delivered the opinion of the Court.

We granted certiorari to decide two questions of law.1 As framed by petitioners, they are:

“1. Whether the National Environmental Policy Act requires federal agencies to include in each environmental impact statement: (a) a fully developed plan to miti*336gate environmental harm; and (b) a ‘worst case’ analysis of potential environmental harm if relevant information concerning significant environmental effects is unavailable or too costly to obtain.

“2. Whether the Forest Service may issue a special use permit for recreational use of national forest land in the absence of a fully developed plan to mitigate environmental harm.” Pet. for Cert. i.

Concluding that the Court of Appeals for the Ninth Circuit misapplied the National Environmental Policy Act of 1969 (NEPA), 83 Stat. 852, 42 U. S. C. §4321 et seq., and gave inadequate deference to the Forest Service’s interpretation of its own regulations, we reverse and remand for further proceedings.

I

The Forest Service is authorized by statute to manage the national forests for “outdoor recreation, range, timber, watershed, and wildlife and fish purposes.” 74 Stat. 215, 16 U. S. C. §528. See also 90 Stat. 2949, 16 U. S. C. § 1600 et seq. Pursuant to that authorization, the Forest Service has issued “special use” permits for the operation of approximately 170 Alpine and Nordic ski areas on federal lands. See H. R. Rep. No. 99-709, pt. 1, p. 2 (1986).

The Forest Service permit process involves three separate stages. The Forest Service first examines the general environmental and financial feasibility of a proposed project and decides whether to issue a special use permit. See 36 CFR § 251.54(f) (1988). Because that decision is a “major Federal action” within the meaning of NEPA, it must be preceded by the preparation of an Environmental Impact Statement (EIS). 42 U. S. C. § 4332. If the Service decides to issue a permit, it then proceeds to select a developer, formulate the basic terms of the arrangement with the selected party, *337and issue the permit.2 The special use permit does not, however, give the developer the right to begin construction. See 36 CFR § 251.56(c) (1988). In a final stage of review, the Service evaluates the permittee’s “master plan” for development, construction, and operation of the project. Construction may begin only after an additional environmental analysis (although it is not clear that a second EIS need always be prepared) and final approval of the developer’s master plan. This case arises out of the Forest Service’s decision to issue a special use permit authorizing the development of a major destination Alpine ski resort at Sandy Butte in the North Cascade Mountains.

Sandy Butte is a 6,000-foot mountain located in the Okanogan National Forest in Okanogan County, Washington. At present Sandy Butte, like the Methow Valley it overlooks, is an unspoiled, sparsely populated area that the District Court characterized as “pristine.” App. to Pet. for Cert. 20a. In 1968, Congress established the North Cascades National Park and directed the Secretaries of the Interior and Agriculture to agree on the designation of areas within, and adjacent to, the park for public uses, including ski areas. 82 Stat. 926, 930, 16 U. S. C. §§90, 90d-3. A 1970 study conducted by the Forest Service pursuant to this congressional directive identified Sandy Butte as having the highest potential of any site in the State of Washington for development as a major downhill ski resort.3 App. to Pet. for Cert. 23a.

*338In 1978, Methow Recreation, Inc. (MRI), applied for a special use permit to develop and operate its proposed “Early Winters Ski Resort” on Sandy Butte and an 1,165-acre parcel of land it had acquired adjacent to the National Forest. The proposed development would make use of approximately 3,900 acres of Sandy Butte; would entice visitors to travel long distances to stay at the resort for several days at a time; and would stimulate extensive commercial and residential growth in the vicinity to accommodate both vacationers and staff.

In response to MRI’s application, the Forest Service, in cooperation with state and county officials, prepared an EIS known as the Early Winters Alpine Winter Sports Study (Early Winters Study or Study). The stated purpose of the EIS was “to provide the information required to evaluate the potential for skiing at Early Winters” and “to assist in making a decision whether to issue a Special Use Permit for downhill skiing on all or a portion of approximately 3900 acres of National Forest System land.” Early Winters Study 1. A draft of the Study was completed and circulated in 1982, but release of the final EIS was delayed as Congress considered including Sandy Butte in a proposed wilderness area. App. to Pet. for Cert. 26a. When the Washington State Wilderness Act of 1984 was passed, however, Sandy Butte was excluded from the wilderness designation,4 and the EIS was released.

*339The Early Winters Study is a printed document containing almost 150 pages of text and 12 appendices. It evaluated five alternative levels of development of Sandy Butte that might be authorized, the lowest being a “no action” alternative and the highest being development of a. 16-lift ski area able to accommodate 10,500 skiers at one time. The Study considered the effect of each level of development on water resources, soil, wildlife, air quality, vegetation, and visual quality, as well as land use and transportation in the Methow Valley, probable demographic shifts, the economic market for skiing and other summer and winter recreational activities in the Valley, and the energy requirements for the ski area and related developments. The Study’s discussion of possible impacts was not limited to on-site effects, but also, as required by Council on Environmental Quality (CEQ) regulations, see 40 CFR §1502.16(b) (1987), addressed “off-site impacts that each alternative might have on community facilities, socio-economic and other environmental conditions in the Upper Methow Valley.” Early Winters Study 1. As to off-site effects, the Study explained that “due to the uncertainty of where other public and private lands may become developed,” it is difficult to evaluate off-site impacts, id., at 76, and thus the document’s analysis is necessarily “not site-specific,” id., at 1. Finally, the Study outlined certain steps that might be taken to mitigate adverse effects, both on Sandy Butte and in the neighboring Methow Valley, but indicated that these proposed steps are merely conceptual and “will be made more specific as part of the design and implementation stages of the planning process.” Id., at 14.

The effects of the proposed development on air quality and wildlife received particular attention in the Study. In the chapter on “Environmental Consequences,” the first subject discussed is air quality. As is true of other subjects, the discussion included an analysis of cumulative impacts over sev*340eral years resulting from actions on other lands as well as from the development of Sandy Butte itself. The Study concluded that although the construction, maintenance, and operation of the proposed ski area “will not have a measurable effect on existing or future air quality,” the off-site development of private land under all five alternatives — including the “no action” alternative — “will have a significant effect on air quality during severe meteorological inversion periods.” Id., at 65. The burning of wood for space heat, the Study explained, would constitute the primary cause of diminished air quality, and the damage would increase incrementally with each of the successive levels of proposed development. Ibid. The Study cautioned that without efforts to mitigate these effects, even under the “no action” alternative, the increase in automobile, fireplace, and wood stove use would reduce air quality below state standards, but added that “[t]he numerous mitigation measures discussed” in the Study “will greatly reduce the impacts presented by the model.” Id., at 67.

In its discussion of air-quality mitigation measures, the EIS identified actions that could be taken by the county government to mitigate the adverse effects of development, as well as those that the Forest Service itself could implement at the construction stage of the project. The Study suggested that Okanogan County develop an air quality management plan, requiring weatherization of new buildings, limiting the number of wood stoves and fireplaces, and adopting monitoring and enforcement measures.5 In addition, the *341Study suggested that the Forest Service require that the master plan include procedures to control dust and to comply with smoke management practices.6

In its discussion of adverse effects on area wildlife, the EIS concluded that no endangered or threatened species would be affected by the proposed development and that the only impact on sensitive species was the probable loss of a pair of spotted owls and their progeny. Id., at 75. With regard to other wildlife, the Study considered the impact on 75 differ*342ent indigenous species and predicted that within a decade after development vegetational change and increased human activity would lead to a decrease in population for 31 species, while causing an increase in population for another 24 species on Sandy Butte. Ibid. Two species, the pine marten and nesting goshawk, would be eliminated altogether from the area of development. Ibid.

In a comment in response to the draft EIS, the Washington Department of Game voiced a special concern about potential losses to the State’s largest migratory deer herd, which uses the Methow Valley as a critical winter range and as its migration route. Id., at Appendix D (letter of November 18, 1982). The state agency estimated that the total population of mule deer in the area most likely to be affected was “better than 30,000 animals” and that “the ultimate impact on the Methow deer herd could exceed a 50 percent reduction in numbers.” Ibid. The agency asserted that “Okanogan County residents place a great deal of importance on the area’s deer herd.” Ibid. In addition, it explained that hunters had “harvested” 3,247 deer in the Methow Valley area in 1981, and that, since in 1980 hunters on average spent $1,980 for each deer killed in Washington, they had contributed over $6 million to the State’s economy in 1981. Because the deer harvest is apparently proportional to the size of the herd, the state agency predicted that “Washington business can expect to lose over $3 million annually from reduced recreational opportunity.” Ibid. The Forest Service’s own analysis of the impact on the deer herd was more modest. It first concluded that the actual operation of the ski hill would have only a “minor” direct impact on the herd,7 but then recog*343nized that the off-site effect of the development “would noticeably reduce numbers of deer in the Methow [Valley] with any alternative.” Id., at 76. Although its estimate indicated a possible 15 percent decrease in the size of the herd, it summarized the State’s contrary view in the text of the EIS, and stressed that off-site effects are difficult to estimate due to uncertainty concerning private development. Ibid.

As was true of its discussion of air quality, the EIS also described both on-site and off-site mitigation measures. Among possible on-site mitigation possibilities, the Study recommended locating runs, ski lifts, and roads so as to minimize interference with wildlife, restricting access to selected roads during fawning season, and further examination of the effect of the development on mule deer migration routes.8 Off-site options discussed in the Study included the use of zoning and tax incentives to limit development on deer winter range and migration routes, encouragement of conservation easements, and acquisition and management by local *344government of critical tracts of land.9 As with the measures suggested for mitigating the off-site effects on air quality, the proposed options were primarily directed to steps that might be taken by state and local government.

Ultimately, the Early Winters Study recommended the issuance of a permit for development at the second highest level considered — a 16-lift ski area able to accommodate 8,200 skiers at one time. On July 5, 1984, the Regional Forester decided to issue a special use permit as recommended by the *345Study.10 App. to Pet. for Cert. 63a. In his decision, the Regional Forester found that no major adverse effects would result directly from the federal action, but that secondary effects could include a degradation of existing air quality and a reduction of mule deer winter range. Id., at 67a. He therefore directed the supervisor of the Okanogan National Forest, both independently and in cooperation with local officials, to identify and implement certain mitigating measures. Id., at 67a-70a.

Four organizations (respondents)11 opposing the decision to issue a permit appealed the Regional Forester’s decision to the Chief of the Forest Service. See 36 CFR § 211.18 (1988). After a hearing, he affirmed the Regional Forester’s decision. Stressing that the decision, which simply approved the general concept of issuing a 30-year special use permit for development of Sandy Butte, did not authorize construction of a particular ski area and, in fact, did not even act on MRI’s specific permit application, he concluded that the EIS’ discussion of mitigation was “adequate for this stage in the review process.” App. to Pet. for Cert. 59a.

Thereafter, respondents brought this action under the Administrative Procedure Act, 5 U. S. C. §§701-706, to obtain judicial review of the Forest Service’s decision. Their principal claim was that the Early Winters Study did not satisfy *346the requirements of NEPA, 42 U. S. C. §4332.12 With the consent of the parties, the case was assigned to a United States Magistrate. See 28 U. S. C. § 636(c). After a trial, the Magistrate filed a comprehensive written opinion and concluded that the EIS was adequate. App. to Pet. for Cert. 20a. Specifically, he found that the EIS had adequately disclosed the adverse impacts on the mule deer herd and on air quality and that there was no duty to prepare a “worst case analysis” because the relevant information essential to a reasoned decision was available. Id., at 39a-44a. In concluding that the discussion of off-site, or secondary, impacts was adequate, the Magistrate stressed that courts apply a “rule of reason” in evaluating the adequacy of an EIS and “take the uncertainty and speculation involved with secondary impacts into account in passing on the adequacy of the discussion of secondary impacts.” Id,., at 38a. On the subject of mitigation, he explained that “[m]ere listing ... is generally inadequate to satisfy the CEQ regulations,” but found that “in this EIS there is more — not much more — but more than a mere listing of mitigation measures.” Id., at 41a. Moreover, emphasizing the tiered nature of the Forest Service’s decisional process, the Magistrate noted that additional mitigation strategies would be included in the master plan, that the Forest Service continues to develop mitigation plans as further information becomes available, and that the Regional Forester’s decision conditioned issuance of the special use permit on execution of an agreement between the Forest Service, the State of Washington, and Okanogan County concerning mitigation. Id., at 41a-42a, 45a.

Concluding that the Early Winters Study was inadequate as a matter of law, the Court of Appeals reversed. Methow Valley Citizens Council v. Regional Forester, 833 F. 2d 810 *347(CA9 1987). The court held that the Forest Service could not rely on “ ‘the implementation of mitigation measures’ ” to support its conclusion that the impact on the mule deer would be minor, “since not only has the effectiveness of these mitigation measures not yet been assessed, but the mitigation measures themselves have yet to be developed.” Id., at 817. It then added that if the agency had difficulty obtaining adequate information to make a reasoned assessment of the environmental impact on the herd, it had a duty to make a so-called “worst case analysis.” Such an analysis is “‘formulated on the basis of available information, using reasonable projections of the worst possible consequences of a proposed action.’ Save our Ecosystems, 747 F. 2d, at 1244-45 (quoting 46 Fed. Reg. 18032 (1981)).” Ibid.

The court found a similar defect in the EIS’ treatment of air quality. Since the EIS made it clear that commercial development in the Methow Valley will result in violations of state air-quality standards unless effective mitigation measures are put in place by the local governments and the private developer, the Court of Appeals concluded that the Forest Service had an affirmative duty to “develop the necessary mitigation measures before the permit is granted.” Id., at 819 (emphasis in original) (footnote omitted). The court held that this duty was imposed by both the Forest Service’s own regulations and § 102 of NEPA. Ibid. It read the statute as imposing a substantive requirement that “ ‘action be taken to mitigate the adverse effects of major federal actions.’” Ibid. (quoting Stop H-3 Assn. v. Brinegar, 389 F. Supp. 1102, 1111 (Haw. 1974), rev’d on other grounds, 533 F. 2d 434 (CA9), cert. denied, 429 U. S. 999 (1976)). For this reason, it concluded that “an EIS must include a thorough discussion of measures to mitigate the adverse environmental impacts of a proposed action.” 833 F. 2d, at 819. The Court of Appeals concluded by quoting this paragraph from an opinion it had just announced:

“‘The importance of the mitigation plan cannot be overestimated. It is a determinative factor in evaluat*348ing 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.’” Id., at 820 (quoting Oregon Natural Resources Council v. Marsh, 832 F. 2d 1489, 1493 (CA9 1987), rev’d, post, p. 360).

II

Section 101 of NEPA declares a broad national commitment to protecting and promoting environmental quality. 83 Stat. 852, 42 U. S. C. §4331. To ensure that this commitment is “infused into the ongoing programs and actions of the Federal Government, the act also establishes some important ‘action-forcing’ procedures.” 115 Cong. Rec. 40416 (remarks of Sen. Jackson). See also S. Rep. No. 91-296, p. 19 (1969); Andrus v. Sierra Club, 442 U. S. 347, 350 (1979); Kleppe v. Sierra Club, 427 U. S. 390, 409, and n. 18 (1976). Section 102 thus, among other measures

“directs that, to the fullest extent possible ... 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 *349action should it be implemented.” 83 Stat. 853, 42 U. S. C. §4332.

The statutory requirement that a federal agency contemplating a major action prepare such an environmental impact statement serves NEPA’s “action-forcing” purpose in two important respects. See Baltimore Gas & Electric Co. v. Natural Resources Defense Council, Inc., 462 U. S. 87, 97 (1983); Weinberger v. Catholic Action of Hawaii/Peace Education Project, 454 U. S. 139, 143 (1981). It ensures that the agency, in reaching its decision, will have available, and will carefully consider, detailed information concerning significant environmental impacts; it also guarantees that the relevant information will be made available to the larger audience that may also play a role in both the decisionmaking process and the implementation of that decision.

Simply by focusing the agency’s attention on the environmental consequences of a proposed project, NEPA ensures that important effects will not be overlooked or underestimated only to be discovered after resources have been committed or the die otherwise cast. See ibid.; Kleppe, supra, at 409. Moreover, the strong precatory language of § 101 of the Act and the requirement that agencies prepare detailed impact statements inevitably bring pressure to bear on agencies “to respond to the needs of environmental quality.” 115 Cong. Rec. 40425 (1969) (remarks of Sen. Muskie).

Publication of an EIS, both in draft and final form, also serves a larger informational role. It gives the public the assurance that the agency “has indeed considered environmental concerns in its decisionmaking process,” Baltimore Gas & Electric Co., supra, at 97, and, perhaps more significantly, provides a springboard for public comment, see L. Caldwell, Science and the National Environmental Policy Act 72 (1982). Thus, in this case the final draft of the Early Winters Study reflects not only the work of the Forest Service itself, but also the critical views of the Washington State Department of Game, the Methow Valley Citizens Council, and *350Friends of the Earth, as well as many others, to whom copies of the draft Study were circulated.13 See Early Winters Study, Appendix D. Moreover, with respect to a development such as Sandy Butte, where the adverse effects on air quality and the mule deer herd are primarily attributable to predicted off-site development that will be subject to regulation by other governmental bodies, the EIS serves the function of offering those bodies adequate notice of the expected consequences and the opportunity to plan and implement corrective measures in a timely manner.

The sweeping policy goals announced in § 101 of NEPA are thus realized through a set of “action-forcing” procedures that require that agencies take a “‘hard look’ at environmental consequences,” Kleppe, 427 U. S., at 410, n. 21 (citation omitted), and that provide for broad dissemination of relevant environmental information. Although these procedures are almost certain to affect the agency’s substantive decision, it is now well settled that NEPA itself does not mandate particular results, but simply prescribes the necessary process. See Strycker’s Bay Neighborhood Council, Inc. v. Karlen, 444 U. S. 223, 227-228 (1980) (per curiam); Vermont Yankee Nuclear Power Corp. v. Natural Resources Defense Council, Inc., 435 U. S. 519, 558 (1978). If the adverse environmental effects of the proposed action are adequately identified and evaluated, the agency is not constrained by NEPA from deciding that other values outweigh the environmental costs. See ibid.; Strycker’s Bay Neighborhood Council, Inc., supra, at 227-228; Kleppe, supra, at 410, n. 21. In this *351case, for example, it would not have violated NEPA if the Forest Service, after complying with the Act’s procedural prerequisites, had decided that the benefits to be derived from downhill skiing at Sandy Butte justified the issuance of a special use permit, notwithstanding the loss of 15 percent, 50 percent, or even 100 percent of the mule deer herd. Other statutes may impose substantive environmental obligations on federal agencies,14 but NEPA merely prohibits uninformed — rather than unwise — agency action.

To be sure, one important ingredient of an EIS is the discussion of steps that can be taken to mitigate adverse environmental consequences.15 The requirement that an EIS contain a detailed discussion of possible mitigation measures flows both from the language of the Act and, more expressly, from CEQ’s implementing regulations. Implicit in NEPA’s demand that an agency prepare a detailed statement on “any adverse environmental effects which cannot be avoided should the proposal be implemented,” 42 U. S. C. *352§ 4332(C)(ii), is an understanding that the EIS will discuss the extent to which adverse effects can be avoided. See D. Mandelker, NEPA Law and Litigation § 10:38 (1984). More generally, omission of a reasonably complete discussion of possible mitigation measures would undermine the “action-forcing” function of NEPA. Without such a discussion, neither the agency nor other interested groups and individuals can properly evaluate the severity of the adverse effects. An adverse effect that can be fully remedied by, for example, an inconsequential public expenditure is certainly not as serious as a similar effect that can only be modestly ameliorated through the commitment of vast public and private resources. Recognizing the importance of such a discussion in guaranteeing that the agency has taken a “hard look” at the environmental consequences of proposed federal action, CEQ regulations require that the agency discuss possible mitigation measures in defining the scope of the EIS, 40 CFR § 1508.25(b) (1987), in discussing alternatives to the proposed action, § 1502.14(f), and consequences of that action, § 1502.16(h), and in explaining its ultimate decision, § 1505.2(c).

There is a fundamental distinction, however, between a requirement that mitigation be discussed in sufficient detail to ensure that environmental consequences have been fairly evaluated, on the one hand, and a substantive requirement that a complete mitigation plan be actually formulated and adopted, on the other. In this case, the off-site effects on air quality and on the mule deer herd cannot be mitigated unless nonfederal government agencies take appropriate action. Since it is those state and local governmental bodies that have jurisdiction over the area in which the adverse effects need be addressed and since they have the authority to mitigate them, it would be incongruous to conclude that the Forest Service has no power to act until the local agencies have reached a final conclusion on what mitigating measures *353they consider necessary.16 Even more significantly, it would be inconsistent with NEPA’s reliance on procedural mechanisms — as opposed to substantive, result-based standards — to demand the presence of a fully developed plan that will mitigate environmental harm before an agency can act. Cf. Baltimore Gas & Electric Co., 462 U. S., at 100 (“NEPA does not require agencies to adopt any particular internal decisionmaking structure”).

We thus conclude that the Court of Appeals erred, first, in assuming that “NEPA requires that ‘action be taken to mitigate the adverse effects of major federal actions,”’ 833 F. 2d, at 819 (quoting Stop H-3 Assn. v. Brinegar, 389 F. Supp., at 1111), and, second, in finding that this substantive requirement entails the further duty to include in every EIS “a detailed explanation of specific measures which will be employed to mitigate the adverse impacts of a proposed action,” 833 F. 2d, at 819 (emphasis supplied).

*354Ill

The Court of Appeals also concluded that the Forest Service had an obligation to make a “worst case analysis” if it could not make a reasoned assessment of the impact of the Early Winters project on the mule deer herd. Such a “worst case analysis” was required at one time by CEQ regulations, but those regulations have since been amended. Moreover, although the prior regulations may well have expressed a permissible application of NEPA, the Act itself does not mandate that uncertainty in predicting environmental harms be addressed exclusively in this manner. Accordingly, we conclude that the Court of Appeals also erred in requiring the “worst case” study.

In 1977, President Carter directed that CEQ promulgate binding regulations implementing the procedural provisions of NEPA. Exec. Order No. 11991, 3 CFR 123 (1977 Comp.). Pursuant to this Presidential order, CEQ promulgated implementing regulations. Under § 1502.22 of these regulations — a provision which became known as the “worst case requirement” — CEQ provided that if certain information relevant to the agency’s evaluation of the proposed action is either unavailable or too costly to obtain, the agency must include in the EIS a “worst case analysis and an indication of the probability or improbability of its occurrence.” 40 CFR § 1502.22 (1985). In 1986, however, CEQ replaced the “worst case” requirement with a requirement that federal agencies, in the face of unavailable information concerning a reasonably foreseeable significant environmental consequence, prepare “a summary of existing credible scientific evidence which is relevant to evaluating the . . . adverse impacts” and prepare an “evaluation of such impacts based upon theoretical approaches or research methods generally accepted in the scientific community.” 40 CFR § 1502.22(b) (1987). The amended regulation thus “retains the duty to describe the consequences of a remote, but potentially severe impact, but grounds the duty in evaluation of scientific opinion rather *355than in the framework of a conjectural ‘worst case analysis.’” 50 Fed. Reg. 32237 (1985).

The Court of Appeals recognized that the “worst case analysis” regulation has been superseded, yet held that “[t]his rescission . . . does not nullify the requirement . . . since the regulation was merely a codification of prior NEPA case law.” 833 F. 2d, at 817, n. 11. This conclusion, however, is erroneous in a number of respects. Most notably, review of NEPA case law reveals that the regulation, in fact, was not a codification of prior judicial decisions. See Note, 86 Mich. L. Rev. 777, 798, 800-802, 813-814 (1988). The cases cited by the Court of Appeals ultimately rely on the Fifth Circuit’s decision in Sierra Club v. Sigler, 695 F. 2d 957 (1983). Sigler, however, simply recognized that the “worst case analysis” regulation codified the “judicially created principl[e]” that an EIS must “consider the probabilities of the occurrence of any environmental effects it discusses.” Id., at 970-971. As CEQ recognized at the time it superseded the regulation, case law prior to the adoption of the “worst case analysis” provision did require agencies to describe environmental impacts even in the face of substantial uncertainty, but did not require that this obligation necessarily be met through the mechanism of a “worst case analysis.” See 51 Fed. Reg. 15625 (1986). CEQ’s abandonment of the “worst case analysis” provision, therefore, is not inconsistent with any previously established judicial interpretation of the statute.

Nor are we convinced that the new CEQ regulation is not controlling simply because it was preceded by a rule that was in some respects more demanding. In Andrus v. Sierra Club, 442 U. S., at 358, we held that CEQ regulations are entitled to substantial deference. In that case we recognized that although less deference may be in order in some cases in which the “ ‘administrative guidelines’ ” conflict “ ‘with earlier pronouncements of the agency,”’ ibid, (quoting General Electric Co. v. Gilbert, 429 U. S. 125, 143 (1976)), substantial deference is nonetheless appropriate if there appears to have *356been good reason for the change, 442 U. S., at 358. Here, the amendment only came after the prior regulation had been subjected to considerable criticism.17 Moreover, the amendment was designed to better serve the twin functions of an EIS — requiring agencies to take a “hard look” at the consequences of the proposed action and providing important information to other groups and individuals. CEQ explained that by requiring that an EIS focus on reasonably foreseeable impacts, the new regulation “will generate information and discussion on those consequences of greatest concern to the public and of greatest relevance to the agency’s decision,” 50 Fed. Reg. 32237 (1985), rather than distorting the decisionmaking process by overemphasizing highly speculative harms, 51 Fed. Reg. 15624-15625 (1986); 50 Fed. Reg. 32236 (1985). In light of this well-considered basis for the change, the new regulation is entitled to substantial deference. Accordingly, the Court of Appeals erred in concluding that the Early Winters Study is inadequate because it failed to include a “worst case analysis.”18

*357IV

The Court of Appeals also held that the Forest Service’s failure to develop a complete mitigation plan violated the agency’s own regulations. 833 F. 2d, at 814, n. 3, 819, and n. 14. Those regulations require that an application for a special use permit include “measures and plans for the protection and rehabilitation of the environment during construction, operation, maintenance, and termination of the project,” 36 CFR § 251.54(e)(4) (1988), and that “[e]ach special use authorization . . . contain . . . [t]erms and conditions which will. . . minimize damage to scenic and esthetic values and fish and wildlife habitat and otherwise protect the environment,” § 251.56(a)(1)(h). Applying those regulations, the Court of Appeals concluded that “[sjince the mitigation ‘plan’ here at issue is so vague and undeveloped as to be wholly inadequate, . . . the Regional Forester’s decision to grant the special use permit could be none other than arbitrary, capricious and an abuse of discretion.” 833 F. 2d, at 814, n. 3. We disagree.

The Early Winters Study made clear that on-site effects of the development will be minimal and will be easily mitigated. For example, the Study reported that “[i]mpacts from construction, maintenance and operation of the proposed ‘hill’ development on National Forest land will not have a measurable effect on existing or future air quality,” Early Winters Study 65, and that “[t]he effect development and operation of the ski hill would have on deer migration should be minor,” id., at 76. Given the limited on-site effects of the proposed *358development, the recommended ameliorative steps — which, for example, called for “prompt re vegetation of all disturbed areas,” id., at 69, and suggested locating “new service roads away from water resources and fawning cover,” id., at 16— cannot be deemed overly vague or underdeveloped.

The Court of Appeals’ conclusion that the Early Winters Study’s treatment of possible mitigation measures is inadequate apparently turns on the court’s review of the proposed off-site measures. Although NEPA and CEQ regulations require detailed analysis of both on-site and off-site mitigation measures, see, e. g., 40 CFR § 1502.16(b) (1987), there is no basis for concluding that the Forest Service’s own regulations must also be read in all cases to condition issuance of a special use permit on consideration (and implementation) of off-site mitigation measures. The Forest Service regulations were promulgated pursuant to a broad grant of authority “to permit the use and occupancy of suitable areas of land within the national forests . . . for the purpose of constructing or maintaining hotels, resorts, and any other structures or facilities necessary or desirable for recreation, public convenience, or safety,” 16 U. S. C. §497, and were not based on the more direct congressional concern for environmental quality embodied in NEPA.19 See H. R. Rep. No. 99-709, pt. 1, p. 2 (1986). As is clear from the text of the permit issued to MRI, the Forest Service has decided to implement its mitigation regulations by imposing appropriate controls over MRI’s actual development and operation during the term of the permit.20 It was surely not unreasonable for the Forest *359Service in this case to have construed those regulations as not extending to actions that might be taken by Okanogan County or the State of Washington to ameliorate the off-site effects of the Early Winters project on air quality and the mule deer herd. This interpretation of the agency’s own regulation is not “plainly erroneous or inconsistent with the regulation,” and is thus controlling. Bowles v. Seminole Rock & Sand Co., 325 U. S. 410, 414 (1945). See also Lyng v. Payne, 476 U. S. 926, 939 (1986); Udall v. Tallman, 380 U. S. 1, 16-17 (1965).

V

In sum, we conclude that NEPA does not require a fully developed plan detailing what steps will be taken to mitigate adverse environmental impacts and does not require a “worst case analysis.” In addition, we hold that the Forest Service has adopted a permissible interpretation of its own regulations. The judgment of the Court of Appeals is accordingly reversed, and the case is remanded for further proceedings consistent with this opinion.

It is so ordered.

concurring.

I write separately to highlight the Court’s observation that “one important ingredient of an EIS is the discussion of steps that can be taken to mitigate adverse environmental consequences.” Ante, at 351.