Opinion for the Court filed by Circuit Judge SILBERMAN.
Appellees challenged the Federal Highway Administration’s approval of plans to replace the Woodrow Wilson Memorial *117Bridge. The district court held that the Administration violated the National Environmental Policy Act and the National Historic Preservation Act. We reverse.
I.
The Woodrow Wilson Memorial Bridge is a microcosm of the Washington, D.C. metropolitan area’s traffic congestion problems. Built in 1961, the six-lane structure carries the Capital Beltway over the Potomac River, connecting the City of Alexandria, Virginia, to Prince George’s County, Maryland; originally intended to serve as a Washington bypass for interstate travelers, it became increasingly used by commuters as the region’s population grew. As a result, traffic volume on the Bridge has increased to over 160,000 vehicles per day, more than twice the capacity the structure was designed to accommodate; congestion is particularly acute during peak hours, where the configuration of an eight-lane Beltway feeding into a six-lane bridge — in addition to steadily increasing local traffic in the surrounding communities — has produced one of the worst rush-hour “bottlenecks” in the region. These congestion problems have created harmful collateral consequences: the heavy volume on the Bridge has contributed to an accident rate nearly double that of similar facilities in the region, and has expedited the deterioration of the Bridge’s structure to the point where the Bridge is projected to be structurally unsound by 2004.
Efforts to replace the Bridge began over ten years ago, when the Federal Highway Administration, in cooperation with its coordinate agencies in Maryland, Virginia, and the District of Columbia, began examining alternative approaches to solving the Bridge’s capacity and structural problems. The Administration began to study the potential effects of rebuilding the Bridge on the surrounding communities early in the project’s development, commissioning surveys of historic and archaeological resources in areas likely to be affected by the projects. The Commission also started the process, mandated by the National Environmental Policy Act (NEPA), 42 U.S.C. § 4321 et seq. (1994), of considering the environmental impacts of alternative project designs. In 1991 the Administration issued a draft Environmental Impact Statement (EIS) for public comment; this statement suggested and compared five proposals for replacing the Bridge. Each of the alternatives in the draft proposed expanding the river crossing from six to twelve lanes, and included a similar expansion of the five-mile Beltway corridor approaching the river crossing from the east and west.1
Reaction to the draft was less than enthusiastic; the Administration was criticized for assessing inadequately the environmental and cultural impacts of its • proposal, and for failing to coordinate its work with that of interested governmental agencies and community groups. By its own admission concerned that “a region-wide consensus about the new bridge had not been reached,” the Administration went back to the drawing board. In response the Administration organized a “Coordination Committee” composed of elected and administrative officials from the region to enhance community and intergovernmental cooperation. The Committee revisited the entire process of developing alternative Bridge designs, ultimately soliciting and considering over 350 proposals from interested individuals and organizations, and increased the Administration’s public outreach efforts in affected communities. In the meantime, pursuant to its obligations under section 106 of the National Historic Preservation Act, 16 U.S.C. § 470f (1995 & Supp.), and section 4(f) of the Department of Transportation Act, 49 U.S.C. § 303 (1997), the Administration continued to assess the project’s potential *118impacts on historic, archaeological, and cultural resources in the area.
In 1997, the Administration issued its Final Environmental Impact Statement (the “Final EIS”). The Final EIS gave detailed consideration to eight alternative proposals (seven “build” alternatives and a baseline “no build” alternative), comparing them on a range of criteria including vehicle capacity, cost, and extent of environmental impacts. As was the case with the draft each of the “build” alternatives scrutinized in the Final EIS had twelve lanes; each alternative also had a lane configuration that separated local and express traffic, and contained a lane dedicated for High Occupancy Vehicle usage. The critical difference among the proposed alternatives was the type of river crossing; the seven “build” alternatives included a range of tunnel and bridge designs. Although the Final EIS discussed narrower eight- and ten-lane options, it did not afford them full treatment as formal “alternatives” because the Administration concluded, on the basis of traffic projections, that narrower river crossings would fall short of meeting the Bridge’s long-term traffic needs. Among the eight options the Administration designated a “Preferred Alternative” that would replace the Bridge with two parallel six-lane drawbridges (one drawbridge for eastbound and one for westbound traffic) clearing the Potomac’s navigational channel by seventy feet at their highest points. The Administration also included in the Final EIS a sixty-page “Section 4(f) Evaluation” identifying and offering plans to mitigate the effects of the Preferred Alternative and all other build alternatives on public parks, wildlife refuges, and historic sites.
After a brief comment period the Administration approved the Preferred Alternative in a Record of Decision and submitted, as is required by section 106 of the National Historic Preservation Act, a Memorandum of Agreement evidencing the Administration’s cooperation with state historic preservation officers in identifying historic sites that might be impacted. The Memorandum identified and offered mitigation plans for several historic sites, but it also noted that the Administration had not yet identified properties to be used for “construction staging, dredge disposal, wetland mitigation, or other ancillary activities” during the period of the Bridge’s construction.
The City of Alexandria filed an action in the district court challenging the Administration’s approval of the project, and the district court permitted three Alexandria-based organizations that opposed the Administration’s proposed alternative (collectively the “Alexandria Coalition” or “appellees”) to intervene as plaintiffs. The City alleged that the Administration had violated a host of regulatory provisions, including the National Environmental Policy Act, section 106 of the National Historic Preservation Act, and section 4(f) of the Department of Transportation Act.2 After both sides had filed for summary judgment the City of Alexandria settled its claim with the Administration, leaving the Alexandria Coalition as the only remaining plaintiffs.
The district court ruled in favor of the Alexandria Coalition. See City of Alexandria v. Slater, 46 F.Supp.2d 35 (D.D.C.1999). The court concluded that the Administration had violated NEPA by not affording detailed consideration to a ten-lane river crossing as a “reasonable alternative” in the Final EIS, and that the Final EIS’ treatment of the temporary environmental impact of the construction phase of the project was too cursory to satisfy NEPA. Relying upon our recent decision in Corridor H Alternatives, Inc. v. Slater, 166 F.3d 368 (D.C.Cir.1999), the district court also determined that the Administration had violated section 106’s re*119quirement that an agency “take into account” the effects of a proposed project on protected historic properties by postponing the identification of the sites that were to be used for construction-related “ancillary activities.” Because an agency must complete the section 106 identification process before it can satisfy section 4(f)’s requirement that an agency use “all possible planning to minimize harm” to historic sites, the court concluded that the Administration had necessarily failed to comply with section 4(f) as well. The district court remanded the project to the Administration; the Administration appealed, as it is entitled to do. See Occidental Petroleum Corp. v. SEC, 873 F.2d 325, 330 (1989) (when district court remand obliges agency to take further actions under an arguably incorrect legal standard an immediate appeal is appropriate).
II.
A.
The National Environmental Policy Act’s mandate “is essentially procedural,” Vermont Yankee Nuclear Power Corp. v. Natural Resources Defense Council, Inc., 435 U.S. 519, 558, 98 S.Ct. 1197, 55 L.Ed.2d 460 (1978); the statute requires that agencies assess the environmental consequences of federal projects by following certain procedures during the decision-making process. See Citizens Against Burlington, Inc. v. Busey, 938 F.2d 190, 193-94 (D.C.Cir.1991). Before approving a project, an agency must prepare a “detailed statement ... [on] the environmental impact of the proposed action, any adverse environmental effects which cannot be avoided should the proposal be implemented, [and] alternatives to the proposed action.” 42 U.S.C. § 4332(2)(C)(i)-(iii). These general prescriptions are given sharper focus in the Council on Environmental Quality’s regulations,3 which require agencies to prepare environmental impact statements; at the “heart of the environmental impact statement” is the requirement that an agency “rigorously explore and objectively evaluate” the projected environmental impacts of all “reasonable alternatives” for completing the proposed action. 40 C.F.R. § 1502.14.
Appellees argue, and the district court agreed, that the Administration violated NEPA by failing to deem a ten-lane bridge a “reasonable alternative” in the Final EIS. They observe that a ten-lane bridge would constitute a significant improvement over the existing six-lane structure, and would reduce congestion with considerably less impact on environmental and cultural resources than each of the twelve-lane alternatives compared by the Administration. In addition to having a narrower river crossing, appellees point out that a ten-lane alternative would have a smaller construction “footprint” along the entire five-mile stretch of the Beltway that will be under construction, and would require smaller interchanges at each of the four points of access to the Beltway in the project corridor. The Administration responds that the ten-lane alternative favored by appellees was excluded after studies determined that it did not meet the traffic capacity needs of the project. The Administration also argues that the difference between the environmental impacts of the two projects is less than appellees suggest; a ten-lane bridge would impact only 1.6 fewer acres of parkland and 12.9 fewer acres of natural resources over the entire length of the project corridor, and *120would have an identical impact on cultural resources.
How are the merits of appellees’ argument to be assessed? After all, the phrase “reasonable alternative,” standing alone, offers no guidance to a reviewing court. Something can only be an “alternative” by reference to something else; “the term ‘alternatives’ is not self-defining.” Vermont Yankee, 435 U.S. at 551, 98 S.Ct. 1197. The Council on Environmental Quality, for its part, does little to clarify the baseline against which a “reasonable alternative” is to be measured; its regulations at times appear to contrast the “alternatives” to the “proposal,” suggesting that the range of reasonable alternatives are to be selected by reference to the project implemented. See 40 C.F.R. § 1502.14. But that approach would seem to bias the process. See, e.g., Calvert Cliffs’ Coordinating Comm., Inc. v. U.S. Atomic Energy Comm’n, 449 F.2d 1109, 1114 (D.C.Cir.1971). And even if we were to understand an “alternative” to be defined by reference to the proposal actually selected, our interpretive task would hardly be easier, as “the adjective ‘reasonable’ is no more self-defining than the noun that it modifies.” Citizens Against Burlington, 938 F.2d at 195.
We have resolved this difficulty by evaluating an agency’s choice of “reasonable alternatives” in light of the objectives of the federal action; as then-judge Thomas put it in Citizens Against Burlington, “[t]he goals of an action delimit the universe of the action’s reasonable alternatives.” Id. But that approach of course requires that we first consider whether the agency has reasonably identified and defined its objectives. The agency’s choice of alternatives are, then, evaluated in light of these stated objectives; an alternative is properly excluded from consideration in an environmental impact statement only if it would be reasonable for the agency to conclude that the alternative does not “bring about the ends of the federal action.” Id. We engage in both of these inquiries — whether an agency’s objectives are reasonable, and whether a particular alternative is reasonable in light of these objectives — with considerable deference to the agency’s expertise and policy-making role. Id. at 196.
The district court’s opinion suggests that the Administration improperly defined its objectives, criticizing the Administration for narrowing its choice of alternatives “based on a set of criteria that focused primarily on transportation and safety issues.” City of Alexandria, 46 F.Supp.2d at 44. This description of the Administration’s objectives is an accurate one; while the “Statement of Purpose and Need” in the Final EIS references several objectives (including protecting the environment), it focuses on the region’s traffic needs. But it hardly follows that the Administration violated NEPA. As mentioned above, NEPA’s injunction that agencies consider the environmental impacts of “all reasonable alternatives” does not substantively constrain an agency’s choice of objectives; to the contrary, it is those very objectives that provide the point of reference for a determination whether an alternative is “reasonable” in the first place. By suggesting that the Administration violated NEPA because it did not sufficiently prioritize environmental goals, the district court subtly- — and impermissibly — transformed a procedural statute into a substantive one. See Baltimore Gas & Elec. Co. v. Natural Resources Defense Council, Inc., 462 U.S. 87, 97, 103 S.Ct. 2246, 76 L.Ed.2d 437 (1983) (“Congress in enacting NEPA ... did not require agencies to elevate environmental concerns over other appropriate considerations.”) The proper question to ask at the outset of a NEPA inquiry is not whether the Administration focused on environmental goals but rather — as we noted — whether its stated objectives were reasonable. It seems rather obvious to us that it is not unreasonable in articulating its objectives for an agency to “focus primarily on transportation and safety issues” *121when replacing a massively congested and structurally unsound bridge. Cf. Corridor H, 166 F.3d at 374 (affirming the Administration’s rejection of highway alternatives that did not meet the transportation and safety needs of the region).
More in keeping with our precedent, the district .court also determined that a ten-lane alternative was reasonable — and therefore should have been given greater attention — in light of these objectives. The district court arrived at this conclusion by characterizing the Administration as “articulat[ing] the problem as one of addressing the future transportation needs of the region.” “Such a broad statement of purpose and need,” the district court explained, “hardly provides an unequivocal basis for eliminating ten-lane alternatives from consideration.” City of Alexandria, 46 F.Supp.2d at 44. This might be so, had the Administration truly characterized its objectives in such general terms. But it did not. As is required by statute, see 23 U.S.C. § 109(b), the Administration instead focused specifically — in its Statement of Purpose and Need and elsewhere — on the traffic needs that will exist twenty years after the project’s approval, and its analyses based on 2020 traffic projections demonstrate that a ten-lane bridge would be insufficient. The Administration’s studies show that appellees’ preferred design (a ten-lane configuration without an HOV lane) would be able to accommodate less than half of the per-hour capacity of the Administration’s preferred alternative, causing peak-hour traffic queues of significantly greater length and extended duration; accident rates would also be markedly higher on a ten-lane structure.
The district court ignored this data, instead focusing exclusively on an Administration study showing that a ten-lane bridge would be able to accommodate up to 295,000 vehicles per day, a number only slightly smaller than the projected daily traffic flow on the Bridge in 2020. City of Alexandria, 46 F.Supp.2d at 44. But that study apparently assumed an even flow of traffic throughout the day (which, of course, is unrealistic). Whatever the total number of vehicles that will cross in a 24-hour period, the relevant question is how long during peak commuting hours it will take to cross the bridge. Appellees also do not seriously challenge the Administration’s findings, instead protesting that these studies establish little more than the “truism ... that a ten-lane bridge would carry somewhat less traffic than a twelve-lane bridge.” It is not apparent to us why this proposition has less force in the case because it is a “truism.”
Appellees’ more fundamental argument is that, regardless of its shortcomings in satisfying future traffic needs, we must hold a ten-lane bridge to be a reasonable alternative in light of our statement in Natural Resources Defense Council, Inc. v. Morton that an agency should not “disregard alternatives merely because they do not offer a complete solution to the problem.” 458 F.2d 827, 836 (D.C.Cir.1972). Appellees overread Morton. In that case an environmental group challenged the Secretary of the Interior’s proposed sale of oil and gas leases to submerged lands in the Gulf of Mexico; the Secretary sought to sell these properties as part of a cross-agency effort, initiated by the President, to increase American energy supplies. We held that the Secretary’s environmental impact statement violated NEPA because it failed to consider alternatives outside of the Department of the Interior’s jurisdiction; we also noted that the agency could not exclude alternatives “supplying only part of the energy that the lease sale would yield.” Id. at 836. This broad articulation of “reasonable alternatives” was compelled by the national scope of the problem being addressed: ‘When the proposed action is an integral part of a coordinated plan to deal with a broad problem, the range of alternatives that must be evaluated is broadened.” Id. at 835.
*122Morton thus stands for the same proposition as Citizens Against Burlington-. namely, that a “reasonable alternative” is defined by reference to a project’s objectives. Morton explained that, within the context of a coordinated effort to solve a problem of national scope, a solution that lies outside of an agency’s jurisdiction might be a “reasonable alternative”; so might an alternative within that agency’s jurisdiction that solves only a portion of the problem, given that other agencies might be able to provide the remainder of the solution. Such a holistic definition of “reasonable alternatives” would, however, make little sense for a discrete project within the jurisdiction of one federal agency, as we recognized in Morton when we contrasted the Secretary’s action with that of building “a single canal or dam.”4 Id. Concerned with severe traffic conditions in the Capital Region, Congress has authorized the Administration to replace the Woodrow Wilson Memorial Bridge. The Administration has sole responsibility for solving this problem; were it to build a ten-lane bridge, no one else would step in and alleviate the congestion that would result.5 In this context, it is simply a non sequitur to call a proposal that does not “offer a complete solution to the problem” a “reasonable alternative.”
One other point merits brief discussion. In finding a ten-lane alternative reasonable, the district court noted that the Administration only conducted a Clean Air Act conformity analysis for the use of ten lanes on the Bridge. See City of Alexandria, 46 F.Supp.2d at 45. If the Administration only expects ten lanes to be open, the district court reasoned, how can it fail to consider a ten-lane bridge as a reasonable alternative under NEPA? The answer is that the Clean Air Act and NEPA inquiries have different time horizons; while a project must show conformity with the Clean Air Act at the-time it is approved, see 42 U.S.C. § 7506(c)(1) (1995), the consideration of reasonable alternatives under NEPA requires, as mentioned above, an assessment of traffic needs in 2020. Accordingly the Administration did not violate the National Environmental Policy Act by failing to include a ten-lane bridge proposal as a “reasonable alternative” in its Final Environmental Impact Statement.
B.
Once an agency identifies the “reasonable alternatives” to a proposed action, NEPA and Council on Environmental Quality regulations also require an agency to identify the “adverse environmental effects” of each alternative. See 42 U.S.C. § 4332(2)(C)(ii); 40 C.F.R. § 1502.16. The district court found fault with the Administration’s treatment of the temporary “construction impacts” that would arise during the period that the Bridge was being built. Again, we disagree.
The district court focused on the brevity of the “Construction Impacts” section of *123the Final EIS, which covers only four pages and, according to the district court, “is of such a broad and generic nature that it could apply to practically any construction project undertaken by the [Administration].” City of Alexandria, 46 F.Supp.2d at 45. While the Administration’s discussion might have been more thorough, we think the district court’s assessment of the Administration’s treatment of these issues too harsh. The Administration addresses a range of expected construction impacts, including the construction’s likely effect on local traffic, air quality, area noise levels, water quality and wetlands, cultural resources, and visual effects. The level of detail of these assessments varies; it is worth noting, in light of the district court’s focus on the terseness of the Administration’s analysis, that some of the shorter analyses are the most eminently reasonable. Take, for instance, the Administration’s discussion of traffic impacts. The Administration acknowledges that the construction project will affect traffic flow on several Alexandria roadways, and may also cause potential delays in the delivery of emergency services. It also offers a range of mitigation strategies: six lanes of the Bridge will be kept open at all times to minimize rush-hour congestion; some access (even if circuitously routed) will be maintained to all roads and areas; there will be no disruption of marine traffic on the Potomac; the public will be notified of temporary road closings through the news media, the posting of signs, and the creation of a “project activities” hotline. Perhaps appellees would prefer the Administration to set forth in the Final EIS a comprehensive plan detailing precisely which streets will be closed, and which alternative routes will be established, but that is not mandated by NEPA. See, e.g., Robertson v. Methow Valley Citizens Council, 490 U.S. 332, 353, 109 S.Ct. 1835, 104 L.Ed.2d 351 (1989) (“[I]t would be inconsistent with NEPA[ ] ... to demand the presence of a fully developed plan that will mitigate environmental harm before an agency can act.”).
We think the terseness of the Administration’s discussion of construction impacts is justified for other reasons as well. The Administration typically delays the identification of “construction staging” sites— locations used to store materials and equipment during project construction— until the design stage of the project. As will be discussed infra, this practice is permissible under the statute and is arguably required by the Administration’s governing regulations. Since the Administration did not identify the location of these areas, it of course could not identify the accompanying environmental impacts with precision. But this does not mean that the Administration did not consider, on a more general level, what those impacts would be; the Final EIS identifies several potential staging areas, and notes that each of these sites are in “previously disturbed” areas with “minimal natural resources.” The Administration’s brevity is particularly understandable given the numerous regulatory constraints that will limit the extent of construction activities. As the Administration notes, Maryland and Virginia require construction contractors to limit noise levels in “noise sensitive areas adjacent to the project area” to eighty decibels — a noise level comparable to that currently produced by traffic on some stretches of the highway. Similar federal and state regulatory provisions require the mitigation of any short-term construction impacts on wetland and aquatic resources, constrain the emissions of dust from construction-related activities and equipment, and limit the Administration’s selection of construction staging areas. The Final EIS’ reference of these provisions is important, as it indicates the Administration’s awareness of the maximum impact that the construction may cause.
We also note that agencies are enjoined by the Council on Economic Quality to develop environmental impact statements that are “no longer than absolutely necessary” and that discuss impacts “in proportion to their significance.” 40 C.F.R. *124§ 1502.2(b)-(c). The Administration points out that each of the seven “build” alternatives would have similar construction impacts, thus making a detailed discussion of each of their effects redundant. More fundamentally, while the disruption caused by the construction of a project as significant as this one is by no means trivial, it is relatively modest in both scope and duration when compared to the environmental impact of the project as a whole. To be sure, there is a point at which an agency’s analysis ventures from the “tolerably terse to the intolerably mute,” Greater Boston Television Corp. v. FCC, 444 F.2d 841, 852 (D.C.Cir.1970), but we simply do not think that the Administration’s analysis of construction impacts reaches that point.
III.
A.
The district court concluded that the Administration also failed to identify adequately the effect that its preferred alternative will have on historic resources in the project area, as is required under two distinct but overlapping statutes: section 106 of the National Historic Preservation Act and section 4(f) of the Department of Transportation Act. Section 106, like NEPA, is essentially a procedural statute; it requires that agencies “take into account the effect of [an] undertaking on any district, site, building, structure, or object that is included in or eligible for inclusion in the National Register [of Historic Places].” 16 U.S.C. § 470f. To comply with section 106, an agency must consult with state historic preservation officers to ensure that historic properties in the project area are thoroughly identified and the effects that the project will have on them fully assessed. See 36 C.F.R. § 800.4-.5.6 The usual product of this consultation process is a Memorandum of Agreement among the consulting parties signifying agreement upon how the detrimental effects will be “taken into account.” Even where disagreement precludes the completion of a Memorandum of Agreement an agency may implement a project after receiving and considering comments from the Advisory Council on Historic Preservation. See id. at 800.6(c).
On the other hand, section 4(f), unlike the other statutes at issue in this case, imposes a substantive mandate on the Administration: It prohibits the agency from taking an action that “uses” a historic resource unless there is “no prudent and feasible alternative to using that land” and the agency engages in “all possible planning” to “minimize harm” to the sites.7 49 U.S.C. § 303(c); see also Citizens to Preserve Overton Park v. Volpe, 401 U.S. 402, 91 S.Ct. 814, 28 L.Ed.2d 136 (1971). Department of Transportation regulations require the Administration to “make the section 4(f) approval” at the same time that it approves its final EIS or issues its Record of Decision, 23 C.F.R. § 771.135©; the Administration ordinarily complies with this requirement by publishing a separate “Section 4(f) Evaluation” along with its final EIS, which identifies the project’s effects on historic properties in the project area and the efforts the agency has taken to mitigate those effects, In order to comply with 4(f)’s substantive requirements, it is of course necessary first to identify historic sites in the project area; accordingly, we have observed that compliance with section 4(f) is predicated upon completion of the section 106 process. See Corridor H Alternatives, Inc. v. Slater, 166 F.3d 368, 371 (1999).
The central dispute between the parties is not about whether, but about when, the Administration must complete its identifi*125cation of historic properties. The Administration has been “taking into account” the effect of the proposed project on historic sites since the project’s inception, conducting several surveys which led to the identification of 23 National Register-listed or National Register-eligible properties, and 36 underwater or terrestrial archaeological sites in the project area. The Administration also identified and visited each National Register-listed property in Alexandria for the purpose of determining, among other things, the “visual impacts” that various alternative bridge proposals would have on each site. The result was publication of a Memorandum of Agreement and a Section 4(f) Evaluation with or prior to the Administration’s approval of the project; these documents identify seven historic sites that will be affected by the project and another six that may be, and offers plans to minimize and mitigate the project’s impact on these properties.
The district court did not question the overall legitimacy or thoroughness of these studies.8 (Indeed, appellees cannot identify a single historic resource in the project area that the Administration failed to “take into account.”9) Instead, the district court concluded that the Administration violated section 106 by deciding to postpone the identification of sites where it would conduct certain construction-related activities, including construction staging areas (the locations where contractors will store materials and mobilize construction activities), wetland mitigation areas, and dredge disposal sites. While the likely impact of these activities, which the Administration describes as “ancillary,” are minimal when compared to those of the project as a whole, it is at least conceivable that they could ultimately affect section 106 properties. Acknowledging this possibility, but noting that it usually defers the identification of such properties until the “design stage” of a large highway project, the Administration included promissory language in its Memorandum of Agreement binding it to fulfill its section 106 responsibilities when selecting these sites. The district court thought that these prospective terms ran afoul of our recent decision in Corridor H Alternatives, Inc. v. Slater, 166 F.3d 368 (D.C.Cir.1999), in which we held that the Administration could not postpone the entire section 106 process until after it issued its Record of Decision.
We think that district court misconstrued our holding in Corridor H. In that case, the Administration postponed the entire section 106 process for a major highway corridor; its Record of Decision instead adopted a “Programmatic Agree*126ment” dividing the highway into fourteen segments, and promised that it would not begin construction of a particular segment before completing the section 106 process for that segment. We held that this Agreement impermissibly abrogated the Administration’s responsibility to assess the project’s impact on historic properties during the planning stages of the project. See 166 F.3d at 373. But that is not the case here, since the Administration has identified historic properties along the entire project corridor and documented its findings prior to approval in both a Memorandum of Agreement and a Section 4(f) Evaluation. All that has been deferred is the identification of sites that might be impacted by a small number of “ancillary. activities.” This is quite distinguishable from the “Programmatic Agreement” we proscribed in Corridor H.
The Administration did not postpone the identification of these properties “merely to avoid having to complete its 4(f) and 106 analyses,” as the district court said. 46 F.Supp.2d at 47. As the Administration points out, the precise identification of these sites requires “substantial engineering work” that is not conducted until the design stage of the project; indeed the Administration is required to conduct such “final design activities” after it completes its Final EIS. 23 C.F.R. § 771.113(a)(iii). Furthermore, then-existing Council regulations explicitly encouraged flexible, staged planning in the section 106 process. See 36 C.F.R. § 800.3(b) (section 106 procedures “may be implemented ... in a flexible manner”); 36 C.F.R. § 800.3(c) (section 106 regulations should not be interpreted to “prohibit phased compliance at different stages in planning.”). Appellees respond that the Administration could nonetheless “feasibly” identify these sites without doing “final design” plans for the project. But the standard of “feasibility,” while relevant to whether an agency may use 4(f) properties, has no application in determining when the agency must identify them. We think that, particularly where the sites postponed are merely ancillary to the project, section 106 and the identification prerequisites of section 4(f) do not forbid the rational planning process adhered to by the Administration.
B.
We also think that the Administration satisfied section 4(f)’s substantive provisions. Appellees barely bother to argue that the Administration did not comply with section 4(f)(l)’s requirement that it consider all “prudent and feasible alternative[s]” to using protected properties. The reason for this gap in appellees’ otherwise vigorous presentation is obvious enough. For while the Administration is required to give the protection of 4(f) property “paramount importance” in determining whether an alternative is “prudent,” Overton Park, 401 U.S. at 412-13, 91 S.Ct. 814, we have squarely held that an alternative cannot be a prudent one if it does not satisfy the transportation needs of the project. See Citizens Against Burlington, 938 F.2d at 204. In light of this limitation, appellees can only win under section 4(f)(1) if they establish one of two propositions: They must show that a narrower Bridge satisfies the transportation needs of the project, or they must offer a “prudent” project alternative that does not impact the 4(f) properties used by the Administration’s preferred design. The former question we have already resolved in the Administration’s favor, and appellees do not advance an alternative highway route that has a less significant impact on 4(f) properties.
Appellees do argue with greater enthusiasm that the Administration violated section 4(f)(2)’s requirement that the agency engage in “all possible planning” to minimize harm to 4(f) properties, but this argument is equally unpersuasive. To begin at the broadest level of generality, appellees do not question the Administration’s express findings that, among the seven “prudent and feasible” alternatives compared in the Final EIS, the preferred alternative *127“results in the least overall impact to section 4(f) resources.” Cf. Druid Hills Civic Ass’n, Inc. v. FHWA, 772 F.2d 700, 716 (11th Cir.1985) (noting that “section 4(f)(2) requires a simple balancing process which totals the harm caused by each alternate route to section 4(f) areas and selects the option which does the least harm”). At the site-specific level, the Administration made several significant project modifications to avoid or minimize impacts to section 4(f) properties, including altering an interchange design to avoid impacting a schoolground and ehminating the construction of a temporary Beltway overpass to minimize the risk of harm to Freedman’s Cemetery. Where the Administration could identify no feasible and prudent plan for avoiding impact to a 4(f) site, it offered .plans to mitigate that impact; for instance, it proposed substantial improvements to Jones Point Park, arguably the most significant 4(f) property impacted by the project. Further recitation of the Administration’s mitigation efforts is possible, but unnecessary; suffice it to say that, after a thorough review of the record, we have little difficulty concluding that the Administration complied with its responsibilities under section 4(f) of the Department of Transportation Act.10
During the course of our consideration of this case, appellees have attempted to bolster their position by pointing to the opposition of prominent legislators to the project, and by noting the hurdles to ultimate congressional approval that still lie in the Administration’s path. These political impediments are irrelevant to us but they indicate where appellees should concentrate their efforts. We have been admonished by the Supreme Court with respect to the very statute that is at the heart of this case to avoid using its requirements as a vehicle to impose our own judgment. Vermont Yankee, 435 U.S. at 554, 98 S.Ct. 1197. Our obligation is not to further our beau ideal of a bridge design, but merely to ensure that the procedures mandated by these statutes have been complied with. We hold that the Administration has satisfied the requirements of NEPA, the National Historic Preservation Act, and the Department of Transportation Act, and reverse.
So ordered.