Plaintiff-appellee State of Wisconsin and intervening plaintiff-appellee County of Marquette, Michigan, sued federal appellants seeking the preparation of a supplemental environmental impact statement (SEIS) in connection with Project ELF, an extremely low frequency submarine communications system developed by the Navy, which the Navy undertook to reactivate and expand in 1981.1 Plaintiffs contended that the Navy’s original 1977 environmental impact statement prepared at the time the project first originated should have been supplemented because of new information regarding the biological effects of extremely low frequency electromagnetic radiation. After a trial on the merits, the district court agreed with this contention and enjoined the Navy from proceeding with any additional work on Project ELF in Wisconsin or Michigan and from installing receivers in submarines until a supplementary environmental impact statement had been prepared.2 This expedited appeal followed.
I.
On July 1, 1968, the Navy announced plans to construct an extremely low frequency (ELF) submarine communications test facility within the Chequamegon National Forest near the town of Clam Lake in northern Wisconsin. This test facility, consisting of two 14-mile strings of overhead antennae mounted on utility poles and a transmitter situated in a fenced compound, became fully operational in 1969. Between 1969 and 1978, while the Navy operated and tested the capability of this facility, two comprehensive ELF projects were formulated. The first, called Project Sanguine, was to include a 6,300 square mile grid of buried antenna cable and was *415to be operational by 1976. Project Sanguine, however, never went into full-scale development. The second system, Project Seafarer, which was proposed in 1977, initially was to consist of two test facilities: the one near Clam Lake, and the other, with a transmitter station and 130 miles of buried antenna cables, to be situated in the upper peninsula of Michigan. In its final, fully-developed form, the system was to have five surface transmitting stations and 2,400 miles of buried antenna cables extending over a 4,000 square-mile area.
In 1978, after initially supporting the allocation of $20.1 million for Project Seafarer research and development, President Carter decided to postpone Project Seafarer indefinitely. His decision not to proceed with the project was based on “reservations regarding this large 2400-mile antenna network, primarily because of public opposition and the inevitable inconvenience to private landowners as well as its excessive costs.” President Carter, however, did instruct the Secretary of Defense to study the possibility of putting together a smaller, less intrusive, and less costly ELF system. Following the President’s lead, Congress undertook no additional funding of Project ELF, as it came to be known, and the project entered a dormant phase.
In 1981, however, Congress, in its Department of Defense Authorization Act, 94 Stat. 1077, 1081, ordered the Navy to resume research and development for an ELF communications system and directed the President to “submit to the Congress a plan for deployment” of such a system in 1981. In April, 1981, President Reagan advised the Secretary of Defense of his intention to review and decide whether to reactivate Project ELF. The Secretary submitted an ELF proposal to the President on August 13, 1981. On October 8, 1981, President Reagan approved the recommended proposal and advised Congress of his intent to proceed. He ordered that the ELF project should
include upgrading of the existing ELF in Wisconsin, a new transmitter facility of comparable size with 56 miles of antenna[e] in Michigan, and ELF receivers for the submarines. In order to make critical improvements in connectivity to the submarine forces, the Navy should support this decision in a way that will provide an initial operating capability in fiscal year 1985.
As a result of this order, the ELF facility in Wisconsin was reactivated in December, 1981, and resumed broadcasting to submarines, which continues to date.
For each of the pre-1981 ELF proposals, the Navy provided documentation of the environmental effects of each project by preparing an environmental impact statement (EIS) in accordance with the National Environmental Policy Act of 1969. Of specific concern to the public were the possible effects of continuously exposing humans, animals, and plants to extremely low frequency electromagnetic radiation. This concern was acknowledged and addressed by the 1972 EIS and 1975 supplemental EIS prepared in connection with Project Sanguine, and the 1977 EIS prepared in connection with Project Seafarer. The 1977 EIS included a 1977 National Academy of Sciences report commissioned by the Navy studying the biological and human health effects of extremely low frequency electromagnetic radiation. This report summarized the state of knowledge concerning these effects and concluded: “On the basis of the evidence available, the Navy believes no adverse effects on human health or performance will be associated with long-term Seafarer exposure.”
When the decision to reactivate Project ELF in 1981 was made, no additional environmental impact assessment was made. In June, 1983, however, the Navy issued, but did not circulate to the public or to interested federal agencies, an environmental impact, assessment of the upgrading of the Wisconsin facility. In that assessment, the Navy observed that the electromagnetic field intensities produced by the ELF system would not increase by upgrading the Wisconsin facility to full operation. This assessment did not reevaluate the conclusions of the 1977 EIS and National Aca*416demy of Sciences report that ELF systems would produce no adverse biological effects.
In contrast to the prior proposals, Project Sanguine and Project Seafarer, the present ELF proposal is much more modest. The modernization of the Wisconsin facility will not substantially alter the nature of its operations as it has existed since its original implementation in 1969. The proposed Michigan facility will include 56 miles of antennae on utility poles in an F shape and will operate similarly to the Wisconsin facility. The facilities in Wisconsin and Michigan can transmit independently, but if operated synchronously can reach areas not otherwise attainable independently.
II.
The National Environmental Policy Act of 1969 (NEPA), 42 U.S.C. §§ 4321-47, articulates “a national policy [to] encourage productive and enjoyable harmony between man and his environment.” 42 U.S.C. § 4321. Although NEPA establishes “significant substantive goals for the Nation,” Vermont Yankee Nuclear Power Corp. v. Natural Resources Defense Council, Inc., 435 U.S. 519, 558, 98 S.Ct. 1197, 1219, 55 L.Ed.2d 460 (1978), the Act does not attempt to dictate the result of any particular decision; instead, the weighing of the substantive environmental goals is to be undertaken by the executive agencies involved, and the judicially reviewable duties imposed by the Act are “essentially procedural.” Id. The procedural approach to the implementation of substantive NEPA policy is accomplished by the Act’s requirement that agencies prepare a detailed statement concerning environmental consequences — known as an environmental impact statement (EIS) — in connection with “every recommendation or report on proposals for legislation and other major Federal actions significantly affecting the quality of the human environment.”3 42 U.S.C. § 4332(2)(C). The EIS thus ensures adequate consideration of environmental consequences by alerting decision-makers to the nature of those consequences with regard to a particular action. See Andrus v. Sierra Club, 442 U.S. 347, 350, 99 S.Ct. 2335, 2337, 60 L.Ed.2d 943 (1979). In addition, the EIS acts to fully inform the public about agency decisions affecting the environment and facilitates public input into the decision-making process. See Weinberger v. Catholic Action of Hawaii/Peace Education Project, 454 U.S. 139, 143, 102 S.Ct. 197, 201, 70 L.Ed.2d 298 (1981). Finally, the EIS serves as a record for substantive review of challenges for noncompliance with NEPA. See Appalachian Power Co. v. EPA, 477 F.2d 495, 507 (4th Cir.1973).
The preparation of an EIS as a prerequisite to the implementation of a major government project has become commonplace. Although the Supreme Court has emphasized that an initial EIS should be finished at a fixed time, Vermont Yankee, 435 U.S. at 534, 555, 98 S.Ct. at 1207, 1217, on occasion, an original EIS may become inadequate when during the life cycle of a project its scope changes in any substantial way or if new circumstances arise or new information becomes available about previously unsuspected environmental impacts. Recognizing that these situations may sometimes occur, the Council on Environmental Quality (CEQ)4 has set standards determining when an agency must supple*417ment its EIS. Under CEQ regulations, agencies are required to supplement an original EIS if: “(i) [t]he agency makes substantial changes in the proposed action that are relevant to environmental concerns; or (ii) [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. § 1602.9(c)(1)(i) & (ii) (1984).
The determination as to whether a SEIS is required is left to the discretion of the agency, just as the decision whether to file an initial EIS is also committed to the agency’s discretion. However, because the CEQ regulations “do not in themselves provide a suitable standard for reviewing an agency’s decision not to supplement an EIS,” Warm Springs Dam Task Force v. Gribble, 621 F.2d 1017, 1024 (9th Cir.1980) (per curiam), courts have applied the same standard used to review an agency’s determination that an EIS is not required in the first place for a particular agency action. See id.; Monarch Chemical Works, Inc. v. Thone, 604 F.2d 1083, 1087 (8th Cir. 1979). We will do the same.
In this circuit, the standard of review of an agency’s decision not to prepare an EIS is taken from section 706(2)(A) of the Administrative Procedure Act, 5 U.S.C. § 706(2)(A), which provides that a reviewing court must set aside an agency decision if it was “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.”5 See Simons v. Gorsuch, 715 F.2d 1248, 1251 (7th Cir.1983); City of West Chicago v. United States Nuclear Regulatory Commission, 701 F.2d 632, 651 (7th Cir.1983); Assure Competitive Transportation, Inc. v. United States, 635 F.2d 1301, 1308 (7th Cir.1980); Nucleus of Chicago Homeowners Association v. Lynn, 524 F.2d 225, 229 (7th Cir.1975), cert, denied, 424 U.S. 967, 96 S.Ct. 1462, 47 L.Ed.2d 734 (1976). Accord Hanly v. Kleindienst, 471 F.2d 823, 830 (2d Cir. 1972), cert, denied, 412 U.S. 908, 93 S.Ct. 2290, 36 L.Ed.2d 974 (1973). See also Kleppe v. Sierra Club, 427 U.S. 390, 412, 96 S.Ct. 2718, 2731, 49 L.Ed.2d 576 (1976). Thus, we must consider whether the Navy has acted arbitrarily, or capriciously, or abused its discretion in deciding not to file a supplement to the 1977 EIS in light of the new information put forward by plaintiffs. We will uphold a decision under the arbitrary and capricious standard if the decision was based on a consideration of relevant factors, see Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 416, 91 S.Ct. 814, 823, 28 L.Ed.2d 136 (1971), and made on a rational basis, see Bowman Transportation, Inc. v. Arkansas-Best Freight System, Inc., 419 U.S. 281, 285, 290, 95 S.Ct. 438, 441, 444, 42 L.Ed.2d 447 (1974).
III.
The specific mandate imposed by the CEQ regulations requires that the agency prepare a SEIS when “[tjhere are significant new circumstances or information relevant to environmental concerns and bearing on the proposed action or its impacts.” 40 C.F.R. § 1502.9(c)(l)(ii). The regulation itself does not define the term “significant.” Apparently the Council on Environmental Quality was willing to rely on the good faith assessments of the various federal agencies and allow them to determine what information would be sufficiently serious to require the preparation of a full-scale supplement to the original EIS. Be*418cause that assessment must be reviewed by us, however, and in the absence of any administrative interpretation of the term, we must attempt to give the term “significant” some more precise content as it is used in the present context. Cf. Hanly v. Kleindienst, 471 F.2d 823, 830-31 (2d Cir. 1972) (giving content to the. term “significantly” in the context of whether a major federal action will “significantly” affect the quality of the human environment so as to require the preparation of an initial EIS).
An important difference between an agency’s decision whether to file an initial EIS and its decision whether to supplement an EIS is that the decision to supplement is made in light of an already existing, in-depth review of the likely environmental consequences of the proposed action. Thus, it seems clear that the principal factor an agency should consider in exercising its discretion whether to supplement an existing EIS because of new information is the extent to which the new information presents a picture of the likely environmental consequences associated with the proposed action not envisioned by the original EIS. As one court has stated, “[tjhere is no benefit in taking another ‘hard look’ at an action if that view is taken from the same vantage point and overlooks the same environmental panorama.” National Indian Youth Council v. Andrus, 501 F.Supp. 649, 661 (D.N.M.1980), aff'd, 664 F.2d 220 (10th Cir.1981), The issue is whether the subsequent information raises new concerns of sufficient gravity such that another, formal in-depth look at the environmental consequences of the proposed action is necessary. The supplementation process is extensive and an agency’s determination as to when one is or is not needed is entitled to some deference. We hold, therefore, that an agency cannot have acted arbitrarily or capriciously in deciding not to file a SEIS unless the new information provides a seriously different picture of the environmental landscape such that another hard look is necessary.
Normally, the plaintiffs may be expected to bring a NEPA challenge under the regulations when a particular item of new information, although specifically considered by the agency, was thought by the agency to be of insufficient significance to require the filing of a SEIS. See, e.g., Massachussets v. Watt, 716 F.2d 946 (1st Cir.1983) (agency’s 97% reduction in its estimate for quantity of offshore oil was a sufficiently significant new circumstance such that agency was required to issue an EIS supplement before awarding oil lease); California v. Watt, 683 F.2d 1253 (9th Cir. 1982), rev’d on other grounds sub nom. Secretary of the Interior v. California, — U.S.—, 104 S.Ct. 656, 77 L.Ed.2d 295 (1984) (doubling of oil recovery estimate did not require filing of supplementary EIS when there was no reason to believe that environmental harms could increase more than proportionally). Such a challenge to a specific decision not to supplement is to be reviewed in accordance with the Supreme Court’s admonition in Vermont Yankee that “the role of a court in reviewing the sufficiency of an agency’s consideration of environmental factors is a limited one, limited both by the time at which the decision was made and by the statute mandating review.” 435 U.S. at 555, 98 S.Ct. at 1217. Because of the general ongoing duties imposed by the regulation, however, there may be situations in which it is not alleged that a particular item of new information was specifically rejected by the agency, but that an accumulation of new information was ignored by the agency in whole or in part, such that the failure to act on it was an abuse of the agency’s discretion in light of the alleged significance of the information.6 Even this kind of challenge, how*419ever, must be framed in terms of information available during a specific time period, with recognition of the fact that an agency cannot have acted arbitrarily unless it has had a reasonable time to consider the alleged new information. Open-ended chai-lenges framed in general terms cannot suffice.
The Navy contends that consideration of the new material must be limited to that information available in 1981 when the Navy made the decision to reactivate and modernize the ELF system and determined that a SEIS was not necessary. The Navy argues that it was therefore improper for tile district court to evaluate the SEIS issue in terms of the information available as of the date of the combined preliminary injunction and merits hearing held in the fall of 1983. Plaintiffs respond that they alleged the Navy had violated NEPA not only by failing to file a SEIS in connection with its decision to modify the project and not to file a SEIS in 1981, but by failing to evaluate the accumulation of information available subsequently that allegedly rose to a level of significance sufficient to trigger the duty to supplement the 1977 EIS. To resolve this issue, we examine the omplaint.7 Our reading of the complaint finds that it is directed at the Navy’s decision in 1981 to substantially upgrade and expand the existing ELF facilities without including a SEIS in the proposal, and also at the Navy’s failure to prepare a SEIS in light of new information then available allegedly relevant to environmental coneerns. There is nothing alleged to identify any other time. The complaint does not and could not legitimately complain about new information coming to light right up until the time the complaint was filed, and certainly not up until the time of trial, since the Navy could not have acted arbitrarily or capriciously without being given time for appropriate consideration of that infermation. The references in the complaint to the 1981 Navy decision definitely identify the Navy’s particular alleged failure and when it occurred. If it were otherwise, the issue regarding significant new information raised by an open-ended complaint would be premature and make preparation for trial difficult at best. Although the complaint in this ease does not run afoul of these normal definite pleading and litigation requirements, by reason thereof, how-ever, the relevant evidence relating to the alleged new information issue must be limited to that available when the 1981 decicsion to upgrade and expand the facilities without a SEIS was made.8
*420It appears that the trial court relied on fourteen post-1977 items of information raising questions about the human health effects of low frequency fields. Over half of these, however, were not available until after the Navy decisions had been made and should not have been admitted. This later information therefore could not possibly form the basis for finding the Navy in default of its NEPA duties at the time of its 1981 decision. A large part of the other material presented by plaintiffs had already been considered by the Navy in 1977, or at least prior to the modernization proposal. Plaintiffs do not attack the original EIS, which was only four years old. Evaluation of the environmental impact of the project therefore has a firm beginning basis. Our inquiry must be limited to what may have developed after that up until the time of the Navy decision in 1981.
We do not believe that NEPA requires the courts to keep watch over an agency’s shoulder as a supervisor and direct the agency as to if, when, and how it should peliminarily review any new information. Our responsibility is limited to determining whether or not the new information was so significant that for the agency not to act on it was irresponsible, arbitrary, or capricious. It is not enough that the information may be worthy of further inquiry or may be considered important research. Our task is the limited one of determining whether or not the new information presents a seriously different picture of the likely environmental consequences of the proposed action not adequately envisioned by the original EIS, such that the Navy’s failure to act on it was arbitrary or capricious. If it was, then the court may order a SEIS. To make this determination, however, we do not have to resolve the scientific issues on the merits. The district court stated that:
The merit of the new information [available since 1977] has been attested to by the expert witnesses. At trial, expert witnesses for both plaintiffs and defendants identified post-1977 research either as new information worth further inquiry or as important research.
578 F.Supp. at 1362. The district court went on to conclude that “the 1977 environmental impact statement is no longer adequate as a source of information necessary to a rational decision on the relative risks and benefits of Project ELF, because it includes none of the scientific work done since 1977 or any evaluation of that work,” id., and that “[t]he scientific information on biological effects generated since 1977 is significant enough to require careful review by the Navy,” id.
We disagree with this open-ended approach applying a different standard and hence the conclusions stemming from it. An original EIS is no longer “adequate as a source of information necessary to a rational decision on the relative risks and benefits” of a proposed action not because it fails to “include” new information or any “evaluation” of it, but because the new information presents a seriously different picture of the likely environmental harms stemming from the proposed action. To say that new information is significant enough to require “careful review” is not to say that it is so significant that the formal procedures associated with the filing of a SEIS must begin. The latter, not the former, is the court’s responsibility to assess under the NEPA regulations.
The district court itself found that the Navy’s previous, relatively recent conclusion that long-term exposure to the electromagnetic radiation emanating from ELF would not produce adverse biological effects was not invalidated by any of the new information received into evidence. That new information was characterized by the district court as “equivocal.” We believe that this is something less than a finding that the new information was so significant that the Navy’s failure to respond in a reasonable time with a SEIS was arbitrary *421or capricious. The district court also found no evidence that ELF would have any immediate adverse environmental impact. Nonetheless, since a SEIS was ordered and some language in the opinion supports the view that the new information was sufficiently significant to justify a SEIS, we will briefly consider the evidence on that basis, including that which we do not regard to be relevant or admissible under the time frame alleged in the complaint.9 Of course, it is not necessary for our purposes to consider in detail the merits of the scientific allegations and the alleged shortcoming of each piece of information and reach our own scientific decision about the project. That is not our function. We are only to decide whether the evidence reveals that the new information presents a seriously different picture of the environmental impact of the proposed project from what was previously envisioned, such that the Navy acted arbitrarily or capriciously in deciding not to file a SEIS.10
The previous picture of the state of knowledge concerning the biological effects of extremely low frequency electromagnetic radiation was summarized in the 1977 National Academy of Sciences report (filed in connection with the 1977 EIS) as follows:
Field and laboratory research has not revealed to date convincing evidence of impacts on the environment or populations of plants or animals either at planned Seafarer surface electromagnetic intensities, or at lesser levels which would ordinarily be encountered by particular species;
Experimental evidence to date indicates normal offspring of plants, animals, and humans are produced in ELF fields comparable to and even larger than Seafarer fields. No study has shown any teratogenic effects (embryological defects) produced by ELF fields;
There have been few identified tangible ELF effects, and no substantial evidence to date that human exposure to electric and magnetic fields characteristic of Seafarer would affect human health or performance;
No generally-accepted adverse human or ecological effects due to such fields have been observed in homes, workplaces, communities or in the immediate vicinity of electric power transmission corridors ____
*422The new information that plaintiffs contend is significant was offered through the testimony of three scientists, Drs. Wer-theimer, Moulder, and Becker. The Navy supplied two experts of its own, Drs. Straub and Justesen, to analyze this evidence.11 A brief review of the information and the comments of the experts convinces us that the information falls short of that threshold of “significance” at which the duty to prepare a SEIS is triggered. Among the most important new information is that contained in the Electrical Power Research Institute studies at the Batelle Laboratories; frequency and power intensity window research conducted by Black-man and Adey, Delgado, and persons at UCLA; Goodman’s slime mold research; and several epidemiological studies.
The Batelle Laboratories studies, conducted in 1978, 1979, and 1982, involved subjecting minipigs (Hanford miniature swine) to electromagnetic radiation at 60 Hz. Although some of the studies showed that multigenerational exposure to such fields resulted in effects on fertility, birth defects, bone fracture repair, and behavior, most of the studies produced negative results. In addition, Justesen and Straub pointed out that there had been errors in the research and that it was being redone. While these two shared Moulder’s praise for the Batelle research, they considered the results irrelevant to Project ELF because the voltages used to carry out the Batelle tests were far outside the intensities of the fields anticipated at the ELF facilities.
In light of these limitations, we do not believe that they present a picture seriously different from that envisioned by the *4231977 National Academy of Sciences report, which found that “[e]xperimental evidence to date indicates normal offspring of plants, animals, and humans are produced in ELF fields comparable to and even larger than Seafarer fields.” Although the studies do point to the possibility of some teratogenic effects at certain intensities, these intensities are far removed from those associated with Project ELF and therefore do not alter the view of its likely environmental impact.
Similarly, we do not believe the research conducted by Delgado, those at UCLA, and Adey and Blackman focusing on points or “windows” on the spectrum of electromagnetic frequencies or power intensities at which biological effects of low frequency electromagnetic radiation may occur have seriously changed the picture in this area existing in 1977 Delgado s 1982 studies on the effect of extremely low frequency electromagnetic fields on the embryological development of chickens, according to Straub and Justesen, has methodological .. . . ’ , ,, . problems, mvolved pulsed rather than smu- ., , „. ’ , 1 . ...» soidal fields, and used different field mten- ., ttit ti j-i i • ., . , , sities than ELF, thus making it irrelevant , ,, , , , ,, _T°_ „ T to the character of the ELF fields. In jj.,. ,j j i - .i. xt addition, while studies funded by the Navy in 1980 at UCLA on the effect of extremely low frequency electromagnetic radiation on monkey behavior suggested the existence of power intensity and frequency windows, neither these studies nor the more intensive work performed in this area by Black-man and Adey rise to the requisite level of significance. Blackman’s studies explored effects of extremely low frequency electromagnetic radiation on the permeability of chicken brain cells (the studies were published in 1979, 1981, and two in 1982). Adey published an article in 1980 discussing the existence of windows found by studying electromagnetic effects on tissue interactions. Although the experts testified at trial that this work was important, Justesen pointed out that Adey had done work on the phenomenon of windows before 1977, some of which was mentioned in the 1977 National Academy of Sciences report. Thus, while the later studies confirm the phenomenon of windows and indicate that a frequency window may exist at 75 Hz. (Project ELF is to operate at a range of 72-80 Hz.), they do not present new information that seriously alters the environmental picture. Indications of such windows were present in 1977, and, as noted by Straub, the new work is not directly relevant to the Navy’s ELF system because 0f the differences in electric field intensities. Certainly the Navy was within the range of its permissible discretion in con-eluding that the information of power and frequency windows since 1977 was not sufficiently significant to require the preparation of a full-blown SEIS,
Ag for the Navy-sponsored slime mold studies conducted by Goodman and others at the University of Wisconsin-Parkside, reportg iggued jn 1979 and 1983 found gome alterations in bagie cell functions and consumption in response to extremely ]()w fre electromagnetic radiation, . . however botb TugWn and strailb questioned the relevancy of these studies to the human health effects of Project _T ELF, because the field strengths used in the experiments were so much greater than T , TTr those at the Navy s facilities. We see ,, , .. ”othfg m these studies that would make the declsl0n ^ ?T flle ? SEIS+arbltrary or caP™- Tbe has continued to review the resujtsjofr( thf3 experiments and contmues to fund Goodman s researcL
Finally, the epidemiological studies conducted by Wertheimer in 1979 and 1982, and several others, such as Nordstrom’s in 1983, do not seriously change the environmental picture that existed in 1977. Wer-theimer herself testified that her work, in which- she examined the possible association between high-current electric power fields and incidences of cancer in children and adults, was merely “exploratory,” and she acknowledged that it suffered from several methodological weaknesses. Dr. Straub testified that he did not believe the Wertheimer work presented significant new information with regard to the human health effects of Project ELF and found her results “speculative.” Justesen gener*424ally. concurred in this judgment, and the IIT Research Institute, under contract with the Navy to evaluate published writings of scientific interest in electromagnetic field effects, was highly critical of both Wer-theimer’s 1979 and 1982 reports. Most of the other epidemiological studies, such as Nordstrom’s, tend to be of questionable relevance because they involve high voltage transmissions substantially greater than those used by the ELF system. None of them significantly alters the view reported in the 1977 National Academy of Sciences report that “[tjhere have been few identified tangible ELF effects, and no substantial evidence to date that human exposure to electric and magnetic fields characteristic of Seafarer would affect human health or performance.” (Emphasis added.)
In short, all of the evidence, viewed in whole or in part, as of 1981 and even the inadmissible evidence accumulated up until the date the complaint was filed, cannot be deemed to have been of such significance as to find that the Navy was arbitrary or capricious in not responding with a SEIS. We have touched only representative samples of the evidence, but as a whole it is generally unimpressive and cannot be deemed sufficiently significant. To elevate the significance of plaintiffs’ evidence, in whole or in part, relevant or irrelevant, to the required level of significance could only be deemed to be clearly erroneous.
The Navy has not been ignoring this area of concern. It has a continuing review process to monitor the relevant information. The Environmental Review Committee was formed to meet monthly to keep up to date with consideration of new environmental information concerning ELF. In addition, the Navy has an ongoing contract with the Illinois Institute of Technology Research, Inc., a nonprofit consulting firm, to constantly review scientific literature of relevance to ELF and for actually monitoring studies at the ELF facility locations. These represent commendable efforts to comply continually with the spirit of NEPA. Nothing more at this time need be requested of the Navy.
Our view of the statutory framework and the evidence offered does not, of course, close the matter for all time. As possibly significant new information comes to light at some subsequent time, the duty remains on the Navy to evaluate it and if necessary to produce a SEIS in good faith without regard to any investment in the facility that may have developed in the meantime. The general subject remains one of great scientific interest and no doubt will continue to develop. Our holding is without prejudice to any further court challenge under the statute whenever warranted, but the information must be new, be significant, and be relevant to the peculiarities of ELF. Not every new publication by a so-called expert on the general subject of low frequency electromagnetic radiation, or even a stack of new articles, will necessarily meet that test. Were we to require the Navy formally to reassess its proposed action with a SEIS every time some bit of new information appeared, we would be unjustifiably interfering with the Navy’s mission.
We hold that there was no violation of NEPA by the Navy. That being so, the injunction must necessarily be vacated, as we have already done by order, without further consideration. Nevertheless, we proceed to consider the merits of the injunction separately since this panel, although not in unanimous agreement about the underlying NEPA violation, is in unanimous agreement that the injunction was unwarranted.
IV.
The panel is in agreement that even if there had been a NEPA violation, the district court abused its discretion in not undertaking a balancing of the relative harms to the parties before entering the injunction prohibiting the Navy from continuing with Project ELF until a SEIS was filed. The district court, in denying the Navy’s motion for reconsideration of the injunction, concluded that no balancing was *425to be undertaken because the point of NEPA is to assure that federal agencies assess the environmental impact of their proposals before deciding to proceed with an action. The district court found this case analogous to TVA v. Hill, 437 U.S. 153, 98 S.Ct. 2279, 57 L.Ed.2d 117 (1978), the snail darter case, in which the Supreme Court held that the Endangered Species Act mandated injunctive relief despite countervailing equities. Because in the district court’s view allowing the ELF proposal to proceed without first reconsidering its impact, except in certain unusual circumstances,12 would frustrate the purposes of the Act, the district court reasoned that there was a presumption that injunctive relief be imposed for a NEPA violation. It ruled that unless the Navy rebutted the presumption by showing that an injunction would not serve the purposes of the Act, the court did not have to undertake a balancing of the relative harms.
Although the district court’s view is not without some support from TVA v. Hill, we disagree with its conclusions in these particular circumstances. We consider Hill and this case to be distinguishable. There is no presumption mandating an injunction in this type of case. NEPA cannot be construed to elevate automatically its procedural requirements above all other national considerations. Although there is no national defense exception to NEPA, and the Navy does not claim one, the national well-being and security as determined by the Congress and the President demand consideration before an injunction should issue for a NEPA violation.
We begin our analysis of the appropriateness of the injunction here with the Supreme Court’s pronouncement in Wein-berger v. Romero-Barcelo, 456 U.S. 305, 102 S.Ct. 1798, 72 L.Ed.2d 91 (1982), that an injunction “ ‘is not a remedy which issues as of course____’” Id. at 311, 102 S.Ct. at 1802 (quoting Harrisonville v. W.S. Dickey Clay Manufacturing Co., 289 U.S. 334, 337-38, 53 S.Ct. 602, 603-04, 77 L.Ed. 1208 (1933)).
Where plaintiff and defendant present competing claims of injury, the traditional function of equity has been to arrive at a “nice adjustment and reconciliation” between the competing claims, Hecht Co. v. Bowles, [321 U.S. 321, 329, 64 S.Ct. 587, 592, 88 L.Ed. 754 (1944)]. In such eases, the court “balances the conveniences of the parties and possible injuries to them according as they may be affected by the granting or withholding of the injunction.” Yakus v. United States, 321 U.S. 414, 440 [64 S.Ct. 660, 675, 88 L.Ed. 834] (1944). “The essence of equity jurisdiction has been the power of the Chancellor to do equity and to mould each decree to the necessities of the particular case. Flexibility rather than rigidity has distinguished it.” Hecht Co. v. Bowles, supra, at 329 [64 S.Ct. at 592],
In exercising their sound discretion, courts of equity should pay particular regard for the public consequences in employing the extraordinary remedy of injunction. [Citation omitted.] ... The grant of jurisdiction to ensure compliance with a statute hardly suggests an absolute duty to do so under any and all circumstances, and a federal judge sitting as a chancellor is not mechanically obligated to grant an injunction for every violation of law. [Citations omitted.]
Romero-Barcelo, 456 U.S. at 312-13, 102 S.Ct. at 1803. Although the Court recognized that Congress occasionally may act to foreclose the exercise of the usual discretion possessed by a court of equity, such occasions will be rare. The Court explained its decision in TVA v. Hill by pointing to the content of the particular statute *426involved there and the particular facts presented. In Hill, the competing interests—the proposed dam and the snail darter habitat—were clearly defined. The actual balancing of these interests, however, was already prescribed by the specific terms of the relevant statute. The Endangered Species Act contained an unequivocal ban on the destruction of critical habitats for endangered species. Because the proposed dam in Hill would have eliminated such a critical habitat, refusal to enjoin the construction of the dam would have ignored the explicit provisions of the Act. “Congress, it appeared to us, had chosen the snail darter over the dam. The purpose and language of the statute limited the remedies available to the District Court; only an injunction could vindicate the objectives of the Act.” Romero-Barcelo, 456 U.S. at 314, 102 S.Ct. at 1804.
We believe, however, that neither the specific terms of NEPA nor the particular interests involved compel the issuance of the prohibitory injunction here. We find nothing in NEPA that explicitly or “by a necessary and inescapable inference, restrict[s] the court’s jurisdiction in equity.” Id. at 313, 102 S.Ct. at 1803. Although the goal of NEPA is to force agencies to consider the environmental consequences of major federal actions, see Kleppe, 427 U.S. at 410 n. 21, 96 S.Ct. at 2730 n. 21, that goal is not to be achieved at the expense of a total disregard for countervailing public interests. NEPA itself is procedural in nature, see Vermont Yankee, 435 U.S. at 558, 98 S.Ct. at 1219, and the statute recognizes that agencies may decide to subordinate environmental values to other social values with which they sometimes compete. Thus, although the judicial role is to insure that this weighing of competing interests takes place, we must fulfill this role in accordance with a consideration of other social costs, as recognized by the statute itself. That, after all, is the traditional way of applying equitable principles.
The recent trend of the majority of courts is to evaluate competing public interests in fashioning permanent injunctive relief for NEPA violations.
When a court has found that a party is in violation of NEPA, the remedy should be shaped so as to fulfill the objectives of the statute as closely as possible, consistent with the broader public interest____
The court should tailor its relief to fit each particular case, balancing the environmental concerns of NEPA against the larger interests of society that might be adversely affected by an overly broad injunction.
Environmental Defense Fund v. Marsh, 651 F.2d 983, 1005-06 (5th Cir.1981). See also Natural Resources Defense Council, Inc. v. United States Nuclear Regulatory Commission, 606 F.2d 1261, 1272 (D.C.Cir. 1979); Alaska v. Andrus, 580 F.2d 465, 485 (D.C.Cir.), vacated in part on other grounds sub nom. Western Oil & Gas Association v. Alaska, 439 U.S. 922, 99 S.Ct. 303, 58 L.Ed.2d 315 (1978); Conservation Society of Southern Vermont v. Secretary of Transportation, 508 F.2d 927, 933-34 (2d Cir.1974), vacated on other grounds, 423 U.S. 809, 96 S.Ct. 19, 46 L.Ed.2d 29 (1975). We do not mean to suggest, of course, that the objectives of NEPA should not be furthered to the fullest extent possible. We are merely emphasizing that NEPA does not foreclose the application of traditional principles of in-junctive relief. Although a majority of this panel has found that no NEPA violation occurred, we believe that even if a violation had occurred, as argued in the dissent, that the application of those traditional equitable principles in light of the evidence would have strongly militated against the imposition of the prohibitory injunction issued by the district court stopping all prepatory work.
The only irreparable injury under NEPA if the Navy were permitted to proceed with Project ELF, as perceived by the district court, was that it would later lead to biased decision-making by the Navy. It posited that because NEPA establishes a policy requiring decision-makers to take account of environmental values, that policy will not be fully achieved if decision-makers are predisposed to favor a particular outcome *427because of an agency’s investment in the project. Such a predisposition can result once a commitment of resources to a project is undertaken, thus possibly upsetting objective decision-making. See Massachusetts v. Watt, 716 F.2d 946, 952-53 (1st Cir.1983).
This concern is valid, and in an appropriate case this kind of possible effect may merit an injunction. The Navy points out, however, that this interest will be only marginally served by the injunction in an ongoing project such as this one. The new commitment to ELF began in 1981, when the project was reactivated, and reaffirmed the already existing commitment. This commitment, largely in terms of research and development dollars, was well established by the time plaintiffs got around to filing their lawsuit in mid-1983. By contrast, the commitment entailed by the remaining construction effort in Wisconsin and Michigan is relatively small, and thus the injunction’s service to NEPA in preserving unbiased decision-making would be slight. Even if the Navy subsequently makes the mistake of being influenced by its prior resource commitment, the President and the Congress are not without power to bring the Navy back into line, and so may we if the Navy’s actions are arbitrary or capricious. The risk the district court found from resource commitment, the only designated basis for the injunction, is wholly inadequate to enjoin this project.
In addition, the Navy argues that no other NEPA purpose would be served by the injunction.13 It notes that it has already adopted the recommendations of the National Academy of Sciences to minimize the electromagnetic fields generated by the project, and has committed itself to monitoring the potential adverse biological effects of such fields and to terminate the project should such effects become manifest. Finally, the Navy observes that whatever benefits might accrue to the public from another opportunity to comment on the content of the SEIS would not stem from the injunction enjoining the construction activities.
We find these arguments compelling. In addition, we note that even if we had found a NEPA violation here, it could not have been characterized as a blatant violation, and the injunction consequently could not have been justified on the ground that it was needed to prevent future violations. The Navy in this project has fully complied with NEPA in the past, and has publicly committed itself to a continuing monitoring of the physical and biological environmental consequences of Project ELF. The Navy has acted responsibly and not in an arbitrary or capricious manner.
More important, however, is the district court’s failure to balance the weight of the alleged NEPA violation against the harm the injunction would cause the Navy and to this country’s defense. The Navy has emphasized that an ELF submarine communications system is of the highest priority for national defense. The system will allow our strategic submarine fleet to operate with a greatly reduced chance for exposure and minimize the possibility of detection, thus strengthening the most survivable element of this nation’s nuclear deterrent. President Reagan’s directive to deploy Project ELF ordered that it achieve initial operating capability in fiscal year 1985. The Secretary of the Navy also stated in an affidavit that the latest evaluation of Project ELF is that it is essential to the national defense and that any delay in its construction is contrary to national defense interests. We have no basis to ignore those executive representations. At oral argument, counsel for the Navy informed *428us that preparation of a SEIS, which had already begun, could not be completed before April of 1985. Therefore, the delay occasioned by the issuance of an injunction could bring about serious consequences for our national defense. The Soviet Union allegedly is already advanced in this low frequency submarine communications technique. The district court gave no consideration to these serious circumstances.
An order requiring the preparation of a SEIS, of course, may always be appropriate to vindicate the purposes of NEPA should a violation of the CEQ regulations occur. Whether an injunction should also be entered preventing a project from continuing, however, is plainly an additional issue. We disagree with the district court that NEPA presumes that such injunctions should issue. That presumption is unrealistic in these circumstances. We conclude, therefore, that even if there had been a NEPA violation, the district court abused its discretion by failing to consider the degree to which the NEPA interest would, in fact, be served by an injunction, the efficacy of other forms of relief, and the harm to national defense interests that would result. Based on the record already before us, any traditional equitable weighing of interest could only lead to a denial of the issuance of the injunction that we have now vacated.
Even though we recognize and commend the conscientious and thoughtful consideration given by the district court to the unique and important problems in this case, we disagree with its view of what NEPA requires and the significance of the evidence in this case. No SEIS need be prepared at this time, and the injunction is vacated pursuant to our previous order.14
Reversed.
. With an ELF communications system, a submarine is able to receive messages from land while maintaining operational speeds at depths at which it is practically immune from detection. Other land-based submarine communications systems require the submarine to surface completely or rise to a point near the surface at slow speeds and drag an antenna to receive messages, thereby jeopardizing the safety of the crew and vessel.
. Wisconsin v. Weinberger, 578 F.Supp. 1327 (W.D.Wis.1984) order amended on reconsideration, 582 F.Supp. 1489.
. The Supreme Court has observed that the statutory language regarding "proposals for ... major Federal action” means that an EIS need not be prepared unless an agency’s planning ripens into a “specific proposal! ] for action,” Kleppe v. Sierra Club, 427 U.S. 390, 401 n. 12, 96 S.Ct. 2718, 2726 n. 12, 49 L.Ed.2d 576 (1976), or a “specific action of known dimensions,” id. at 402 n. 14, 96 S.Ct. at 2727 n. 14.
. The Council on Environmental Quality, created for the purpose of researching and advising the President on environmental matters, see 42 U.S.C. §§ 4341-47, has issued guidelines for implementing NEPA. These guidelines are to be followed by all federal agencies, see Andrus v. Sierra Club, 442 U.S. 347, 358, 99 S.Ct. 2335, 2341, 60 L.Ed.2d 943 (1979), and the specific CEQ regulations at issue here are explicitly incorporated into Department of Defense and Navy regulations. See 32 C.F.R. § 775.3(b)(1) (1984); 32 C.F.R. § 214 Encl. 1, D(4) (1984).
. Other Circuits have held that an agency’s decision not to issue an EIS will be upheld so long as it is "reasonable.” See Friends of the River v. Federal Energy Regulatory Commission, 720 F.2d 93, 109 (D.C.Cir.1983); Massachusetts v. Watt, 716 F.2d 946, 948 (1st Cir.1983); Warm Springs Dam Task Force v. Gribble, 621 F.2d 1017, 1024 (9th Cir.1980); Minnesota Public Interest Research Group v. Butz, 498 F.2d 1314, 1320 (8th Cir.1974) (en banc). One explanation given for this approach is that an agency’s determination as to whether it must file an EIS for a particular action is not discretionary, but a mandatory duty imposed by NEPA when certain conditions exist, thereby precluding judicial review of that determination under the arbitrary, capricious, and abuse of discretion standard of section 706(2)(A) of the Administrative Procedure Act, 5 U.S.C. § 706(2)(A). See Wyoming Outdoor Coordinating Council v. Butz, 484 F.2d 1244, 1248-49 (10th Cir. 1973).
. Although the duty to supplement may be a continuing one, that duty is practically limited by the nature and life cycle of the proposed project or action and the kind of change or new information involved. A decision to grant a license for the construction of a hydroelectric power plant, for example, would not have to be continually reassessed every time new forecasts for power consumption came to light if such forecasts were speculative and subject to frequent changes. Under such circumstances, it would be unreasonable to require the agency to *419reassess its decision every time the new forecasts were released, for it could "risk immobilizing the agency." Friends of the River v. Federal Energy Regulatory Commission, 720 F.2d 93, 109 (D.C.Cir.1983). “[RJeview and reassessment of [such a licensing] decision, absent sufficient reason, should not continue indefinitely.” Id.
. The c°mplam genera ly alleges that in 1981 the defendants decided to upgrade the Wisconsin facility and put it into operation on a permanent basis in conjunction with another facility in the upper peninsula of Michigan (f 11); that 42 U.S.C § 4321, ei seq., requires that any federal agency proposing "a major federal action significantly affecting the quality of the human environment" include an EIS (f 12); that the defendants plan to upgrade the Wisconsin ELF facility and put it into operation with the Michigan facility which is a major federal action significantly affecting the human environment (f 13); that the defendants originally prepared an EIS in 1977 (fl 14); that 40 C.F.R. § 1502.9(c) requires a SEIS if (i) the agency makes substantial changes in the proposed action relevant to environmental concerns, or (ii) there are significant new circumstances or information relevant to environmental concerns and bearing on the proposed actions fl[ 15); that these requirements were adopted by the Department of Defense and the Navy Off 16 & 17); that defendants have made substantial changes relevant to environ-mental concerns since 1977 (f 18); that there are significant new circumstances and informarelevant tQ environmental concerns beari on gLF proposal (f 19); that the defendants • i * j *1. *. * * 1. / .,• ■. . , j vlolated fadlnS to lncl?de the recl“r“ EIf " SEIS ™ ‘heir proposal to up-Srade the ELF test facihty m Wisconsin and to P«t it mto operation on a permanent full time basls ln conjunction with another ELF transmit-,facdlty be ™nstl;ucted “ Michigan (JT 20); that the defendants have failed to satisfy tbe statute without a SEIS in the current propos(If 21).
. Our reading of this complaint does not mean that a complaint could not be drawn alleging that the Navy acted arbitrarily or capriciously in not preparing a SEIS in connection with the proposal to modify ELF in 1981, and in the *420alternative contending that significant new information had come to light at a certain time thereafter, which the Navy in due course capriciously and arbitrarily avoided. But that complaint is not the complaint brought by plaintiffs here.
. The evidence offered by plaintiffs coming to light after the 1981 Navy decision suffers from the same deficiencies as the new evidence arising prior to that decision and does not justify a finding that the Navy was arbitrary or capricious in not responding with a SEIS to any or all of it. We are not sure that the district court actually found to the contrary that the new information was in fact sufficiently significant to trigger the duty to file a SEIS, although a SEIS was ordered, or only whether the district court found that the new information merely was significant enough to impose upon the Navy the duties of evaluation and explanation, a finding that would not support the ordering of the preparation of a SEIS under our standard of review of an agency’s action. The essence of the violation perceived by the district court appears to flow from the following findings and conclusions:
I conclude that the record does not demonstrate that the Navy fulfilled its duty of conducting a thorough and comprehensive review of the new scientific information on the biological effects of extremely low frequency electromagnetic radiation and the significance of that information.
I find that the record does not demonstrate that the Navy provided an explanation or additional data for its decision not to prepare and file a supplemental environmental impact statement.
In summary, I find and conclude that in proceeding with the reactivation of Project ELF without undertaking a thorough and comprehensive review of the significant new information on biological effects of electromagnetic radiation that has been generated since 1977 the Navy abused its discretion. In so proceeding, the Navy acted in violation of the National Environmental Policy Act.
Id. at 1364.
. The dissent contends that our view of the evidence is “virtually de novo " and “extraordinary.” See infra at 428. This contention mis-characterizes our analysis. We review the record evidence not to refute the specific conclusions of the district court, which were made under a mistaken impression of what NEPA and the CEQ regulations require, but to demonstrate that such evidence falls far short of establishing the level of significance necessary to sustain a finding of arbitrary or capricious behavior on the part of the Navy.
. Although it is not our responsibility to assess the credibility of the witnesses, we note that statements made by plaintiffs’ witnesses, Drs. Wertheimer, Moulder, and Becker, and their credentials could only make their testimony less persuasive than that of the Navy’s witnesses. Dr. Becker, an orthopedic surgeon and professor of orthopedic surgery who has written numerous articles dealing with the biological effects of extremely low frequency electromagnetic radiation, charged that low frequency fields could be responsible in part for numerous diseases, including everything from Legionaires Disease to AIDS. He stated that he regretted that he had not been in a position to stop the electrification of America in the early 1900’s. Although Dr. Becker had previously testified in another proceeding that an intensity of 150 volts per meter was a safe exposure level for low frequency electromagnetic radiation, he refused to ratify his prior testimony; he stated that even though he had given the testimony about the 150 volts per meter previously, he did not, in fact, believe it to be true at the time he testified to it under oath.
Dr. Moulder candidly admitted that none of his previous work was in the field of nonioniz-ing radiation and that others would be better qualified to comment on the new information. He also stated that he was generally unfamiliar with pre-1977 literature in this area, and with the 1977 National Academy of Sciences report in particular, and that he limited his selection of post-1977 materials for discussion at trial to those he thought showed that extremely low frequency fields pose a health hazard, a somewhat biased approach to say the least.
Plaintiffs third expert, Dr. Wertheimer, who testified in part about two of her own studies, is a non-tenured clinical associate professor in the
Department of Preventive Medicine in Biometrics at the University of Colorado Medical Center. She admitted that prior to the two current studies about which she testified, she had not done any research, writing, or field work in the area of electromagnetic radiation and its relationship to human health. Her own work, she candidly conceded, was only exploratory and suffered from several methodological weaknesses.
Testifying for the Navy were Dr. Straub, a physician and scientist with a joint appointment as Associate Chief of Staff of Research at the Veteran’s Administration Hospital in Little Rock, Arkansas, and as Professor of Medicine and Professor of Biochemistry at the University of Arkansas Medical Sciences campus, and Dr. Justesen, Professor of Psychiatry at the University of Kansas School of Medicine and Director of the Behavioral Radiology Laboratories of the Veteran’s Administration Medical Center in Kansas City, Missouri. Dr. Straub has worked for a number of years in the area of nonionizing radiation, has published several articles on the subject, and serves on an advisory committee on biological effects of electromagnetic radiation for the International Electrical & Electronic Engineers Society. Dr. Justesen is president-elect of the Bioelectromagnetic Society (which was identified by Dr. Moulder, one of plaintiffs’ witnesses, as the only scientific organization devoted to the kind of radiation relevant to this case), served as chairman of the International Electrical & Electronic Engineers Society’s Committee on Man and Radiation, and has performed research in the area of extremely low frequency nonionizing electromagnetic radiation.
. The district court noted three situations in which issuance of an injunction would not be required: (1) when there has been a technical violation of NEPA but not of the Act's substantive requirements, so that the agency cannot be said to be proceeding in ignorance of possible adverse environmental consequences; (2) when there are other circumstances making it unlikely that the agency will proceed with the project before the statement is filed; and (3) when the proposed action will not lead to irrevocable commitment of resources so as to foreclose alternative choices.
. The district court itself concluded that there would be no imminent irreparable damage to the environment or human health that would be avoided by entering injunctive relief. The SEIS, of course, would not relate to the Wisconsin and Michigan construction activities themselves, which were adequately reviewed in an Environmental Assessment. Although the SEIS would reconsider the biological effects of extremely low frequency electromagnetic radiation, the injunction does not enjoin the Navy from operating the existing ELF facility in Wisconsin at its current level.
. Plaintiffs filed a motion to stay our order of June 13, 1984, vacating the district court’s injunction, see supra note *, pending disposition of a petition for a writ of certiorari to be filed by plaintiffs. That motion is unanimously denied.