OPINION

Overview

The Price Road Neighborhood Association (“PRNA”), an association of residents living near a proposed freeway interchange, appeals the district court’s grant of summary judgment in favor of the Arizona Department of Transportation (“ADOT”) and the U.S. Department of Transportation (“USDOT”). The PRNA claims that the defendants violated the National Environmental Policy Act (“NEPA”), 42 U.S.C. § 4332 , as well as the procedural requirements mandated by the Council on Environmental Quality (“CEQ”) and the Federal Highway Administration (“FHWA”) regulations by failing to conduct a supplemental Environmental Assessment after the defendants modified the original design of the interchange. We must determine whether NEPA and the regulations require the FHWA to perform a supplemental Environmental Assessment in light of proposed design changes, or whether they authorize the agency first to conduct an environmental reevaluation in order to determine if the new design’s environmental impacts are significant or uncertain and thus warrant further documentation. Because we conclude that a reevaluation is an appropriate vehicle to determine whether a design change will produce significant or uncertain impacts calling for further assessment, we affirm.

The PRNA also contends that the agencies failed to afford adequate public participation opportunities and that the agencies’ conclusions, based on an inadequate reevaluation, were arbitrary and capricious. We conclude that these contentions are without merit.

Background

This dispute arises out of the partial redesign of a federally-funded freeway interchange, the “Price Interchange,” being built by the ADOT at the intersection of U.S. 60, the Superstition Freeway, and S.R. 101, the Price Freeway or the Outer Loop Highway. The interchange is located in a Tempe, Arizona residential area and includes eight interchange ramps.

In 1987-1988, pursuant to NEPA regulations, the ADOT conducted an Environmental Assessment (“EA”) of the interchange proposal. On September 7,1988, the FHWA completed its review of the EA and issued a Finding of No Significant Impact (“FONSI”), indicating the proposal would have no significant effect on the human environment. The preferred design at that time was a four-level, fully-directional interchange, which in- *1508 eluded two below-ground, fully-enclosed tunnels. The tunnels were chosen partially to allay concerns of nearby residents about the original design’s total ramp height of 50-feet above-ground. The tunnels reduced the total ramp height to 25-feet above-ground (one-level).

In January 1995, the ADOT informed the FHWA that it wanted to use two fully-directional loop ramps rather than the two tunnel ramps. The primary reason for the proposed change was cost savings in construction and maintenance. The FHWA required the ADOT to prepare a Change of Access Report (“CAR”) and to conduct an environmental reevaluation in order to determine the continuing validity of the original EA/FONSI.

The ADOT conducted a public meeting on April 5, 1995 to inform the public of the proposed modification. FHWA officials reported significant public opposition to the redesign. Subsequent to this meeting, two citizens, James Peterson and Rick Schuster, submitted alternative designs for the two ramps. The ADOT modified its proposal and adopted a semi-directional ramp design based on the Peterson plan. A second public meeting was held on June 14,1995 to discuss the modified design as well as other alternatives, including the Schuster plan. The Schuster plan proposed 50-foot below-ground, fully-directional, unenclosed tunnel ramps.

On June 11, 1996, the ADOT submitted its final CAR, which included as Appendix B the Environmental Reevaluation, comparing the environmental impacts of the original design with those of the proposed redesign. The preferred design called for two semi-directional, open-air loops, placed between the ground and 25-foot below-ground levels. The overall height of the interchange did not increase. On June 14, 1996, the FHWA approved the CAR and the Environmental Reevaluation, making the following findings:

On October 9, 1996, the PRNA filed a complaint against the USDOT, the ADOT, and certain officials of each agency, seeking declaratory and injunctive relief. The PRNA also filed a motion for a preliminary injunction. The defendants opposed the preliminary injunction and moved for summary judgment.

The district court denied the preliminary injunction, held that extra-record material submitted by the PRNA (an air quality and noise study by their expert, Goddard & Goddard Engineering) was not admissible, and granted summary judgment in favor of the USDOT and the ADOT. The PRNA appeals. The PRNA also filed an emergency motion for injunction pending appeal; this Court denied the motion but ordered the appeal expedited.

Standard of Review

Two standards govern our review of an agency’s NEPA actions. We review factual disputes, which implicate substantial agency expertise, under the arbitrary and capricious standard. Marsh v. Oregon Natural Resources Council, 490 U.S. 360, 376-77 , 109 S.Ct. 1851, 1859-61 , 104 L.Ed.2d 377 (1989); Greenpeace Action v. Franklin, 14 F.3d 1324, 1330-31 (9th Cir.1992). We review legal disputes under the reasonableness standard. Alaska Wilderness Recreation & Tourism Ass’n v. Morrison, 67 F.3d 723, 727 (9th Cir.1995).

Discussion

I

There is no question that the FHWA must determine whether the proposed project change will have significant impacts on the environment. Moreover, there is no question that, if the environmental impacts resulting from the design change are significant or uncertain, as compared with the original de *1509 sign’s impacts, a supplemental EA is required. The dispute, then, centers around what vehicle the FHWA may use to make the initial significance determinations. The NEPA-implementing regulations require supplemental documentation only if a project change meets a certain threshold standard; they further provide a reevaluation procedure to assess the continuing validity of prior environmental impact determinations. We conclude that conducting an environmental reevaluation is an appropriate procedure for the FHWA to determine the significance of the impacts produced by the modified design and whether a supplemental EA is required.

Section 102 of NEPA requires that all agencies prepare an environmental impact statement (“EIS”) for major federal actions “significantly affecting the quality of the human environment.” 42 U.S.C. § 4332 (2)(C). NEPA “does not mandate particular substantive results, but instead imposes only procedural requirements.” Laguna Greenbelt, Inc. v. United States Dep’t of Transp., 42 F.3d 517 , 523 (9th Cir.1994). NEPA requires an agency to take a “hard look” at the potential environmental consequences of proposed projects before taking action. Baltimore Gas & Elec. Co. v. Natural Resources Defense Council, Inc., 462 U.S. 87, 97 , 103 S.Ct. 2246, 2252 , 76 L.Ed.2d 437 (1983). NEPA also imposes a continuing duty to supplement previous environmental documents. Stop H-3 Ass’n v. Dole, 740 F.2d 1442, 1463 (9th Cir.1984).

The CEQ regulations, supplemented by the FHWA regulations, implement NEPA. For a freeway interchange project, such as the Price Interchange, the regulations require the FHWA to prepare an EA. 40 C.F.R. § 1501.4 (b); 23 C.F.R. § 771.115 (c) (requiring EAs in “[a]etions in which the significance of the environmental impact is not clearly established”). Based on the significance determinations made in the EA, the agency must then prepare either an EIS or a FONSI. 40 C.F.R. § 1501.4 (e), (e). Here, the FHWA issued a FONSI.

An agency’s NEPA responsibilities do not end with the initial assessment; supplemental documentation “is at times necessary to satisfy the Act’s ‘action-forcing’ purpose.” Marsh, 490 U.S. at 371 , 109 S.Ct. at 1857-58 . Recognizing NEPA’s goal of focusing government and public attention on the environmental effects of agency action, the Supreme Court has stated:

Id. However, an agency need not start the environmental assessment process anew with every change in a project. See id. at 373 , 109 S.Ct. at 1859 (“[A]n agency need not supplement an EIS every time new information comes to light after the EIS is finalized. To require otherwise would render agency decisionmaking intractable ____”) (footnote omitted). Instead, supplementation is required only under certain circumstances. Although the CEQ regulations do not address when an EA must be supplemented or a FONSI revisited, they do provide that an agency shall supplement an EIS if “[t]he agency makes substantial changes in the proposed action that are relevant to environmental concerns.” 40 C.F.R. § 1502.9 (c)(1)(i). Similarly, the FHWA regulations direct that an EIS shall be supplemented whenever the FHWA determines that “[e]hanges to the proposed action would result in significant environmental impacts that were not evaluated in the EIS.” 23 C.F.R. § 771.130 (a)(1) (emphasis added). The same regulation provides that “[w]here the Administration is uncertain of the significance of the new impacts, the applicant will develop appropriate environmental studies, or, if the Administration deems appropriate, an EA to assess the impacts of the changes, new information, or new circumstances.” Id. § 771.130(c) (emphasis added). Notably, supplemental documentation is only required when the environmental impacts reach a certain threshold-i.e. significant (defined at 40 C.F.R. § 1508.27 ) or uncertain. Therefore, the FHWA must initially determine the significance of the impacts brought about by the proposed change *1510 in order to decide whether supplemental documentation is necessary.

Although neither NEPA nor the CEQ regulations discuss how an agency may make that determination, the FHWA has issued a regulation providing for reevaluation of environmental documents. The regulation provides in relevant part:

23 C.F.R. § 771.129 (c) (emphasis added). If reevaluation reveals that an EA/FONSI remains valid, no additional documentation is required; a supplemental EA is only necessary if an EA/FONSI is no longer valid. Thus, while the FHWA must take a “hard look” at the environmental impacts of project changes, whether a supplemental EA is required depends on the significance of the new impacts. Here, the FHWA and the ADOT, pursuant to this regulation, reevaluated the original EA and FONSI to determine whether its analysis and conclusions remained valid in light of the project change. Finding that “there are no discernible differences in the level of environmental impacts when comparing the original configuration ... with the revised configuration,” the FHWA concluded that supplemental documentation was not necessary. At that point, the FHWA was in full compliance with NEPA and was not required to conduct a supplemental EA See Vine Street Concerned Citizens, Inc. v. Dole, 630 F.Supp. 24 (E.D.Pa.1985) (approving FHWA’s reevaluation to determine impact of proposed convention center on expressway project); cf. Laguna Greenbelt, 42 F.3d at 529 (supplemental EIS not required to assess impact of fires when agency took requisite “hard look,” relying on scientific expertise of two coordinating agencies who concluded no new significant impacts resulted); Animal Defense Council v. Hodel, 840 F.2d 1432, 1439 (9th Cir.1988) (agency decision not to supplement EIS in light of new information was reasonable where agency “carefully considered the information, evaluated its impact, and supported its decision not to supplement with a statement of explanation”), modified, 867 F.2d 1244 (9th Cir.1989).

The PRNA lodges several objections to the FHWA’s using this reevaluation process rather than conducting a supplemental EA It claims, for example, that a supplemental EA is required under the same circumstances as an initial EA — when the significance of the environmental impact is uncertain, see 23 C.F.R. § 771.115 (e); that a supplemental EA is, at minimum, required when an SEIS would be required — if the agency makes substantial changes that are relevant to environmental concerns, 40 C.F.R. § 1502.9 (c)(1)(i), or that result in significant, unevaluated impacts, 23 C.F.R. § 771.130 (a)(1); and that the FHWA improperly substituted a reevaluation for a required supplemental EA. What the PRNA overlooks in each of these arguments is that the environmental reevaluation was used to make the initial significance determination, not to supplant any documentation that would be required if the threshold were met. By addressing the impacts caused by and unique to the redesign in its reevaluation, the agencies have taken the requisite “hard look” at the environmental consequences of its action. NEPA’s goal of “focusing agency attention” to “ensure[ ] that the agency will not act on incomplete information, only to regret its decision after it is too late to correct” has been met. Marsh, 490 U.S. at 371 , 109 S.Ct. at 1857 . To require more would task the agencies with a sisyphean feat of forever starting over in their environmental evaluations, regardless of the usefulness of such efforts.

Thus, we conclude that, when faced with a project change, the FHWA may conduct a reevaluation to determine the significance of the new design’s environmental impacts and the continuing validity of its initial EA A supplemental EA is not automatically required under the regulations, but rather its necessity is dependent upon the findings and conclusions reached by the FHWA through its reevaluation process.

*1511 II

One of the twin aims of NEPA is active public involvement and access to information. See Robertson v. Methow Valley Citizens Council. 490 U.S. 332 . 349, 109 S.Ct. 1835, 1845 , 104 L.Ed.2d 351 (1989). The PRNA argues that the reevaluation process in this case violated explicit statutory and regulatory public participation requirements. The FHWA regulations outline public involvement, generally requiring public hearings or the opportunity to comment on proposed projects during the environmental evaluation process. 23 C.F.R. § 771.111 (h). When a reevaluation is conducted, the regulations provide that “the FHWA and the State highway agency will determine whether changes in the project or new information warrant additional public involvement.” Id. § 771.111(h)(3).

The ADOT held two public meetings, on April 5, 1995 and on June 15, 1995, to discuss the redesign of the Price Road Interchange. Not only did neighborhood residents voice their concerns at those meetings, members of the public actively participated in the design process by submitting alternatives. Moreover, the participation was meaningful. The agencies’ serious consideration of the community’s input is evidenced by the fact that the ADOT adopted a citizen’s alternative design. Athough the agencies did not hold additional meetings while the reevaluation was conducted, that decision did not violate the regulations.

Ill

The PRNA next challenges the adequacy of the reevaluation, contending that the agencies’ determination that the proposed redesign produced “no discernible differences” was arbitrary and capricious. In determining whether an agency’s action is arbitrary and capricious, we must consider whether the agency has taken the requisite “hard look” at the environmental conse1uences of the ProPosed action. Greenpeace, 14 F.3d at 1332 . We must carefully review the record “to ascertain whether the agency deeision is ‘founded on a masoned evaluation of the relevant factors.’ ” Id. (quoting Marsh, 490 U.S. at 373-74 , 109 S.Ct. at 1858-59 (internal quotation omitted)). “Once we are satisfied that an agency’s exercise of discretion is truly informed, “we must defer to that informed discretion.’” Id. (quoting Marsh, 490 U.S. at 377 , 109 S.Ct. at 1861 (internal quotation omitted)). Again, it is important to keep in mind that NEPA does not mandate substantive results, but only imposes procedural requirements.

The PRNA argues that the agencies failed to adequately identify and analyze at least six potentially significant impacts in the environmental reevaluation: safety (including hazardous cargo) and traffic congestion, air quality, noise, visual impact, vibration, and waste material. After carefully reviewing the agencies’ analysis of each of these areas and the flaws alleged by the PRNA, we conclude that the agencies’ decision was founded on a reasoned evaluation of the relevant impacts and was truly informed. The agencies’ actions were not arbitrary and capricious.

We note that the PRNA sought to engage in a battle of the experts with regard to several impacts, including air quality and noise, offering their own studies to contradict those of the agencies. We have consistently rejected such attempts, noting that “when specialists express conflicting views, an agency must have discretion to rely on the reasonable opinions of its own qualified experts even if, as an original matter, a court might find contrary views more persuasive.” Greenpeace, 14 F.3d at 1332 (quoting Marsh, 490 U.S. at 378 , 109 S.Ct. at 1861 (alteration omitted)). The methodology used and conclusions reached by the agencies were supported by adequate scientific data. 1

*1512 Conclusion

A supplemental EA is not required in response to every design change. The NEPAimplementing regulations authorize the FHWA to use the reevaluation procedure to determine whether a project change will cause significant or uncertain environmental impacts. In their reevaluation, the ADOT and the FHWA adequately assessed whether the ramp design would produce significant or uncertain effects and concluded, based on valid scientific opinion, that there was no discernible difference between the impacts of the tunnel design and those of the ramp design. The agencies thus met the requirements of the statute and regulations and fulfilled NEPA’s goal of informed decision-making.

. For these same reasons, the district court did not abuse its discretion in excluding the PRNA's expert studies (the Goddard Report). Initially, the material does not fall within an exception to the rule that judicial review of agency action is limited to the administrative record. Animal Defense Council v. Hodel, 840 F.2d 1432 , 1436-38 (9th Cir.1988), modified, 867 F.2d 1244 (9th Cir. 1989). Moreover, the Goddard Report's contrary conclusions about the significance of the air and noise impacts represent a difference of scientific opinion that does not undercut the agencies’ conclusions, which were based on substantial data and the reasonable opinions of their *1512 qualified experts. See Greenpeace, 14 F.3d at 1332-33 .