concurring in part and dissenting in part:

I concur in parts I, II and III of Judge Wallace’s thoughtful and well-crafted opinion. I respectfully dissent from part IV, however, because I believe that a “worst case” analysis of a major oil spill is necessary at the lease sale stage under NEPA and its relevant implementing regulation, 40 C.F.R. § 1502.22 (1982).

The prime purpose of NEPA in requiring Environmental Impact Statements is to assure that federal decision-makers consider the environmental consequences of their major actions before the decision to act is made. See Kleppe v. Sierra Club, 427 U.S. 390, 409 , 96 S.Ct. 2718, 2729 , 49 L.Ed.2d 576 (1976); Conference Report on NEPA, . 115 Cong.Ree. 40416 (1969). Where some of the consequences are unknown, as they unquestionably are here, and are important to the decision, 1 the Council on Environmental Quality has required that the worst possible consequences be assessed. 40 C.F.R. § 1502.22 (1982). The regulation is thus designed to assure what common sense would in any event dictate: that a decision-maker be given the opportunity to decide against taking action when the benefits to be gained, although substantial, are outweighed by the risk, although small, of a truly catastrophic environmental impact. The weighing and balancing of gains against risks is, of course, the province of the decision-maker. But the decision-maker must be informed of the extent of a possible catastrophe, a worst case, at a time when he or she is free to make an unfettered decision to refrain from an action because the slight risk of immense harm overshadows the potential benefits. I am satisfied that in the present case, that moment occurs no later than the lease sale stage, before sale and execution of any leases.

Prior to sale, the Secretary has absolute discretion to decline to lease an OCS tract. See 43 U.S.C. § 1344 (a). He can therefore decline to lease on the ground that exploration or development will run a small but real risk of immense environmental harm. Once the Secretary leases a tract, however, he loses that freedom, and consequently commits himself to incur such a risk. The reasons why the Secretary loses his freedom upon sale of the leases are both legal and practical.

As a legal matter, the Secretary is allowed to cancel an existing lease for envi *618 ronmental reasons only if he determines that:

(i) continued activity ... would probably cause serious harm or damage to ... [the] environment; (ii) the threat of harm or damage will not disappear or decrease to an acceptable extent within a reasonable period of time; and (iii) the advantages of cancellation outweigh the advantages of continuing such lease or permit in force.

43 U.S.C. § 1334 (a)(2)(A) (emphasis added); 43 U.S.C. § 1351 (h)(1)(D). The requirement of a determination that continued activity “would probably cause serious harm to the environment” is a forceful restriction on the Secretary’s authority. At least under ordinary circumstances, it prohibits cancellation because of the possibility of a major oil spill; as we observed in Southern Oregon Citizens Against Toxic Sprays v. Clark, 720 F.2d 1475 , 1479 (9th Cir.1983), a major oil spill is not a probable occurrence, but rather is “an event of low probability but catastrophic effects.” The effect of the statute, therefore, is that sale of the leases ends the Secretary’s right to call a total halt to exploration and development out of concern over remote environmental catastrophes. 2

The majority opinion views the restrictions on the Secretary’s discretion after leasing as only “minor alterations,” because the statute is not exclusive and the Secretary may by regulation expand his power to suspend or cancel leases. I cannot agree. It is true that the Secretary’s power to suspend operations remains broad, but we have held that the power to suspend is exceeded when the suspension is so open-ended as to amount to a cancellation of a lease. Union Oil Co. of California v. Morton, 512 F.2d 743, 750-51 (9th Cir.1975). Suspension is therefore a temporary remedy and, being temporary, cannot eliminate the possibility of a major oil spill. Only cancellation can do that.

Perhaps the majority opinion is correct in stating that the Secretary by regulation could expand his powers of cancellation, but the proposition is by no means self-evident. In Union Oil Co. of California, 512 F.2d at 750 , we held that the Secretary’s statutory authority to “prescribe ... such ... regulations as he determines to be necessary and proper ... for the conservation of the natural resources of the outer Continental Shelf,” 43 U.S.C. § 1334 (a), did not authorize him to issue a regulation effectively cancelling a lease. In 1978, three years after the decision in Union Oil, Congress amended OCSLA to expand the power of the Secretary to cancel a lease or disapprove exploration or development plans, but the House Report stated that “the Secretary is given authority to disapprove a plan, but only for [the] specified reasons." H.R.Rep. No. 95-590, 95th Cong., 2d Sess. 168 (emphasis added), reprinted in 1978 U.S.Code Cong. & Ad. News 1450, 1574. In any event, the Secretary has not regulated to expand his powers of cancellation; the present regulation tracks the language of 43 U.S.C. § 1334 (a)(2)(A) and permits cancellation only when continued activity under the lease “would probably cause serious harm or damage ... to the ... environment.” 30 C.F.R. § 250.12 (d)(4)(i). The identical language appears in the statute and regulation requiring the Secretary to disapprove development plans because of exceptional environmental circumstances. 43 U.S.C. § 1351 (h)(1)(D); 30 C.F.R. § 250.34-2 (g)(2)(iii)(C) (1982). Our decision should be based on the constraints in existing regulations, which now bind the Secretary. See California v. Block, 690 F.2d 753, 762-63 (9th Cir.1982) (EIS required at first stage of multi-stage project where regulation commits agency to action at later stage).

*619 Even if the majority is correct in concluding that the Secretary is not legally committed upon the sale of leases to a program of exploration and development, the Secretary is committed “as a practical matter.” See California v. Block, 690 F.2d at 761 . Once the leases are sold, immense amounts of money change hands, expensive exploration projects are undertaken, and the Department of Interior and various state agencies plan for the consequences of the lease program. As the First Circuit has stated, “[ejach of these events represents a link in a chain of bureaucratic commitment that will become progressively harder to undo the longer it continues.” Massachusetts v. Watt, 716 F.2d 946, 952 (1st Cir. 1983). Moreover, cancellation of a lease may require the payment of very substantial compensation to the lessee. 43 U.S.C. § 1334 (a)(2)(C). That possibility is bound to be a significant deterrent to cancellation. The combined impact of all of these factors renders the lease sale a practical commitment by the Secretary to a program of exploration and, if oil or gas is discovered, to production. It is therefore at the lease sale stage that the Secretary 3 needs to know the worst environmental consequences that may result from that program.

My conclusion that the worst case analysis of a major oil spill must be considered at the lease sale stage is not inconsistent with the phased nature of OCS development. The possibility of a major oil spill cannot be eliminated merely by later-stage regulation of exploration and development; it can only be eliminated by a total refusal to permit exploration and development. The Secretary’s power to refuse, and thus to avoid the risk of an oil spill, is curtailed after leases are sold. It is therefore “important” to study the worst case effects of a major spill at the lease sale stage. 4

My conclusion is also unaffected by Secretary of the Interior v. California, — U.S. -, 104 S.Ct. 656 , 78 L.Ed.2d 496 (1984). That case held that nothing that occurs at the OCSLA lease sale stage “directly affects” the coastal zone so as to require a review to determine whether the leases are consistent with the state’s management plan adopted pursuant to the Coastal Zone Management Act (CZMA), 16 U.S.C. §§ 1451-64 (1982). Two points distinguish that case from ours. First, as the Supreme Court in Secretary v. California repeatedly emphasized, a state’s power under CZMA is just as great, and as fully exercisable, at the exploration and production stages of offshore development as it is at the lease sale stage. E.g., — U.S. at -, 104 S.Ct. at 671 (“OCSLA expressly provides for federal disapproval of [an exploration] plan that is not consistent with an applicable state management plan____ 43 U.S.C. § 1340 (c)(2).”); id. (“The State can veto [the development and production] plan as ‘inconsistent’ [with its coastal management program]____ 43 U.S.C. § 1351 (d).”). Therefore, entering a lease has little import from a CZMA perspective. It has great import from a NEPA perspective, as I have explained. Second, the Supreme Court’s opinion in Secretary v. California rested in large part on “the lengthy, detailed, and coordinated provisions of CZMA § 307(c)(3)(B), and OCSLA §§ 1344-1346 and 1351.” — U.S. at-, 104 S.Ct. at 672 . The detailed coordination was a principal support for the court’s conclusion that Congress intended to postpone consistency review to the later stages specified by the cross-referenced statutory provisions. Id. There is no such comprehensive cross-referencing between NEPA and OCSLA to suggest an intent by Congress to confine NEPA review to late stages. Indeed, the parties here concede NEPA’s application to lease sales, and the Supreme Court in Sec *620 retary v. California introduced its description of lease sales with the statement that “[Requirements of the National Environmental Protection Act and the Endangered Species Act must be met first.” — U.S. at-, 104 S.Ct. at 670 .

I would therefore hold the unknown consequences of a major oil spill to be “important” to the lease sale decision within the meaning of 40 C.F.R. § 1502.22 (b) (1982), and would require the EIS to include a worst case analysis of its consequences. Once the leases are sold, the risk of such a spill has been taken.

. Like the majority, I view the determinative issue to be whether the information relevant to adverse impacts is "important to the decision” within the meaning of 40 C.F.R. § 1502.22 (b) because that is the way all parties have framed this appeal. Unlike the majority, however, I would adhere to the ruling of Save Our Ecosystems v. Clark, Nos. 83-3908, 3918, 3887 & 3916, (9th Cir. Jan. 27, 1984), that no rational distinction can be made, and none was intended, between the standards of "important to the decision" and "essential to the decision” when the distinction purports to be based on the reason why information is unavailable. Save our Ecosystems blended both standards into one of significance: "If significant information cannot be produced because the costs are exorbitant or the methods beyond the state of the art, a [worst case analysis] must be prepared.” Id.

. The First Circuit has held that the 1978 Amendments to OCSLA did not restrict the Secretary’s powers under the ESA. Conservation Law Foundation of New England, Inc. v. Andrus, 623 F.2d 712, 714-15 (1st Cir.1979). That holding may expand the Secretary’s powers to cancel exploration and development out of concern over the potential impacts on endangered species; the holding, however, does not increase the Secretary’s powers to protect other environmental resources from remote but catastrophic risks.

. The public also needs to know, for informed public participation, like informed decision-making, is a purpose of NEPA. See Save Lake Washington v. Frank, 641 F.2d 1330, 1334 (9th Cir.1981).

. There are many environmental risks other than those of an oil spill that can be eliminated through regulation at the exploration or production stage. I agree with the majority that the phased nature of OCS development makes it unnecessary to study those risks in a worst case analysis at the leasing stage.