environmental impact of the proposed action; the adverse environmental impacts that cannot be avoided; and alternatives to the proposed actions. See 42 U.S.C. 4332(2)(C). /14/ NEPA requires a "detailed statement" of environmental consequences, but it does not require that every subsidiary matter (such as mitigation) be discussed comprehensively, nor does it require -- explicitly or implicitly -- a detailed discussion of mitigation. Obviously, an agency's discussion of the five subjects identified in NEPA can be "detailed" even if the discussion of their subsidiary components, viewed individually, contains varying levels of detail. /15/ The most significant effect on fish -- the loss of spawning grounds for coho salmon and steelhead trout -- had already been mitigated through the construction of the Cole M. Rivers fish hatchery. Elk Creek Supplemental Statement 35-36. The fish hatchery would also compensate for decreases in fish production owing to the expected minor increase in turbidity (see id. at 37). /16/ At trial, respondents' wildlife experts estimated that the inundation would affect approximately 100 out of more than 100,000 deer in the county, reducing their number by about 50. Tr. 19-20, 22 (Jan. 16, 1986) (Werner testimony). Respondents' experts also estimated that reservoir construction would displace 60 elk that were believed to live part-time in the reservoir area, but the project was not expected to reduce this highly migratory population at all (id. at 18-19). These estimates are generally consistent with the information contained in the Elk Creek Supplemental Statement. /17/ The Elk Creek Supplemental Statement's deferral of mitigation details pending completion of current studies at the Applegate Dam is consistent with the view, expressed by the Ninth Circuit as well as other courts, that the planning process is an ongoing endeavor and that agencies should be encouraged to develop and institute mitigation measures as the project develops, even after the environmental analysis is complete. See Village of False Pass v. Clark, 733 F.2d 605, 616 (9th Cir. 1984); Township of Springfield v. Lewis, 702 F.2d 426, 437-438 (3d Cir. 1983); National Indian Youth Council v. Watt, 664 F.2d 220, 225 (10th Cir. 1981); Concerned Citizens on I-190 v. Secretary of Transportation, 641 F.2d 1, 6 (1st Cir. 1981). See also County of Suffolk v. Secretary of the Interior, 562 F.2d 1368, 1378 (2d Cir. 1977), cert. denied, 434 U.S. 1064 (1978). In Township of Springfield, for example, the court of appeals determined that mitigation measures added to a project after circulation of the final environmental impact statement do not trigger a requirement to supplement the statement, in part because the measures were a specification of general measures spelled out in that statement. In fact, the Third Circuit noted that "the responsiveness to environmental concerns demonstrated by these changes reflects a commendable effort to comply with the spirit of NEPA.'" 702 F.2d at 438 (citations omitted). /18/ This obviously is not a case, like Jantzen (see note 10, supra), where the planned implementation of mitigation measures eliminated the need for an environmental impact statement. The Elk Creek Supplemental Statement provided a discussion of the project's significant environmental effects without mitigation and then identified mitigation measures that, by definition, could be expected to reduce those impacts. See Elk Creek Supplemental Statement 26-27, 38. See also Pet. App. 56a-58a, 61a. The Statement's qualitative predictions that implementation of mitigation measures would reduce, to some degree, the adverse impacts are certainly reasonable. /19/ See, e.g., 40 C.F.R. 1501.2, 1502.5(b) (mandating early application of NEPA); 40 C.F.R. 1502.2(a)-(c) (requiring that environmental impact statements be analytic rather than encyclopedic, that impacts be discussed in proportion to their significance, and that statements be concise and "no longer than absolutely necessary to comply with NEPA and with these regulations"); 40 C.F.R. 1502.2(f) (forbidding agencies from committing resources in a manner that will bias the agency's ultimate decision); 40 C.F.R. 1502.20 (encouraging tiered environmental analysis). /20/ This regulation, which remained unchanged from 1978 to 1985, is reproduced in full in Addendum A, page 2a, infra. /21/ It appears that the regulation grew out of the CEQ's early recognition that federal agencies cannot avoid drafting an environmental impact statement simply because information concerning a potential environmental impact is unknown. See Scientists' Inst. for Pub. Information, Inc. v. AEC, 481 F.2d 1079, 1092 (D.C. Cir. 1973). As the CEQ later explained, the worst case regulation was adopted to ensure that agencies would not "disregard uncertainties as having no weight in the balancing process." 50 Fed. Reg. 32,236 (1985). /22/ As the Earl of Gloucester's son, Edgar, lamented (King Lear, Act IV, Scene 1): Who is't can say, "I am at the worst?" I am worse than e'er I was. * * * * * And worse I may be yet. The worst is not, So long as we can say, "This is the worst." /23/ The regulation is reproduced in Addendum A, pages 1a-2a, infra. /24/ "Reasonably foreseeable impacts" are expressly defined to include "impacts which have catastrophic consequences, even if their probability of occurrence is low, provided that the analysis of the impacts is supported by credible scientific evidence, is not based on pure conjecture, and is within the rule of reason" (40 C.F.R. 1502.22(b)). /25/ The court specifically cited Scientists' Inst. for Pub. Information, Inc. v. AEC, 481 F.2d 1079, 1092 (D.C. Cir. 1973); Carolina Envtl. Study Group v. United States, 510 F.2d 796, 799 (D.C. Cir. 1975); and Alaska v. Andrus, 580 F.2d 465, 473-474 (D.C. Cir. 1978), vacated on other grounds, 439 U.S. 922 (1978). /26/ The CEQ was well aware of the case law concerning the "worst case" requirement when it promulgated the new regulation. The CEQ specifically considered whether those cases mandated worst case analysis and concluded that they did not (51 Fed. Reg. 15,625 (1986)): Comment: Case law required worst case analysis prior to adoption of (the worst case regulation). Response: This assertion is incorrect. Case law prior to the adoption of 40 CFR 1502.22 did require agencies to make a "good faith effort * * * to describe the reasonably foreseeable environmental impact(s)" of the proposal and alternatives to the proposal in the face of incomplete or unavailable information, consistent with the "rule of reason" (citation omitted). The "worst case analysis" requirement was a technique adopted by CEQ as a means of achieving the goals enunciated in such case law. The "worst case" requirement itself, however, was clearly a "major innovation." * * * The U.S. Court of Appeals for the Fifth Circuit, interpreting the "worst case analysis" requirement for the first time in a litigation context, recognized that it was an innovation of CEQ. Sierra Club v. Sigler, 695 F.2d 957, 972 (5th Cir. 1983). CEQ has since observed difficulties with the technique of "worst case analysis" and is replacing it with a better approach to the problem of incomplete or unavailable information. A number of commenters requested CEQ to state that the new regulation is "intended to repudiate and overrule the Ninth Circuit decisions on worst case analysis" (51 Fed. Reg. 15,625 (1986)). The CEQ recognized that its new regulation did not "overrule" judicial interpretations of the worst case requirement; instead, it simply rendered them irrelevant. The agency accordingly responded that "(t)he Ninth Circuit opinions are based on the requirements of former Section 1502.22 or agency reflections thereof, and are inapplicable to this revision" (ibid.). /27/ Even the respondents concede (Br. in Opp. 11) that the CEQ's judgment is entitled to "substantial deference." /28/ The ODFW stated that it is "probably true in many cases" that the project would contribute only a small percentage of the flow to the Rogue, but added that "in other situations such as during flood control operations or during the summer when releasing storage there may be periods when Elk Creek provides a significant percentage of the total flow of the Rogue." Supplemental Statement 68 (Comment 6-5). In response, the Corps explained that it was aware that "(d)uring high flow periods, Elk Creek could contribute up to about 25 percent of the flow in the Rogue River * * *. During other periods, the percentage would be considerably lower." Id. at 70 (Response to Comment 6-5). And, as the Corps further noted, this fact had been fully evaluated in the 1974 Water Quality Report, which had been incorporated into the Supplemental Statement by reference and was available in Corps offices for public review (ibid.). /29/ An environmental impact statement "need not achieve unanimity on the desirability of proceeding with the proposed action" (Life of the Land v. Brinegar, 485 F.2d 460, 473 (9th Cir. 1973), cert. denied, 416 U.S. 961 (1974)). Obviously, even the best scientific information is subject to disagreement among experts. The CEQ envisions that the environmental impact statement will bring the bulk of those disagreements to the attention of the decision-maker through the comment and response section. See 40 C.F.R. 1503.1 et seq. Section 1502.22 is strictly limited, by contrast, to incomplete information "essential to a reasoned choice among alternatives" (40 C.F.R. 1502.22). Indeed, if disagreements like that at issue here were to require resort to Section 1502.22's procedures, the process of preparing environmental impact statements would become a potentially interminable exercise in abstract scientific debate. This was not the CEQ's intention and is not a reasonable result. /30/ Both the Corps and the CEQ regulations implementing NEPA instruct an agency to prepare a supplemental environmental impact statement if the agency determines that the new information has "significant" relevance or bearing on the proposed action or its effects. See 33 C.F.R. 230.11(b) (Corps NEPA regulations); 40 C.F.R. 1502.9(c)(ii) (CEQ regulations). /31/ As this Court observed in Vermont Yankee Nuclear Power Corp.: "Administrative consideration of evidence . . . always creates a gap between the time the record is closed and the time the administrative decision is promulgated (and, we might add, the time the decision is reviewed). . . . If upon the coming down of the order litigants might demand rehearings as a matter of law because some new circumstance has arisen, some new trend has been observed, or some new fact has been discovered, there would be little hope that the administrative process could ever be consummated in an order that would not be subject to reopening." 435 U.S. at 554-555 (quoting ICC v. Jersey City, 322 U.S. 503, 514 (1944)). /32/ As the Seventh Circuit noted (Wisconsin, 745 F.2d at 417 n.5), a number of courts -- including the Ninth Circuit -- do not follow the APA standard in these circumstances; instead, they evaluate the agency's decision whether to prepare or supplement an environmental impact statement based on whether the decision is "reasonable." See, e.g., Pet. App. 8a; Enos v. Marsh, 769 F.2d 1363, 1372 (9th Cir. 1985); Massachusetts v. Watt, 716 F.2d 946 (1st Cir. 1983); Minnesota Public Interest Research Group v. Butz, 498 F.2d 1314, 1320 (8th Cir. 1974) (en banc); Wyoming Outdoor Coordinating Council v. Butz, 484 F.2d 1244, 1248-1249 (10th Cir. 1973). The Fifth Circuit apparently originated that practice in Save Our Ten Acres v. Kreger, 472 F.2d 463 (1973), a case decided prior to Vermont Yankee Nuclear Power Corp. and Kleppe, reasoning that the "spirit of (NEPA) would die aborning" if the APA standard were applied (472 F.2d at 466). One may debate how significant the difference is between the two standards; the fact remains that the APA expressly specifies review under the "arbitrary (or) capricious" standard. /33/ Furthermore, there is no need for a less deferential standard. As the Second Circuit explained (Hanly, 471 F.2d at 829-830): We see no reason for application of a different standard here since the APA standard permits effective judicial scrutiny of agency action and concomitantly permits the agencies to have some leeway in applying the law to factual contexts in which they possess expertise. /34/ That staff biologists' internal memorandum does not represent an official position of ODFW or the State of Oregon. See Addendum, infra, 9a. /35/ Under this theory, increased winter water temperatures would cause the salmon eggs to hatch early -- in February, rather than March or April -- when there is a diminished food supply for young fish and when the "fry" are more susceptible to harm from spring flooding. The Cramer Memorandum estimated (Addendum, infra, 3a) "that an increase of only 1 degree C in river temperature during January would decrease survival of spring chinook fry by 60-80%." /36/ The Memorandum noted (Addendum, infra, 4a) that the "release of cool water from Lost Creek Dam has decreased the incidence of disease among spring chinook, but an unprecedented 76% of the fall chinook in 1979 and 32% in 1980 were estimated to have died before spawning." /37/ The authors of the ODFW Report had themselves drawn into question the accuracy of the model by substantially revising the predicted correlation between temperature increases and fish mortality. The Report predicted that an increase of 1 degree Celsius in the mean maximum January river temperature would decrease the fry survival rate by 60 to 80 percent. See note 35, supra. The authors subsequently concluded that a 1 degree Celsius temperature increase would decrease the fry survival rate by 30 to 45 percent. See Addendum, infra, 9a. /38/ "In most cases, the results indicate that Rogue River temperatures with Elk Creek Dam should be about the same or slightly cooler. The only periods that showed a discernible increase were parts of July, August, and December 1979 which showed a maximum temperature increase of about 0.6 degrees C" (Addendum, infra, 10a). /39/ The Cramer Memorandum's suggestion that Elk Creek reservoir releases would adversely affect angling during a two-month period apparently was based on complaints from fishermen rather than specific scientific data. See Addendum, infra, 4a. The Corps observed that its studies on this matter were inconclusive (id. at 11a). The Corps also disagreed with the Cramer Memorandum's suggestion that flow augmentation would dewater spawning redds and noted, in response to the Memorandum's suggestion that salmon runs had decreased, that the 1985 run "was one of the best on record" (id. at 11a-12a). The Cramer Memorandum made only passing mention of turbidity (id. at 4a). The Corps included in its Elk Creek SIR a discussion of the effect of timber clearing on turbidity, noting that "(a)lthough there appears (sic) to have been higher turbidity values at Elk Creek during the early 1980's as a result of logging, watershed recovery appears to have occurred to reduce turbidity levels back to those of the 1970's when primary studies were conducted" (id. at 12a). The Corps further noted that "turbidity levels will to a large extent be controllable through the operational flexibility provided by the dam's multilevel intake tower" (ibid.). /40/ The first expert, a professor at the University of Washington School of Fisheries, "filed an extensive and highly critical review" of the ODFW Report (Pet. App. 22a). As Judge Wallace explained, the expert found that the Report "contained 'considerable statistical inaccuracies, over-extension of statistical methods, and undue biological speculation'; reported that other conclusions of the study, though apparently valid as far as they went, were not statistically significant; disagreed with some of the study's interpretations of causation; questioned the strength of the study's data and the accuracy of the models it used in reaching its conclusions; dismissed some statements from the report as 'spurious' and others as 'speculation . . . (that was) unfounded'; and accused the authors of the study of throwing out data that did not fit their conclusions" (Pet. App. 22a-23a). See A.R. Doc. No. 112. The second expert, who was less negative in his assessment, stated that the methodology "was 'not likely to yield conclusive results' and that the (ODFW's) model was based on 'limited data over only a few years,' which substantially increased the likelihood of 'chance relationships'" (Pet. App. 23a). See A.R. Doc. No. 114. /41/ The Corps also addressed logging influences on turbidity in the Elk Creek SIR, Addendum, infra, 12a. /42/ The Corps geologist further testified that Corps specialists had conducted a two-week investigation of the proposed reservoir site in 1979 to investigate erosion potentials and potential for long-term turbidity production (Tr. 72-73 (Jan. 16, 1986)). /43/ Indeed, at the hearing, the author of the SCS Report agreed that the "determining factor" for turbidity control at Elk Creek would be "proper management" (Tr. 50 (Jan. 16, 1986) (Borine testimony)). /44/ Cf. Anderson v. City of Bessemer City, 470 U.S. 564, 574-575 (1985) ("Duplication of the trial judge's efforts in the court of appeals would very likely contribute only negligibly to the accuracy of fact determination at a huge cost in diversion of judicial resources."). /45/ On remand, the agency would be obligated to prepare a supplemental statement providing further elaboration on the subject of cumulative impacts. See note 6, supra. APPENDIX