We are once again faced with environmental challenges to the proposed construction by the State of Hawaii of the remaining portion of Interstate Route H-3.1 In this skirmish, the appellants2 challenge on numerous grounds the appellees'3 approval of H-3, alleging violations of the National Environmental Policy Act of 1969, 42 U.S.C. Secs. 4321-4347 (1976 & Supp. V 1981) (NEPA), the Endangered Species Act of 1973, 16 U.S.C. Secs. 1531-1543 (1982) (ESA), the Department of Transportation Act of 1966, 49 U.S.C. Secs. 1651-1660 (1976 & Supp. V 1981) (DOTA), the Federal-Aid Highway Act of 1966, 23 U.S.C. Secs. 101-157 (1982) (FAHA), and various implementing regulations. The appellants appeal the District Court's Findings of Fact and Conclusions of Law, Stop H-3 Association v. Lewis, 538 F. Supp. 149 (D.Hawaii 1982), which denied many of their claims for declaratory and injunctive relief and which dissolved the injunctions against construction of H-3 that had been in place since 1972. The appeal is timely, and we have jurisdiction to consider the appeal under 28 U.S.C. Sec. 1291 (1982) and 28 U.S.C. Sec. 1292(a)(1) (1982). We affirm in part and reverse in part.

The principal issue in this appeal is whether the Secretary of Transportation (Secretary) has complied with section 4(f) of DOTA, 49 U.S.C. Sec. 1653(f), and section 18 of FAHA, 23 U.S.C. Sec. 138. (Both statutes, which essentially are identical,4 are hereinafter referred to simply as "section 4(f).")

Section 4(f) is part of Congress' response to the growing public concern over the preservation of our Nation's natural beauty. See Citizens to Preserve Overton Park v. Volpe, 401 U.S. 402, 404, 91 S. Ct. 814, 817, 28 L. Ed. 2d 136 (1971). In section 4(f), Congress has determined that the preservation of our parklands should be given major consideration in connection with all proposed highway construction programs that are to receive financial aid from the federal government. The statute provides, in declaring national policy, that "special effort should be made to preserve the natural beauty of the countryside and public park and recreation lands ...." The statute further provides that:

[T]he Secretary [of Transportation] shall not approve any project or program which requires the use of any publicly owned land from a public park ... of national, State, or local significance ... unless (1) there is no feasible and prudent alternative to the use of such land, and (2) such program includes all possible planning to minimize harm to such park ... resulting from such use.

23 U.S.C. Sec. 138; 49 U.S.C. Sec. 1653(f). It is obvious that the requirements of section 4(f) are stringent. Moreover, the implementing regulations promulgated by the Secretary pursuant to section 4(f) (4(f) regulations) require the Secretary to prepare and circulate a statement (4(f) statement) that must examine the highway's proposed use of parkland in light of the requirements of section 4(f). See 23 C.F.R. Sec. 771.19 (1980). The 4(f) regulations specifically require the 4(f) statement to analyze alternatives to the use of the parkland to determine whether the alternatives are feasible and prudent. See id.

In its proposed configuration, H-3 will use land from two public parklands: (1) Ho'omaluhia Park,5 a major regional park; and (2) Pali Golf Course Park,6 one of Oahu's most challenging and heavily used public golf courses. Because of H-3's use of the parklands, 4(f) statements were prepared in 1971 (approved by the Secretary in 1974) for Pali Golf Course Park and in 1979 (approved by the Secretary in 1980) for Ho'omaluhia Park. In response to the District Court's order, 538 F. Supp. at 184, the Pali Golf Course Park Section 4(f) Statement was supplemented in 1983. See Fed. Highway Admin., U.S. Dep't of Transp., Highways Div., State of Hawaii Dep't of Transp., Final Second Supplement to the Interstate Route H-3 Environmental Impact/4(f) Statement (1982), lodged with this Court on July 7, 1983.

All of the above mentioned 4(f) statements conclude that there is no feasible and prudent alternative to the use of Ho'omaluhia Park or to the use of Pali Golf Course Park.7 The Secretary concurred in that conclusion and the District Court held that the Secretary properly found there is no feasible and prudent alternative to the use of Ho'omaluhia Park.8 See 538 F. Supp. at 181, 183. The District Court also held that the Secretary reasonably rejected certain of the alternatives to the use of Pali Golf Course Park.9 The appellants challenge the Secretary's rejection of the alternatives to the use of Pali Golf Course Park and Ho'omaluhia Park as being unsupported by the record. They challenge the District Court's holdings on the same ground, as well as on the ground that the District Court's decision was made upon an erroneous application of Citizens to Preserve Overton Park v. Volpe, 401 U.S. 402, 91 S. Ct. 814, 28 L. Ed. 2d 136 (1971). After a thorough, probing, and in-depth review of the administrative record, we agree with the appellants' contentions in respect to the "Makai Realignment" alternative and the "No Build" alternative, and, accordingly, we reverse.10

As to all of the Secretary's section 4(f) determinations at issue in this case, the standard of judicial review is whether the Secretary's decision was "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law." Administrative Procedure Act, 5 U.S.C. Sec. 706(2)(A) (1982); Overton Park, 401 U.S. at 416, 91 S.Ct. at 823. While the Secretary's decisions are entitled to a presumption of regularity, that presumption does not "shield his action[s] from a thorough, probing, in-depth review." Overton Park, 401 U.S. at 415, 91 S.Ct. at 823. The reviewing court is required to consider whether:

1. The Secretary acted within the scope of his authority (not at issue in this case).

2. The Secretary properly construed his authority to approve the use of parkland as limited to situations where none of the alternatives to such use are feasible and prudent.

3. The Secretary could have reasonably believed that in the case under review there are no feasible and prudent alternatives.

4. The Secretary's decision was based on a consideration of the relevant factors.

5. The Secretary made a clear error of judgment.

6. The Secretary's action followed the necessary procedural requirements (not at issue in this case). Id. at 415-17, 91 S.Ct. at 823-24. See also Stop H-3 Association v. Coleman, 533 F.2d 434, 445 (9th Cir.) (the court, in reviewing the Secretary's decision, "must satisfy itself that the Secretary evaluated the highway project with the mandates of section 4(f) clearly in mind"), cert. denied, 429 U.S. 999, 97 S. Ct. 526, 50 L. Ed. 2d 610 (1976).

In its consideration of the Secretary's determination, the reviewing court must draw upon Overton Park's definition of a "feasible and prudent alternative":11 parklands may be "used" for highway purposes only if "there [are] truly unusual factors present in [the] case," if "feasible alternative routes involve uniquely difficult problems," or if "the cost or community disruption resulting from alternative routes [reach] extraordinary magnitudes." 401 U.S. at 413, 416, 91 S. Ct. at 822, 823.12

Moreover, the reviewing court should consider the full administrative record of the agency's action, id. at 420, and if the record fails to show a sufficient basis for the Secretary's decision, the 4(f) determination must be overturned, see id. In addition, the "reviewing court, in dealing with a determination or judgment which an administrative agency alone is authorized to make, must judge the propriety of such action solely by the grounds invoked by the agency. If those grounds are inadequate or improper, the court is powerless to affirm the administrative action by substituting what it considers to be a more adequate or proper basis." Securities & Exchange Commission v. Chenery Corp., 332 U.S. 194, 196, 67 S. Ct. 1575, 1577, 91 L. Ed. 1995 (1947).

In reviewing the District Court's decision affirming the Secretary's action, this Court should apply the same standards of review used by the District Court. See Southeast Alaska Conservation Council v. Watson, 697 F.2d 1305, 1312 (9th Cir.1983). The District Court's review is accorded no particular deference, because the District Court, limited to the administrative record, is in no better position to review the Secretary's action than is the Court of Appeals. See Arizona Past & Future Foundation v. Lewis, 722 F.2d 1423, 1425-26 (9th Cir.1983); Asarco, Inc. v. Environmental Protection Agency, 616 F.2d 1153, 1161 (9th Cir.1980). Thus, this Court may review the administrative record and determine for itself whether the Secretary's action was arbitrary, capricious, or an abuse of discretion. Southeast Alaska Conservation Council, 697 F.2d at 1312.

Bearing in mind the specified legal standards, we now turn to the appellants' contentions.

The appellants contend: (1) that the Ho'omaluhia Park Section 4(f) Statement, the Ho'omaluhia Park Section 4(f) Determination,13 and the rest of the administrative record fail to provide a basis for the Secretary to conclude that no feasible and prudent alternative exists to H-3's use of Ho'omaluhia Park, and (2) that the District Court erred in holding that the Secretary could have reasonably reached such a conclusion. Two of the rejected alternatives, the Makai Realignment and the No Build alternative, would have no impact upon either Ho'omaluhia Park or Pali Golf Course Park. The appellants argue that these alternatives have not been shown to be imprudent. We agree.

In the Makai Realignment, H-3 would be realigned to turn northward before reaching Ho'omaluhia Park and would follow the existing alignment of Likelike Highway and Kamehameha Highway from the Kaneoke Interchange to the Halekou Interchange. The H-3 traffic would merge with the Likelike Highway traffic, and, along Kamehameha Highway, H-3 would be on a viaduct with at-grade frontage roads underneath to permit cross-corridor movement for local residents.

As above noted, this alternative would avoid all use of both Ho'omaluhia Park and Pali Golf Course Park. The Secretary nonetheless rejected this alternative because:

[I]t would require the dislocation of one church, four businesses and 31 residences adjacent to Likelike and Kamehameha Highways; increase noise, air quality and visual impacts to residences in the general vicinity; require additional costs due to the need for the viaduct structure ($42 million additional); and require construction to lesser design geometric standards.

Ho'omaluhia Park Section 4(f) Determination, at 3.

The appellants argue that the above listed reasons do not represent the "unique problems," the "truly unusual factors," or the "cost or community disruption [reaching] extraordinary magnitudes" required by Overton Park. Indeed, the District Court specifically found that the displacements resulting from the Makai Realignment were not, by themselves, sufficient to justify use of the parkland. See 538 F. Supp. at 180. Nevertheless, the District Court found that the Secretary could reasonably have believed that the sum of the listed factors rendered the alternative imprudent. See id. The court based its finding on the proposition that "Overton Park does not bar considering whether all of the difficulties posed by an alternative route, taken together, render that alternative imprudent." Id.14 The only "factor" the District Court discussed was "construction to lesser geometric standards." See id. After a painstaking and thorough review of the record, we conclude that we cannot affirm the District Court's decision because the reasons for finding the Makai Realignment imprudent advanced by the Secretary in his 4(f) determination do not satisfy the stringent Overton Park standards that we must apply.

The first three reasons need not long detain us, for these are displacements that one would normally expect might happen in following Overton Park. The dislocation of one church, four businesses and thirty-one residences no doubt is a community disruption of some magnitude. We do not believe, however, that this disruption is of the "extraordinary" magnitude required by Overton Park.15 In Overton Park the Supreme Court stated:

[S]ince people do not live or work in parks, if a highway is built on parkland no one will have to leave his home or give up his business. Such factors are common to substantially all highway construction. Thus, if Congress intended these factors to be on an equal footing with preservation of parkland, there would have been no need for the statutes.... But the very existence of the statutes indicates that protection of parkland was to be given paramount importance.

Overton Park, 401 U.S. at 412-13, 91 S.Ct. at 821-22 (footnote omitted). We find that the Secretary could not have reasonably concluded that the community displacements resulting from the Makai Realignment rose to the level required by Overton Park.

Likewise, the increased cost of $42 million (1978 dollars) is not a cost of extraordinary magnitude, especially in light of the projected total cost of H-3--$386 million (1979 dollars), see NHV-SEIS, vol. I, at 37. This is not to say that $42 million is not a considerable sum of money; however, when the taking of parkland is involved, "cost is a subsidiary factor in all but the most exceptional cases." Coalition for Responsible Regional Development v. Brinegar, 518 F.2d 522, 526 (4th Cir.1975). We hold that the Secretary could not have reasonably concluded that the increased cost of the Makai Realignment was of the "extraordinary magnitude" required by Overton Park.16

As to the third reason--increased noise, air quality and visual impacts to residences in the general vicinity--there is nothing in the record to show that this factor represents a disruption of extraordinary magnitude. Overton Park amply made clear that only in the most exceptional cases may parkland be taken solely to prevent highways from adversely affecting areas that are already developed. See Overton Park, 401 U.S. at 412-13, 91 S.Ct. at 821-22. We are not convinced that the Secretary could have reasonably concluded that this is one of those exceptional cases.

This brings us to the fourth and final reason--the only reason that we find even somewhat troubling--"lesser design geometric standards." "Lesser design geometric standards" has been translated in the context of this case to mean "safety considerations." See 538 F. Supp. at 180.

At the outset, we note that there appears to be a dearth of case law that specifically addresses safety issues as they relate to the taking of 4(f) land.17 Nonetheless, in Overton Park the Supreme Court emphatically stated:

[The defendants] contend that the Secretary should weigh the detriment resulting from the destruction of parkland against the cost of other routes, safety considerations, and other factors, and determine on the basis of the importance that he attaches to these other factors whether, on balance, alternative feasible routes would be "prudent." ...

[N]o such wide-ranging endeavor was intended .... [I]f Congress intended these factors to be on an equal footing with preservation of parkland there would have been no need for the [4(f) ] statutes .... [T]he very existence of the statutes indicates that protection of parkland was to be given paramount importance.

Overton Park, 401 U.S. at 411-13, 91 S.Ct. at 821-22 (footnote omitted) (emphasis supplied).

It seems obvious to us that safety considerations, since they so directly involve human life, warrant extremely close scrutiny when determining whether such considerations satisfy the Overton Park standards. Neither a court nor an agency should weigh lightly the potential risk to human life an alternative might pose. On the other hand, undue deference to a prior pronouncement that an alternative is undesirable because of safety considerations would transform such a pronouncement into a "talisman." For these reasons, there is a need for an especially "thorough, probing, and in-depth review" when safety issues are presented for review. In the case at hand, we have conducted just this sort of review, and we find that the record before the Secretary could not have provided a sufficient basis for him to conclude reasonably that the safety considerations of the Makai Realignment were "truly unusual factors," that they reflected "unique problems," or that they represented cost or community disruption reaching "extraordinary magnitudes."

The Ho'omaluhia Park Section 4(f) Statement, upon which the Secretary relied, see 538 F. Supp. at 178-79, mentions "safety considerations" a scant three times:

....

(e) Traffic movements will be complex due to the high volume of H-3 traffic to be funneled into Likelike Highway and the short distance between the Kaneohe Interchange, Kaheliki Interchange, and Kamehameha Highway.

(f) Undesirable curves for H-3/Likelike Highway movements. Design speeds on the through route will have to be reduced from 55 MPH to 30 MPH to negotiate the ramp curves safely and comfortably.

....

(j) ... The loop ramp configuration at Kaneohe Interchange is required, because of grade differences, and is unusual for a through highway connection. The confusing configuration, coupled with the high volume of merging and weaving traffic from H-3 plus Likelike Highway (3,300 v.p.h. on H-3 plus 3,300 v.p.h. on Likelike) is undesirable because of safety considerations.

Ho'omaluhia Park Section 4(f) Statement, at 20-21.

Two points should be noted regarding the traffic density figures cited in the Ho'omaluhia Park Section 4(f) Statement. First, the capacity of the Likelike Highway is 3650 vph (vehicles per hour) and the capacity of H-3 is projected to be 3900 vph. See NHV-SEIS, vol. III, app. B, at 29-30. Therefore, even at peak rush hour, the two highways will be operating at less than capacity. See id. at 31. Second, the 3300 vph figure was based on an outdated population projection for the Windward side of 150,000; the most recent official population projections forecast a Windward side population of between 125,700 and 138,500. See 538 F. Supp. at 166.

The District Court, in finding that the Secretary properly rejected the Makai Realignment as imprudent, expressly relied upon the "safety considerations" of this alternative: "In particular, the Makai Realignment would necessitate reducing design speeds on the through route from 55 MPH to 30 MPH and require an unusually complex and unsafe ramp configuration." Id. at 180.

A close examination of the record, however, reveals that the above mentioned speed reduction would be necessary only at one interchange, and then only to negotiate the "exit" ramp safely--a total distance of less than one mile. See, e.g., NHV-SEIS, vol. I, at fig. III-9. In other words, the ramp configuration is not per se unsafe as suggested by the District Court--the traffic on H-3 merely would have to slow down to use the "exit" ramp.18 The conclusion that the ramp configuration, or the rest of the Makai Realignment for that matter, is not per se unsafe is further supported by the fact that there is nothing in the record to indicate that the Makai Realignment in any way fails to meet the minimum criteria for safety standards set forth by the Federal Highway Administration at 23 C.F.R. Sec. 625.3(a)(3) (1979) (Am. Ass'n of State Highway & Transp. Officials, Geometric Design Standards for the National System of Interstate and Defense Highways (1967)).

In addition, the Likelike Highway, a little over one mile from its proposed interchange with H-3, contains a curve of 40 mph design speed. See Fed.Highway Admin., U.S. Dep't of Transp., Region 9 Staff Analysis: Interstate H-3 and Existing Trans-Koolau Highway Alternatives 13 (1979) [hereinafter referred to as "Region 9 Staff Analysis"]. The current analyses do not investigate adequately the extent to which this may reduce the difference in speeds between the merging traffic from H-3 and the traffic on the Likelike Highway.

Moreover, the Pali Highway contains "substandard curves" on its Windward portion. See id. at 11. This, when considered together with the existence of a 40 mph design speed curve on the Windward portion of the Likelike Highway, indicates to us that, in the absence of further evidence suggesting a contrary conclusion, the existence of a ramp on H-3 with a curve of 30 mph design speed does not represent an "unusual situation" or a community disruption of "extraordinary magnitude." In other words, since trans-Koolau commuters currently are faced with travelling on highways that contain curves similar to, if not worse than, the curved ramp on H-3, we do not see that the H-3 curve will extraordinarily disrupt the community or will present an unusual situation such that the taking of 4(f) parklands is warranted.

The record, then, paints the following picture. Emptying H-3's traffic onto a curved off-ramp, decreasing the speed of that traffic from 55 mph to 30 mph, and merging that traffic with the traffic on the Likelike Highway probably raises more safety concerns than the less circuitous route of H-3's recommended alignment. The Makai Realignment, like other highways commonly in use at present, probably presents a safety risk of some magnitude. The question is, however, whether this situation presents a safety risk of the magnitude required by Overton Park. The problem is that the record does not illustrate what magnitude of risk this alternative in fact poses19 and, consequently, does not support adequately the Secretary's conclusion that the alternative is imprudent because of safety considerations.20 In other words, the record in this case does not provide a sufficient basis for the Secretary to have reasonably concluded that the safety considerations of the Makai Realignment were of such a magnitude as to overcome the paramount importance given to the protection of parkland. See Overton Park, 401 U.S. at 412-13, 91 S.Ct. at 821-22. Moreover, we note that the District Court did not rely on safety reasons, per se, as the basis for upholding the Secretary's rejection of the Makai Realignment. Rather, it found it necessary to cumulate safety concerns with other unrelated factors in order to do so. Thus, even accepting the District Court's view of the record, the safety concerns would be insufficient to warrant affirmance of the Secretary's findings.

In conclusion, we have examined against the record the four reasons advanced by the Secretary, and we find that the specified reasons, even when amalgamated, are insufficient to support a determination that the Makai Realignment is imprudent. Therefore, because the Secretary could not have reasonably believed that no feasible and prudent alternative exists to the use of Ho'omaluhia Park, his approval of H-3 was an abuse of discretion. See id. at 415-17, 91 S. Ct. at 823-24. Accordingly, we must reverse the District Court's judgment affirming the Secretary's action and remand to the District Court for that court to remand to the Secretary for a more comprehensive 4(f) determination considering sufficiently the Makai Realignment.

We stress that we do not find, nor is it this Court's role to find, that the Makai Realignment is in fact a feasible and prudent alternative. We obviously do not possess the technical expertise of roadbuilders, and we should not interfere in the technical processes of building roads. At the same time, however, it remains our solemn responsibility to insure that those with technical expertise exercise it in accordance with the laws of the United States and the public welfare. See Citizens to Preserve Overton Park v. Volpe, 432 F.2d 1307 at 1318 (6th Cir.1970) (Celebrezze, J., dissenting). The record before us simply does not demonstrate that the stringent requirements of section 4(f), as defined in Overton Park and its progeny, have been satisfied. Until those requirements are satisfied, we cannot allow our Nation's sacred parklands to be taken or used.

The No Build alternative entails not constructing the portion of H-3 that runs between the Halekou Interchange and the Halawa Interchange. See 538 F. Supp. at 180; Ho'omaluhia Park Section 4(f) Determination, at 4.

The record discloses that the Secretary based his rejection of the No Build alternative on four reasons:21

1. The alternative would require that thirty-one additional buses be purchased to meet year 2000 trans-Koolau travel demand at a total purchase cost of $3.3 million (1977 dollars) and annual operating costs, defrayed by fares, of $1.19 million (1977 dollars). See Ho'omaluhia Park Section 4(f) Determination, at 4; Ho'omaluhia Park Section 4(f) Statement, at 24.

2. The alternative would result in traffic congestion and increased delays experienced by Windward commuters. Ho'omaluhia Park Section 4(f) Determination, at 4.

3. The alternative would result in increased safety hazards on Likelike and Pali Highways which would directly affect Kalihi Valley and Nuuanu Valley residents. Id.

4. The costs of providing increased bus service which will not effectively reduce the congestion on the existing highways are documented in the supplement to the Interstate H-3 EIS. Id.

We hold that these four reasons, when viewed against the record, do not establish that the Secretary could reasonably conclude that the No Build alternative must be rejected as imprudent.

First, we can discern no basis in the record for the Secretary to conclude reasonably that the purchase and operation of thirty-one additional buses would cause "unique problems" or that the cost would be of "extraordinary magnitude." Moreover, under 23 U.S.C. Sec. 103(e)(4) (1982), the cost could be met by transferring funds already allocated for H-3 construction to purchase and operate the buses. The record, however, lacks an adequate analysis of the effect the transferability of such funds has on the prudence of the No Build alternative. Cf. Benton Franklin Riverfront Trailway & Bridge Committee v. Lewis, 701 F.2d 784, 790-91 (9th Cir.1983) (section 4(f) determination that no feasible and prudent alternatives exist held invalid because there was no consideration of the "potential of federal funds").

Second, we are not wholly convinced that the record clearly demonstrates that the increased congestion or commuter delays projected for the year 2000 would be so unusual or extraordinary that the No Build alternative must be rendered imprudent. The Pali Highway currently is operating during peak hour at 3000 vph, with the Likelike operating during peak hour at 3100 vph. NHV-SEIS, vol. III, app. B, at 30. There is, therefore, a current unused peak hour capacity of 1200 vph available to meet future demand. See id. Indeed, the NHV-SEIS reveals that, if H-3 is not built, the projected year 2000 peak hour demand (7300 vph) can be met by the present combined capacity of the Pali Highway (3650 vph) and the Likelike Highway (3650 vph).22 See id. at 29-31.

In addition, the population projections for Windward Oahu originally used for H-3 planning were revised downward in 1978 as part of the changes in the Oahu General Plan that redirected growth from Windward Oahu to the central plain of Leeward Oahu. See 538 F. Supp. at 166-67. These revisions were made before the Ho'omaluhia Park Section 4(f) Statement was prepared. Rather than analyze the projected congestion and commuter delays in light of the revised population projections, however, the Secretary and the FHWA chose to ignore the revised projections and to continue to use the pre-1978 figures. See id. Even the District Court noted that "[i]t would have been wiser ... to have considered whether [H-3] would still be viable if [the new] population goals are met." Id. at 167.

The question the Secretary failed to address, then, is whether the current unused capacity of trans-Koolau highways combined with the lower growth projections for Windward Oahu will prevent year 2000 congestion and commuter delays from becoming "truly unusual factors," from becoming "most unusual situations," or from disrupting the community to an "extraordinary" degree. In our view, this is a question that must be answered before determining that the No Build alternative is imprudent.

Moreover, the Region 9 Staff Analysis states that the rush hour capacity of the Pali and Likelike Highways could be increased by fifty percent simply by banning trucks during rush hour. Region 9 Staff Analysis, at 45. This information does not appear to have been included in any analysis of current or future congestion.23

Finally, our review of the Region 9 Staff Analysis reveals that of the sixteen alternatives studied therein, two non-H-3 (no build) alternatives have nearly identical projected operational characteristics as H-3: (1) the same peak hour volume to capacity ratio, (2) the same projected congestion, and (3) similar projected "accidents & severity." Id. at 14. We note also that these two non-H-3 alternatives offer the following advantages over H-3: (1) they have environmental impact ratings that are over three times as favorable as H-3's, (2) they increase the use of mass transit,24 and (3) they cost only one-thirtieth of the cost of H-3. Id. There is no analysis in the record as to why these two alternatives should be rejected--or why they are any less prudent than H-3 in terms of congestion. In conclusion, we are not convinced that the No Build alternative must be rejected as imprudent because of traffic congestion and increased commuter delays.

Third, except for the following excerpt, the "increased safety hazards" are not substantiated or discussed in the record: "[The increased traffic on the Likelike and Pali Highways] is resulting in increased congestion and safety hazards on Likelike and Pali Highways which will have a direct effect on the residents of the valleys through which these two facilities traverse." Ho'omaluhia Park Section 4(f) Statement, at 24. This bald statement does not seem to us to provide sufficient support for the Secretary's conclusion that the No Build alternative is imprudent because of safety considerations. On the other hand, the Region 9 Staff Analysis indicates that all of the various permutations of the No Build alternative pose more safety hazards than H-3 poses. The problem is that the Region 9 Staff Analysis does not specifically mention the Nuuanu and Kalihi Valleys in this respect; in addition, for the majority of the No Build permutations, the increase in safety hazards is slight--up one rating from "minor" to "moderate" "accidents & severity." Region 9 Staff Analysis, at 14. And, if the revised population figures are used, projected congestion will likely be decreased with a concomitant decrease in projected safety hazards. We are not convinced that the present record sufficiently supports a conclusion that the No Build alternative is imprudent because of the "increased safety hazards to residents of the Nuuanu and Kalihi Valleys."

Fourth, we are uncertain as to the significance, if any, of the Secretary's fourth "reason"--that "[t]he costs of providing increased bus service which will not effectively reduce the congestion on the existing highways have been documented in the supplement of the Interstate H-3 EIS." The costs of the increased bus service indeed are documented in the supplemental EIS. As above discussed, however, the relatively modest cost of the increased bus service does not provide a sufficient basis for the Secretary to conclude that the No Build alternative is imprudent. As to the failure to reduce congestion, we note that at least four of the rejected No Build alternative's permutations (including the alternative of banning trucks at peak rush hour) are projected to have identical levels of congestion as that of H-3. Id. The inference is that if the No Build alternative should be held imprudent because of its failure to reduce congestion, so should H-3 for the same reason.25 Moreover, the inference from the record is that year 2000 traffic demand can be met by increased bus service alone. See Ho'omaluhia Park Section 4(f) Statement, at 23-24; NHV-SEIS, vol. I, at 272. And, again, if the lower revised population figures and the banning of trucks are considered, the effectiveness of increased bus service in reducing congestion likely would be enhanced.

In conclusion, then, it is our view that the present record does not support a determination that the No Build alternative must be rejected as imprudent. We emphasize that, as with the Makai Realignment, we do not hold that the No Build alternative is, in fact, reasonable and prudent. We merely hold that the record before us does not demonstrate that the stringent requirements of section 4(f) have been satisfied. Therefore, we reverse the District Court's judgment affirming the Secretary's action and remand to the District Court for that court to remand to the Secretary for a more comprehensive 4(f) determination considering sufficiently both the Makai Realignment and the No Build alternative.

A. FACTUAL BACKGROUND AND PROCEDURAL HISTORY

The appellants also challenge the appellees' compliance with the ESA.

On March 9, 1978, the Federal Highway Administration (FHWA), pursuant to the ESA, initiated formal consultation with the United States Fish and Wildlife Service (USFWS) with regard to the potential impact of the H-3 project on the Oahu Creeper.26 The Oahu Creeper is an extremely rare species of bird; it was officially listed as endangered on October 13, 1970. See 50 C.F.R. Sec. 17.11 (1980). As part of the consultation process, FHWA authorized several avifaunal surveys to produce current information on the Creeper. The surveys were conducted by Dr. Robert Shallenberger and confirmed the presence of Creepers in the North Halawa Valley.

Based on these surveys, USFWS issued its biological opinion, dated September 11, 1978. The opinion discusses Dr. Shallenberger's surveys and concludes with the following statement:

In essence, we have very little data for providing an opinion, but feel it would be unreasonable to request [an additional] study which would be unlikely to provide definitive results. We must, therefore, assume the Oahu Creeper would be like most species in that a highway would not split a population.

Based on the available information, which we grant is weak, it is our opinion the proposed project is not likely to jeopardize the continued existence of the Oahu Creeper.

Also, in a biological opinion on the Achatinella, dated October 7, 1981, USFWS states that FHWA need not reinitiate consultation on the Creeper.

In district court, the appellants challenged the adequacy of the biological opinion and charged that FHWA must reinitiate consultation with USFWS. In its Order Denying Plaintiffs' Motion for Partial Summary Judgment as to the Twenty-Sixth Cause of Action, the District Court ruled that the facts showed: (1) USFWS could reasonably conclude that the H-3 project is not likely to jeopardize the continued existence of the Creeper, and (2) based on 50 C.F.R. Sec. 402.04 (1980),27 no further consultation was required. In addition, the District Court limited the scope of review at trial to the issue of: "Whether or not the USFWS has requested that [appellees] reinitiate formal consultation regarding the Oahu Creeper ...."

At trial Dr. Shallenberger testified to the presence of the Oahu Creeper in North Halawa Valley, to the significance of the valley to the survival of the species, and that the H-3 freeway "is likely" to jeopardize the Creeper's continued existence. The appellants attempted to use Dr. Shallenberger's testimony to show that FHWA's decision to grant location and design approval for H-3 was arbitrary and capricious in light of H-3's potential impact on the Creeper. The appellees moved to strike the testimony of Dr. Shallenberger on the ground that USFWS had informed the appellees that further consultation on the Creeper was not required. The court granted the motion and, ultimately, ruled that FHWA had complied with the ESA.

The appellants allege that it was improper for the District Court to strike Dr. Shallenberger's testimony, since it precluded inquiry into FHWA's substantive decision to grant location and design approval for H-3. They argue that, by limiting the scope of review to the procedural issue of consultation with USFWS, the District Court failed to examine FHWA's decision under the proper legal standard. Thus, the appellants urge this Court to conclude that FHWA's grant of location and design approval was arbitrary, capricious, an abuse of discretion, and not in accordance with the Endangered Species Act on the ground that FHWA failed to make a rational determination, based on the best available scientific data, that H-3 is or is not likely to jeopardize the existence of the Creeper.

The appellees respond that the District Court properly struck Dr. Shallenberger's testimony because the testimony did not offer any information that had not already been considered by the USFWS before issuing the biological opinion. The appellees argue that the administrative record, on its face, provides adequate support for the conclusions reached in the biological opinion, and, therefore, there was no reason to go beyond the administrative record to consider Dr. Shallenberger's testimony. Finally, the appellees contend that they have complied with the requirements of the ESA, that they properly deferred to the USFWS's biological opinion, and that, therefore, their decision to grant location and design approval for H-3 was not arbitrary and capricious.

Under the ESA, FHWA has a duty to "insure" that its action "is not likely to jeopardize"28 the continued existence of the Creeper. See 16 U.S.C. Sec. 1536(a)(2) (1982). To this end, FHWA is required to consult with an expert agency (here, the USFWS). See 50 C.F.R. Sec. 402.01 (1980). Using the "best scientific and commercial data available," the expert agency is required to issue a biological opinion to FHWA. See 16 U.S.C. Sec. 1536(a)(2).

Here, since FHWA has complied with the consultation requirements, the question is whether FHWA's decision to rely on USFW's biological opinion in granting location and design approval for H-3 was valid. FHWA's decision is subject to review under the "arbitrary and capricious" standard of the Administrative Procedure Act, 5 U.S.C. Sec. 706(2)(A) (1982). See Village of False Pass v. Clark, 733 F.2d 605, 609-10 (9th Cir.1984). Thus, the issue for review is whether the decision was based on a consideration of the relevant factors and whether there has been a clear error of judgment. See Overton Park, 401 U.S. at 416, 91 S.Ct. at 823.

In the case at hand, the District Court did not review FHWA's decision under the "arbitrary and capricious" standard. Rather, the District Court limited its review to the substantive issue of the adequacy of USFWS's biological opinion and the procedural issue of consultation with USFWS; the court did not examine FHWA's decision to rely on the biological opinion. The court limited its review in this manner apparently because the appellants' complaint, in respect to this cause of action, was couched in language that challenged USFWS's substantive compliance with the ESA and language that challenged FHWA's procedural compliance with the ESA. Appellants made clear at trial, however, that the gravamen of the cause of action was a challenge to FHWA's substantive compliance with the ESA--a challenge that the District Court should have evaluated under the "arbitrary and capricious" standard. Thus, it appears the District Court erred by not applying the proper legal standard.

Remand is not required on this issue, however, because we are able to determine from the record that as a matter of law FHWA's decision to rely on USFWS's biological opinion was not "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law." See Southeast Alaska Conservation Council, 697 F.2d at 1312 ("[T]he appellate court may review the administrative record and determine for itself whether the [agency's action] was arbitrary, capricious, or an abuse of discretion."); cf. Asarco, Inc. v. Environmental Protection Agency, 616 F.2d at 1161 ("District court review of agency action is generally accorded no particular deference, because the district court, limited to the administrative record, is in no better position to review the agency than the Court of Appeals."). It is clear that FHWA complied with all initial consultation obligations and relied on an opinion issued by an expert agency. On its face, this does not seem to be a "clear error of judgment." Furthermore, while the information used to form the opinion admittedly was weak, the expert agency determined on two separate occasions that no further inquiry was necessary.

The appellants are correct when they argue that FHWA cannot abrogate its responsibility to decide whether it has taken all possible action to insure that H-3 is not likely to jeopardize the continued existence of the Creeper. See National Wildlife Federation v. Coleman, 529 F.2d 359, 371 (5th Cir.), cert. denied, 429 U.S. 979, 97 S. Ct. 489, 50 L. Ed. 2d 587 (1976). But, no improper abrogation of responsibility occurred here. Dr. Shallenberger's testimony may challenge the conclusions contained in the biological opinion, but that testimony offered no information that had not already been evaluated by the expert agency. Cf. 50 C.F.R. Sec. 402.04(h)(1) (1980) (reinitiation of consultation required when new information comes to light). In these circumstances, it was not unreasonable for FHWA to rely on the expert agency's opinion in deciding not to make a separate opinion regarding the Creeper. Thus, FHWA's ultimate conclusion that, based on the best available scientific data, H-3 is not likely to jeopardize the existence of the Creeper clearly was grounded on "a consideration of the relevant factors" and, not being unreasonable as a matter of law, was not a "clear error of judgment." We hold, therefore, that the appellees have complied with the mandate of the ESA.

III. NATIONAL ENVIRONMENTAL POLICY ACT AND FEDERAL-AID HIGHWAY ACT

The appellants also contend that the appellees have not complied sufficiently with NEPA or, in addition to the alleged noncompliance with section 4(f), with FAHA. We are not persuaded by the appellants' arguments in respect to this contention and will only briefly discuss the pertinent issues.

The appellants challenge the adequacy of the Environmental Impact Statement (EIS) on three grounds: (1) that the EIS inadequately assess the secondary (socio-economic) effects of H-3, (2) that the EIS inadequately analyzes whether H-3 is consistent with local land use plans, and (3) that the EIS must be supplemented to include discussions of significant new information. The EIS for the North Halawa Valley alignment of H-3 consists of the 1972 EIS, 1973 EIS Preface, and the NHV-SEIS.

In our Circuit, a district court's finding that an EIS is adequate will be reversed only if based upon an erroneous legal standard or upon clearly erroneous findings of fact. Save Lake Washington v. Frank, 641 F.2d 1330, 1334 (9th Cir.1981). The district court's review of an EIS also is limited:

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. Sec. 706(2)(D), which directs courts to set aside an agency action if taken 'without observance of procedure required by law ....'

Id. (citations omitted). Under this standard, the court employs a "rule of reason" that inquires: (1) whether the EIS contains "a reasonably thorough discussion of the significant aspects of the probable environmental consequences," Village of False Pass v. Clark, 733 F.2d at 613; Trout Unlimited, Inc. v. Morton, 509 F.2d 1276, 1283 (9th Cir.1974); and (2) whether the EIS's "form, content and preparation foster both informed decision-making and informed public participation," Village of False Pass v. Clark, 733 F.2d at 613; California v. Block, 690 F.2d 753, 761 (9th Cir.1982). Once satisfied that the agency has taken this procedural and substantive "hard look" at environmental consequences in the EIS, see Kleppe v. Sierra Club, 427 U.S. 390, 410 n. 21, 96 S. Ct. 2718, 2730 n. 21, 49 L. Ed. 2d 576 (1976); Village of False Pass v. Clark, 733 F.2d at 613, the court's review is at an end.

The role of the reviewing court under NEPA, then, differs fundamentally from the role of the reviewing court under section 4(f). While the mandate of section 4(f) essentially is prohibitory, the mandate of NEPA essentially is procedural. The court's review, therefore, is much more limited under NEPA than under section 4(f). In the case at hand, the reviewing court should consider only whether, under the Council on Environmental Quality regulations implementing NEPA, the H-3 EIS satisfies the above specified "hard look" standard.

The Council on Environmental Quality (CEQ), established under 42 U.S.C. Sec. 4342 (1976), promulgates uniform, mandatory regulations for implementing the procedural provisions of NEPA. See Andrus v. Sierra Club, 442 U.S. 347, 357, 99 S. Ct. 2335, 2340, 60 L. Ed. 2d 943 (1979); Village of False Pass v. Clark, 733 F.2d at 613; 40 C.F.R. Sec. 1515.2 (1982). Under CEQ regulation 40 C.F.R. Sec. 1500.8(a)(3)(ii) (1978), an EIS must assess and discuss the secondary (socio-economic) effects of the project in question.29 In the case at hand, the District Court found that the EIS adequately discusses the socio-economic impacts of H-3. 538 F. Supp. at 166. Based on our review of the relevant case law and the record, we cannot conclude that the District Court's finding was "clearly erroneous."

The appellants allege that the EIS fails to assess adequately the secondary effects of H-3 on population growth, public services, and community cohesion and stability of Windward Oahu. The appellants essentially are arguing that the EIS discussion of secondary impacts lacks sufficient detail and meaningful supporting data. To support this argument, the appellants cite two cases, City of Davis v. Coleman, 521 F.2d 661 (9th Cir.1975), and Coalition for Canyon Preservation v. Bowers, 632 F.2d 774 (9th Cir.1980), in which EIS's were found inadequate. In both cases, however, the EIS's were not nearly as detailed as the EIS in the case at hand. Here, while there are some "general" discussions and "assumptions" in the EIS, and while it may have been preferable to consider the secondary impacts in more detail, it is our view that the EIS contains reasonably sufficient data for decisionmakers to take the requisite "hard look" at the environmental consequences of H-3 and to reach subsequently an environmentally-informed and independent conclusion about H-3.

First, an "H-3 Socio-Economic Study" was prepared and circulated in 1973. The 1973 study briefly discusses the socio-economic impacts of H-3 on Windward Oahu. The appellants charge that the 1973 study is obsolete in light of the new Oahu General Plan.

Second, even if portions of the 1973 study are obsolete, the NHV-SEIS, prepared after the new Oahu General Plan went into effect, adequately updates the 1973 study. The NHV-SEIS contains a reasonably thorough discussion of H-3's secondary impacts in light of the planning changes that have occurred. See NHV-SEIS, vol. I, at 118-29.

Third, the NHV-SEIS, in a reasonably thorough fashion, discusses socio-economic phenomena in Windward region-wide terms, as well as island-wide terms. See, e.g., id. at xv-xxiv, 45-55, 100-13, 118-29, 316-17.

Fourth, the NHV-SEIS relies upon, among other things, the conclusions and data developed by the City and County of Honolulu in connection with the updating of the Oahu General Plan in 1977. See id. at 13-17, 49 (table III-3), 100-13.

On the other hand, the District Court found that the appellees put forth contradictory assertions in respect to the ability of the General Plan to control H-3 induced growth. See 538 F. Supp. at 166. The appellants are correct when they point out that this may reflect a less than complete evaluation of H-3's secondary impacts. Nonetheless, NEPA only requires a "reasonably thorough discussion" that "fosters informed decisionmaking," not a "complete evaluation." Therefore, it is our view that the District Court was not "clearly erroneous" in finding that the EIS assesses and discusses adequately H-3's socio-economic impacts.

The regulations of both the Department of Transportation (DOT) and CEQ require the EIS to analyze the relationship of H-3 to local land use plans and to discuss how H-3 "may conform or conflict with the objectives and specific terms" of land use plans, policies, and controls for the area. See 23 C.F.R. Sec. 771.18(h) (1978); 40 C.F.R. Sec. 1500.8(a)(2) (1978). If a conflict or inconsistency exists, the EIS "should describe the extent of reconciliation and the reason for proceeding notwithstanding the absence of full reconciliation." 23 C.F.R. Sec. 771.18(h) (1978); 40 C.F.R. Sec. 1500.8(a)(2) (1978).

On January 18, 1977, the City and County of Honolulu adopted a revised Oahu General Plan, which became law on February 2, 1977. See NHV-SEIS, vol. I, at 100. The obsolete Oahu General Plan had envisioned large-scale development and population growth, including a deep draft harbor, for Windward Oahu. The new 1977 Oahu General Plan altered significantly the planning objectives for Windward Oahu. See 538 F. Supp. at 165. The new Plan envisions limited growth for the region and calls for a reduction in the proportion of Oahu's population living in the region. Id. The appellants strongly urge that: (1) H-3 is inconsistent with the population objectives and policies of the 1977 Plan,30 (2) the inconsistencies are not resolved in the EIS, and, therefore, (3) the EIS is inadequate.

Indeed, the portion of the EIS that we find the most troubling is its analysis of the 1977 General Plan. For example, the appellees, in both their H-3 Travel Demand Analysis and Region 9 Staff Analysis, use outdated Windward region population projections, apparently assuming that the 1977 General Plan population goals will not be met. See 538 F. Supp. at 166-67. As the District Court points out, "this assumption contradicts [the appellees'] assertion that growth will be limited by the General Plan." Id. This would seem to indicate that the appellees may not have reasonably concluded that H-3 is consistent with the 1977 Plan. Moreover, in the NHV-SEIS the appellees assert several times that H-3 was planned only in response to the projected and desired growth pattern in Windward Oahu that appears in the 1977 General Plan. See, e.g., NHV-SEIS, vol. I, at 87, 90, 103. The H-3 freeway, however, was under development when the obsolete 1964 General Plan was in effect, see, e.g., 1973 EIS Preface, at 5-1 to 5-12, app. B, at 69, and at that time the appellees asserted that H-3 was consistent with the then current planning and growth policies, see id. app. B, at 211-292. It seems incongruous to us, therefore, that the appellees assert that H-3 is only a response to the 1977 General Plan when H-3 originally was designed to help implement the 1964 General Plan vision of an urbanized, industrial Windward Oahu.

Nonetheless, our role is not that of a "super-planner," see 538 F. Supp. at 164, and, under NEPA, we are not allowed to substitute our judgment for that of the agency concerning the wisdom of a proposed action. See California v. Block, 690 F.2d at 761. Our role is limited to insuring that the appellees have taken a "hard look" at H-3's environmental consequences. The NHV-SEIS contains a fairly detailed discussion of H-3's relationship to state and city land use plans, policies, controls, goals, and objectives. See NHV-SEIS, vol. I, at 85-114, 126-29. Furthermore, the relationship between H-3 and the 1977 Plan specifically is discussed. Id. at 100-10. And, one of the terms upon which the appellee-Secretary's concurrence in the EIS was conditioned was:

That the Hawaii Department of Transportation will cooperate with the City and County of Honolulu in monitoring land use and development trends on Windward Oahu, including the impact of H-3 on such trends. Hawaii DOT will cooperate with the City and County in the implementation of measures proposed to achieve General Plan objectives for Windward Oahu.

Office of the Secretary, U.S. Dep't of Transp., Concurrence Memorandum 3 (Nov. 21, 1980). See also Fed. Highway Admin., Decision Memorandum 4 (Dec. 5, 1980) ("the conditions required by the Office of the Secretary ... are accepted and will be implemented by FHWA and the Hawaii DOT").

Thus, while it may have been preferable to include a comparison study of the interrelationship between H-3, the 1964 Plan, and the 1977 Plan, we cannot conclude in these circumstances that the District Court's decision was "clearly erroneous" when it found that the EIS contains adequate information for the relevant decisionmakers to reach a reasoned conclusion concerning H-3's conformity with land use planning on Oahu.

a. Standard of Review

A federal agency has a continuing duty to gather and evaluate new information relevant to the environmental impact of its actions, even after release of an EIS. Warm Springs Dam Task Force v. Gribble, 621 F.2d 1017, 1023-24 (9th Cir.1980) (Warm Springs Dam II ). The CEQ regulations require that agencies "prepare supplements to either draft or final environmental impact statements if ... [t]here are significant new circumstances or information relevant to environmental concerns and bearing on the proposed action or its impacts." 40 C.F.R. Sec. 1502.9(c)(1)(ii) (1980).

An agency's decision not to supplement an EIS will be upheld if it was reasonable. Warm Springs Dam II, 621 F.2d at 1024. When new information comes to light the agency must consider it, evaluate it, and make a reasoned determination whether it is of such significance as to require implementation of formal NEPA filing requirements. Id. Reasonableness depends on the environmental significance of the new information, the probable accruacy of the information, the degree of care with which the agency considered the information and evaluated its impact, and the degree to which the agency supported its decision not to supplement with a statement of explanation or additional data. Id.

The appellants allege that the 1977 General Plan, revised population projections, and 1980 census data are "significant new information" requiring EIS supplementation.

b. 1977 New General Plan

The appellants assert that the "complete reversal" of Windward planning goals, as embodied by the 1977 Plan, is a "significant new circumstance" requiring further supplementation of the 1973 EIS. They argue that the shift of planned population centers from Windward Oahu to Ewa should be specifically addressed in a supplement to the EIS.

This argument has little merit. The 1977 General Plan is not "new information" that has come to light after release of the EIS. The relationship between H-3 and the 1977 General Plan objectives and policies is addressed in some detail in the NHV-SEIS, a supplement to the 1973 EIS. See NHV-SEIS, vol. I, at 100-10. We have already concluded that the EIS adequately considers the 1977 Plan; the appellees' decision not to supplement further the EIS was reasonable. See Warm Springs Dam II, 621 F.2d at 1024.

c. Revised Population Projections

As above noted, in 1978 the State of Hawaii Department of Planning and Economic Development (DPED) revised its year 2000 population projections downward; DPED reduced the Oahu population projection from 1,039,000 persons to 917,400 for the year 2000. Applying the 1977 General Plan distribution percentages to this new figure yields a target Windward Oahu population of between 125,700 and 138,500. See 538 F. Supp. at 166.

The NHV-SEIS fails to discuss in detail the 1978 projections, even though they were issued nearly two years before the NHV-SEIS was approved. See id. The appellees, however, did consider the 1978 projections in their Region 9 Staff Analysis. See Region 9 Staff Analysis, at app. B. They apparently concluded that older projections, predicting a Windward Oahu population of 150,500, were more valid. See NHV-SEIS, vol. I, at xviii; 538 F. Supp. at 167.

The District Court found, under 40 C.F.R. Sec. 1500.8(a)(1) (1978),31 that the appellees met their NEPA obligations by identifying the population and growth assumptions used to justify H-3. See 538 F. Supp. at 166-67.

The appellants argue that the 18,000-person reduction in population figures is "significant new information" because federal decisionmakers regarded population projections and goals important when they approved the EIS. See, e.g., Office of the Secretary, U.S. Dep't of Transp., Concurrence Memorandum 2 (Nov. 21, 1980) ("Since construction of H-3 is likely to stimulate development on Windward Oahu, and could contribute to pressures for development in excess of that contemplated in the General Plan, monitoring of population growth and of traffic levels on H-3 will be needed to assure that development objectives of the plan are not exceeded."). The appellants also allege that the NHV-SEIS fails to state any reason why a discussion of the reduction in projected growth was omitted from the EIS.

The appellees correctly respond that both the "outdated" and 1978 projected population figures are actually a range of projections and that the NHV-SEIS discusses the changing population projections and their ranges. See NHV-SEIS, vol. I, at xvii-xviii. Moreover, the EIS explains, albeit briefly, the appellees' reasons for selecting the population projections used in the EIS. See id. at xviii.

On this record, the appellants simply have not shown that the appellees have violated the Warm Springs Dam II standards; accordingly, we will not disturb the appellees' decision not to supplement further the EIS on account of the revised population projections.

d. 1980 Census Data

The appellants argue that the 1980 census data is "significant new information" because it shows that Windward Oahu is growing faster than called for in the 1977 General Plan, and that, therefore, the growth-inducing impact of H-3 virtually will insure that General Plan growth limits will not be met. With regard to supplementing the EIS, this argument has little merit for several reasons.

First, the appellants have not persuasively shown the significance of the 1980 census data in terms of new environmental impact. Cf. Citizens Committee Against Interstate Route 675 v. Lewis, 542 F. Supp. 496, 554-56 (S.D.Ohio 1982) (census data does not require supplementation of EIS).

Second, the appellants' argument assumes, without showing, that H-3 will induce growth beyond that envisioned in the General Plan.

Third, as above noted, the Secretary's concurrence in the EIS was conditioned upon the Hawaii Department of Transportation's cooperation with the City and County of Honolulu "in their implementation of measures proposed to achieve General Plan objectives for Windward Oahu." See Office of the Secretary, U.S. Dep't of Transp., Concurrence Memorandum 3 (Nov. 21, 1980). See also Fed. Highway Admin., Decision Memorandum 4 (Dec. 5, 1980) (FHWA's and Hawaii DOT's acceptance of Secretary's conditions).

For these reasons, the appellees acted reasonably in not supplementing the EIS on account of the 1980 census data.

B. FAHA--ADEQUACY OF LOCATION/DESIGN REPORTS

Under FAHA, a request for location or design approval must be accompanied by reports and other documents that discuss, inter alia, the anticipated economic, social, and environmental effects of the proposed action and alternatives under consideration. 23 C.F.R. Secs. 790.9(c), 790.8(b)(2)(i) (1978). The H-3 Location/Design Study Report, NHV-SEIS, 1973 EIS Preface, and 1972 EIS were submitted as evidence of compliance with this requirement. The appellants challenge the adequacy of these reports with respect to socio-economic impacts.

The appellants allege that: (1) the FAHA requirements cannot be satisfied by incorporating the EIS documents by reference; and (2) even if incorporation by reference is acceptable, the EIS is inadequate and needs to be supplemented by a detailed, current socio-economic study. The first contention has no merit. The CEQ regulations clearly permit the type of incorporation by reference to which appellants object. See 23 C.F.R. Sec. 790.8(b)(2)(iv) (1978). As to the second contention, we have already concluded that the documents adequately discuss the secondary impacts of H-3.

The District Court's rulings in respect to the ESA, NEPA, and the portions of FAHA other than section 4(f) are affirmed. The District Court's Order dissolving the injunctions against construction of H-3, however, is reversed. On remand, the District Court must enjoin construction of the entire highway as proposed until such time that the Secretary can demonstrate his full compliance with section 4(f) as the statute applies to Ho'omaluhia Park and has made a determination in harmony with the statutory requirements.32 Our decision does not affect any injunctions the District Court has not dissolved that originated in the district court.

The judgment of this Court shall issue forthwith, and no Petition for Rehearing will be entertained. See Fed.R.App.P. 2.

AFFIRMED IN PART, REVERSED IN PART, AND REMANDED.

I concur in the portions of the majority opinion that discuss the Makai Realignment, the Endangered Species Act of 1973, and the National Environmental Policy Act of 1969 (NEPA) together with the Federal-Aid Highway Act of 1968 (FAHA). I cannot concur, however, in the majority's unnecessary analysis of the No Build alternative in connection with Ho'omaluhia Park.

The majority interprets "no build" to mean not building any of Interstate Route H-3 not already completed. The portion of H-3 that affects Ho'omaluhia Park, however, runs only between the Kaneohe and Halekou interchanges. Common sense suggests that the No Build alternative concerned with Ho'omaluhia Park covers this short segment, not all of H-3. See, e.g., Citizens' Committee for Environmental Protection v. United States Coast Guard, 456 F. Supp. 101, 119-20 (D.N.J.1978). The Stop H-3 Association (the Association) advanced this argument as an alternative position in the district court. The Secretary of Transportation (the Secretary), accepting the Association's suggestion, established to the district court's satisfaction that this No Build alternative essentially duplicated the Makai Realignment. See Stop H-3 Association v. Lewis, 538 F. Supp. 149, 180 (D.Hawaii 1982). Properly analyzed, questions about the No Build alternative should therefore fall completely within our discussion of the Makai Realignment. The majority's discussion beyond that point, although termed a holding, is actually dictum based on similarly unnecessary portions of the district court's opinion and the section 4(f) statement prepared for Ho'omaluhia Park that posited and rejected a No Build alternative covering all of H-3. See, e.g., id. Nevertheless, I conclude that by focusing solely on this definition of a No Build alternative, the majority errs both in law and in not adopting a common sense approach to the issue.

For every proposed project, a first great question is whether or not to undertake it. The second, equally important question after deciding to undertake a project is which way to construct it. Under FAHA,

the Secretary shall not approve any program or project which requires the use of any publicly owned land from a public park ... unless (1) there is no feasible and prudent alternative to the use of such land, and (2) such program includes all possible planning to minimize harm to such park ....

23 U.S.C. Sec. 138 (section 138); accord 49 U.S.C. Sec. 303 (section 303) (recodifying and amending section 4(f) of the Department of Transportation Act of 1966 (DOTA), 49 U.S.C. Sec. 1653(f) (1976), repealed by Pub.L. No. 97-449, Sec. 7(b), 96 Stat. 2444). As the plain language indicates, this statute chiefly prescribes which way to build the project. The legislative histories of section 138 and the 4(f) predecessor to section 303 shed little additional light on the meaning of this language. See, e.g., S.Rep. No. 1410, 89th Cong., 2d Sess., reprinted in 1966 U.S.Code Cong. & Ad.News 2800, 2837-38, 2840, 2844 (section 138); Conf.Rep. No. 2236, 89th Cong., 2d Sess., reprinted in 1966 U.S.Code Cong. & Ad.News 3448, 3450 (section 4(f)). Thus, as the Supreme Court observed in Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 412 n. 29, 91 S. Ct. 814, 821 n. 29, 28 L. Ed. 2d 136 (1971) (Overton Park ), "we must look primarily to the statutes themselves to find the legislative intent." On their faces, sections 138 and 303 do not bar any particular project. Instead, they require the Secretary to adjust projects to avoid the use of park land. If building a project and avoiding the use of park land prove irreconcilable aims, the statutes allow the project to go forward. This scheme clearly regulates the way to build a project rather than whether to undertake it.

But what is a project? Under 23 U.S.C. Sec. 101(a), " 'project' means an undertaking to construct a particular portion of a highway ...." Under 23 U.S.C. Sec. 105(a), a program consists of "proposed projects for the utilization of the funds apportioned." Consistent with these terms, "a 'program for projects' usually presents for approval one or more stages or part [sic] of the work necessary ultimately to be completed to result in the actual finished construction of a highway." Movement Against Destruction v. Volpe, 361 F. Supp. 1360, 1380 (D.Md.1973) (per curiam) (two judge court), aff'd, 500 F.2d 29 (4th Cir.1974). FAHA thus focuses on approval of alternative plans for highways. Other statutes of environmental protection have a much broader scope. NEPA, for example, clearly allows consideration of alternatives to highways besides other highways. See, e.g., 42 U.S.C. Sec. 4332(2)(A), (C). For purposes of FAHA, however, only highways, not other modes of transportation, represent alternatives. Cf., e.g., Airport and Airway Development Act of 1970, Sec. 12(b), Pub.L. No. 91-258, tit. I, Sec. 12(b), 84 Stat. 221, repealed by Pub.L. No. 97-248, tit. V, Sec. 523(a), 96 Stat. 695 (explicitly requiring consideration of alternate forms of transportation); see also 49 U.S.C. Sec. 2201(b) (statement of purpose for "various modes of transportation" in airport planning); but see 49 U.S.C. Sec. 2208(b)(5) (equivalent of sections 138 and 303). Although I recognize statements to the contrary exist, e.g., D.C. Federation of Civic Associations v. Volpe, 459 F.2d 1231, 1239 (D.C.Cir.1971), cert. denied, 405 U.S. 1030, 92 S. Ct. 1290, 31 L. Ed. 2d 489 (1972), the language, intent, and scope of FAHA lead me to conclude that Congress took the choice of planning a highway as a given under section 138. Congress did not envision that the Secretary would have to reconsider that initial choice of project in protecting park lands.

Section 303 covers any "transportation program or project," 49 U.S.C. Sec. 303(c) (emphasis added), not just highway building, but accepting the initial choice of project as a given still seems appropriate. See, e.g., Monroe County Conservation Council, Inc. v. Volpe, 472 F.2d 693, 700 (2d Cir.1972). In Coalition for Canyon Preservation v. Bowers, 632 F.2d 774 (9th Cir.1980) (Bowers), we faced a situation similar to the construction of H-3: the proposed construction of a four-lane highway in upper Montana. Examining the adequacy of a section 4(f) review of the project, we remanded because the alternative of an improved two-lane road had not received consideration. Id. at 784-85. We did not suggest the Secretary consider a complete No Build alternative, although, as in every case, that choice was available. Cf. id. at 785 n. 5 (No Build was not one of the five alternatives considered in the 4(f) statement); see also Louisiana Environmental Society, Inc. v. Coleman, 537 F.2d 79, 85 (5th Cir.1976) (rejecting out of hand a No Build alternative to an entire bridge project). As Bowers shows, we have not previously given project-wide No Build alternatives the status the majority would.

The decision in Maryland Wildlife Federation v. Lewis, 560 F. Supp. 466 (D.Md.1983), provides yet another argument why project-wide No Build alternatives have no place in analyses under sections 138 and 303. There, the district court found a complete No Build alternative to a freeway in western Maryland imprudent because it would not satisfy the general purposes of the Appalachian Regional Development Act, 40 U.S.C. app. Sec. 2. 560 F. Supp. at 473-74. The No Build alternative to H-3 proposed by the majority would run afoul of similarly general purposes in FAHA. See, e.g., 23 U.S.C. Sec. 101(b). In both cases, however, the statutory purposes are so general that they add no more than an extra measure of federal legitimacy to a project decision already made. This indicates that project-wide No Build propositions do not actually provide "prudent alternatives" for purposes of sections 138 and 303. The choice whether to undertake a project stands as a condition precedent to application of these sections. Other laws such as NEPA guide that first choice and include consideration of complete No Build alternatives. Sections 138 and 303 simply regulate the way the government may implement a chosen project.

As a final example, in Overton Park the Supreme Court analyzed a proposed interstate route by implicitly accepting that the choice to build a highway lay beyond any review made under sections 4(f) and 138. As proof, consider that any project-wide No Build alternative is, by definition, "feasible" under sections 303(c)(1) and 138. The Supreme Court, however, limited the question of feasible alternatives to alternative highway routes: "For this exemption to apply, the Secretary must find that as a matter of sound engineering it would not be feasible to build the highway along any other route." 401 U.S. at 411, 91 S.Ct. at 821. Nowhere in the opinion does the Court even imply that a complete No Build option, or some substitute mode of transportation, would ever represent a feasible or prudent alternative for purposes of sections 4(f) and 138 after the choice of a highway as the project.

Large highway projects assisted by federal funding are plain examples of cooperative federalism. Cf., e.g., Hodel v. Virginia Surface Mining & Reclamation Association, Inc., 452 U.S. 264, 289, 101 S. Ct. 2352, 2366, 69 L. Ed. 2d 1 (1981) (Surface Mining Control and Reclamation Act of 1977). The majority, by proposing the complete rejection of an H-3 highway as an alternative to the use of Ho'omaluhia Park by one portion of the proposed route, stands the supposition of an initial project choice in sections 138 and 303 on its head. This improperly interferes with the cooperative system regulated by those statutes. The district court also erred, in my judgment, by confusing the purposes of sections 138 and 303 with the initial decision to build a highway, see 538 F. Supp. at 180. A proper reading of those sections, however, shows Congress intended them to regulate which way a government constructed a project, not whether a government constructed a project at all. For this reason, I do not concur in the majority's unnecessary discussion of their No Build alternative.