New England Coalition on Nuclear Pollution v. United States Nuclear Regulatory Commission, Public Service Company of New Hampshire, Intervenors

COFFIN, Chief Judge. This latest episode in our continuing review of agency decisions concerning the proposed nuclear power plant at Seabrook, New Hampshire, is before us on four petitions for review of decisions of the Nuclear Regulatory Commission (NRC or Commission). Each petition has been brought by the New England Coalition on Nuclear Pollution (NECNP). We have granted motions to intervene made by the Commonwealth of Massachusetts and the Seacoast Anti-Pollution League and Audubon Society of New Hampshire (SAPL). NECNP and Massachusetts briefed the same issues (and both are included in “petitioner” as used herein), and SAPL briefed a separate set. The petitions name the NRC as respondent. Public Service Company of New Hampshire (PSCO) and New Hampshire Voice of Energy intervened, by permission of the court, on respondent’s side. PSCO, as lead applicant for a consortium of several New England utilities, applied in March, 1973, for a permit to build a two unit nuclear generating station at Sea-brook, New Hampshire. Evidentiary hearings were held in 1975 and 1976 before the NRC’s Atomic Safety and Licensing Board (Licensing Board). The Licensing Board issued its initial decision granting construction permits by a divided vote in June, 1976. Public Service Co. of New Hampshire, 3 N.R.C. 857 (A.S.L.B. 1976). In 1977, after the permits had been suspended for a period, 5 N.R.C. 39 (A.L.A.B. 1977), due to decisions made by the Environmental Protection Agency (EPA), 1 the Atomic Safety and Licensing Appeal Board (Appeal Board) affirmed the issuance of the permits. 6 N.R.C. 33 (1977). The full Commission affirmed the Appeal Board in January of 1978. 7 N.R.C.-(1978). Other aspects of the case’s procedural history will be developed as the issues concerning them arise. Petitioners assert that the NRC failed to carry out its responsibilities as to a variety of issues under the Atomic Energy Act, 42 U.S.C. §§ 2011 et seq., the National Environmental Policy Act (NEPA), 42 U.S.C. §§ 4321 et seq., and the Administrative Procedure Act (APA), 5 U.S.C. §§ 551 et seq. We will discuss each issue and its factual background separately, organized according to the Act under which it arises. THE ATOMIC ENERGY ACT 1. The NRC Siting Regulations Under the Atomic Energy Act “no license may be issued ... if, in the opinion of the Commission, the issuance of a license . would be inimical to the common defense and security or to the health and safety of the public.” 42 U.S.C. § 2133(d). The Act authorizes the Commission to “prescribe such regulations or orders as it may deem necessary ... to govern any activity authorized pursuant to this Act, including standards and restrictions governing the design, location, and operation of facilities used in the conduct of such activi *91 ty, in order to protect health and to minimize danger to life or property.” 42 U.S.C. § 2201(i)(8). The regulations adopted, codified at 10 C.F.R. part 100 (Reactor Site Criteria), attempt to accomplish the statutory mandate, inter alia, 2 through criteria designed to assure a safe separation between the reactor and surrounding population even in the event of a hypothetical “major accident . that would result in potential hazards not exceeded by those from any accident considered credible.” 10 C.F.R. § 100.-11(a) n.l. Immediately around the reactor the applicant must control an “exclusion area” “in which the reactor licensee has the authority to determine all activities including exclusion or removal of personnel and property. . . . ” 10 C.F.R. § 100.3(a). This area must be large enough so that someone at the perimeter of the area for the first two hours of an accident would not receive more than the “once in a lifetime . dose for radiation workers which . may be disregarded in the determination of their radiation exposure status.” 10 C.F.R. § 100.11(a)(1) & n.2. Immediately surrounding the circular exclusion area there must be a concentric band, labelled a “low population zone”, with a sufficiently small number of residents so “that there is a reasonable probability that appropriate protective measures could be taken in their behalf in the event of a serious accident.” 10 C.F.R. § 100.3(b). This band must be wide enough so that an individual at its perimeter during the entire period of the accident (assumed to be roughly 30 days) would not receive more than the once in a lifetime radiation dosage. 10 C.F.R. § 100.11(a)(2). Finally, the regulations require that the “Population Center Distance”, “the distance from the reactor to the nearest boundary of a densely populated center containing more than about 25,-000 residents”, 10 C.F.R. § 100.3(c), be “at ¡east one and one-third times the distance from the reactor to the outer boundary of the low population zone.” 10 C.F.R. § 100.-11(a)(3). The dispute in this case turns on the proper determination of the population center distance. The Licensing Board found that Portsmouth, New Hampshire, 12 miles away, was the nearest “densely populated center containing more than about 25,000 residents”. Since the low population zone radius was 1.5 miles, the Board found that the plant complied with the regulations. NECNP argued to the Appeal Board that the nearby towns of Seabrook, Hampton Falls, and Hampton 3 and the Hampton/Seabrook beach with its large summer population should be considered as one single population center. The Appeal Board rejected this argument as to the towns because it decided that the purpose of the population center distance was “to insure that the cumulative exposure dose to the population as a whole is kept within bounds in the event of a postulated major accident.” Given this interpretation, drawn from language in 10 C.F.R. § 100.11(a)(3) and from the requirement of a “center” and a straight line distance rather than the circular areas focused on for the exclusion area and low population zone, the Appeal Board determined that the semi-circle of towns 4 was so spread out that the large population mentioned in the regulations would not be threatened to the required extent. This is so because the regulations, concerned about hazards produced by the worst credible accident, assume that the radiation released in an accident would, because of constant wind direction, all be blown over the densely populated center, whereas the whole population described above would be threatened only if the wind were to shift. In that event any particular location would be exposed to less than the *92 contemplated radiation dose. 5 Therefore, any damage to the genetic pool would still be within the bounds suggested by 10 C.F.R. § 100.11(a)(3). The Appeal Board reached a different conclusion with respect to the beach population. On particular days during the summer the beach is the most densely populated place in New Hampshire, with up to 37,000 people on a projected peak day in 1980. The Appeal Board was troubled because these people were not technically “residents” of the beach, but recognizing the mandate to apply the regulations flexibly, see 10 C.F.R. § 100.2(b), it held that the beach was the nearest population center. The nearest point of the beach was 1.67 miles from the reactor; therefore, the low population zone radius could be no more than 1.25 miles. It is undisputed that the plant meets this reduced low population zone requirement. 6 Therefore, the Appeal Board upheld the determination that the plant met the siting safety criteria. 6 N.R.C. at 42-54. The NRC declined to review this decision. We affirm. The Supreme Court has held that the Commission’s interpretation of its regulations is controlling so long as it is reasonable and consistently applied. Northern Indiana Public Service Co. v. Porter County Chapter of the Izaak Walton League of America, Inc., 423 U.S. 12, 15, 96 S.Ct. 172, 46 L.Ed.2d 156 (1975). We find that the Appeal Board acted reasonably in deciding that the three towns should not be “Thus, for example, if 25,000 individuals were to be found in a population center located in one general direction from the plant, and each were to receive a dose of ‘y’ rem if exposed to the radioactive cloud for the entire period of its passage after an accident, the total population dose would be 25,000y. The same overall dose would not be encountered by two groups of 12,500 people living at the same distance but in significantly different directions from the plant. Assuming that, during the entire period of the radioactive cloud (assumed by the regulations to be 30 days), the wind shifted in such a fashion that each of the two groups was exposed to the plume for an interval of time sufficient to give it one-half of the maximum dose, the total population dose would be only 12,500y (i. e., 12,500 X ‘Ay + 12,500 X ‘Ay). In other words, a given number of persons who are congregated in one area (and thus might receive the maximum dose occasioned by the accident) are more significant, in terms of the rationale behind the population center concept, than the same number of persons located in varying directions from the facility, all other things being equal.” Public Service Co. of New Hampshire, 6 N.R.C. 33, 49-50 (A.L.A.B.1977). considered as a single population center, and petitioners have not alleged that the decision is inconsistent with other decisions of the NRC. The NRC, not this court, is entrusted with the task of making sure that nuclear power is safe. Our job is to see that the NRC performs that task in accordance with law. It is enough that we find that the NRC did make this decision in accordance with the relevant statutes and regulations. 2. Financial Qualifications The Atomic Energy Act requires “[e]ach application for a license hereunder . [to] state such information as the Commission, by rule or regulation, may determine to be necessary to decide such of the technical and financial qualifications of the applicant ... as the Commission may deem appropriate for the license.” 42 U.S.C. § 2232(a). The Commission’s regulations require the applicant to provide “[information sufficient to demonstrate to the Commission the financial qualifications of the applicant to carry out, in accordance with the regulations . . ., the activities for which the permit or license is sought. If the application is for a construction permit, such information shall show that the applicant possesses the funds necessary to cover estimated construction costs and related fuel cycle costs or that the applicant has reasonable assurance of ob *93 taining the necessary funds, or a combination of the two.” 10 C.F.R. § 50.33(f). Petitioner argues that the record does not support the NRC’s conclusion that PSCO is financially qualified to build the plant. 7 A great deal of evidence was presented on this issue, and it is true that the evidence does not reveal that PSCO’s financial outlook is rosy. PSCO must raise about twice its current assets in order to complete Seabrook. PSCO’s stock is selling at below book value. Moody’s has lowered PSCO’s bond rating from A to Baa, the lowest rating of any utility under the jurisdiction of the NRC. But we do not review this record de novo. Three levels of the NRC have already decided 8 on the basis of that record. The PSCO has demonstrated the requisite “reasonable assurance” that it can raise enough money to carry out the project. Once before PSCO raised almost twice its assets; there is expert testimony that PSCO can raise the money in the capital market, albeit at a high price; and the NRC found that the New Hampshire Public Utilities Commission is likely to support the project with favorable rate decisions. 9 We think there is substantial evidence to support the NRC’s conclusion, 10 therefore we will not disturb it. THE NATIONAL ENVIRONMENTAL POLICY ACT The Act gives the NRC complete discretion to decide what financial qualifications are appropriate. The regulations require only a “reasonable assurance”. We will not second guess the NRC as to its interpretation of the level of proof that standard requires. 3. Adequacy of the Final Environmental Impact Statement NEPA requires copies of a detailed environmental impact statement, together with the comments and views of appropriate agencies, to “accompany the proposal [for major Federal actions significantly affecting the quality of the human environment] through the existing agency review process.” 42 U.S.C. § 4332(2)(C). In order to comply with NEPA the NRC requires applicants for a construction permit to file an environmental report. 10 C.F.R. § 51.20. The NRC staff then prepares a draft environmental impact statement (DES) and distributes it with requests for comments. 10 C.F.R. §§ 51.22-51.25. After comments have been received, but before hearings on the application have been held, the staff prepares a final statement (FES) and distributes it in the same manner as the DES. 10 C.F.R. § 51.26. Both the DES and FES together with any comments received then accompany the application through the NRC review process. ' 10 C.F.R. § 51.26(d). The issue in this case arises because between the time the FES was prepared and the time the Licensing Board held its hearing the EPA decided that the originally proposed cooling water intake tunnel location was not acceptable. The EPA required PSCO to move the intake from 3,000 to 7,000 feet offshore in order to minimize the environmental impact. The NRC required PSCO to revise its environmental report but chose not to redo its own impact state *94 ments. Instead, the Licensing Board went ahead with the hearing, evaluating the 7,000 foot location, and issued initial decisions granting the construction permit. In doing so, the Licensing Board relied on 10 C.F.R. § 51.52(b)(3) which provides that: “an initial decision of the presiding officer may include findings and conclusions which affirm or modify the content of the final environmental impact statement prepared by the staff. To the extent that findings and conclusions different from those in the final environmental statement prepared by the staff are reached, the statement will be deemed modified to that extent and the initial decision will be distributed as provided in § 51.26(c).” We must decide whether this procedure, on the facts of this case, satisfied the requirement of NEPA. We have no trouble finding that it did. The Supreme Court has held that the provision of NEPA quoted above does not affect the time when the statement must be prepared; rather, it says what must be done with the statement when it is prepared. Aberdeen & Rockfish R. R. Co. v. SCRAP, 422 U.S. 289, 320, 95 S.Ct. 2336, 2356, 45 L.Ed.2d 191 (1975). 11 As the Court pointed out, “the time at which the agency must prepare the final ‘statement’ is the time at which it makes a recommendation or report on a proposal for federal action.” Id. (emphasis omitted); 42 U.S.C. § 4332(2)(C). In that case the ICC was required to prepare a statement, at the earliest, after an oral hearing when the agency issued a decision. Similarly, in our case, the earliest recommendation or report of the NRC, as distinguished from one by its staff, 12 was the Licensing Board’s initial decision. At that point, of course, the FES was in its final form, modified by the decision, and discussed the new intake location. The District of Columbia Circuit has reached a similar conclusion, upholding application of 10 C.F.R. § 51.52(b)(3) to modify a final environmental impact statement in compliance with NEPA. Citizens for Safe Power, Inc. v. NRC, 173 U.S.App.D.C. 317, 320, 524 F.2d 1291, 1294 & n. 5 (1975). That conclusion was not mere dictum since the court evidently considered the issue, and had the court come out the other way, it would have had to take some action other than simply dismissing the petition for review. Nor do we consider it significant that the modification in that case was a stipulation entered into by the parties. The parties’ agreement is not binding on the NRC since the NEPA requirement exists to protect the public at large and other interested agencies as well as those actually before the deciding agency. Even if we did not reach this conclusion, we would not remand the case because the procedures followed by the NRC satisfied the spirit of NEPA. A report based on the 3,000 foot location had been circulated. The EPA had decided that the new location would have a smaller impact on the aquatic environment than would the original location. 13 To that extent, the NRC was entitled to rely on the EPA conclusion. See fuller discussion in section 6 below. Thus, although the FES did not discuss the new location (which was not chosen until after the FES was prepared), the FES did consider a stronger case against the plant. Moreover, all the testimony at the hearing was directed to the new location and the Board’s decision was in reference to the new location and based on full information. See Porter County Chapter of the Izaak Walton League of America, Inc. v. AEC, 533 F.2d 1011, 1019 (7th Cir. 1976), cert. denied, 429 U.S. 945, 97 S.Ct. 366, 50 L.Ed.2d 316 (1976). *95 4. Alternative Site Comparisons Two issues arise out of the Commission’s handling of its NEPA mandate to prepare a detailed statement on “alternatives to the proposed action”. 42 U.S.C. § 4332(2) (C)(iii). The Commission decided that alternatives had not been adequately reviewed and remanded to the Licensing Board to conduct the required studies. See 5 N.R.C. 503 (1977). It instructed that panel first that “the test to be employed in assessing whether a proposed site is to be rejected in favor of any of the alternative sites considered . . . [is] whether an alternate site is obviously superior to the site the applicant had proposed.” Id. at -. Second, the Commission instructed, “in comparing construction costs of the proposed site and at alternate sites, actual completion costs should be used.” Id. at -. NECNP argues that these two standards violate NEPA, impermissibly skewing the mandated comparison in favor of the proposed site not only by insisting on clear superiority, but also by giving effect to money already expended. “The requirement for a thorough study and a detailed description of alternatives ... is the linchpin of the entire impact statement.” Monroe County Conservation Council, Inc. v. Volpe, 472 F.2d 693, 697-98 (2d Cir. 1972). But the requirement for thorough study does not determine the result of the comparison. The ultimate decision is left to the discretion of the agency. “Neither [NEPA] nor its legislative history contemplates that a court should substitute its judgment for that of the agency as to the environmental consequences of its actions. . . . The only role for a court is to insure that the agency has taken a ‘hard look’ at the environmental consequences; it cannot ‘interject itself within the area of discretion of the executive as to the choice of the action to be taken.’ ” Kleppe v. Sierra Club, 427 U.S. 390, 410 n. 21, 96 S.Ct. 2718, 2730, 49 L.Ed.2d 576 (1976) (citations omitted). We will not be able to determine whether or not the Commission has in fact taken the “hard look” until the ultimate decision on remand within the agency is before us. On this record the outcome of the alternate site comparison is left open, and we are limited to deciding whether the obvious superiority and completion cost standards, on their faces, conflict with the Commission’s duty to perform the required studies. 14 We find no such conflict. The obvious superiority standard, as it is explained in the Commission’s opinion, says nothing about whether or how the required studies will be performed. Rather it goes to what the Commission will do with findings that the studies will generate. The standard is designed to guarantee that a proposed site will not be rejected in favor of a substitute unless, on the basis of appropriate study, the Commission can be confident that such action is called for. Given the necessary imprecision of the cost-benefit analyses involved and the fact that the proposed site will inevitably have been subjected to far closer scrutiny than any alternate site, we cannot say that it is unreasonable to insist on a high degree of assurance that the extreme action of denying an application is appropriate. This is especially so since NEPA does not require that a plant be built on the single best site for environmental purposes. All that NEPA requires is that alternative sites be considered and that the effects on the environment of building the plant at the alternative sites be carefully studied and factored into the ultimate decision. The second standard requires the Licensing Board to consider the actual cost of completing the proposed site as compared to the cost of building at an alternative site. In other words, the money already sunk into the proposed site, where construction has been permitted pending final approval, will weigh to the benefit of that site when alternatives are considered. Like the Commission itself, we are troubled by this pro *96 cedure which encourages the applicant to sink as much money into the proposed site as possible even as the preliminary review of that site continues. As a result, the likelihood that an alternative site will be selected as obviously superior declines and the pressure against denying the application increases. At some point the Commission could well find itself seriously prejudiced in trying to carry out its NEPA mandate. We have to be careful, however, to separate out two distinct problems. One is the legitimacy of comparing sites on the basis of the facts actually existing — the actual cost of completing the plant if it is to be built at all. This is not only reasonable but necessary if the Commission is to fulfill its responsibility of protecting the public interest in the construction of atomic energy facilities. The public’s interest in the development of less expensive sources of power should not be ignored. In the abstract situation, the standard is a realistic way of dealing with existing circumstances, for instance the fact that certain preliminary costs must go into preparation and studies of the site before the Commission can even consider it. The problem becomes serious where, as here, the Commission has allowed the applicant to go beyond the relatively minor, necessary investigatory expenditures and to sink large sums into actual construction. That raises the second problem, the advisability of allowing construction permits to issue before final approval of the site. If the Commission is careful about granting such permits, and if the Commission wisely uses its power to stay such grants, situations in which sunk costs predetermine comparative site analyses, rendering them a meaningless form, can be avoided. Other courts have decided that the Commission may give an already developed site an advantage when comparing it to alternate sites. 15 Aeschliman v. NRC, 178 U.S.App.D.C. 325, 335, 547 F.2d 622, 632 n. 20 (1976), rev’d on other grounds sub nom. Vermont Yankee Nuclear Power Corp. v. NRDC, 435 U.S. 519, 98 S.Ct. 1197, 55 L.Ed.2d 460 (1978); Porter County Chapter of the Izaak Walton League of America, Inc. v. AEC, 533 F.2d 1011, 1017 n. 10 (7th Cir. 1976), cert. denied, 429 U.S. 945, 97 S.Ct. 366, 50 L.Ed.2d 316 (1976); Union of Concerned Scientists v. AEC, 163 U.S.App.D.C. 64, 79, 499 F.2d 1069, 1084 n. 37 (1974). We think this principle properly extends to our situation. The cases that NECNP cites to support its position are cases concerning injunctions to stop construction. Arlington Coalition v. Volpe, 458 F.2d 1323, 1332 (4th Cir.), cert. denied sub nom. Fugate v. Arlington Coalition, 409 U.S. 1000, 93 S.Ct. 312, 34 L.Ed.2d 261 (1972); Environmental Defense Fund v. TVA, 468 F.2d 1164, 1183-84 (6th Cir. 1972). That situation is not now before us. The Commission (in allowing its Appeal Board’s actions to stand) stayed construction at various times during the long course of these licensing proceedings. We are confident that the Commission is aware of the problem continuing construction can pose and will deal with the issue responsibly. Review of any final decision the Commission reaches on motions to stay construction will, of course, be available before a court of appeals. Neither the “obviously superior” nor the “completion cost” standard can be used as a facade for shirking NEPA duties, but neither is either standard in and of itself violative of NEPA. It remains to be seen whether the Commission carries out its review of alternate sites under these standards in conformance with NEPA’s “hard look” requirement. 5. Need for Power The Commission, interpreting its duty under NEPA in this case, held that “[a] nuclear plant’s principal ‘benefit’ is of course the electric power it generates. Hence, absent some ‘need for power,’ justi *97 fication for building a facility is problematical.” 6 N.R.C. 33, 90 (A.L.A.B. 1977), quoting Duke Power Co., 4 N.R.C. 397, 405 (A.L.A.B. 1976). The issue before us is whether there is sufficient evidence in the record to support the Appeal Board’s conclusion that there is a need for the power that Seabrook would generate. The Appeal Board’s decision is in two parts. First, it found that sometime in 1985 or 1986 the energy Seabrook could generate would be needed in order for the applicant utilities to be able to meet their power demands at all times. The Board rejected this basis for establishing a need for the plant for the period the plant is scheduled to be in operation before 1985, however, because the Licensing Board had failed to allow intervenors to cross-examine certain witnesses at a reopened hearing on the issue of projected power needs. The procedural error did not undermine the finding that Seabrook’s power would be needed after 1985 because the Appeal Board determined that there was no dispute as to that period. During the period until 1985 the energy Seabrook could generate would be needed, the Board went on to hold, in order to substitute for power the applicant now generates using fossil fuels, principally oil and coal. This conclusion does not rest on projections of demand for power. Rather, it reflects a determination that New England, heavily dependent on foreign oil, would derive substantial benefit from developing alternative sources of energy in order to decrease costs and assure a more certain supply of fuel. Petitioners challenge both parts of this decision. As to the first part, with respect to the years beginning in 1985, NECNP suggests that it did dispute the projected need for additional power. We have reviewed the parts of the record pointed to by NECNP, and we cannot find any statement, unequivocal or otherwise, suggesting that the additional power Seabrook would generate would never be needed in New England. NECNP may have disputed the question of whether Seabrook would be needed at any specific future date, but it seems to have conceded that the evidence supported a conclusion that demand for power in New England was increasing. See, e. g., NECNP Brief in Support of Exceptions, September 17,1976, p. 24, filed with the Appeal Board. The Appeal Board was certainly justified in reading that brief, id., as adopting the conclusion of Dr. Salo, a member of the Licensing Board who dissented and would have found that the need for Seabrook had not been adequately démonstrated in that there was “considerable doubt [about] the need for Seabrook in the near future (i. e., before 1985 or 1986).” 3 N.R.C. 857, 947 (A.S.L.B. 1976) (Salo, dissenting). A legitimate inference from his statement is that the need for power after 1985 or 1986 was established. Regardless of the position NECNP might have taken, the Appeal Board was justified in concluding, as it implicitly did, that the record as it stood justified finding a need for power generated by Seabrook as of 1985 and that nothing that might have been discovered on the cross-examination that had been denied would have undermined such a finding. In any event, we are certainly not prepared to say that on the record the Appeal Board chose to consider, 16 it was error to conclude that the power generated by Seabrook would not be needed by 1985. With respect to the second part of the decision, reliance on a “substitution theory” to support the need for Seabrook in the years 1981-1984, petitioner makes two arguments. First, NECNP challenges the Board’s cost comparison because it uses allegedly inaccurate, out-dated figures and because it considers only costs of operation, not of construction. The Board’s opinion, 6 N.R.C. at 97, identifies the locations in the record of the cost data on which it relied, including updated data. We are in no position to decide that NECNP’s data are more *98 accurate. Nor are we in a position to dictate at what point the Board may stop updating its data in order to reach a final decision. Furthermore, we do not think it would be unreasonable to base the cost comparison on operating costs given that the need for the additional power, and thus the need to construct a plant, at some point in the future had already been established. See id. But in the alternative the Board stated that the nuclear plant would be a preferable substitute even if construction costs were considered. Id. Second, petitioners argue that the Board failed to consider adequately alternative ways to substitute for oil fueled plants. Our view of the record indicates, to the contrary, that the Licensing Board heard evidence on and considered the alternatives proposed, unanimously concluding that they would not be viable substitutes for Sea-brook. See 3 N.R.C. at 903-07 and 939. The Appeal Board review of the record “indicate[d] that, even taken in combination, the numerous considered alternative energy sources would not be a better choice than the nuclear facility.” 6 N.R.C. at 99. We cannot substitute our own conclusions or preferences for those of the Commission. We find no basis on which to reverse the Commission’s conclusion that the applicant PSCO had demonstrated a need for Sea-brook. 6. Treatment of EPA Findings The Commission decided that it could accept as conclusive the Environmental Protection Agency’s findings with respect to the aquatic impact of the Seabrook plant’s once-through cooling system. 17 Before Seabrook can be put into operation it must get a license from the Environmental Protection Agency to discharge thermal pollution into the waters around Seabrook. The EPA may not issue such a license unless it first determines that the plant’s design can assure “the protection and propagation of a balanced, indigenous population of shellfish, fish, and wildlife in and on the body of water into which the discharge is to be made.” 33 U.S.C. § 1326(a). If the EPA should ultimately reach that conclusion 18 the Federal Water Pollution Control Act specifically states that NEPA does not “authorize any Federal agency authorized to license or permit the conduct of any activity which may result in the discharge of a pollutant into the navigable waters to review any effluent limitation or other requirement ... or the adequacy of any certification under section 401 of this Act . . . .” 33 U.S.C. § 1371(c)(2)(A). Petitioner argues that under NEPA neither the FWPCA nor the EPA’s special expertise in matters of water quality permit the NRC to adopt the EPA findings without an independent inquiry into the effect Seabrook would have on the aquatic environment. There is no dispute that NEPA requires the NRC to factor any anticipated marine pollution into its cost-benefit analysis of the Seabrook application. The NRC accepts this duty but argues that it is justified in refusing to reach an independent judgment about matters determined by the EPA because those matters are committed by law to the special expertise of the EPA and because repetitious adjudication of the issue would be wasteful. The NRC argues, and we agree, that it can properly limit its concern to deciding whether permits should be issued given the aquatic impact as determined by EPA and other environmental impacts as determined by the NRC. The NRC did not shirk its NEPA duties. It performed a wide-ranging study of Sea-brook’s environmental impact. In so doing, it obeyed its FWPCA duties by deciding to accept as dispositive EPA determinations concerning one aspect of the overall environmental impact. We cannot agree with petitioner that this course of action unfairly *99 deprives them of a chance for input. If any party chose not to appear before the EPA, it was not for lack of opportunity to do so. NECNP and Massachusetts do not suggest that they have evidence to present that was not ably presented to the EPA by SAPL. Clearly what they seek is a second forum in which to present their case in hopes of improving their chance of success. Fairness does not require that they be accorded such an opportunity. NEPA does not require that the NRC offer such an opportunity. Therefore, we hold that the NRC may rely on EPA findings made .in the course of determining whether to issue a discharge permit. 7. Reviewability PSCO has suggested that the remaining issues are not reviewable because intervenor, SAPL, which has raised them before us, failed to present the issues to the full Commission after the Appeal Board’s decision. The Commission does not join in this argument. The Appeal Board is authorized to render final decisions. See 10 C.F.R. §§ 2.785(a) and 2.770. An appeal to superior agency authority is not a prerequisite to reviewability absent an agency rule requiring an appeal before the agency action becomes final. 5 U.S.C. § 704. The NRC has recently adopted a procedure allowing parties to seek review of Appeal Board decisions before the Commission, but, as the Commission stated at oral argument, this rule is not mandatory, and the Commission does not view it as a prerequisite to review by a court. See 10 C.F.R. § 2.786. Therefore, these issues, which were presented to and decided by the Appeal Board, are properly before us, and we will proceed to decide them. 8. Decommissioning and Tourism SAPL challenges the adequacy of the Commission’s NEPA review of two aspects of Seabrook’s impact on the environment — the effect of decommissioning the plant at the end of its projected 30 to 40 year useful life and the effect on the area’s tourist industry. Certainly NEPA does require the NRC to assess these impacts, at least once a party has raised the issues. The fact is, however, that the Licensing Board did hear testimony on these questions, and it did consider these questions in its opinion, 3 N.R.C. at 881-82 and 884. We are singularly ill-equipped, and we would have to go beyond what we perceive to be our scope of review, to find that the Commission did not meet its NEPA obligation. We fully agree with SAPL that these are important issues, and we have some sympathy for SAPL’s argument that the Commission could have given these issues a more thorough examination. But we are not about to tell the Commission what the proper number of witnesses or the proper length or depth of a hearing must be for any given issue. Clearly the Commission gave these issues the consideration it felt they were due. There are no allegations that the Commission refused to accept proffered testimony. In short, we hold that the Commission did give these issues a hard enough look under the circumstances. THE ADMINISTRATIVE PROCEDURE ACT 9. Replacement of the Licensing Board Chairperson After extensive hearings had been held before the three person Licensing Board, its chairperson decided to leave the NRC to take a job with another federal agency. The day after he announced his departure, a new chairperson was appointed and the hearings continued. Both before the Commission and before us SAPL has argued that the APA, 5 U.S.C. § 554(d), requires an agency in such a situation “to declare a mistrial and order a de novo hearing, at least as to those issues in which the credibility of the witnesses was contested.” Section 554(d) requires that “[t]he employee who presides at the reception of evidence . . . shall make the recommended decision or initial decision . . , unless he becomes unavailable to the agency.” Clearly the Licensing Board member in this case had become unavailable to the *100 Commission. 19 The question we must decide is whether the APA dictates what action an agency must take when the presiding employee becomes unavailable. The APA expresses a strong preference for having decisions made by someone who has heard the evidence presented. See Gamble-Skogmo, Inc. v. FTC, 211 F.2d 106, 114 (8th Cir. 1954). As the legislative history indicates, the presiding employee’s opinion is of special consequence “to the extent that material facts in any case depend on the determination of credibility of witnesses as shown by their demeanor or conduct at the hearing.” See id. at 115. On that basis the Eighth Circuit decided that an agency, in a hearing governed by § 554(d), 20 would have to hold a de novo hearing unless “it fairly could be said that a credibility evaluation from hearing and seeing the witnesses testify was unnecessary . . .” Id. We regard the decision under this standard to be a question of procedure and evidence committed to the discretion of the agency. In this case we cannot say that the agency abused its discretion. Though credibility of the conflicting experts must play a central role in the Licensing Board decision, that credibility is a function of logical analysis, credentials, data base, and other factors readily discernible to one who reads the record. SAPL has not demonstrated that this is an issue that turns on conflicting eyewitness reports or evaluations of the witnesses’ demeanor or conduct. The Commission’s interest in not holding new hearings is obvious and significant. The process of evaluating an application to construct a nuclear power plant is long enough without having to repeat it. Therefore, we uphold the decision to allow a new person to step in to take the rest of the evidence and to participate in the decision. Each of the petitions for review is dismissed. Footnotes [1] For a discussion of the EPA’s role in the Seabrook licensing process, see SAPL v. Costle, 572 F.2d 872 (1st Cir. 1978 ). [2] Other factors to be considered in evaluating proposed sites are listed in 10 C.F.R. § 100.10. [3] The combined population of the three towns, including transients on a weighted basis, would exceed 25,000 by 1985. [4] The towns are spread out in an arc surrounding the site from the southwest to the northeast. The population of the towns is interspersed by salt marshes and sparsely populated areas. [5] The Appeal Board gives the following example: [6] The radius of the low population zone is a variable that can be affected by the design features of the reactor. In this case Seabrook was already “overdesigned” for the 1.5 mile distance, so that it still met the requirements at 1.25 miles. [7] Petitioner also argued that the Commission was not following its own regulations because it expressed doubts about the value of the financial qualifications inquiry. The Commission, however, clearly stated that it was following the existing regulations in this case, leaving for future proceedings the resolution of issues concerning the proper content of the regulations. [8] The decision of the full Commission is the one before us for review. We need not concern ourselves with whether or not the Licensing Board and Appeal Board may have used different standards resolving this issue. [9] Clearly the likelihood of future New Hampshire Public Utilities Commission decisions favorable to PSCO is relevant to the applicant’s ability to obtain the necessary funds and is a proper subject for NRC consideration. We note that the NRC is not bound by this decision should circumstances change in the future or should predictions not be borne out. [10] We are not impressed by the suggestion that we have no role to play in reviewing this issue. The regulations do set out specific substantive standards for the applicant to meet. The Commission’s decision pursuant to these standards is subject to review under 5 U.S.C. § 706(2). [11] The Court in SCRAP specifically disapproved the cases on which NECNP suggests we rely. 422 U.S. at 321 n. 20, 95 S.Ct. 2336. [12] Functionally, the staff is a party at the hearing, it takes its position as an adversary, and the Commission does not adopt any position officially until after the hearing. [13] We do not mean to imply approval or disapproval of the substance of the EPA decision, which we have not reviewed. See SAPL v. Costle, 572 F.2d 872, 883 (1st Cir. 1978). [14] We are treating the Commission’s adoption of these standards as a final agency action subject to review, 5 U.S.C. § 704, because it is the Commission’s last word on an issue of law, and were we to disagree with that view, much lost effort could be avoided. [15] On the other hand an already developed site can enjoy no advantage in comparison to the option of abandoning the project altogether. E. g., Aeschliman v. NRC, 547 F.2d 622, 632 n. 20 (D.C.Cir. 1976), rev’d on other grounds sub nom. Vermont Yankee Nuclear Power Corp. v. NRDC, 435 U.S. 519, 98 S.Ct. 1197, 55 L.Ed.2d 460 (1978). [16] There is no challenge before us of the Board’s decision to disregard testimony concerning the need for power. That testimony, if considered, of course, would have tended to bolster the Board’s conclusions. [17] For a full discussion of the design of the plant and the issues before the EPA see SAPL v. Costle, 572 F.2d 872 (1st Cir. 1978). [18] The case is now back before the EPA on remand from this court. SAPL v. Costle, supra. If the EPA should reach a different conclusion, the NRC will, of course, be able to adjust its action accordingly. [19] We reject SAPL’s suggestion that the departing member had some kind of duty to clear his decision with the parties in the case or that the NRC should be held bound to have tried to prevent his departure. [20] Because we find no violation of § 554(d), we need not decide whether this aspect of § 554(d) was intended to apply in initial licensing cases.