Gregory W. Sample, Asst. Atty. Gen., with whom James E. Tierney, Atty. Gen., and Kay R.H. Evans, Asst. Atty. Gen., Augusta, Me., were on brief, for the State of Maine, amicus curiae.

Jonathan B. Hill, with whom John P. Schnitker, Dow, Lohnes & Albertson, Washington, D. C., Bruce W. Chandler, and Marden, Dubord, Bernier & Chandler, Waterville, Me., were on brief, for the Pittston Co., intervenor.

Rosanne Mayer, Atty., Dept. of Justice, with whom Carol E. Dinkins, Asst. Atty. Gen., Land and Natural Resources Div., Donald W. Stever, Jr., Atty., Dept. of Justice, Washington, D. C., and Susan Studlien, Atty., E. P. A., Boston, Mass., were on brief, for U. S. E. P. A., respondent.

Wayne S. Henderson, Boston, Mass., for New England Legal Foundation, et al., intervenor.

Before COFFIN, Chief Judge, BOWNES and BREYER, Circuit Judges.

COFFIN, Chief Judge.

In these three consolidated appeals petitioners challenge the final decision of the EPA Administrator to issue a National Pollutant Discharge Elimination System (NPDES) permit to the Pittston Company pursuant to § 402 of the Clean Water Act, 33 U.S.C. § 1342. The permit authorizes the Pittston Co. to construct and operate a 250,000 barrel per day oil refinery and associated deep water terminal at Eastport, Maine, in accordance with specified effluent limitations, monitoring requirements, and other conditions. Petitioners contend that EPA's actions violated the National Environmental Policy Act (NEPA), 42 U.S.C. § 4321 et seq., the Endangered Species Act, 16 U.S.C. § 1531 et seq., and the Clean Water Act, 33 U.S.C. § 1251 et seq.

Pittston proposes to construct an oil refinery and marine terminal in Eastport, Maine, a relatively pristine area of great natural beauty near the Canadian border. The area is known for being the foggiest on the East Coast, experiencing some 750-1000 hours of fog a year; daily tides approximate twenty feet. The plan contemplates that crude oil shipments will arrive several times a week in supertankers, or Very Large Crude Carriers (VLCCs), as long as four football fields, or slightly less than a quarter of a mile. The tankers will travel through Canadian waters1 around the northern tip of Campobello Island, where the Roosevelt Campobello International Park is located, see 16 U.S.C. § 1101 et seq., down Head Harbor Passage to a refinery near Eastport where they will be turned and berthed. Numerous barges and small tankers will carry the refined product from Eastport to destination markets in the Northeast.

The protracted procedural history of this case begins in April 1973, when Pittston applied to the Maine Board of Environmental Protection (BEP) for permission to locate the refinery in Eastport. After public hearings, the BEP approved the proposal under the Maine Site Location of Development Law, 38 M.R.S.A. § 481 et seq., subject to a number of pre-construction and pre-operation conditions designed primarily to reduce the risk of oil spills. Pittston subsequently filed an application with EPA to obtain an NPDES permit, and submitted an Environmental Assessment Report to aid EPA in its duty to prepare an Environmental Impact Statement (EIS) pursuant to NEPA. See 33 U.S.C. § 1371(c)(1); 42 U.S.C. § 4332(2)(C). EPA promulgated a draft EIS recommending issuance of the permit as conditioned by the Maine BEP, held a joint public hearing with the Army Corps of Engineers in Eastport, and received approximately 600 responses during a public comment period. In September 1977, the Maine Department of Environmental Protection certified, under § 401(a)(1) of the Clean Water Act, 33 U.S.C. § 1341(a)(1), that the proposed discharge would satisfy the appropriate requirements of state and federal law. In June 1978, the final EIS was issued, again recommending that the permit be issued pursuant to the BEP conditions.

Several months later, the National Marine Fisheries Service (NMFS) of the Department of Commerce and the Fish and Wildlife Service (FWS) of the Department of Interior initiated consultations with EPA concerning the proposed refinery's impact on endangered species-the right and humpback whales, and the northern bald eagle, respectively-under § 7 of the Endangered Species Act (ESA), 16 U.S.C. § 1536. In November, the NMFS issued a threshold determination that there were insufficient data to conclude that the project was not likely to jeopardize the continued existence of the endangered whales. In December, the FWS concluded that the project was likely to jeopardize the bald eagle. In light of these opinions and of the value of the natural resources in the Eastport area as noted in the EIS, EPA's Region I issued a notice of determination to deny Pittston's application for an NPDES permit in January 1979. Pittston thereafter sought an adjudicatory hearing and administrative review of this decision.2

Prior to the hearing, extensive consultation between EPA, NMFS, FWS, and Pittston took place to consider mitigation measures proposed by Pittston. In May, NMFS concluded on the basis of the best scientific data available that EPA was unable to comply with the statutory mandate that it "insure that (the project) is not likely to jeopardize the continued existence of" endangered whales. 16 U.S.C. § 1536(a)(2). In June FWS reaffirmed its previous determination that the refinery was likely to jeopardize the bald eagle. EPA Region I amended its decision to include these new findings.

The adjudicatory hearing took place over five weeks in January and February of 1980. More than fifty witnesses testified and were cross-examined; several hundred exhibits were introduced. In January 1981, the ALJ rendered EPA's Initial Decision, overturning EPA Region I and ordering that the NPDES permit issue. He concluded that the EIS was adequate to comply with NEPA, and that no supplemental EIS was necessary; that the risk of oil spills was "minute" and that the refinery was therefore not likely to jeopardize any endangered species; and that the conditions imposed by the Maine BEP, and assumed by the EIS, were not required to be conditions of the federal permit. Petitioners subsequently sought review before the EPA Administrator, and also moved to reopen the record to admit a recent study showing an increased number of endangered whales in the Eastport region. Both motions were denied, and in September 1981 EPA Region I issued the NPDES permit to the Pittston Company. Petitioners now seek review in this court pursuant to § 509(b)(1)(F) of the Clean Water Act, 33 U.S.C. § 1369(b)(1)(F).

It is now well settled that there are two aspects to a court's review of agency action subject to the requirements of NEPA:

"First, the court makes a substantive review of the agency's action to determine if such action is arbitrary and capricious under the Administrative Procedure Act, 5 U.S.C. § 706. This substantive review, although conducted on the basis of the entire administrative record, is quite narrow in scope. The court should only assure itself that the agency has given good faith consideration to the environmental consequences of its actions and should not pass judgment on the balance struck by the agency among competing concerns.

Second, a reviewing court must assess the agency's compliance with the duties NEPA places upon it. These duties are 'essentially procedural'. The primary procedural mechanism embodied in NEPA is the requirement that an agency prepare 'a detailed statement' discussing, inter alia, 'alternatives to the proposed action', 42 U.S.C. § 4332(2)(C). Requiring an agency to discuss alternatives within the EIS serves numerous goals. The detailed statement aids a reviewing court to ascertain whether the agency has given the good faith consideration to environmental concerns discussed above, provides environmental information to the public and to interested departments of government, and prevents stubborn problems or significant criticism from being shielded from internal and external scrutiny." Grazing Fields Farm v. Goldschmidt, 626 F.2d 1068, 1072 (1st Cir. 1980) (citations & footnote omitted). See also Silva v. Lynn, 482 F.2d 1282, 1283-84 (1st Cir. 1973).

In order to weigh the benefits of the project against the potential environmental costs, the EIS contained an analysis of the justification for the project and the anticipated economic benefits. The project was deemed consistent with a longstanding federal policy of encouraging the construction of domestic refining capacity in order to promote national security. New England, heavily dependent on imported oil, had no regional refining capacity. The project was designed to accommodate VLCCs, thus taking advantage of the cost savings offered by economies of scale. Constructing a refinery in the United States rather than abroad had the additional advantage of retaining jobs and investments in this country. Finally, the project was particularly attractive because it was designed to handle high sulfur crude oil and refine it into low sulfur fuels, thus facilitating compliance with new environmental standards. Such a refinery was "of an entirely different design" than most existing domestic refineries, which were built to handle domestic and steadily depleting sources of low sulfur crude.

Petitioners argue that the EIS was faulty because it failed to consider the possibility of conservation and the use of alternative fuels instead of the construction of additional oil refining capacity. We note first that petitioners failed to raise this concern in a meaningful way during the comment period. In any case, it is clear that at the time the EIS was drafted, there was a forecasted need for the type of refinery planned by Pittston, and that it was federal policy to encourage the construction of such refineries. Nor are we persuaded by petitioners' argument that the discussion in the EIS of the need for the project is "totally outdated and of no present use." Even accepting their contention-based, we might add, primarily on statements of energy policy under President Carter, which might themselves be considered outdated-that there is no longer a strong need for additional refining capacity, it remains uncontested that there is still a demand for domestic refineries capable of processing high sulfur crude into low sulfur products. Given this continuing national and regional need, we see no need to supplement the EIS in order to accommodate the most recent data and federal policy shifts. Cf. New England Coalition on Nuclear Pollution v. NRC, 582 F.2d 87, 96-98 (1st Cir. 1978).

C. Adequacy of Consideration of Alternatives

Petitioners contend that the EIS failed to discuss adequately a number of alternatives to the proposed refinery at Eastport. First, they argue that EPA erred by conducting a less searching analysis of alternatives to this privately sponsored project than it would have had the project been publicly funded. Second, they urge that EPA unreasonably limited its consideration of alternative sites to three locations in Maine. Finally, they allege that EPA's comparison of the various sites was inadequate.

EPA's evaluation of alternatives was explicitly based on the premise that its role in reviewing privately sponsored projects "is to determine whether the proposed site is environmentally acceptable", and not, as in the case of a publicly funded project, "to undertake to locate what EPA would consider to be the optimum site for a new facility." Therefore, EPA considered its purpose in this case to be to search for alternatives "that would be substantially preferable from an environmental standpoint." EPA concluded that "(t)his different purpose affects the extent of the information on alternatives necessary to make a decision."

We are unable to fault EPA's reasoning. Petitioners concede that the substantive standard-"substantially preferable"-was correctly stated. Cf. New England Coalition on Nuclear Pollution v. NRC, 582 F.2d at 95-96 ("obvious superiority"). No purpose would be served by requiring EPA to study exhaustively all environmental impacts at each alternative site considered once it has reasonably concluded that none of the alternatives will be substantially preferable to the proposed site. Moreover, the guideline adopted by EPA to limit its study of alternatives appears, in this case, to be consistent with the "rule of reason" by which a court measures federal agency compliance with NEPA's procedural requirements. See, e.g., Grazing Fields Farm, 626 F.2d at 1074; Massachusetts v. Andrus, 594 F.2d 872, 884 (1st Cir. 1979).

EPA's choice of alternative sites was focused by the primary objectives of the permit applicant, the Pittston Co. Pittston stated that its basic consideration was to find a port with deep water near shore in order to accommodate VLCCs. Only by using such supertankers could Pittston take advantage of economies of scale, thereby making the project economically feasible. Therefore, after Pittston had reviewed and rejected a number of sites lacking such deep water, EPA limited its consideration to the only ports providing deep water access. Three alternative areas in Maine were considered: Portland, Machias, and Penobscot/Blue Hill.3

EPA's duty under NEPA is to study all alternatives that "appear reasonable and appropriate for study at the time" of drafting the EIS, as well as "significant alternatives" suggested by other agencies or the public during the comment period. In order to preserve an alternatives issue for review, it is not enough simply to make a facially plausible suggestion; rather, an intervenor must offer tangible evidence that an alternative site might offer "a substantial measure of superiority" as a site. See Seacoast Anti-Pollution League v. NRC, 598 F.2d 1221, 1228-33 (1st Cir. 1979).

In light of this standard, petitioners' argument that EPA erred by restricting its consideration to alternative sites in Maine must fail, because they did not suggest any reasonable alternatives to EPA during the comment period. One alternative-a monobuoy off the mid-Atlantic coast-was raised for the first time at the adjudicatory hearing, too late for inclusion in the EIS. Although petitioners now contend that EPA should reasonably have been aware of such an alternative earlier, their citation to a 1976 study by the Office of Technology falls far short of persuading us; nor have they explained their failure to bring the report to EPA's attention in a timely manner. Another possibility, an offshore monobuoy in New England, was rejected by EPA in spite of its potential environmental benefits because of fierce public opposition to similar proposals off the coast of Massachusetts and New Hampshire. We cannot say that EPA acted unreasonably in concluding that such an option was not feasible. In short, we are not convinced that EPA failed to consider all alternatives which were feasible and reasonably apparent at the time of drafting the EIS.

Petitioners next urge that EPA's consideration of these alternative sites was inadequate. The EIS contained comparative analysis of the effects of the proposed project on air quality, water quality, present land and sea uses, terrestrial and aquatic flora and fauna, and aesthetics at the various sites. The Portland area was eliminated by Pittston because its shallow channel is unable to accommodate VLCCs and suitable land for a refinery site and marine terminal was not available. EPA considered an offshore monomooring system near Portland, but rejected it due to the vulnerability of the proposed location to the elements and the chronic spills associated with monomooring which would interfere with the nearby fishing and recreation industries.4 The Machias site was considered substantially similar to Eastport from an environmental perspective, but was eliminated by Pittston because suitable land was unavailable. EPA also noted that the harbor at Machias was more exposed to wind and weather than that at Eastport, thus making a tanker approach more hazardous. Heavy tourism at Penobscot/Blue Hill made a refinery undesirable; the area is also a center for Maine's lobster, clam and fishing industry. Finally, the tanker approach at the area is quite long with numerous islands, increasing the risk of mishap close to shore and inhabited areas. Consequently, EPA concluded that none of the alternative sites would provide a significantly greater degree of environmental protection than the Eastport site. Having carefully reviewed the record, we cannot say that EPA's consideration of these sites was inadequate, or that its conclusion to reject them was arbitrary and capricious.

We defer our consideration of additional NEPA issues raised by petitioners-the adequacy of the risk spill analysis in the EIS, and the need for a supplemental EIS-until after our discussion of the risk of oil spills in the context of the Endangered Species Act.

As noted earlier, EPA Region I originally issued a notice of determination not to issue the permit based on the opinion of the NMFS and the FWS that EPA could not insure that the project was not likely to jeopardize the right and humpback whales, and the bald eagle, respectively. The ALJ, in his Initial Decision, rejected these biological opinions and held that the project was not likely to jeopardize the continued existence of these species. While administrative review was being sought, the National Oceanic and Atmospheric Administration (NOAA) moved to reopen the record, proffering a 1980 study indicating the presence in the Eastport region during the summer of a significant portion of the north Atlantic right whale population. The acting Administrator of EPA, assuming arguendo the validity of the study, concluded that the new information was not significant because of the ALJ's supportable finding that any risk of a major oil spill was minute. At the same time, he summarily affirmed the initial decision, and EPA Region I subsequently issued the NPDES Permit to Pittston.

The obligation imposed on EPA by section 7(a)(2) of the ESA, 16 U.S.C. § 1536(a)(2) is to "insure that any action authorized, funded, or carried out ... is not likely to jeopardize the continued existence of any endangered species." An action would "jeopardize" the species if it "reasonably would be expected to reduce the reproduction, numbers, or distribution of a listed species to such an extent as to appreciably reduce the likelihood of the survival and recovery of that species in the wild." 50 C.F.R. § 404.02 (1980).

Although the 1979 Amendments to ESA softened the obligation on an agency from requiring the agency to "insure" the species would not be jeopardized to requiring the agency to "insure" that jeopardy is not "likely", Pub.L.No.96-159, § 4(1)(C), 93 Stat. 1225, 1226 (1979), the legislative intent was that the Act "continues to give the benefit of the doubt to the species." H.Conf.Rep.No.96-697, 96th Cong., 1st Sess. 12, reprinted in (1979) U.S.Code Cong. & Ad.News, 2557, 2572, 2576. Agencies continue to be under a substantive mandate to use "all methods and procedures which are necessary", TVA v. Hill, 437 U.S. 153, 185, 98 S. Ct. 2279, 2297, 57 L. Ed. 2d 117 (1978) (quoting 16 U.S.C. §§ 1531(c), 1532(2), emphasis added by the court), "to prevent the loss of any endangered species, regardless of the cost." Id. at 188 n.34, 98 S. Ct. at 2299 n.34 (emphasis in original). The Act does, however, create a special "exemption" procedure (not at issue here, see note 2, supra ) designed to allow necessary actions even if they threaten the loss of an endangered species. See 16 U.S.C. §§ 1536(g), (h).

An agency's duty to consult with the Secretary of Commerce or Interior, depending on the particular endangered species, does not divest it of discretion to make a final decision that "it has taken all necessary action to insure that its actions will not jeopardize the continued existence of an endangered species". National Wildlife Federation v. Coleman, 529 F.2d 359, 371 (5th Cir. 1976). The consultation process, however, is not merely a procedural requirement. Not only is a biological opinion required of the Secretary of Commerce or Interior, "detailing how the agency action affects the species or its critical habitat", 16 U.S.C. § 1536(b), but the 1979 Amendments to ESA require that in fulfilling its consultation duty and in insuring the absence of likelihood of jeopardy "each agency shall use the best scientific and commercial data available." 16 U.S.C. § 1536(a)(2). Moreover, the legislative history emphasizes that "(c)ourts have given substantial weight to these biological opinions as evidence of an agency's compliance" with the Act, that "(t)he Amendment would not alter this state of the law or lessen in any way an agency's obligation" under § 7, and that a federal agency which "proceeds with (an) action in the face of inadequate knowledge or information ... does so with the risk that it has not satisfied the standard of" § 7(a)(2). H.Conf.Rep. at 12, reprinted in (1979) U.S.Code Cong. & Ad.News at 2576. See also H.R.Rep.No.95-1625, 95th Cong., 2d Sess. 12, reprinted in (1978) U.S.Code Cong. & Ad.News 9453, 9462.

In reviewing an agency's decision after consultation our task is "to ascertain whether 'the decision was based on a consideration of the relevant factors and whether there has been a clear error of judgment.' " National Wildlife Federation v. Coleman, 529 F.2d at 372 (quoting Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 416, 91 S. Ct. 814, 823-824, 28 L. Ed. 2d 136 (1971)). We must also inquire into whether the ALJ "followed the necessary procedural requirements." Overton Park, 401 U.S. at 417, 91 S.Ct. at 824.

C. The Administrative Law Judge's Initial Decision

The FWS biological opinion concerning the risk of jeopardy to the bald eagle began with its identification of Cobscook Bay (adjacent to the proposed Eastport refinery) as the most productive of three areas essential to conservation of the species in Maine and the northeastern United States. The specific threats included air pollution containing mercury emissions and increased acidification of lakes dangerously adding to the already high concentrations of heavy metals in eagle eggs and the food chain; the intrusion of economic development and human population; and a number of problems caused by oil spills, including the mortality of embryos and young eagles, reduction in the fish and bird food supply of eagles, fouling of wings, and ingestion problems.

The ALJ found that the FWS claim as to mercury emissions rested on an assumed daily emission of 200 grams. After a review of the evidence, he determined that, on a "worst case" basis, there might be a daily emission of 17.6 grams. He concluded that this amount, as well as negligible amounts of lead and vanadium, would not affect the eagle's food chain. The ALJ also found that refinery emissions would have no important impact on the acidity of lakes, one year's exposure to such emissions being equivalent to that in six hours of rainfall. The threats based on human activity were found to be inconsequential in view of the difficulty of access to nesting areas, the demonstrated tolerance of human presence by eagles, the recreation-inhibiting inclement spring weather during the time of greatest eagle sensitivity, and proposed mitigation measures.

We cannot say that these findings were not adequately supported, that the relevant factors were not considered, or that the ALJ made a clear error of judgment.5 But there was one additional finding: that although a significant oil spill would have an adverse impact on eagles and their reproduction, the risk of such a spill was "very small or minute", so that the species was not jeopardized. The validity of this finding will be considered below.

The NMFS biological opinion singled out right and humpback whales as being subject to adverse impact, their population being limited and their migratory pattern placing them in the Eastport area during spring and summer. The anticipated harm from oil spills included illness from ingestion, skin irritation, fouling of baleen plates, and contamination of food.

The ALJ accepted an estimate of a total north Atlantic right whale population of between 70 and 100, and a humpback whale population of 2000 or more. He concluded, based on the combination of the brief periods when the whales were in Eastport waters and, given the navigational safeguards and restrictions to be imposed on Pittston by both the State of Maine and the Coast Guard, the low probability of a massive oil spill, that there was no reasonable likelihood that the continued existence of the two whale populations would be jeopardized.

The low risk of spills was also central to EPA's determination not to reopen the record to receive a 1980 New England Aquarium study estimating that a minimum of 48 right whales, or approximately one half the total population, had been in or near the proposed tanker approaches to Eastport in that year. Repeating the analysis relied upon by the ALJ, the EPA acting Administrator concluded that the new study was not sufficiently probative to open the record. In his words, the "(a)bsence of risk, rather than the absence of whales" underlay the ALJ's decision, and his own decision to affirm the granting of the permit.

We now proceed to outline the ALJ's reasoning leading to his finding, crucial to both ESA issues, of the unlikelihood of a significant oil spill.

The ALJ's conclusion that the risk of a major oil spill was minute was based primarily on three items of evidence. First, the ALJ relied heavily on assurances from the Coast Guard which, after reviewing the testimony of Pittston's witnesses before the BEP and other data, wrote EPA on March 28, 1977, that the channel in Head Harbor Passage was "adequate for safe navigation by 250,000 DWT tankers" if four conditions were met. These conditions were

"(1) that the channel passage area depths, configurations and current data shown on nautical charts and surveys be confirmed by hydrographic survey, (2) provision for a navigation system wherein the existence and movement of all traffic in the area could be monitored, communicated with and scheduled, (3) provision for means to control movement of tankers in the event of steering and/or propulsion failure during transit and (4) development and strict adherence to an operating procedure for tanker passage."

In response to a request by the Council on Environmental Quality that the Coast Guard assist Pittston in carrying out "real time simulation" studies6 in order to ascertain the precise conditions for safe navigation prior to granting the permit, Rear Admiral Fugaro of the Coast Guard responded in August 1977 that it could not divert scarce resources until "final clearance had been granted for construction of a refinery ... (so that) no possibility exists that these efforts may be wasted." After explaining the Coast Guard's "function in port development",7 he concluded:

"The Coast Guard feels it can be premised that tank vessels can safely navigate the channel approaches to Eastport under certain conditions-and the Coast Guard fully intends to determine those conditions and see to their implementation. Although we will continue to work closely with the Pittston Company, the Environmental Protection Agency, and the State of Maine and other affected and concerned groups as required and within available resources, we feel that further delay of the project for the purpose of studying the issue of channel adequacy appears unjustified at this time. In that any major Federal action taken to implement operational restrictions and control procedures would necessarily be the subject of an additional EIS, I believe that both the spirit and letter of NEPA are well served." (emphasis added)

Subsequently, on December 31, 1979, the Coast Guard clarified its position as to item 2 in its March 28, 1977, letter (provision for a navigation system) by saying that notwithstanding the capability of any navigation system, there would be some meteorological conditions, e.g., fog producing poor visibility, which would preclude safe transit.

Second, the ALJ found confirmation of the Coast Guard's assurance in the computer simulation studies of Dr. Eda, who concluded that a loaded 250,000 DWT tanker could maintain a trajectory close to a desired track in Head Harbor Passage without tug assistance in a 60 knot wind. Although these studies could not account for the human factor, i.e., could not test any difficulty on the part of the human pilot in perceiving the location, heading and rate of change of heading of the ship, the ALJ understood there was "an encouraging correlation" between computer simulation and actual sea trial. The ALJ accepted Dr. Eda's statement that "for obtaining an overall perspective of the suitability of a particular channel for ship traffic of specific sizes under particular conditions, off-line computer studies are adequate."

Also cited with approval by the ALJ was a second study by Frederick R. Harris, Inc. premised on provision for a more adequate turning basin for the VLCCs than an earlier study which had approved the project subject to severe restrictions and "a high order of seamanship and prudence." This study, the ALJ found, deemed the proposed approach "satisfactory for the type and size of vessels specified providing navigational aides are installed, and providing recommended operational procedures were followed." These included tug assistance from entry into channel, lighted buoys and radar reflectors, an electronic guidance system involving land based radar and electronic range finders, confining berthing and deberthing to slack tide, limiting Head Harbor transit to daylight or clearly moonlit hours, proscribing entrance to the Passage if visibility is less than a mile, and barring tankers awaiting a berth from anchoring in Eastport waters.

Finally, the ALJ made rather minute review of testimony concerning prevailing currents and cross-currents, fog, wind, and duration of oil spill effects, concluding in general that currents were not excessive for shipping, that the expected presence of fog was not so great as to bar shipping during most of the time, that winds were in general within tolerable limits, and that the effects of large known oil spills had not been long lasting over a period of years.8

We have set forth in some detail and full strength all of the strands of the decision of the ALJ because we conclude that, in light of EPA's duty to insure that the project is unlikely to jeopardize endangered whales or eagles, the ALJ's failure to require, at a minimum, that "real time simulation" studies be done to assure the low risk of an oil spill prior to granting the permit violated his duty to "use the best scientific ... data available."9 Given the Supreme Court's statement that the ESA is designed to prevent the loss of any endangered species, "regardless of the cost", TVA v. Hill, 437 U.S. at 188 n.34, 98 S. Ct. at 2299 n.34, we cannot see how the permit can issue when real time simulation studies, which EPA, the State of Maine, and the Coast Guard all view as being necessary to a final determination of safety, are to be delayed until the Coast Guard has adequate funds to undertake them.

We begin with the linchpin-the Coast Guard opinion. From what we have reported above, we think it quite clear that the Coast Guard was not purporting to do a risk analysis. It was, in effect, signifying its willingness to accept the problem of devising procedures to minimize navigation risks for vessels of certain characteristics transiting via Head Harbor Passage to Eastport. That this is a correct reading is confirmed by the testimony of Rear Admiral Fugaro, who candidly stated of the Coast Guard opinion that "(i)t's not designed to provide a risk analysis." His letter to EPA, which we have quoted, makes clear that he expected any set of Coast Guard orders and procedures to go through the EIS process.10

This was also the understanding of Wallace Stickney, the EPA Region One Director responsible for drafting the EIS, who viewed the Coast Guard evaluation not as describing "what the actual risks were intrinsically", but as purely a comparison to other supertanker ports. We see the Coast Guard "assurances" as falling short of what Coast Guard Admiral Barrow, relied on by the ALJ when he rejected the use of world-wide statistics relating to oil spills (see note 12, infra ), prescribed: "(A)ny comprehensive and meaningful oil spill study for the development of spill probability and expected spill size must be concerned with site specific factors such as tanker fleet composition, density, navigation systems, route characteristics, operational conditions, regulatory regimes etc."

We cannot presume to know what issues may be posed as the result of real time simulation studies, or, for that matter, real sea trials by VLCCs under ballast. Risks of collisions or grounding may be identified whose assured prevention may entail costs unacceptable to Pittston or measures involving other environmental intrusions or, simply, unacceptable risks which may persist despite the most stringent and expensive procedures and equipment. That those further studies are conceded to be vital is demonstrated by the following testimony of EIS drafter Stickney:

"Q. So, basically, then you decided that that (results of real time simulation studies) wasn't information that was needed to determine whether this refinery should be built or not?

A. We felt the information was needed and that if the facility failed the real time simulation study it would never be built.

Q. It will be too late, will it not, if that study, for example, shows some problems that you haven't anticipated in the final EIS, it will be too late for EPA to say at that point, now the weighing of risk versus benefit is different than we originally thought? Will it not be too late for that?

A. No, sir."

We see absolutely no justification for issuing an NPDES permit before a closer and feasible risk assessment is made.11

Additionally, we note the Coast Guard's requirements of a hydrographic survey to make sure that the depth figures on the navigation chart fairly represent the entire length, width, and depth of the channel, face of pinnacles and outcroppings, so that VLCCs with draft beginning at 65 feet may pass without danger of grounding during the lowest of tides. Never, we suspect, has there been occasion to make sure that the bottom is from 70 to 90 or 100 feet under low water at all times of the year at all points beneath a broad channel seven miles long in Head Harbor Passage. Should the hydrographic survey reveal embarrassing obstructions, this fact and ways of dealing with it must receive the most careful scrutiny.

The other grounds relied on by the ALJ leading to his conclusion of small or minute risk are even less persuasive than the Coast Guard undertaking. Dr. Eda's computer simulations were avowedly valuable for obtaining "an overall perspective of ... suitability"; they could not approach even a rough approximation of risk, nor could they account for human error in confronting diverse weather conditions. The second Harris study merely pronounced a route "acceptable" if fairly rigorous conditions were complied with, but none of these conditions were incorporated in the federal permit. Finally, the ALJ's analyses of current, wind, fog, and duration of spills gave only general assurance that prudence, procedures, and equipment can, most of the time and absent human error, compensate for difficult conditions of tide, current, fog, wind and weather.

We stress that our disagreement with the ALJ does not involve challenging his credibility judgments, although we do not share his view that the "overwhelming weight" of evidence pointed to the feasibility of safe transit.12 Were the issue whether, by a preponderance of the evidence, it had been established that VLCCs could make the transit through Head Harbor Passage to Eastport with reasonable safety, the ALJ's decision might be accepted. But the issue is a harder one: whether, after using the best data available, it is established that the risk of significant oil spills from the proposed tanker traffic is so small as to insure that there is no likelihood of jeopardizing the two endangered species. All witnesses have agreed that real time simulation studies would contribute a more precise appreciation of risks of collision and grounding. We think the same could be said of a hydrographic survey of the depth of the channel, and perhaps of trial runs by VLCCs in ballast. If so, such methodologies obviously represent as yet untapped sources of "best scientific and commercial data".

It may very well be that, after conducting real time simulation studies and any other tests and studies which are suggested by the best available science and technology, the most informed judgment of risk of a major oil spill will still have a large component of estimate, its quantitative element being incapable of precise verification. But at least the EPA will have done all that was practicable prior to approving a project with such potentially grave environmental costs.

We also conclude, for many of the same reasons, that the real time simulation studies and other new data must be the subject of a supplemental EIS, both to assess the magnitude of risk and, if acceptable, to establish appropriate conditions of navigation. The testimony quoted above demonstrates that EPA and the Coast Guard have acknowledged the need for such a supplemental EIS on this issue. See also Alaska v. Andrus, 580 F.2d 465, 477-80 (D.C.Cir.), other portion of holding vacated on other grounds sub nom. Western Oil & Gas Ass'n v. Alaska, 439 U.S. 922, 99 S. Ct. 303, 58 L. Ed. 2d 315 (1978). The EIS itself recognizes that "real time simulation studies ... will help to settle the navigation (safety) issue." Given the importance of the studies to the crucial issue of the risk of oil spills, NEPA provides an additional ground for overturning the issuance of a permit until the studies have been conducted, circulated, and discussed. See NRDC v. Callaway, 524 F.2d 79, 92 (2d Cir. 1975); Silva v. Lynn, 482 F.2d at 1287-88.

Petitioners' final argument is that the ALJ erred by ruling that conditions imposed on the project by the Maine BEP under state law are not incorporated into the federal NPDES permit. They allege that the certification issued by the State of Maine pursuant to § 401(a)(1) of the Clean Water Act, 33 U.S.C. § 1341(a)(1), though making no explicit mention of the conditions previously imposed by the Maine BEP, incorporated these terms by implication. Therefore, these requirements must also be "a condition on any Federal license or permit." § 401(d), 33 U.S.C. § 1341(d). The State of Maine, as amicus curiae, makes a somewhat different argument. It argues that the prior certification is irrelevant, because the proposal as approved by the state had been substantially modified by the ALJ. But the state contends that it has been denied its right to certify the new proposed discharge, and therefore the NPDES permit is invalid. Respondents argue that the state has waived its right to certify the proposed modified discharge because it failed to intervene in the hearing before the ALJ or to certify the discharge within 30 days of receiving notice that the prior proposal had been amended. 40 C.F.R. § 125.32(e)(8)(v) & (vi) (1978). They also urge that the ALJ's finding that the prior state certification did not incorporate the BEP conditions was not clearly erroneous, and must be upheld.

The ALJ considered testimony and evidence to determine whether the state certification implicitly incorporated the conditions previously imposed by the Maine BEP. Contrary to respondents' contention, he found as a factual matter that "the conditions of the Maine BEP Order are conditions precedent to the effectiveness of" the state's certification. He further ruled, however, as a matter of law, that § 401(d) of the Act precludes the state from including in its certification requirements of state law which do not relate to "water quality standards, effluent limitations or schedules of compliance." See § 301(b)(1)(C), 33 U.S.C. § 1311(b)(1)(C). Finally, he concluded that "conditions of the Maine BEP relating to test runs with tankers prior to delivering oil, limiting the size of tankers ..., requiring real time simulation studies, stating times and conditions of navigation of Head Harbor Passage, and other matters unrelated to water quality may not legally be regarded as part of the State of Maine's Sec. 401 certification, irrespective of the intention of the issuer of the certification."

Petitioners argue, with some force, that the conditions listed above are related to water quality, since they are designed to minimize the risk of an oil spill which would severely impair water quality. We believe that the ALJ made a more fundamental error by seeking to determine which requirements of state law were appropriately affixed to the state's certification. Section 401(a) of the Clean Water Act empowers the state to certify that a proposed discharge will comply with the Act and "with any other appropriate requirement of State law." Any such requirement "shall become a condition on any Federal license or permit." § 401(d). EPA has interpreted this provision broadly to preclude federal agency review of state certification. "Limitations contained in a State certification must be included in a NPDES permit. EPA has no authority to ignore State certification or to determine whether limitations certified by the State are more stringent than required to meet the requirements of State law." EPA, Decision of the General Counsel No. 58 (March 29, 1977); see also Decision of the General Counsel No. 44 (June 22, 1976). The NPDES regulations state that "(r)eview and appeals of limitations and conditions attributable to State certification shall be made through the applicable procedures of the State and may not be made" through the procedures established in the federal regulations. 40 C.F.R. § 124.55(e) (1981). The courts have consistently agreed with this interpretation, ruling that the proper forum to review the appropriateness of a state's certification is the state court, and that federal courts and agencies are without authority to review the validity of requirements imposed under state law or in a state's certification. See United States Steel Corp. v. Train, 556 F.2d 822, 837-39 & n.22 (7th Cir. 1977); Lake Erie Alliance v. U.S. Army Corps of Engineers, 526 F. Supp. 1063, 1074 (W.D.Pa.1981); Mobil Oil Corp. v. Kelley, 426 F. Supp. 230, 234-35 (S.D.Ala.1976).

Our conclusion that EPA lacked authority to review the conditions imposed by the State of Maine is also supported by the statutory scheme of the Clean Water Act. Section 511(c)(2) of the Act, 33 U.S.C. § 1371(c)(2), makes clear that "(n)othing in the National Environmental Policy Act ... shall be deemed to authorize any Federal agency ... to review any effluent limitation or other requirement established pursuant to this Act or the adequacy of any certification under section 401 of this Act." (emphasis added). Section 510 of the Act, 33 U.S.C. § 1370, specifically preserves the right of a state to "adopt or enforce ... any requirement respecting control or abatement of pollution", even if it is more stringent than those adopted by the federal government. Finally, it is clear that even in the absence of state certification, EPA would be bound to include in the federal permit "any more stringent limitations ... established pursuant to any State law or regulations (under authority preserved by section 510)." § 301(b)(1)(C), 33 U.S.C. § 1311(b)(1)(C); see United States Steel Corp., 556 F.2d at 837-39; Decision of the General Counsel No. 44, at 5.

The regulations cited by respondents do not compel a different result. The 1978 regulation cited by EPA-which provided that failure to certify a proposed permit, within thirty days after the state is notified that the permit has been modified, "shall be deemed a waiver of such certification rights"-was no longer in force in 1980 when the decision to modify the proposal was made by the ALJ. The new regulation, 40 C.F.R. § 124.55(d) (1981), states that "(a) condition in a draft permit may be changed during agency review in any manner consistent with" state certification without requiring recertification. This regulation clearly does not authorize EPA to amend a permit in a manner inconsistent with state certification by deleting conditions imposed by the state during the certification process. Although the new regulations also require the state to cite to state law when imposing more stringent conditions on a draft permit, 40 C.F.R. § 124.53(e)(1) (1981), and to indicate the extent to which the condition can be relaxed without violating state law, 40 C.F.R. § 124.53(e)(2) (1981), it would be inequitable to hold that the state has waived its rights here by failing to comply with these requirements when no similar requirements were in force in 1977 when state certification took place. Since the state at no time waived its rights to certify the proposed discharge, and the ALJ lacked authority to exclude the previously imposed state conditions from the federal permit, these conditions must be included in any NPDES permit for the Pittston project to be issued in the future, unless the conditions are modified according to law. See 40 C.F.R. § 124.55(b) (1981).

Accordingly, we vacate EPA's decision to issue the NPDES permit to Pittston, and remand the case to EPA to conduct further proceedings consistent with this opinion. EPA's jeopardy determination under the Endangered Species Act must be reconsidered in light of the results of real time simulation studies and a hydrographic survey of Head Harbor Passage, and any other studies, such as the 1980 whale study by the New England Aquarium, which EPA determines to be necessary to meet its statutory obligation to use the best scientific data available. If, in light of the studies, EPA decides to recommend approval of the project, this proposal shall be the subject of a supplemental EIS relating to the conditions of navigation necessary to minimize the risk of oil spills. Finally, the conditions imposed by the State of Maine in its certification of the proposed discharge must be included in any federal permit unless the conditions are subsequently modified according to law.13

So ordered.