OPINION OF THE COURT
WEIS, Circuit Judge.Finding that circumstances warranted an exception to its general prohibition against floodplain development, New Jersey granted permission for construction of dwelling units in a seaside community. The federal Environmental Protection Agency later agreed to grant funds for the construction of an indispensable sewage treatment plant in the area, but only on the condition that no hookups be permitted to the proposed residences. In view of the record in this case and because Congress has encouraged state and local regulation of coastal areas, we conclude that EPA acted arbitrarily in imposing the hookup restriction in defiance of the state and local action. Accordingly, we vacate the judgment entered in favor of EPA and remand for further proceedings.
Plaintiff Cape May Greene, Inc. sought injunctive and declaratory relief against the restrictive condition, but the district court denied relief and entered summary judgment against plaintiff. Other claims against non-federal defendants and cross-claims were ultimately terminated, and plaintiff appeals.
The plaintiff, a land developer, owns most of two parcels of land, designated as tracts A and B, in Cape May City, New Jersey. The City is part of the Cape May region located at the southern tip of New Jersey. The Atlantic Ocean borders the region on the east and south, the Delaware Bay lies to the west, and to the north the mainland is cut off by a canal and the Cape May Harbor.
The region encompasses some 5700 acres. More than 3,600 acres, or 64% of the area, consist of wildlife habitats, wetlands, public open space, dunes, and beaches, all of which are closed to development by state and local regulation. Thirty-one percent of the area has been developed and provided with sewers. The controversy in this case centers on plans to install sewers in a part of the remaining 5%, or 297 acres. Tracts A and B constitute 196 acres, or 3% of the region.
Tract A consists of 47 acres surrounded by existing structures and paved streets; 5.7 acres are already developed. Tract B contains 149 acres and is bounded on the west and north by housing and on the east by a United States Coast Guard Receiving Center. Fifteen acres of that parcel are developed.
In 1979, the plaintiff developer applied to the New Jersey Department of Environmental Protection for a permit to construct 244 residential units in tract A. The state agency reviewed the developer’s proposal in accordance with the Coastal Area Facility Review Act, N.J.Stat.Ann. §§ 13:19-1 to -21 (West 1979 & Supp.1982). That Act, which is the New Jersey management plan for regulation of the coastal area, had been approved by the federal government pursuant to the Coastal Zone Management Act of 1972, 16 U.S.C. §§ 1451-1464 (1976 & Supp. IV 1980). The New Jersey agency reviewed such factors as flood hazard possibilities, air and water quality, traffic volume, road access, and the effect on environmentally sensitive areas. In 1980, a permit was approved, conditioned on the availability of sewage hookups to the housing units.
As the New Jersey agency was aware, the developer expected that a proposed regional sewage disposal plant would service the new housing. The treatment plant had been under consideration for some years, but the federal EPA had indicated it might restrict sewer connections to the plant.
The existing waste water treatment plant is owned and operated by the City of Cape May. Constructed in 1958, it has a capacity of 3.0 m.g.d.1 The plant is unsatisfactory *182because of the high level of pollutants it discharges into Delaware Bay. The need to improve the plant’s efficiency became apparent in the summer of 1975 when the beaches in this resort area had to be closed because of pollution. In the following year, the New Jersey Department of Environmental Protection imposed a ban on any further hookups to the plant. The ban was lifted in November 1978 after plans to rehabilitate the facility were undertaken.
The Cape May County Municipal Utilities Authority proposed to construct a new, more efficient and slightly larger (3.2 m.g.d.) system on the site of the existing plant and applied to EPA for a matching funds grant. EPA is authorized to grant funds for the construction of sewage facilities under the Water Pollution Control (Clean Water) Act, 33 U.S.C. §§ 1251-1376 (1976 & Supp. IV 1980).
In October 1978, EPA informed the Authority that it would fund a disposal plant whose capacity could service the existing population and projected growth in the area, but not any development within the floodplain or environmentally sensitive areas of the Cape May region.2 To that end, EPA proposed a plant capacity of 3.0 m.g.d. The Authority accepted the reduction in capacity, and in January 1979, EPA determined that no significant detrimental environmental impacts would result from the proposed facility. Based on that determination, EPA preliminarily decided not to prepare an environmental impact statement.3
The agency invited comments from the public and interested entities. Included in the EPA announcement was a statement that the agency had recommended changes that reduced the capacity of the system to service flood-hazard areas. The notice continued, “However, we cannot do local land use planning for the municipalities in the region and some treatment capacity could very well be used up by development in floodplains and other sensitive areas unless local measures are adopted.”
Several environmental groups protested the EPA action and complained that the agency was not doing enough to prevent development in environmentally sensitive and floodplain areas. In response to one of these communications, the EPA Regional Director wrote:
“We continue to believe that it is the primary function of local government to provide the ‘fined [sic] tuned’ growth control through land use measures. A sewer hookup ban as a condition to the grant would, we believe, be neither appropriate nor effective in providing the kind of ‘guarantee’ against future development in environmentally sensitive areas that you seek.”
In a reply to another environmental group, an EPA official wrote:
“We fully support your view that controls over land uses in environmentally sensitive or critical areas should be required, but we feel that this should be done by the affected municipalities. The EPA is not the empowered agency with jurisdiction to change, enforce or control local land use.”
*183Despite these statements, EPA changed its position and told the Authority, in June 1979, that the grant would be conditioned on a ban against sewer hookups to designated lots in the floodplain and environmentally critical areas.4 In September 1979, EPA advised that it would be receptive to fully justifiable exceptions to the policy and would accept redevelopment of lots that previously had structures.
A number of the municipalities to be serviced by the new facility adopted resolutions protesting the EPA action, arguing that land use planning and controls were matters reserved for determination by state and local governments. They pointed out that the critical areas had already been removed from development by ordinances and the municipalities were in compliance with all provisions of the federal flood insurance program.
After extensive consultations with the affected municipalities, the Authority submitted an amendment to the disposal plant plan in January 1980 and supplemented it in April of that year. In brief, the Authority’s proposal excluded sewer hookups in environmentally sensitive areas, such as beaches, dune complexes, intermittent stream corridors, bogs, and fresh water wetlands. These areas had already been designated by the state coastal management plan as prohibited for development. Although, with the exception of wetlands, the state plan applied only to housing projects consisting of 25 or more units, the Authority’s proposal would extend the ban on hookups to individual lots as well.
In addition, the prohibition against hookups would be extended to critical wetland and upland habitat areas, consistent with local plans and the state coastal management program. The net result would be that any new construction in the region would be forced to take place within or adjacent to areas that were previously developed and already had available the infrastructure for transmission of waste water to the disposal plant. With respect to tracts A and B, the Authority pointed out that the value of the storm, sanitary and water service infrastructure in place was approximately $1,423,000. Moreover, both of these tracts had been identified as development areas in the Cape May County comprehensive plan and the state’s Coastal Zone Management Plan.
Summarizing, the Authority stated that “[a]ll recognized environmentally sensitive areas are excluded from future sewer connections under this proposal and any new structures which would be allowed within the floodprone areas must, by existing ordinances, comply with FEMA [Federal Emergency Management Agency] safety regulations.”
In response, EPA agreed to allow hookups only from tracts that were less than 3.4 acres in size, were surrounded by existing structures, and had all infrastructure for municipal services in place. Such development was characterized as “fill in.” Since most of tracts A and B did not meet these criteria, the EPA action was in effect a denial of the Authority’s proposal. Yielding to the necessity of having the long-delayed project get under way, the Authority reluctantly, and under protest, accepted the EPA demands. The agency then issued its approval for a grant of $529,087 toward the total cost of $705,499.
In a published report of its Cape May environmental review, EPA stated it would *184take steps to designate the critical and floodplain areas as unsuitable for septic tanks. It also stated, “EPA will deny permits for any package treatment plant that is proposed to serve development that is delineated as ‘nonsewerable’ in the approved 201 facilities plan.” Thus, the agency announced its intention to ban, not only sewer hookups, but any other means of waste water disposal as well.
The developer then asked the district court to declare the restrictive condition void as beyond EPA’s authority, and to prohibit the Authority from agreeing to the grant condition.5 The district court, however, granted summary judgment in favor of EPA, finding the grant condition to be reasonable and in accordance with the agency’s authority. The court read Executive Order 11,988, 42 Fed.Reg. 26951, reprinted in 42 U.S.C. § 4321 note at 820 (Supp. IV 1980), as exhorting EPA to minimize floodplain development as far as possible and to supply the necessary authorization for the grant condition.
The plaintiff also contended that allowing sewer connections to tracts of 3.4 acres or less, while denying them to larger areas, was an equal protection violation. The court found that the agency’s action met the rational basis test. In addition, the court held that Tucker Act jurisdiction on the plaintiff’s taking of property claim rested with the Court of Claims. See 28 U.S.C. § 1346(a)(2) (Supp. IV 1980). Finally, a Tenth Amendment claim was dismissed since there was no regulation of “states as states.”
I.
Preliminarily we must address a jurisdictional issue, the finality of the district court’s judgment.
The district court entered summary judgment against the plaintiff and in favor of the EPA as well as several other defendants on December 3, 1981. Similar judgments were granted to other defendants in February and April 1982. Finally, on May 11, 1982, summary judgment was granted in favor of the last remaining defendant, the New Jersey State Department of Environmental Protection. Thus, as of that date, summary judgment had been entered against the plaintiff on all of its claims and in favor of all the defendants.
Plaintiff filed a notice of appeal from the summary judgments against it on May 19, 1982. There remained on the record, however, a cross-claim filed by the Authority against the federal defendants, and they moved to dismiss the appeal. On July 22, 1982, a stipulation was filed in the district court by all parties dismissing the cross-claim without prejudice. The federal defendants then filed an amended jurisdictional statement in their brief on July 29, 1982, asserting, “This court has jurisdiction over these appeals pursuant to 28 U.S.C. § 1291. The orders of the district court dispose of all claims by all parties.”
At the time the plaintiff filed its notice of appeal, the only aspect of the litigation remaining open was the cross-claim between two of the defendants on a collateral matter.6 Strictly speaking, however, at the time the notice was filed, the judgment of the district court was not final. See Fed.R. Civ.P. 54(b). It did not become final until after the parties had filed their briefs in this court, but before we had taken any action on the merits.
A similar jurisdictional issue was presented in Pireno v. New York Chiropractic Ass’n., 650 F.2d 387 (2d Cir.1981), aff’d, *185Union Labor Life Insurance Co. v. Pireno, - U.S. -, 102 S.Ct. 3002, 73 L.Ed.2d 647 (1982), where an appeal was filed after judgment had been entered for two of three defendants, but before judgment was entered in favor of the third. At the time of disposition in the court of appeals, the action of the district court had become final. The court of appeals treated the premature notice as being effective on the date the judgment was entered in favor of the last remaining defendant and took jurisdiction of the appeal.
The same result was reached in Jetco Electronic Industries, Inc. v. Gardiner, 473 F.2d 1228, 1231 (5th Cir.1973), a case with similar facts. There the court cited the admonition of Gillespie v. U.S. Steel Corp., 379 U.S. 148, 152, 85 S.Ct. 308, 310, 13 L.Ed.2d 199 (1949), that “practical, not technical considerations are to govern the application of principles of finality.” But see U.S. v. Taylor, 632 F.2d 530 (5th Cir.1980).
In Richerson v. Jones, 551 F.2d 918, 922 (3d Cir.1977), we said that “a premature appeal taken from an order which is not final but which is followed by an order that is final may be regarded as an appeal from the final order in the absence of the showing of prejudice to the other party.” (emphasis in original) See also Hodge v. Hodge, 507 F.2d 87, 89 (3d Cir.1975); Eason v. Dickson, 390 F.2d 585 (9th Cir.), cert. denied, 392 U.S. 914, 88 S.Ct. 2076, 20 L.Ed.2d 1373 (1968).
In Dawson v. Chrysler Corp., 630 F.2d 950 (3d Cir.1980), cert. denied, 450 U.S. 959, 101 S.Ct. 1418, 67 L.Ed.2d 383 (1981), and Tilden Financial Corp. v. Palo Tire Service, Inc., 596 F.2d 604 (3d Cir.1979), we held that Federal Rule of Civil Procedure 54(b) certification filed after a notice of appeal had been docketed was adequate to confer appellate jurisdiction. This body of case law, therefore, supports the proposition that we have jurisdiction to consider the appeal.
However, there remains for consideration the question of whether Griggs v. Provident Consumer Discount Co., - U.S. -, 103 S.Ct. 400, 74 L.Ed.2d 225 (1982) requires a contrary result. We conclude it does not.
In Griggs, the appellant filed a notice of appeal before the disposition of a post-trial motion under Fed.R.Civ.P. 59. The Supreme Court held there was no appellate jurisdiction because under Fed.R.App.P. 4(a)(4), as amended in 1979, the “notice of appeal was not merely defective; it was a nullity,” and hence it was as if “no notice of appeal were filed at all.” Id. at -, 103 S.Ct. at 403.
Rule 4(a)(4) explicitly cites the instances in which a premature notice of appeal “self-destructs” as those in which motions have been filed in the district court under Fed.R. Civ.P. 50(b), 52(b) and 59. See 9 J. Moore, B. Ward & J. Lucas, Moore’s Federal Practice § 204.12[1] at 4-65 and n. 17 (2d ed. 1982). That situation is not present here.
The 1979 amendments to Rule 4(a)(2) expressly give effect to a premature notice of appeal taken after the announcement of a decision, but before entry of judgment — the result we reached four years earlier in Hodge v. Hodge. The Advisory Committee’s notes accompanying the 1979 amendments state that the new Rule 4(a)(2) is intended to
“extend to civil cases the provisions of Rule 4(b), dealing with criminal cases, designed to avoid the loss of the right to appeal by filing the notice of appeal prematurely. Despite the absence of such a provision in Rule 4(a), the courts of appeal quite generally have held premature appeals effective, [citations omitted]
The proposed amended rule would recognize this practice, but make an exception in cases in which a post trial motion has destroyed the finality of the judgment. See Note to Rule 4(a)(4) below.”
Thus, the Rules contemplate that the prohibition against giving effect to premature notices of appeal shall be confined to the specific instances cited in Rule 4(a)(4). Only that exception was ruled upon by the Supreme Court in Griggs — not the situation presented here. Accordingly, we conclude that Griggs does not apply to the facts at hand and we have before us an appealable order.
*186II.
We turn to the merits of the appeal. The developer contends that EPA’s actions are inconsistent with the Coastal Zone Management Act of 1972, 16 U.S.C. §§ 1451-1464 (1976 & Supp. IV 1980), and the National Flood Insurance Act of 1968, 42 U.S.C. §§ 4001-4127 (1976 & Supp. IV 1980). Moreover, EPA is said to have exceeded its authority under both the National Environmental Policy Act of 1969, 42 U.S.C. §§ 4321-4347 (1976 & Supp. III 1980), and Executive Order 11,988. The developer concludes by arguing that EPA did not correctly appraise the topography of the land in determining that tracts A and B would serve as water storage and wave dissipation areas in the event of flooding.7
EPA asserts that the grant condition was reasonable and that its authority extends to deterrence of development in floodplains. It argues that its action is within the scope of Executive Order 11,988, and a federal environmental standard more demanding than one adopted by a state in its management plan may be enforced.
The parties’ arguments are wide ranging, and it is helpful to narrow the issues to those presented in this case. There is no dispute about environmentally sensitive sites such as sand dunes, bogs, marshes, wetlands, or wildlife habitats. All areas of this description in the Cape May region are already protected by the local and state land use plans, and no objection is raised to incorporating sewer restrictions for these places into the grant conditions. The controversy is limited to the restriction as applied to floodplain locations whose sole function is said to be storage of flood waters and wave energy dissipation. No plant, animal or marine life, or other ecological considerations, are advanced as additional values which must be considered with respect to tracts A and B.
In addition, the need to improve the efficiency of the existing sewage disposal plant has existed for years and that circumstance is unrelated to the floodplain. The current pollution problem must be alleviated, even apart from whether development takes place in the parcels at issue.
Narrowing the matter even more are the facts that state and local plans have previously found the disputed area suitable for development, and that sewers and water lines are adjacent to the tracts and available for service. Further, local building restrictions mandate compliance with federal flood protection insurance standards. In short, the local and state governments would allow some regulated development in tracts A and B.8 EPA would allow none because its prohibition against sewer hookups is as effective a ban against residential building as can be devised, particularly when considered in conjunction with its announced policy that no septic tanks or package treatment systems will be allowed in the area.
EPA does not contend that the resulting land use control is simply an unavoidable by-product of the grant condition. Rather, it has openly stated that its aim is to prohibit housing in the floodplain. Nor does EPA argue that the restriction is needed to insure the efficiency of the sewage plant. Thus, the agency has reversed its earlier *187view that land use controls must come from the state and local governments, and has asserted authority that it previously disclaimed.
It is also clear that EPA is using its power to regulate grants under the Clean Water Act to accomplish matters not included in that statute. Although it also cites the National Environmental Policy Act, the agency relies primarily on the Executive Order to support its action. Essentially, the conflict here centers on federal agency action not explicitly required by statute and contrary to state and local legislation in a field where congressional intervention has been hesitant and tentative.
Land use planning has traditionally been considered a matter of local concern and Congress has not been hospitable to demands that it preempt the field.9 See Biderman v. Morton, 497 F.2d 1141, 1144 (2d Cir.1974). (In enacting the Fire Island National Seashore Act, 16 U.S.C. § 459e (1976 & Supp. V 1981), “Congress carefully avoided interfering with the power of the municipalities on the Seashore to enact zoning ordinances or grant zoning variances.”) In 1972, Congress did tentatively touch a toe in the water when the Coastal Zone Management Act was passed. Although the statute is designed to incorporate national environmental policies into land use decisions, the approach is one of encouragement, rather than mandate.
The legislative history of the Coastal Zone Management Act is emphatic in stating this guideline:
“[The Act] has as its main purpose the encouragement and assistance of States in preparing and implementing management programs to preserve, protect, develop and whenever possible restore the resources of the coastal zone of the United States. * * * There is no attempt to diminish state authority through federal preemption. The intent of this legislation is to enhance state authority by encouraging and assisting the states to assume planning and regulatory powers over their coastal zones.”
Sen.Rep. No. 753, 92nd Cong., 2d Sess., reprinted in 1972 U.S.Code Cong. & Ad.News 4776.
State participation in the program is voluntary. Funding is offered for the preparation and implementation of state plans which meet federal approval. These plans must incorporate designated policy interests, but are administered by the states and it is they who make the development and local use decisions.
New Jersey’s plan was submitted to and approved by the federal government. See 45 Fed.Reg. 71640 (Oct. 29, 1980); 43 Fed. Reg. 51829 (Nov. 7,1978). The state, therefore, has accepted the congressional invitation to regulate coastal areas with due regard for national policies. The management plan, as defined by the statute, is to set forth “objectives, policies, and standards to guide public and private uses of lands and waters in the coastal zone.” 16 U.S.C. § 1453(g).10 The Act provides that it shall not supersede, modify or repeal existing laws, id. § 1456(e)(2), but also states that *188“[e]ach federal agency conducting or supporting activities directly affecting the coastal zone shall conduct or support those activities in a manner which is, to the maximum extent practicable, consistent with approved state management programs.” Id. § 1456(c)(1).
The role assigned to EPA by the Clean Water Act, 33 U.S.C. §§ 1251-1376, is to reduce the discharge of pollutants into the nation’s waterways and coastal areas. In carrying out that task, EPA is authorized to grant funds for the construction of sewage treatment plants. In addition to this direct and primary obligation, EPA asserts a secondary or indirect role, along with other federal agencies, in the protection of the environment under the National Environmental Policy Act, 42 U.S.C. §§ 4321-4347, and Executive Order 11,988.
The purpose of the National Environmental Policy Act is to “provid[e] a statement of national environmental goals, policies, and procedures,” and to impose on federal agencies a responsibility to consider the consequences of their actions on the environment. Sen.Rep. No. 296, 91st Cong., 1st Sess. 14. All agencies of the federal government are required to “identify and develop methods and procedures ... which will insure that presently unquantified environmental amenities and values may be given appropriate consideration in decision-making along with economic and technical considerations.” 42 U.S.C. § 4332(2)(B) (emphasis added).
The Act is not a mandate to pursue environmental policies to the exclusion of all others, but is rather a congressional “reordering of priorities so that environmental costs and benefits will assume their proper place along with other considerations.” Calvert Cliffs’ Coordinating Committee, Inc. v. U.S. Atomic Energy Commission, 449 F.2d 1109, 1112 (D.C.Cir.1971). An agency need not, “in selecting a course of action, ... elevate environmental concerns over other appropriate considerations.” Strycker’s Bay Neighborhood Council, Inc. v. Karlen, 444 U.S. 223, 227, 100 S.Ct. 497, 499, 62 L.Ed.2d 433 (1980). In short, the National Environmental Policy Act requires a balancing between environmental costs and economic and technical benefits. Calvert Cliffs’ Coordinating Committee, Inc. v. U.S. Atomic Energy Commission, 449 F.2d at 1113. See also Environmental Defense Fund v. Corps of Engineers, 470 F.2d 289, 298 (8th Cir.1972), cert. denied, 412 U.S. 931, 93 S.Ct. 2749, 37 L.Ed.2d 160 (1973).
The National Environmental Policy Act does not expand the jurisdiction of an agency beyond that set forth in its organic statute, see Gage v. Atomic Energy Commission, 479 F.2d 1214 (D.C.Cir.1973); Kitchen v. Federal Communications Commission, 464 F.2d 801 (D.C.Cir.1972), and the Supreme Court has characterized “its mandate to the agencies [as] essentially procedural.” Vermont Yankee Nuclear Power Corp. v. Natural Resources Defense Council, Inc., 435 U.S. 519, 558, 98 S.Ct. 1197, 1219, 55 L.Ed.2d 460 (1978). “Even the most enthusiastic commentators of the Act agree that NEPA does not confer unlimited power on the agencies, does not delegate vast areas of congressional authority, and does not initiate a free-for-all among agencies for new programs and domains.” F.R. Anderson, Jr., “The National Environmental Policy Act” in Federal Environmental Law 291 (Environmental Institute, E. Dolgin & T. Guilbert eds., West Publishing Co. 1974).11
Thus, the National Environmental Policy Act provides little, if any, support for an agency taking substantive action beyond that set forth in its enabling act. EPA’s reliance on the National Environmental Policy Act is general. Because it has authority under the Clean Water Act to condition grants and NEPA mandates a consideration of environmental effects, EPA argues that it can condition funding for sewage *189plants on terms designed to avert environmental consequences that pose risks to human health and safety. The agency does not contend, indeed it could not in light of the statute’s “essentially procedural” nature, that the Act dictates any particular substantive policy authorizing EPA to impose land use controls or ban floodplain development.
The other ground on which EPA relies is Executive Order 11,988, issued on May 25, 1977, in part “to avoid the direct or indirect support of floodplain development whenever there is a practicable alternative.” 42 Fed.Reg. 26951. The order was based on the National Environmental Policy Act, the National Flood Insurance Act of 1968, 42 U.S.C. §§ 4001-4127, and the Flood Disaster Protection Act of 1973, Pub.L. 95-128, 91 Stat. 1144, 1145, codified at 42 U.S.C. §§ 4003, 4106 (Supp. IY 1980). The order was prompted to some extent by the unsatisfactory federal experience with losses under the Flood Insurance Programs. In addition, federal agencies had not properly observed flood protection precautions for their own installations, despite the fact that these measures had been required for state and nongovernmental structures.12
Executive Order 11,988 was designed to apply to federal facilities, as well as those constructed for other entities through the use of federal funds. Essentially, the Order requires federal agencies to avoid taking action in a floodplain wherever there is a practicable alternative and to minimize the harm to floodplains that might be caused by any agency action.13
Two aspects of this case touch on floodplain policy. The sewage plant itself is to be located in the floodplain for reasons of economy. The existing plant is there and some of its structures are to be integrated into the new facility. Largely for that reason, EPA concluded that other sites would not be practical alternatives and decided that an exception to its ban on action within the floodplain was appropriate.
The second aspect is only peripheral to the plant itself and that is the prohibition on hookups to the proposed residences within the floodplain. EPA’s ban was not directed at the operator of the sewage plant or its appurtenances, but rather to customers who would be serviced by the plant— people who are not directly subject to EPA authority. The agency persisted in its position, even though the state plan required developers to comply with federal flood insurance protection standards, thus reducing the government’s loss exposure, one of the principal objects of the Executive Order. Not satisfied with the state’s limitations, however, EPA went further and adopted a zero growth approach, the ultimate in minimization. Congress has never taken such a drastic step in regulating overall floodplain development. When Congress did act with respect to specific regions, the coastal barrier areas, it made its position absolutely clear.14
*190We do not meet the issue in this case of whether EPA lacked all authority to take the action it did, but decide the case on a narrower ground.15 Our brief review of the National Environmental Policy Act and the Executive Order authority relied on by EPA is but a backdrop to our consideration of whether EPA’s action was “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” 5 U.S.C. § 706(2)(A) (1976).
The Supreme Court has stated that “[t]o make this finding the court must consider whether the [agency] decision was based on a consideration of the relevant factors and whether there has been a clear error of judgment.” Citizens to Preserve Overton Park v. Volpe, 401 U.S. 402, 416, 91 S.Ct. 814, 823, 28 L.Ed.2d 136 (1971). Yet, “the concept of ‘arbitrary and capricious’ review defies generalized application and demands, instead, close attention to the nature of the particular problem faced by the agency.” Natural Resources Defense Council, Inc. v. Securities Exchange Commission, 606 F.2d 1031, 1050 (D.C.Cir.1979). Thus, the scope of judicial review necessarily admits of some flexibility and the stringency of our inquiry will depend on, and often vary according to, the variety of factors presented by a particular case.
If an agency’s action is clearly within its statutory authority, then the arbitrary and capricious standard focuses on the factual issues. When, however, there is some doubt about the agency’s compliance with statutory constraints, that factor may throw a somewhat different light on the factual evaluation. As agency action moves toward the gray area at the outer limits of statutory authority, the arbitrary and capricious nature of the action may be more evident. For that reason, we have discussed the agency’s asserted sources of power. Another shadow is cast when agency action, not clearly mandated by the agency’s statute, begins to encroach on congressional policies expressed elsewhere.16
We conclude that there are a number of factors which demonstrate that the EPA action was arbitrary. The most obvious is the agency’s failure to give sufficient weight to the congressional admonition in the Coastal Zone Management Act that, to the “maximum extent practicable,” federal *191actions are to be consistent with the state’s management plan.
EPA contends that “the consistency requirement of the statute only contemplates that federal agencies will not support activities in the coastal zone which are prohibited by the state’s plan.” (Appellee’s Brief at 31). In essence, EPA is arguing, not that its action is consistent, but that the consistency requirement does not apply to its action since rather than allowing prohibited development, it seeks to prohibit allowed development. EPA also points out that a request by the New Jersey Department of Environmental Protection for mediation of the dispute resulting from EPA’s action was denied by the Secretary of Commerce. See 16 U.S.C. § 1456(h).
The Secretary’s decision to deny mediation and EPA’s argument here both rest on a regulation promulgated by the National Oceanic and Atmospheric Administration, the federal agency charged with administering the Coastal Zone Management Act. That regulation reads:
“When Federal agency standards are more restrictive than standards or requirements contained in the State’s management program, the Federal agency may continue to apply its stricter standards (e.g., restrict project development or design alternatives notwithstanding permissive management program policies) .... ”
15 C.F.R. § 930.39(d) (1982).
The regulation is contained in Subpart C of NOAA’s regulations on “Federal Consistency with Approved Coastal Management Programs.” Subpart C deals with “Consistency for Federal Activities.” The regulations expressly provide that “[t]he term ‘federal activity’ does not include ... the granting of Federal assistance to an applicant agency (see Subpart F of this part).” Id. § 930.31(c).
It is Subpart F, which deals with “Consistency for Federal Assistance to State and Local Governments,” that is pertinent to EPA’s action here and that section contains no counterpart to the regulation cited by EPA. Subpart F states that “[njotwithstanding State agency consistency for the proposed project, the Federal agency may deny assistance to the applicant agency.” Id. § 930.96(a). The regulations thus make a distinction between “federal activity” and “federal assistance to a non-federal activity.”
When federal assistance is provided for what is' essentially a state or local activity, the congressional preference for having policies initiated at the state level must be respected. Consistency to the maximum extent practicable with the state’s determination is at the heart of the statutory scheme of encouraging, but not directing, state management of the coastal areas. The congressionally mandated consistency requirement becomes even more compelling where, as here, the federal agency seeks to reach beyond the local activity it is funding and impose a federal standard on private activity traditionally subject only to state and local regulation. In short, the inconsistency of EPA’s action with the state’s plan is a factor to be considered in determining whether that action was arbitrary.
EPA should not have defied the state’s decision to allow development in the limited areas of the floodplain at issue here. The same national interest in floodplains that EPA purports to uphold has been incorporated into the New Jersey coastal management plan. In compliance with the directive of the Coastal Zone Management Act that “[t]he management program provided for adequate consideration of the national interest involved in the siting of facilities,” 16 U.S.C. § 1455(c)(8), New Jersey adopted a policy that “[i]n general, coastal development is discouraged in flood-hazard areas.” State of New Jersey Coastal Management Program — Bay and Ocean Segment (FEIS), Part II, Chap. 4, Sec. 5.23.2(a) at 161 (effective September, 1978). That policy was based on a review of the Flood Disaster Protection Act, the National Flood Insurance Act and Executive Order 11,988 and a recognition that “the national interest in these areas is to avoid the long and short term adverse impacts associated *192with the occupancy and modification of floodplains.” Id., Chap. 6 at 190.
The developer’s proposal for development of Tract A was approved as an exception to the general policy because all structures would be elevated one foot above the base flood level and would not increase flood damage potential by obstructing flood waters. Thus, the state had not acted irresponsibly.
In addition, the state decision to allow development in tract A was based to some extent on a local need to encourage building close to other developed areas, rather than at scattered sites away from the population center or near environmentally sensitive areas. In essence, this was simply “fill in” on a somewhat larger scale than EPA proposed. The difference between the EPA position and the state plan in this respect was one of degree. EPA wished to limit “fill in” to lots of 3.4 acres; the state plan looked to a larger area.
A related economic concern of the local municipality was its investment of over a million dollars in roads, sewers and water mains to the area. The plan to use the existing infrastructure is tied'in with the local governments’ aversion to scattered site development in areas where no such facilities had been constructed.
Under the EPA restrictions, houses may be built on lots that are at an elevation of more than 10 feet above mean sea level. These lots comprise 7.7 acres in tract A and 22 acres in tract B. The record reveals no efforts by EPA to have serious discussions about the concept of cluster zoning in the area, which would allow development but at the same time reduce the number of structures on land lower than 10 feet above sea level.
We note also that EPA’s assertion that the tracts under scrutiny would provide flood water storage and wave energy dissipation have been sharply attacked and have little record support.17
It is significant also that the compromise plan which the Authority submitted to the EPA provided for a tightening of local control. Although the provisions of the coastal management plan apply to development of 25 or more lots, the Authority proposed to extend the restrictions to individual lots. Thus, the building limitations which had previously been applicable only to large developments would be applied to all.
When passing the Coastal Zone Management Act, Congress was cognizant of the valuable contribution local governments can make to responsible management of the coastal areas.
“Local government does have continuing authority and responsibility in the coastal zone. * * * Whenever local government has taken the initiative to prepare commercial plans and programs which fulfill the requirements of the Federal and coastal state zone management legislation, such local plans and programs should be allowed to continue to function under the state management program.”
Sen.Rep. No. 753, 92nd Cong., 2d Sess., reprinted in 1972 U.S.Code Cong. & Ad.News 4776, 4779. The Authority’s plan seems especially appropriate here in that it not only meets the requirements, but extends the reach, of the coastal management legislation.
After weighing all the factors, we are persuaded that EPA acted arbitrarily and contrary to law in refusing to accept the Authority’s compromise. ' That proposal was fully in accord with the state’s management plan and the Coastal Zone Management Act. Even if it be conceded that EPA had the power to enforce the land use re*193striction, a question we do not decide, the agency’s action was in excess of that required under the circumstances.
We recognize the legitimate interest in limiting development of floodplains and that, under other circumstances, EPA’s actions might be sustainable. But the circumstances here lead us to hold that the district court erred in entering judgment for the defendants. Accordingly, the judgment of the district court will be vacated and the case will be remanded for further proceedings consistent with this opinion.
. “M.g.d.” means “million gallons daily” and refers to the maximum amount of waste water a sewage facility is equipped to treat on a daily basis. Thus, the existing Cape May sewage plant has a capacity of 3 million gallons of waste water a day.
. The term “floodplain” refers to the lowland and relatively flat areas of land adjoining inland and coastal waters. See Ex. Order No. 11,988, sec. 6(c), 42 Fed.Reg. 26951 (May 25, 1977), reprinted in 42 U.S.C. § 4321 note at 820 (Supp. IV 1980). Floodplains are basically those lower areas of land that flood waters will flow to first and recede from last. Based on historical studies of prior flooding and statistical analyses of terrain and water flow, the Federal Insurance Administration has prepared Flood Insurance Rate Maps that identify those areas of a community that, on the average, are likely to be flooded once every 100 years (i.e., a one percent chance of flooding in a given year).
. The National Environmental Policy Act, 42 U.S.C. §§ 4321-4347 (1976 & Supp. 1980), requires all federal agencies to prepare an environmental impact statement on “major Federal actions significantly affecting the quality of the human environment.” Id. § 4332(2)(C). An agency may dispense with the EIS requirement where the initial environmental assessment reveals no significant effects on the environment. See 40 C.F.R. §§ 1500.4(q), 1508.13 (1981). EPA’s decision to forego an EIS is not at issue in this appeal.
. The special condition attached to the grant read, in pertinent part:
“In order to implement Executive Order 11988, in conjunction with non environmentally sensitive and non wildlife habitat flood plain, and the Endangered Species Act, the Federal Wildlife Coordination Act, and the National Environmental Policy Act, and regulations thereunder in connection with environmentally sensitive and wildlife habitat areas, the grantee agrees that for a period of 50 years from the date of the FNSI/EA no sewer hook-up or other connections to the sewage treatment system included in the scope of this grant will be allowed or permitted so as to allow the discharge of wastewater from any building, facility, or other construction on any parcel of land within any critical wildlife habitat or environmentally sensitive area, or within the 100 year floodplain, as those areas are delineated in the approved facilities plan amendment....”
. Also named as defendants were the regional director of EPA, two section chiefs, the Authority, and the New Jersey Department of Environmental Protection. Although listed as a defendant in the district court and an appellee in this appeal, the Authority agrees with the plaintiffs position. When the case was before the district court for summary judgment, the New Jersey Department of Environmental Resources wrote that it tentatively concluded that EPA had failed to act consistently with the Coastal Zone Management Act.
. The Authority’s cross-claim against EPA requested a declaratory judgment determining the effective date of the grant condition. As such, the claim was irrelevant to plaintiffs action challenging EPA’s authority for imposing the condition.
. The developer also challenges EPA’s authority to deny permits for septic tanks and package treatment plants. The agency argues that the permit challenge is not ripe for review because no application for a permit has yet been made. In view of our resolution of the case, we do not reach this issue.
. Most of parcels A and B lie within the “100-year floodplain,” or “Zone A” as depicted in the Flood Insurance Rate Maps for the Cape May region. Hence, by virtue of the local government’s compliance with flood insurance regulations promulgated by the Federal Emergency Management Administration, any construction of residential structures in the tracts must “have the lowest floor (including basement) elevated to or above the base flood level.” 44 C.F.R. § 60.3(1), (2) (1980). As the base flood level for the Cape May region has been determined to be 10 feet above mean sea level, any structures built in the tracts will have their lowest floor elevated above that level. The permit for development of residential structures in tract A, issued by the New Jersey Department of Environmental Protection, was conditioned on all first habitable floor levels being 11 feet or more above mean sea level.
. “[Tjhere is a considerable body of opinion, well represented in Congress, that in most cases land use decisionmaking should be a state and local, not a federal concern.” B.H. Holmes, “Federal Participation in Land Use Decisionmaking at the Water’s Edge — Floodplains and Wetlands,” 13 Nat. Resources Law., 351, 352 (1980-81). In the early 1970’s, Congress rejected numerous attempts to pass a national land use policy act. Although the bills introduced would merely have made federal funds available for the states to implement their own programs, “opponents considered it to be an attempt to indirectly impose federal environmental policies on the local zoning process.” Id.
. The 1980 amendments to the Coastal Zone Management Act state that the management plan should include provisions for the protection of floodplains and minimizing the loss of life or property caused by “improper development in flood-prone areas.” 16 U.S.C. §§ 1452(2)(A), (B). The House Committee Report emphasized that this clarification of national policy did not represent a new program requirement. H.R.Rep. No. 1012, 96th Cong., 2d Sess. 16, reprinted in 1980 U.S.Code Cong. & Adm.News 4362, 4364. The approved New Jersey plan has incorporated the protection of floodplain areas as one of its policies for coastal management.
. See also Cramton and Berg, “On Leading a Horse to Water: NEPA and the Federal Bureaucracy,” 71 Mich.L.Rev. 511, 521 (1973) (“[NEPA] should not be interpreted as granting to every federal agency a roving commission to defend the environment wherever its writ runs, irrespective of the nature of the subject matter before it.”)
. Executive Order 11,988 was preceded by Executive Order 11,296, 31 Fed.Reg. 10663 (1966), which directed federal agencies to “preclude the uneconomic, hazardous or unnecessary use of floodplains” in discharging their responsibilities for federal lands and facilities. Part of the impetus for Executive Order 11,988 was the failure of federal agencies to adequately comply with Executive Order 11,296. See Holmes, supra, 13 Nat. Resources Law. at 364-66.
. The Water Resources Council has published guidelines for agency implementation of Executive Order 11,988. See 43 Fed.Reg. 6030 (Feb. 10, 1982). “The guidelines do not intend to prohibit floodplain development in all cases, but rather to create a consistent government policy against such development under most circumstances.” Id. at 6033. The guidelines’ interpretation of Executive Order 11,988 is not inconsistent with the policy in New Jersey’s coastal management plan which generally discourages building in floodplain areas.
. The Coastal Barriers Resource Act, Pub.L. No. 97-348, 96 Stat. 1653 (approved Oct. 18, 1982) establishes the coastal barrier resources system, consisting of bay barriers, tombolos, barrier spits and barrier islands within speci- ' fled areas of the Atlantic and Gulf coasts. Further federal assistance, with certain limited exceptions, for development within or access to those areas is banned. The Act also prohibits flood insurance for any new construction or substantial improvements of structures within the system after October 1, 1983. The stated purpose of the legislation is “to minimize the loss of human life, wasteful expenditure of fed*190eral revenues, and the damage to fish, wildlife, and other natural resources ... by restricting future federal expenditures and financial assistance which have the effect of encouraging development of coastal barriers.” Id. § 2(b). This statute demonstrates that when Congress intends to adopt a zero growth approach for an area by withdrawing all financial incentive to development, it does so expressly. Congress has never taken such an approach to floodplain development generally.
. In City of New Brunswick v. Borough of Milltown, 686 F.2d 120 (3d Cir.1982), petition for cert. filed, 51 U.S.L.W. 3321 (U.S. Oct. 12, 1982) (No. 82-662), we held that EPA has authority to withhold federal grant funds for a sewage plant because one of the municipalities to be serviced by the plant had not adopted a system of user’s chdrges as required by the Clean Water Act. See 33 U.S.C. § 1284(b)(1) (Supp.1982). The EPA action there was based upon an express provision of the Clean Water Act that “ensures that every facility built with federal financial assistance will be economically self-sufficient.” 686 F.2d at 126. Hence, unlike here, EPA’s action was directly related to its primary obligation under the statute.
. See Southern S.S. Co. v. N.L.R.B., 316 U.S. 31, 47, 62 S.Ct. 886, 894, 86 L.Ed. 1246 (1942) (“Frequently the entire scope of Congressional purpose calls for careful accommodation of one statutory scheme to another, and it is not too much to demand of an administrative body that it undertake this accommodation without excessive emphasis upon its immediate task.”); Oglala Sioux Tribe of Indians v. Andrus, 603 F.2d 707 (8th Cir.1979) (agency decision was arbitrary and capricious where based upon regulations that violated congressional policy expressed in another statute); Zabel v. Tabb, 430 F.2d 199, 209 (5th Cir.1970), cert. denied, 401 U.S. 910, 91 S.Ct. 873, 27 L.Ed.2d 808 (1971), (“Governmental agencies in executing a particular statutory responsibility ordinarily are required to take heed of, sometimes effectuate and other times not thwart other valid statutory governmental policies.”) See also, L. Jaffe, Judicial Control of Administrative Action 590 (1965) (“[T]he statute under which an agency operates is not the whole law applicable to its operation. An agency is not an island entire of itself. It is one of many rooms in the magnificent mansion of the law. * * * Thus in the review of administrative actions a court may appeal to criteria of validity which have no specific locus in the [agency’s] statute.”)
. In formulating its condition, EPA simply adopted the 100-year floodplain standard and the designations provided by the Flood Insurance Rate Maps for the Cape May region. There was no independent evaluation of the topography of the tracts at issue or findings as to whether the tracts would in fact store flood waters and dissipate wave energy. Thus, the administrative record contains no response to plaintiffs claim that tract A is actually higher than the surrounding land. Nor is there any indication that EPA gave serious consideration to plaintiffs proposal to design the tract B development so as to improve wave energy dissipation capabilities.