The plaintiffs and the intervenor challenge the validity of a permit issued by the United States Army Corps of Engineers (the “Corps”) to Paul Fornell to build a facility for “transloading” coal from trucks to barges on the Mississippi River at Warsaw, Illinois. The plaintiffs and the inter-venor argue that the Corps should not have issued the permit without preparing an Environmental Impact Statement, that the Corps improperly conducted its public welfare review and its consideration of alternatives to the proposed facility and that the district court improperly excluded certain evidence in its approval of the permit. We agree with the district court that the Corps could properly decline to prepare an Environmental Impact Statement, and we find that the district court could properly exclude extrinsic testimony in its review of the Corps’ decision. However, because the administrative record does not permit us to *635conclude that the Corps adequately evaluated the economics of and alternatives to the proposed facility, we affirm in part and vacate and remand in part.
I.
On November 19, 1981, Paul Fornell applied to the Rock Island District of the Corps for a permit to construct and operate a facility in Warsaw, Illinois that would transload coal from trucks to barges on the Mississippi River. The proposed facility consists of an access road from the highway, a dumphouse in which tractor-trailer rigs dump coal into an underground hopper, a conveyor belt beneath the hopper that carries the coal underground to the river bank and then, supported by piers, some 300 feet into the river, and a hinged-boom loading chute on a dock where barges are moored and filled with the coal. Fornell (the “applicant”) intends to trans-load some 485,000 tons of coal per year, trucking it from the Freeman United Coal Mining Company mine (the “Freeman Mine”) near Industry, Illinois to his facility, then towing it in barges to Muscatine Power and Water Company (“Muscatine Power”) in Muscatine, Iowa. The applicant estimates this operation will require approximately 129 truck round-trips per day and 58 barge round-trips per year during the operating season of March 15 to December 1.
The proposed site of the facility lies within the Warsaw Historic District (part of the National Register of Historic Places) and includes a noteworthy brewery complex, dating from 1868. The site lies near the Fort Edwards State Historic Monument— the westernmost frontier post in the War of 1812 and now a scenic outlook — and abuts a portion of the Great River Road, a National Scenic Highway along the Mississippi River. The proposed truck route between the Freeman Mine and this facility would pass through three state and private nature reserves totaling some 900 acres, and through a Western Illinois University life-science field station. These preserves are sanctuaries for many species of wildlife and serve as a major wintering area for bald eagles, with at least 50 eagles present in each recent winter and more than 450 in the winter of 1978-79. See Administrative Record (“AR”) at 679.
The Corps Rock Island District Engineer, in order to address these and other concerns, conducted a public hearing in Warsaw on March 18, 1982, solicited opinions from many local, state and federal agencies and distributed a draft Environmental Assessment to approximately 630 individuals and institutions for comments. After conducting the hearing and reviewing the hundreds of comments and opinions in support of or in opposition to the proposed facility, the district engineer concluded that it was not in the public interest to issue the permit. Because the Governor of Illinois favored the proposal, however, Corps regulations required that the proposal be evaluated by the Corps Division Engineer in Chicago. The division engineer reviewed the record from the district and solicited additional information from the public and government agencies. Upon his review the division engineer determined that an Environmental Impact Statement was not required and that the permit was in the public interest. On January 10, 1984, the permit was granted with several special conditions designed to mitigate adverse impacts of the facility. Among these special conditions were requirements that the facility not operate between December 1 and March 15, that 200 new trees be planted as visual screening, that no coal storage or backhauling occur on the site, that no barges be fleeted on the permit area and that the applicant repair and restore the brewery building on the site.
In May of 1984, the plaintiffs filed this action in the district court. Plaintiffs are Warsaw residents who live across the street from the proposed site of the coal-loading facility. The people of the State of Illinois, by the attorney general, sought and were granted intervention as of right on January 4, 1985. A bench trial was held, and on November 27, 1985, the district court suspended the permit and remanded the case to the Corps to evaluate the proposal’s impact on the Great River Road, and thereafter to reevaluate environmental consequences and rebalance the public interest. On remand the Corps did that much and more, enlarging the administrative record with some reevaluation of alternatives (by recalculating mileages from the Freeman Mine) and with some findings on recreational and scenic impact. A hearing was held on March 24, 1986 to evaluate compliance with the remand. On April 30, 1986, the district court entered judgment affirming the Corps’ decision to issue the permit. This appeal followed.
*636II.
The Corps is, of course, responsible for evaluating proposed construction projects in navigable waters of the United States. See River and Harbors Appropriation Act of 1899, 33 U.S.C. § 403 (1982). And in that connection it must comply with the National Environmental Policy Act of 1969 (“NEPA”), 42 U.S.C. §§ 4321 to 4347 (1982). Regulations construing NEPA and applying it to the Corps, 40 C.F.R. §§ 1500 to 1508, 33 C.F.R. § 230, outline the procedure the Corps must follow in evaluating a proposed facility. Before issuing a permit the Corps must prepare an Environmental Assessment (an “EA”) to determine whether the proposed facility “could have significant effects on the environment” and therefore require a more comprehensive Environmental Impact Statement (an “EIS”). 33 C.F.R. § 230, App. B(8)(a). Typically the EA is a brief interdisciplinary evaluation of the need for a proposed action, the likely environmental effects of a proposal, the alternatives to a proposal and a list of agency, interest group and public comments received regarding the proposal. See 33 C.F.R. § 230.9(c). In addition to the EA and its Finding of No Significant Impact (a “FONSI”) or its requirement of an EIS, where a proposal “involves unresolved conflicts concerning alternative uses of available resources,” the Corps must “study, develop, and describe appropriate alternatives to recommended courses of action.” NEPA, 42 U.S.C. § 4332(2)(E); see River Road Alliance v. United States Army Corps of Engineers, 764 F.2d 445, 452-53 (7th Cir.1985). cert. denied, — U.S. -, 106 S.Ct. 1283, 89 L.Ed.2d 590 (1986).
Ultimately, the Corps must undertake a general “public interest review” to decide whether a permit should issue. See 33 C.F.R. § 320.4(a). In this general review, the Corps must evaluate a proposal’s overall impact on the public interest, balancing “benefits which reasonably may be expected to accrue from the proposal ... against its reasonably foreseeable detriments,” considering all relevant factors, including more than twenty specific subjects of concern detailed in 33 C.F.R. § 320.4(a). In reviewing this public interest determination by the Corps, it is not our role to second-guess. We merely consider whether the Corps followed required procedures, evaluated relevant factors and reached a reasoned decision. As this circuit has noted, whether we dub the standard one of “reasonableness” or one to determine if a decision was an “abuse of discretion,” our role is to decide whether the Corps exceeded its decision-making authority in issuing the permit. See River Road Alliance, 764 F.2d at 449-50; Wisconsin v. Weinberger, 745 F.2d 412, 417 (7th Cir.1984). As has often been noted, we must see to it that the agency took a “hard look” at the environmental factors implicated and based its decision on a rational consideration of relevant factors. 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); Wisconsin v. Weinberger, 745 F.2d at 417; Sierra Club v. Peterson, 717 F.2d 1409, 1413 (D.C.Cir. 1983).
Here the Corps produced an administrative record of some 2,600 pages and ultimately issued a permit hedged about by 23 special conditions designed to mitigate adverse environmental impacts. The district court engaged in extensive review of the factual record, considered plaintiffs’ several challenges in some detail and remanded the case once for further evaluation before concluding that the Corps’ decision to issue the permit should be upheld. But this has not stilled the plaintiffs’ or the attorney general’s objections.
As a threshold matter, we do not agree with Fornell’s suggestion that this court no longer has jurisdiction because construction of the facility has begun. See Fornell Brief at 38-41. The plaintiffs and the in-tervenor claim that the permit was improperly issued. The injuries that flow from this unsanctioned issuance, they claim, occur both in the construction and in the operation of the facility. If plaintiffs and the attorney general were to succeed in showing that the permit had to be revoked or modified, we presumably could order that the facility be dismantled, altered or *637operated differently. See Columbia Basin Land Protection Ass’n v. Schlesinger, 643 F.2d 585, 591 n. 1 (9th Cir.1981); Burbank Anti-Noise Group v. Goldschmidt, 623 F.2d 115, 116 (9th Cir.1980), cert. denied, 450 U.S. 965,101 S.Ct. 1481, 67 L.Ed.2d 614 (1981). Thus this appeal is not moot.
A. The Finding of No Significant Impact
Plaintiffs argue both that the Corps improperly determined that an Environmental Impact Statement was unnecessary and, what amounts to the same thing, that its FONSI was inadequately justified by the Environmental Assessment. Appellants’ Brief at 9. We agree with the district court that the Corps’ FONSI and its decision not to require an EIS were not arbitrary or capricious and evidenced a “hard look” at environmental factors. See Abbema v. Fomell, No. 84 C 1153, mem. op. at 19 (C.D.Ill. Nov. 27, 1985) (“Nov. 27 mem. op.”); mem. op. at 12 (C.D.Ill. April 30, 1986) (“April 30 mem. op.”).
Obviously, what comprises a “significant” impact is hard to define. See, e.g., River Road Alliance, 764 F.2d at 450 (describing difficulties). Nonetheless, the regulations provide a list of relevant factors. See 40 C.F.R. § 1508.27. Here the Corps substantially evaluated possible threats to public health and safety, see, e.g., AR at 890-93, 2542 & 2563-65; 40 C.F.R. § 1508.-27(b)(2), unique aspects of the proposed site such as nearby parks or cultural resources, see, e.g., AR at 1044-45 & 2565; 40 C.F.R. § 1508.27(b)(3), whether special cumulative effects are likely, see, e.g., AR at 3-21, 679-88 & 2463-67; 40 C.F.R. § 1508.-17(b)(7), the degree to which historic districts and landmarks would be adversely affected, see, e.g., AR at 1061-69, 1273-1357, 1957-62 & 2563-64; 40 C.F.R. § 1508.27(b)(8), and the degree to which endangered species might be threatened, see, e.g., AR at 34-62, 677-88 & 2561-62; 40 C.F.R. § 1508.27(b)(9). The Corps gathered comments from the U.S. Fish and Wildlife Service, the Environmental Protection Agency, the National Advisory Council on Historic Preservation, the Federal Highway Authority, the Governor of Illinois, the Illinois Department of Transportation, the Illinois State Historic Preservation Office and the mayor and city council of Warsaw, outlining their concerns about the impacts of the proposed facility. The Corps fashioned twenty-three special conditions to attach to the permit to mitigate adverse environmental effects. Thus, there is a total ban on construction and operating activity between December 1 and March 15 during the peak eagle season. The permit includes permission to cut down only four trees and, in fact, orders that two-hundred new trees be planted. The tying of barges to trees is prohibited. In addition, work must stop immediately upon the discovery of new cultural, archaeological or historical features. See AR at 2378-80. Like the district court, we might have pursued certain factors differently, such as the effect of truck noise, the impact on nature preserves abutting the proposed route and the unrestricted hours of operation. The Corps, however, undertook its assessment properly; it considered the relevant factors and it made rational decisions about the likely impacts of the proposal. The numerous special conditions that the Corps imposed on the permit also suggest that the Corps took an adequately detailed and critical look at environmental impacts and concluded that no significant effects would follow either issuance or denial of the permit. We must conclude, therefore, that in reaching its FONSI, the Corps did not act arbitrarily or abuse its discretion. Cf. Boles v. Onton Dock, 659 F.2d 74 (6th Cir.1981) (affirming issuance of permit for barge coal-loading facility without preparation of EIS); James v. Tennessee Valley Authority, 538 F.Supp. 704 (E.D.Tenn. 1982) (same).
B. The Issuance of the Permit
A Finding of No Significant Impact does not end the Corps’ responsibilities. Relative absence of significant environmental effect does not translate directly into overall social benefit. The Corps must follow two distinct, if parallel, guidelines in issuing a permit after a FONSI. First, *638section 102(2)(E) of NEPA, 42 U.S.C. § 4332(2)(E), requires all federal agencies to “study, develop, and describe appropriate alternatives to recommended courses of action in any proposal which involves unresolved conflicts concerning alternative uses of available resources.” And second, Corps regulations require that a permit shall issue only after a general “public interest review” determines that the benefits outweigh the detriments of a proposal. 33 C.P.R. § 320.4(a). The plaintiffs and the attorney general argue that the Corps did not adequately weigh various factors in striking the public interest balance and in considering alternatives.
First, the plaintiffs challenge the Corps’ treatment of local zoning and land-use issues, claiming that the site of the proposed facility was never described precisely enough to evaluate these issues. See Appellants’ Brief at 16-22. The Corps must, of course, consider zoning and land-use issues, but this consideration is only a part of striking the general balance. See 33 C.F.R. § 320.4(e) and (j)(2). The Corps is not a zoning enforcement agency; rather, its role is to take into account “[historic, cultural, scenic, and recreational values,” which are often “reflected” in local land-use and zoning laws. Id. The Corps considered these factors extensively and had an adequate basis for believing that the project could be brought into compliance with local law. See Nov. 27 mem. op. at 6-7, 11-12.
The plaintiffs also contend that the Corps inadequately considered the proposed facility’s effect on a local road (part of the national Great River Road that traces the Mississippi River) and its surroundings. See Appellants’ Brief at 22-33; 23 U.S.C. 148. However, not only did the Federal Highway Authority approve the proposed use of the road as consistent with the Great River Road project, see AR at 2390, but the Corps received much information about nature preserves and wildlife stations along the road, including winter eagle roosts. See, e.g., AR at 34-63, 677-88, 2561-66. We have found that the Corps’ FONSI was proper, including considerations affecting the road and its environs. We cannot say the Corps’ consideration of effects on the road was arbitrary, capricious or contrary to law.
The plaintiffs further argue that the Corps did not properly weigh the effects of the proposed operation, such as truck noise, coal dust, barge loading and convey- or operation, on the human environment. Indeed, no quantitative noise study was done. And no quantitative analysis or modeling of atmospheric conditions and potential coal-dust leakage was done. But the Corps, based on its experience, on visits to the site and on expert testimony, evaluated what the effects of noise would be on humans and animals, and applied its experience and considered expert testimony with respect to the possibility of coal-dust leakage. See Transcript (“Tr.”) at 78-83, 105-06, Abbema v. Fomell (C.D.Ill. July 23, 1985, morning session); Nov. 27 mem. op. at 9-10. And the size, speed, noise and emissions of the trucks are regulated by the state. The Corps did not weigh these concerns arbitrarily or capriciously.
The most serious challenges to the Corps’ handling of its public welfare review involve economics and the related study of alternatives as required by NEPA. 42 U.S.C. § 4332(2)(E). At the outset we note that the evaluation of “alternatives” mandated by NEPA is to be an evaluation of alternative means to accomplish the general goal of an action; it is not an evaluation of the alternative means by which a particular applicant can reach his goals. In the current proposal the general goal is to deliver coal from mine to utility. See AR at 2559-60 (Final EA). In some discussion of alternatives to the proposal, the Corps has suggested that an alternative may not be feasible at least partly because the applicant does not own the necessary land or perhaps cannot gain access to it. See, e.g., AR at 1072, 1073 (Preliminary Case Report). The fact that this applicant does not now own an alternative site is only marginally relevant (if it is relevant at all) to whether feasible alternatives exist to the applicant’s proposal. This is particularly *639true because an existing facility in Quincy, Illinois is presently transloading the mine’s coal from truck to barge.
In evaluating the “economics” of a project the Corps need not produce microe-conomic models of the applicant’s venture to predict its behavior for the forseeable future. But, on the other hand, the Corps may not finesse any serious economic analysis by finding merely, that “the economic viability of this proposed project must be ultimately determined in the free market.” AR at 2552 (Division Engineer Findings of Fact). Although competitive forces may have the last word, Corps regulations require that the Corps initially conduct an overall public welfare review. 33 C.F.R. § 320.4(a). Where, as here, a proposal’s benefits are entirely economic and its costs environmental, the Corps must make at least a minimally reliable effort to establish economic benefit. Obviously, the Corps is not an expert in every business seeking a permit, but, as guardian of the public welfare, it must credibly attempt to appraise economic benefit.
Here the Corps did undertake a serious review of alternatives to the proposal and did consider significant amounts of economic data. Superficially at least there appears to be record evidence to support the Corps’ decision that no preferable alternatives exist and that the economic benefits “slightly” outweigh any adverse consequences of the proposal. See AR at 2547-53 (Division Engineer Findings of Fact). Generally that would end the matter, and we would agree with the district court that the Corps’ decision was not arbitrary or capricious and would uphold the issuance of the permit. However, the attorney general, in particular, contends that the Corps based its conclusions on evidence so inadequate and misleading that we must undertake a closer scrutiny. The plaintiffs and the intervenor have a point to the extent that the Corps’ conclusions must find some reasonable support in the record. See Bethlehem Steel v. EPA, 638 F.2d 994, 1004 (7th Cir.1980) (agency must create record adequate to permit review). The Corps certainly may utilize reports and facts derived from outside reports and sources, see 33 C.F.R. § 230, App. B; Save Our Wetlands, Inc. v. Sands, 711 F.2d 634, 643 (5th Cir.1983), but the Corps is responsible for the independent verification of specifically challenged information obtained from applicants or outside consultants. See 33 C.F.R. § 230, App. B(3) and B(8)(6); Save Our Wetlands, 711 F.2d at 643; Sierra Club v. United States Army Corps of Engineers, 701 F.2d 1011, 1035 (2d Cir.1983); Johnston v. Davis, 698 F.2d 1088, 1095 (10th Cir.1983). If the Corps bases its conclusions on entirely false premises or information, even when its attention is specifically directed to possible defects in its information, we would have difficulty describing its' conclusions as reasoned; we would have to call them arbitrary and capricious. See Sierra Club v. United States Army Corps of Engineers, 701 F.2d at 1035 (“a decision made in reliance on false information, developed without an effort in objective good faith to obtain accurate information, cannot be accepted as a ‘reasoned decision’ ”); Johnston v. Davis, 698 F.2d at 1095.
The district court approved the Corps’ public interest review of economic factors, stating: “an economic examination was made ... costs were balanced against benefits, and ... the conclusion reached by the Corps was based on study and analysis.” Nov. 27 mem. op. at 14. The court acknowledged the plaintiffs’ claims that the Corps must verify challenged information, but found that the Corps did conduct “some small amount of independent research” that indicated that the City of Warsaw and the State of Illinois supported the project as economically beneficial. The court found that the Corps “believed that the ultimate conclusion of the Bums/McDonnell Report [Burns II, infra ] would have been justified even if a different [transportation] route had been utilized.” Id. We do not agree that the Corps may respond to specific challenges to reports underlying its economic analysis merely by stating generally that the affected city and state are in favor of a proposal. *640Specific challenges to a report used by the Corps in its public interest review require specific responses or a determination that the report is not being relied upon in its challenged aspects. Here we think the district court misapplied River Road Alliance, 764 F.2d at 452-53, when it construed that case to validate the Corps’ treatment of alternatives without responding in any but the most superficial fashion to the challenges brought by the plaintiffs. See Nov. 27 mem. op. at 15-16. When the Corps receives particularized objections to material upon which it importantly relied in its review, the Corps must undertake some independent effort to verify or discredit the challenged material. No doubt there are several coal docks along the Mississippi River from which Freeman Mine coal may be shipped to Muscatine Power, and competition may ultimately identify the most efficient point of shipment. This does not mean that, in issuing a permit for a new loading facility, the Corps is relieved of all responsibility for verifying economic information justifying the new construction.
1. Use of 1980 Burns & McDonnell Report
In 1980 the Bums & McDonnell Engineering Co. prepared a report (“Bums I”) for Muscatine Power that considered alternative ways to transport coal from the Freeman Mine to Muscatine Power. The basic conclusion of this report was that the optimal transportation arrangement would be to truck coal to an existing but unused dock in Warsaw (not the facility proposed by Fornell), transload coal to barges and tow the barges upriver to Muscatine Power. See AR at 1505-07 (Burns I). Comparing a facility in Keokuk, Iowa, with the existing dock in Warsaw, Burns I concluded that as much as $810,000 per year could be saved in transportation costs by using the existing Warsaw dock rather than the Keokuk location. See id. at 1501. The division engineer in his findings supporting the issuance of the permit to Fornell cited “the potential of as much as $810,000 annual savings” to electricity consumers. Id. at 2192 (Division Engineer Statement of Findings). However, this savings was calculated on the basis of an existing facility in Warsaw and obviously could not be applied directly to validate a new facility that will cost $2 million to build. Several commen-tors took issue with the savings figure, but the division engineer responded: “I have reviewed the economic data provided by the applicant and that data appears reasonable for the purpose of determining the overall economics of the proposed project.” Id. at 2193. It is simply unreasonable to estimate potential cost savings based upon a report analyzing an entirely different facility without explaining why or to what extent the data is applicable to the proposed facility.
Later the Corps reconsidered the estimated savings, with attention to the cost of constructing the proposed facility, by using another report (Burns II) prepared by Bums & McDonnell Engineering Co. at the request of Fornell. See AR at 131-47 (Burns II). This report, which compared the ostensible advantages of the proposed facility over a currently operating facility in Quincy, had other problems that we shall explore below.
2. Comparison of Quincy Facility and Proposed Facility
In both its public interest review and its consideration of alternatives the Corps was required to evaluate the “no-build” alternative. Muscatine Power currently arranges for coal to be shipped from the Freeman Mine through a “temporary” facility on the Mississippi River at Quincy, some 35 miles downriver of Warsaw. A review of the record reveals that the Corps never adequately evaluated Quincy, the “no-build” alternative. The only comments directed to Quincy were based on false data or unexplained assumptions.
In this connection, the Rock Island District of the Corps prepared a Preliminary Case Report (the “PCR”) in October 1982, regarding the proposed facility. See AR at 1029-74. The PCR’s treatment of the Quincy alternative was essentially incorporated in the final EA. See id. at 2560. The PCR posits the need for some coal-loading *641facility: “[t]here is no question as to the need for facilities of this kind along the Mississippi River,” id. at 1050, and then evaluates alternative sites, including Quincy. The analysis of Quincy as an alternative site is drawn nearly verbatim from Fornell’s letter of June 9, 1982 in support of his application. See id. at 1183-89. Fornell’s letter and the PCR both suggest that “the economic consequences to be borne by the [electricity] consumers do not justify this location [Quincy] as a permanent operation.” Id. at 1058, 1183. The economic consequences are then listed as higher truck costs, based on Quincy’s location allegedly farther from the mine, and higher barge costs, based on Quincy’s location farther along the river from Muscatine Power. Plaintiffs and other opponents of the proposed facility argue, however, that the road mileage used was substantially wrong and that a shorter river haul would not result in lower barge costs from Warsaw. See Appellants’ Brief at 36. No one now argues that the mileage figure that was used (round trip mine-to-Quincy 19.4 miles farther than mine-to-Warsaw) is right. In addition, the Corps apparently took no steps to verify that barge rates between Quincy and Muscatine would actually be higher than rates between Warsaw and Muscatine. Thus the economic analysis was apparently based on inaccurate mileage and unverified rate estimates. Cf. Tr. at 83-85, July 23, 1986, morning session. More disturbing for purposes of our review, these deficiencies were brought specifically to the attention of the Corps by commentors. See, e.g., AR at 1702-03, 1829-31, 1892-95, 2068-75, 2081-82.
In April 1983, apparently at Fornell’s request, Burns & McDonnell Engineering prepared the Burns II report comparing the proposed facility to the existing Quincy dock. This report concluded that the proposed facility would save Muscatine Power $0.55 per ton of coal transported, thus some $266,750 per year. See AR at 141-46. Burns II stated that its assumptions included the fact that the Quincy facility is 13.5 miles farther from the mine than is the proposed facility (the difference between 62.5 and 49 miles). Id. at 143. Again, no one now contends that these mileage figures are right. We cannot ascertain precisely how Burns II derives its $0.55 per ton conclusion, but it appears to be based in considerable part on mileage figures. At the time of the report both the mileage and the related analysis of barge costs were challenged as incorrect, but the Corps offered no explanation, and we cannot tell from the record what assumptions were made. In addition, Burns II suggested that the projected $266,750 annual savings would be passed directly on to electricity consumers, resulting in an estimated $2.22 savings on each customer’s average monthly bill (a 6% to 10% reduction). See id. at 134-35. We find it remarkable that Bums II concludes that a $0.55 per ton savings in transportation costs will save each Musca-tine customer more than $2 per month in electric bills.
In June 1983, the Corps’ Rock Island District Engineer recommended that the permit be denied because it was not in the public interest. See AR at 22. The district engineer commented that there was no “overwhelming need” to approve, the proposed facility “when other alternatives are available, and in fact, functioning at the present time in providing Illinois coal to customers.” AR at 231. As noted, because the Governor of Illinois supported the proposed facility, the issue, pursuant to regulations, was forwarded to the division engineer. See 33 C.F.R. § 325.8.
In December 1983, the division engineer released an Environmental Assessment. This EA refers to the Preliminary Case Report as containing the “detailed analysis” of alternatives to the proposed facility. AR at 2201. The EA states that “[a] savings of approximately $810,000 annually on transportation costs could be realized by Muscatine Power and Water Company, if the loading facility was located in Warsaw.” Id. at 2205. This figure of $810,-000, of course, came from Burns I, comparing an existing Warsaw dock to a Keokuk location. The division engineer does not explain why the more recent and more rele*642vant Burns II figure of $266,750 is not discussed. The division engineer also enumerates economic benefits Warsaw would receive if Quincy business were shifted into Warsaw, but recognizes that “jobs created at Warsaw could be lost at Quincy for no net gain to the state.” Id. at 2206. In his Statement of Findings the division engineer concludes that the economics of the proposed facility compared with other possibilities “appear[ ] to be slightly positive.” Id. at 2193.
And in August 1985, after hearings in this case had been conducted in the district court, the Corps revised the economics of the proposed facility using mileage figures different from those used previously — specifically amending Burns I and the comparisons included in the PCR, using a 60-mile rather than a 49-mile distance from the proposed facility to the mine. See id. at 2589-2602 (Fomell Permit Case Economics in Review). The alleged $810,000 annual advantage over Keokuk was thus reduced to $356,400, and the alleged annual advantage over Quincy in track costs was reduced from $519,415 to $165,791. The plaintiffs challenge even these revised figures as inaccurate. They contend that the Corps corrected only one-half of its error. Because the PCR calculation was based on round trips, a 22-mile adjustment was necessary, not an 11-mile adjustment. And the plaintiffs repeat their assertion that Bums I is not relevant to the economics of the proposed facility. We agree that the 11-mile adjustment apparently went only halfway. And we too doubt the relevance of Bums I. The Corps’ obliviousness of these details (which were offered to support the major argument for the project) extended over a period of years and occurred despite specific criticisms pointing out deficiencies. Even though we think a Warsaw loading facility may very well be justified, we cannot approve a public interest review entirely indifferent to the facts. The Corps has a duty to ensure the accuracy of information that is important to the decision it is making, at least when obvious errors are brought clearly to its attention.
3. Evaluation of Alternatives to the Proposed Facility
As is well known, the Corps must evaluate alternatives to the proposed facility in deciding whether or not a permit should issue. See 42 U.S.C. § 4332(2)(E); 33 C.F.R. § 320.4(a). Here the Corps received substantial information about alternatives, but plaintiffs again argue that much of the information was inaccurate and unverified and that this court should remand to ensure that proper consideration is given to options that might mitigate adverse (though “nonsignificant”) environmental effects.
Fomell supplied the Corps with information on six alternatives to his proposed facility and suggested why each was less appropriate than the proposed facility. See AR at 1178-90 (June 9, 1982 Fornell letter). Fornell’s data suggested that each of the six (including an existing but unused facility in Warsaw itself) was either less economical, less environmentally benign, or both. The Corps does not appear to have conducted any substantial investigation of alternatives on its own. See Nov. 27 mem. op. at 14-16; Tr. at 83-85, July 23, 1986, morning session. The Corps may rely on reports prepared by outsiders or applicants, but as we have noted, when such information is specifically and credibly challenged as inaccurate, the Corps has an independent duty to investigate. The Fornell letter seems to rely on inaccurate information, especially with respect to relative distances used to compare transportation costs among alternative sites. See, e.g., AR at 1892-95, 2068-75, 2081-82. As in its consideration of a “no-build” alternative, the Corps in considering alternative sites relies upon a record replete with important factual inconsistencies and ambiguities that the Corps did not attempt to resolve. This does not constitute a “hard look” at alternatives. Rather, it evidences blind reliance on material prepared by the applicant in the face of specific challenges raised by opponents.
*643III.
Fornell and the Corps urge that this court’s decision in River Road Alliance validates the Corps’ heavy reliance on the Fornell report. See River Road Alliance, 764 F.2d at 452-53. But in that decision we suggested that if challengers of the permit “were prepared to shoulder the burden” of specifying some factors overlooked in the applicant’s report, then the Corps must respond. Id. In this case, the plaintiffs and the intervenor direct our attention to a record teeming with specific factual challenges to reports tendered to the Corps. In these circumstances it is the Corps’ responsibility to verify in a reasonable way the data on which it relies.
We therefore believe it is necessary to vacate the grant of the permit and remand for reconsideration of the economic factors involved in the public interest review and for a further consideration of the economics of the alternatives. This does not mean that every Environmental Assessment containing factual inaccuracies will have to be redone. In this case the Corps apparently faced a close decision, evidenced by the disagreement between the district and division engineers and by the vigorous public and scientific debate. The alleged benefits were economic and related to relative transportation costs, yet our review of the record reveals no consistent analysis of what these benefits might actually amount to. An overall public welfare review cannot be deemed reasonable when the economic half of the balance is obscured by a record of miscalculations followed by recalculations apparently intended only to bolster a decision already made. We, of course, urge that any action required by remand be undertaken and completed promptly since final action on this project has already been unconscionably delayed.
We do not remand merely to require a longer record and more paper. We appreciate that “[ujltimately, of course, it is not better documents, but better decisions that count.” 40 C.F.R. § 1500.1(c). But on this record we cannot determine that the Corps’ consideration of alternatives and the public welfare was a reasoned one.
IV.
The plaintiffs contend that the district court erred in refusing to admit testimony from four expert witnesses who would have challenged the adequacy of the administrative record. See Appellants’ Brief at 42-47. We agree that in certain cases admission of new evidence is proper, even desirable, in assisting the court in its evaluation of the sufficiency of an administrative record. Cf. County of Suffolk v. Secretary of the Interior, 562 F.2d 1368, 1384-85 (2d Cir.1977), cert. denied, 434 U.S. 1064, 98 S.Ct. 1238, 55 L.Ed.2d 764 (1978). But the district court has substantial discretion to conclude that any such evidence is helpful, or necessary, or cumulative. In a review of an administrative decision, the administrative record is the focus. The district court here did not abuse its discretion in deciding not to admit testimony of these four witnesses where the administrative record already contained significant material from each of the four.
We therefore affirm in part and vacate and remand in part for further proceedings consistent with this opinion.
Affirmed In Part, Vacated In Part And Remanded.