The plaintiffs, Save Lake Washington (“SLW”) and numerous private citizens residing on Lake Washington, appeal from the district court’s judgment vacating a preliminary injunction and denying entry of a permanent injunction against the construction of certain docking facilities on the lake for use by oceangoing vessels in the service of the National Oceanic and Atmospheric Administration (“NOAA”). We affirm.
I. BACKGROUND
The NOAA is an agency within the Department of Commerce which administers a variety of research programs dealing mainly with meteorology and with marine resources. The NOAA Pacific Northwest headquarters, located in Seattle, is currently spread out over seven locations in and around Seattle. Ever since the establishment of the headquarters in Seattle, NOAA officials have been concerned with the necessity from an economic standpoint to consolidate operations at a single facility. After searching the Puget Sound area for suitable locations in the early 1970’s, the NOAA fixed its attention upon a portion of the former Sand Point Naval Air Station located on freshwater Lake Washington, immediately to the east of Seattle. The Sand Point site currently lacks docking facilities of any type.
The proposed consolidation of facilities at Sand Point would require the dredging of the offshore waters and the construction of piers adequate to service the NOAA’s fleet of oceangoing research vessels. At present, NOAA vessels dock at Lake Union. Access to Lake Washington from Puget Sound would be accomplished by entering the Lake Washington Ship Canal from Shilshole Bay, heading east through Lake Union and the Montlake Cut and then onto the lake itself. While the passage of ocean vessels from Puget Sound to Lake Union is not an uncommon occurrence, such vessels rarely enter Lake Washington. Lake Washington is currently devoted primarily to recreational use by local residents.
In January 1975, the NOAA published a draft environmental impact statement relating to the relocation project in an attempt to comply with the National Environmental Policy Act' (“NEPA”), 42 U.S.C. § 4332 et seq. Following public commentary on the draft EIS, NOAA published a final EIS in January 1976. The final EIS concluded that location of all facilities at Sand Point would represent the most cost-effective alternative to the various sites *1333which had been under consideration, including one “split-site” plan which would have located the NOAA’s land-based facilities at the Sand Point site, with the dock facilities located elsewhere.
In December 1977, SLW and the other plaintiffs, concerned over NOAA’s apparent intent to berth oceangoing vessels at Sand Point, filed their complaint in the present suit seeking injunctive and declaratory relief against the construction of ocean vessel berths at Sand Point. The plaintiffs alleged inter alia that NOAA’s final EIS did not comply with the requirements of NEPA, and that the NOAA had violated the Coastal Zone Management Act of 1972, 16 U.S.C. § 1451. Shortly thereafter, plaintiffs moved for a preliminary injunction. On March 15, 1978, the district court adopted a magistrate’s report recommending that a preliminary injunction issue. The court ruled that the final EIS was deficient in its failure to discuss adequately alternative sites and for failing to discuss increased navigational hazards which would result from the passage of large vessels through Montlake Cut to Lake Washington.
In response to the district court’s evaluation of the deficiencies contained in the final EIS, the NOAA prepared a supplemental environmental impact statement. In November 1978, the Director of the NOAA issued a decision memorandum in which he approved consolidation of the Seattle facilities at Sand Point and set out his reasons for doing so. The Director was aided in his decision by a document titled the “NOAA Western Regional Center Option Assessment,” prepared by his staff, and presumably by the supplemental EIS.
On January 9, 1979, the NOAA filed a motion to vacate the preliminary injunction, which was again referred to a magistrate, along with plaintiffs’ motion to certify questions of state law to the Supreme Court of Washington. On April 30, 1979, the district court, acting upon the magistrate’s recommendation, vacated the preliminary injunction and granted the NOAA’s motion for summary judgment. This appeal followed.
In January 1981, during the pendency of this appeal, the NOAA substantially altered its plans for consolidation of its facilities at Sand Point.1 Due to budgetary constraints, the NOAA decided to reduce the number and size of piers planned for Sand Point and to provide moorage space at that site only for periodic research staging and transient vessel visits. The revised project will require moorage space to accommodate two NOAA vessels, rather than twelve vessels as originally planned. The permanent berthing facilities for NOAA’s oceangoing fleet will now be located outside of Lake Washington rather than at Sand Point.
II. NEPA
SLW raises three contentions under its NEPA-based challenge to the Sand Point project:
1. The supplemental EIS fails to address adequately the navigational risks associated with Sand Point;
*13342. The supplemental EIS fails to acknowledge and respond to responsible opposing views;
3. The decision to berth oceangoing vessels on Lake Washington is reviewable under the “arbitrary and capricious” standard.
We examine each of these arguments separately, in light of the NOAA’s current plans.
A. Standard of Review
Sitting as an appellate court reviewing an environmental impact statement which has received a clean bill of health from the district court, we deal with now-familiar principles. The district court’s finding that an EIS is adequate will be reversed if based upon an erroneous legal standard or upon clearly erroneous findings of fact. Warm Springs Dam Task Force v. Gribble, 565 F.2d 549, 551, 552 (9th Cir. 1977). The task of the district court is also a limited one. Judicial review of an EIS covers only the issue of whether NEPA’s procedural requirements have been met, and whether the EIS performs its primary function of presenting the decision-maker with an environmentally-informed choice. The correct standard is provided in the Administrative Procedure Act, 5 U.S.C. § 706(2)(D), which directs courts to set aside an agency action if taken “without observance of procedure required by law...” See Lathan v. Brinegar, 506 F.2d 677, 693 (9th Cir. 1974) (en banc); see also, Lange v. Brinegar, 625 F.2d 812 (9th Cir. 1980); Matsumoto v. Brinegar, 568 F.2d 1289, 1290 (9th Cir. 1978).
The narrow standards of review applicable at both the trial and appellate court levels leave the judiciary with an extremely narrow substantive role, if any. In Vermont Yankee Nuclear Power Corp. v. NRDC, 435 U.S. 519, 98 S.Ct. 1197, 55 L.Ed.2d 460 (1978), the Supreme Court defined the following role for federal courts in NEPA eases:
“NEPA does set forth significant substantive goals for the Nation, but its mandate is essentially procedural.... It is to insure a fully informed and well-considered decision, not necessarily a decision the judges of the Court of Appeals or of this Court would have reached had they been members of the decisionmaking unit of the agency. Administrative decisions should be set aside in this context, as in every other, only for substantial procedural or substantive reasons as mandated by statute, .., and not simply because the court is unhappy with the result reached.” (citations omitted.)
Id. at 558, 98 S.Ct. at 1219.
As the Supreme Court recently noted in Strycker’s Bay Neighborhood Council v. Karlen, 444 U.S. 223, 100 S.Ct. 497, 62 L.Ed.2d 433 (1980), a federal court may look only to insure that the agency in question has “considered the environmental consequences” of its action. 444 U.S. at 227, 100 S.Ct. at 499-500. NEPA does not require that the agency elevate environmental concerns over legitimate non-environmental considerations. Id. It requires only that they be brought to the agency’s attention and that they receive fair consideration. In the words of an earlier Supreme Court decision, the agency must take a “hard look” at the environmental consequences. Kleppe v. Sierra Club, 427 U.S. 390, 410, n.21, 96 S.Ct. 2718, 49 L.Ed.2d 576 (1976).
Guidelines in evaluating the adequacy of an EIS are well established. An EIS need set forth only those alternatives necessary to permit a “reasoned choice.” See, e. g., Brooks v. Coleman, 518 F.2d 17, 19 (9th Cir. 1975). The twofold purpose of an EIS is to “.. . (1) provide decision-makers with an environmental disclosure sufficiently detailed to aid in the substantive decision whether to proceed with the project in the light of its environmental consequences, and (2) make available to the public, information of the proposed project’s environmental impact and encourage public participation in the development of that information.” Trout Unlimited v. Morton, 509 F.2d 1276, 1287 (9th Cir. 1974).
*1335With these considerations in mind, we proceed to SLW’s challenges of the supplemental EIS.
B. Navigational risks
SLW charges that the supplemental EIS has not cured one of the original EIS’s more glaring deficiencies, the failure to examine adequately the navigational risks associated with the passage of large ships through the Montlake Cut and onto Lake Washington. Specifically, SLW alleges that the risks have not been researched adequately in that no tests for slow-speed stopping of large vessels have been conducted, and that the adverse environmental impacts of an oil spill have not been fully investigated. Both the magistrate’s report and the district court opinion found these contentions to be without merit. We affirm.
The district court acknowledged that more research could be performed to determine with greater precision the slow-speed stopping distances of NOAA vessels, but concluded that available research was sufficient to apprise the decision-maker of the risks involved. The supplemental EIS contains a lengthy appendix which describes in some detail a study of navigational risks likely to be encountered in berthing at Sand Point. Included in this thorough discussion of the problems associated with Sand Point is full-speed stopping data provided by the vessel’s manufacturer. The NOAA includes an estimate of slow-speed stopping distances for its two largest vessels. The EIS further recommends as a portion of the “mitigation package” that slow-speed tests be conducted in order to provide additional information for vessel captains. The purpose of the proposed research was to mitigate the navigational risks by making such data available to the vessel captains, and not to postpone necessary preliminary research. In more general terms, SLW attacks the risk assessment for its failure to employ a “worst case” analysis. On the contrary, as the magistrate’s report found, both the potential magnitude and frequency of accidents is presented at some length. The EIS specifically notes that a “worst case” analysis was used, and nothing in the data presented suggests to the contrary.
SLW’s criticism of the oil spill discussion centers mainly on the statement’s brief reference to the toxic effects of a spill on flora and fauna, without a more detailed description of the effect of a spill on different types of aquatic life. The SEIS lists the “maximum credible oil spill” that could occur if each of its vessels were involved in a collision, calculated by adding together the fuel capacity of the vessel’s two largest adjacent fuel tanks. The SEIS does not contain any data on the size of spill that would be likely to result from a collision, nor does it discuss the effects of such a spill in detail; it states only that spills would be “difficult to contain and to clean up” and that fuels used by NOAA ships “are toxic to both flora and fauna.”
We need not decide whether this limited treatment of the topic would have been adequate had the NOAA proceeded with its original plans to berth its fleet at Sand Point. As we have noted above, the NOAA has abandoned this portion of its plan and now intends to berth its fleet outside of Lake Washington. The safety record of vessels traversing the Lake Washington Ship Canal, and the greatly reduced number of transits through the Montlake Cut that will be necessary now that the NOAA has altered its plans, convinces us that the chance of a major oil spill is remote. A discussion of “remote and conjectural consequences,” while possibly desirable, is simply not required by NEPA. Sierra Club v. Hodel, 544 F.2d 1036, 1039 (9th Cir. 1976).
C. Failure to acknowledge responsible opposing viewpoints
SLW’s second general area of attack is upon the alleged failure of the supplemental EIS to take into account opposing view*1336points. This contention breaks down into two stages. SLW first charges that viewpoints of certain National Ocean Survey officers as to the risks of passage from Puget Sound to Lake Washington have not been acknowledged. Secondly, SLW points to opposing viewpoints of land use officials in the Lake Washington area.
SLW notes that a number of officers employed by the NOS, the agency responsible for navigating NOAA’s vessels, have voiced serious doubts about the chances of passing the vessels through the waterways to Lake Washington on a routine basis without a significant chance of a serious accident occurring. SLW points, for example, to a statement by an NOS officer who piloted a large NOAA vessel through the passage, and described his experience as a “tense, risky trip.” This, and similar comments by experienced sea captains familiar with NOAA ships and with the waters involved, led the author of the risk assessment to conclude in a separate document that the risks associated with berthing vessels at Sand Point were “marginally acceptable,” a comment not included in the supplemental EIS.
SLW’s argument that the risk assessment contained in the supplemental EIS unnecessarily disparages these comments is without merit. As the magistrate’s report notes, the risk assessment clearly identifies both the sources and the substance of these comments. The account of the passage is laid out in a footnote. It is apparent that the NOAA considered the officers’ criticisms in the supplemental EIS. The information presented was sufficient to provide the NOAA with a reasoned basis for making its decision.
SLW has further argued that the supplemental EIS does not take into account the published views of several state and local land use officials that the presence of oceangoing ships on Lake Washington would be inconsistent with certain land use goals established for the area, including the Lake Washington Regional Shoreline Goals and Policies. The primary complaint is that specific views were not represented in the supplemental EIS. Instead, summaries of local officials’ views were included. The district court found that these summaries adequately conveyed the substance of the criticisms. We again see no reason to disturb the court’s finding. The summaries sufficiently apprised the decision-maker of the opposition of certain officials to the project, and their reasons. Viewing the EIS as a whole, the agency had an adequate basis for evaluating the concerns of the land use jurisdictions affected by the Sand Point project. We also agree with the magistrate’s report that failure to summarize the views of the President of the Seattle City Counsel and the letter of opposition from the Town of Hunt’s Point is not fatal. Again, the EIS has adequately taken into account similar views.
D. The decision to locate at Sand Point
SLW invites us to overturn the decision to locate at Sand Point as being arbitrary and capricious. In so doing, SLW in effect asks that we evaluate as a substantive matter the decision itself. This we cannot do. In another context, this court has stated that: “[ujnless the agency decision was so arbitrary and capricious as to amount to bad faith, the court cannot review the substantive decision of the agency.” Daly v. Volpe, 514 F.2d 1106, 1108 (9th Cir. 1975). We cannot weigh the relative merits of one proposal against another, nor can we attempt to compare beneficial environment effects. We may only examine the decision for the requisite good-faith consideration of environmental facts as expressed in the EIS.
As to the initial contention that the Director of the NOAA erroneously relied upon the Option Assessment prepared by his staff, we note that the Option Assessment was made available for public comment along with the supplemental EIS and the Decision Memorandum. Consequently, an opportunity for public review and comment *1337on the Option Assessment was provided. We see no basis for overturning the Administrator’s decision on the basis of his use of the Option Assessment.
• We also are not persuaded that the alleged mischaracterization in the Decision Memorandum of the aquatic effects of permanently berthing the NOAA vessels at Lake Union as “moderate” instead of “minimal” or “low” as described in the supplemental EIS requires us to set aside the Director’s decision. We concur with the district court’s finding that the allegedly erroneous description had a de minimis effect, if any, upon the decision to berth the vessels at Sand Point instead of Lake Union.
SLW’s other arguments require only brief responses. There was an adequate basis for determining that the construction costs favored Sand Point. Property acquisition costs would appear to support a decision to place some or all facilities at Sand Point. With regard to SLW’s argument that the Director did not consider the advantages associated with the Lake Union split-site alternative, we note that a split-site alternative has now apparently been adopted by the NOAA. Moreover, the Decision Memorandum summarizes briefly the extended discussion in the Option Assessment dealing with the Lake Union proposal. Based upon the content of the Decision Memorandum, we cannot say that the Director failed to make a good-faith, reasoned choice. The Decision Memorandum is not a document required by NEPA and need not contain a detailed discussion of all potential alternatives. We review it here only to satisfy ourselves that the Director did not act in bad faith.
We reiterate our very limited role in the process of reviewing the Director’s decision. The ailtimate conclusion that Sand Point is an appropriate site for NOAA facilities may well be a blunder, but we have served our purpose under NEPA by assuring that it was a “knowledgeable blunder.” Matsumoto v. Brinegar, supra, 568 F.2d at 1290. There was in the record a sufficient basis for the Director’s decision. We find that the Sand Point location decision was neither arbitrary nor capricious.
III. THE COASTAL ZONE MANAGEMENT ACT
SLW urges as a separate ground for enjoining construction of the pier the failure of the NOAA to comply with certain provisions of the Coastal Zone Management Act of 1972, 16 U.S.C. § 1451 et seq., and applicable regulations. Specifically, SLW charges that the NOAA has violated the “federal consistency” provisions of the Act, 16 U.S.C. § 1456(c)(1), (c)(2), and (c)(3), by failing to conform the project to local land use plans for the Lake Washington shoreline region, and by failing to obtain a “substantial development permit” for the project.
Because we affirm the district court’s finding that the NOAA was entitled to rely on an apparent determination by the Washington Department of Ecology that the proposed development was consistent with the Washington Shoreline Management Act, and that the substantial development permit issue was not raised in a timely fashion, we will assume both that the plaintiffs here had standing under the Act, and that a claim for relief can be stated under the Act by a private litigant. See Town of North Hempstead v. Village of North Hills, 482 F.Supp. 900, 905 (E.D.N.Y. 1979). We defer resolution of these issues for a case where they are more squarely presented.
The purpose of the Act is to encourage the development of state coastal management programs by providing 80% of the funding necessary for such programs. Where a state program meets with federal approval, the Act mandates that any federal activity which directly affects the coastal zone of a state must “to the maximum extent practicable” be consistent with the state’s coastal management program.
*1338Washington became the first state to enact a comprehensive management program when it passed the Shoreline Management Act of 1971, R.C.W. 90.58 et seq. Under the Shoreline Management Act, the responsibility for enacting master programs lies with local governments, subject to approval by the Washington Department of Ecology (“WDOE”). In order to aid in coordinated planning among the various jurisdictions which line Lake Washington’s shores, the WDOE created the Lake Washington Region. See WAC, Chapter 173-28. The Region eventually • adopted the Lake Washington Regional Shoreline Goals and Policies from which local jurisdictions were expected to develop their shoreline master programs. The Regional Goals and Policies contained an unequivocal statement against the expanded moorage of oceangoing ships on the lake. Following completion of the regional policy statement, the City of Seattle published the Seattle Shoreline Master Program (“SSMP”), which did not include an express prohibition against moorage of large vessels. The SSMP map also provided for the location of the NOAA headquarters at Sand Point, though it does not clearly provide for the construction of any docking facilities.
SLW argues that because the federal Coastal Zone Management Act requires federal projects affecting the shoreline to be consistent with state and local programs and because the Lake Washington Regional Goals and Policies so clearly prohibit the moorage of oceangoing ships on Lake Washington, this court must find that the NOAA contravened the federal consistency provisions of the Act by attempting to develop a docking facility for large ships at Sand Point. The SSMP, it is urged, must be read in a manner consistent with the regional planning policies even though it does not expressly prohibit moorage of ocean vessels on the lake.
Both federal and state law provide a mechanism for determining whether a proposed federal project is “consistent” with state shoreline management programs. Under the regulations implementing the Coastal Zone Management Act, the federal agency must determine in the first instance whether a proposed project is consistent with the state program. 15 C.F.R. § 930.37. In the event that the federal agency concludes that its proposed project is consistent with the state management program, it must then submit a “consistency determination” to the appropriate state agency describing the activity, its effects, and data to support the conclusion that the project is consistent. 15 C.F.R. § 930.39. If the state agency fails to respond within 45 days, then the federal agency is entitled to “presume” that the project has been deemed consistent with the state program. 15 C.F.R. § 930.-41(a). The WDOE has promulgated somewhat similar guidelines which provide that if the WDOE determines that a project is consistent, then “nothing more will be done.”
In January 1977, the NOAA forwarded a consistency determination on the Sand Point project to both the WDOE and the Seattle Department of Community Development, the civic agency responsible for administration of the SSMP. At the time, the federal regulations on consistency determinations were not in force. According to one affidavit filed by D. Rodney Mack, the supervisor of the Shorelands Division of the WDOE, the WDOE accepted “without comment” the NOAA’s consistency determination. Mack’s affidavit reveals that the potentially harmful effects of mooring ocean vessels on Sand Point were considered by the WDOE, but were rejected because of “certain differences” between the NOAA’s ships and regular private, commercial oceangoing craft. The . Director of the Seattle Department of Community Development responded positively to the consistency determination in a letter dated October 10, 1977.
The district court found that the NOAA was entitled to rely upon the WDOE’s apparent agreement with the de*1339termination that the Sand Point development was consistent with the Shoreline Management Act, despite certain indications that the WDOE had mistakenly assumed that the NOAA was not subject to the Coastal Zone Management Act. We concur with the district court. Where procedures to resolve potential federal-state disagreements over matters affecting the jurisdiction of both have been established, we should be reluctant to set aside determinations made pursuant to those procedures absent a compelling reason to do so. The Mack affidavit discloses that regard was given to the merits of the Sand Point project. While we mean to express no opinion on the substantive issue of consistency, we affirm the district court’s finding that the NOAA’s reliance was justified.
SLW’s second contention under the Coastal Zone Management Act is that NOAA failed to obtain a “substantial development permit” as required by the Shoreline Management Act. See R.C.W. 90.58.-140. Any doubt that a federal agency is required to obtain a substantial development permit prior to undertaking a project in the State of Washington is dispelled by the state regulations implementing the Shoreline Management Act. WAC 173-14-062 provides inter alia that:
“The permit system shall be applied in the following manner to federal agencies on lands meeting the criteria of the Shoreline Management Act and the department for shorelines of the state.
u * * *
“(e) The permit system shall apply to developments undertaken on lands not federally owned but under lease, easement, license, or other similar federal property rights short of fee ownership, to the federal government.”
The proposed pier construction apparently extends into state waters and therefore is subject to the permit system.
The district court refused to reach the merits of SLW’s substantial development permit argument because it found the claim to be untimely. SLW’s original complaint alleged only that the proposed project was inconsistent with the Lake Washington Regional Shoreline Goals and Policies and with various state shoreline management programs. The first specific mention of the substantial development permit issue in the record appears in a memorandum addressed to the magistrate on February 7, 1979. The magistrate’s report, filed on February 15, makes no mention of the substantial development permit issue. The point was more fully argued in a pair of subsequent memoranda addressed to the district judge, with a rejoinder to the contention filed by the NOAA on April 9, 1979. The district court found that it would be unfair to consider the argument because it had been raised for the first time over one year after the NOAA had undertaken to cure the defects in the first EIS.
SLW argues that the complaint was sufficient to put the NOAA on notice of the development permit issue, and that the issue was tried by consent of the parties under Fed.R.Civ.P. 15(b). We disagree with both of SLW’s attempts to revive the issue, and affirm the district court.
The complaint was insufficient to put NOAA on notice of the precise argument that the permit was required. As the briefs in this case have amply demonstrated, an allegation of “inconsistency” under the Coastal Zone Management Act is a wide-open argument which involves an examination of a number of different federal, state, and local regulatory schemes. Moreover, the complaint does not allege a violation of WAC 173-14-062. Consequently, even under our liberal standards of pleading, the complaint was inadequate to give notice of the substantial development permit issue.
We are unpersuaded that Rule 15(b) required that the district court reach the merits of the issue. Rule 15(b) provides in pertinent part that: “[w]hen issues not raised by the pleadings are tried by express *1340or implied consent of the parties, they shall be treated in all respects as if they had been raised in the pleadings.” Although Rule 15(b) applies in its express terms only to the actual trial of issues, it has been applied to questions raised in pretrial motions. See Bobrick Corporation v. American Dispenser Co., 377 F.2d 334 (9th Cir. 1967). The amendment to conform to issues actually tried lies within the sound discretion of the trial judge. See, e.g., Cole v. Layrite Products Company, 439 F.2d 958, 961 (9th Cir. 1971). It is clear that an attempted amendment causing prejudice to the defendant can be rejected by the trial court. See Gonzales v. United States, 589 F.2d 465, 469 (9th Cir. 1979).
The district judge found prejudice here in that the development permit issue was not raised until well over a year after the preliminary injunction had been issued and NOAA had gone to the expense of conforming its EIS to the district court’s order. The February 7 memo gave NOAA very little time to respond prior to the magistrate’s report. Given the length of this litigation and the ample opportunity which SLW had to press the development permit issue, we find no abuse of discretion. We also note that while NOAA did not formally object to the raising of the question before the district judge, neither did SLW formally move to amend its pleadings. The district court acted properly.
IV. CONCLUSION
The judgment of the district court is
AFFIRMED.
. On February 3, 1981, we ordered this case withdrawn from submission and requested the parties to brief the question of whether the NOAA’s revision of its plans rendered the controversy moot. At oral argument, we had sua sponte raised the possibility that continuing construction at the Sand Point site had mooted the appeal. The parties have submitted a stipulated construction schedule, and the NOAA has submitted a supplemental statement regarding the project’s current scope. These documents convince us that this appeal continues to present a live controversy. According to the schedule, dredging and shoreline excavation for the purpose of constructing the piers has been substantially completed. If we were examining only the potential damage from the dredging, then this appeal likely would be moot since irreversible change has already occurred. We note, however, that the piers are not scheduled to be available for occupancy until June 1982, and that SLW has alleged substantial environmental damage from the mere presence of oceangoing vessels on Lake Washington. Because all phases of the project are interrelated, and because the project still requires construction and the presence of some oceangoing ships on Lake Washington, we find that this appeal is not moot.