Full opinion text
ORDER BUTLER, Chief Judge. I. Introduction This action again brings before this court a contest between real estate development along the coast of Baldwin County, Alabama, and preservation of a federally-protected endangered species, Peromyscus polionatus ammobates, commonly known as the Alabama Beach Mouse. In its current procedural posture, the case presents a narrow issue (which in the nature of such matters is replete with acronyms): whether plaintiffs are entitled to a preliminary injunction in relation to their claim that the United States Fish and Wildlife Service (“FWS”) violated the National Environmental Policy Act (“NEPA”) in granting two Incidental Take Permits (“ITPs”) to developers without requiring preparation of an Environmental Impact Statement (“EIS”). The ITPs at issue involve 196.5 acres of the remaining beach mouse habitat, and allow two developers to take (harm or kill) beach mice both during construction of several high-rise condominium towers and other structures on Fort Morgan Peninsula, Baldwin County, Alabama, and through the degradation and/or destruction of its habitat. These projects and the permits are discussed in more detail below. Plaintiffs bring this action pursuant to the National Environmental Policy Act (“NEPA”), 42 U.S.C. §§ 4321, et seq., and the Administrative Procedure Act (“APA”), 5 U.S.C. §§ 551, et secy. Plaintiffs seek a preliminary injunction prohibiting the “take” of any ABM and the destruction of any of its habitat by Gulf Highlands LLC and Fort Morgan. Paradise Joint Venture (“Developers” or “In-tervenors”) in reliance on the ITPs issued for the proposed developments on Fort Morgan Peninsula in Baldwin County, Alabama. On April 30, 2002, the Court held a hearing on the plaintiffs’ Motion for Preliminary Injunction. All parties, including intervenors, were represented. The government and the intervenors presented selected documents which formed part of the administrative record. The government also offered live testimony of two witnesses, David Flemming, an ecological services supervisor in the Atlanta Regional Office of the FWS, and Will McDearman. Mr. Flemming supervised the ITP review in this case. Mr. McDearman was involved with preparation of the ITP analysis for two prior ITPs which this court overturned in Sierra Club v. Babbitt. Since 1998, he has been working as a biologist for the FWS field office in Jackson, Mississippi, and thus had no role in addressing the two current ITPs. In addition, each of the developers submitted affidavits addressing the amount of economic harm should the court grant a preliminary injunction, the value in monetary terms of the land set aside as a “preserve” for the beach mouse under the HCP, and other matters. The court has been previously called upon to wade into similar conflicts over the Alabama Beach Mouse. In one such recent case, Sierra Club v. Babbitt, 15 F.Supp.2d 1274 (S.D.Ala.1998), this court remanded two other ITPs which had been issued by the FWS for development projects oh the Fort Morgan Peninsula in Baldwin County. " The court held that the lack of current, reliable population data, lack of data on distribution of the Alabama Beach Mouse within its range, and lack of an estimate of the minimum viable population size rendered the FONSI arbitrary and capricious, in the absence of any basis for determining how the proposed action would affect the ability of the Alabama Beach Mouse to survive or recover. In the instant case, the FWS has again issued two ITPs for development on 196.5 acres of privately owned land on the Fort Morgan Peninsula of Baldwin County. The agency- asserts that it has avoided the failings previously found by this court. The FWS does not now claim to have sufficient additional data on total population and population density and distribution within the range to allow a determination of the effect of the developments on beach mouse viability; the agency has not determined the minimum population needed for the continued viability and recovery of the species. Instead, the agency asserts that it can approve the permits based primarily upon a habitat-based analysis. However, the record does not demonstrate that the agency has any clearer idea what effect the project will have on the beach mouse under this approach than it did under the incomplete population-based approach previously criticized by the court. Indeed, the Environmental Assessment issued by the FWS acknowledges that, under its habitat-based analysis, the agency still does not know what effect the loss of optimal habitat is likely to have on the species. Tab 7(EA) §§ 4.4.4, 5.1.1. Nor has the Service cogently identified, under its habitat-based analysis, the minimum habitat requirements for the continued survival-let alone recovery-of the species. It thus has not considered such minimum requirements in connection in the decision that the ITPs at issue would not cause a “significant” impact on the environment, so as to require more detañed analysis of the decision through preparation of an Environmental Impact Statement. Despite this lack of data, the agency entered a finding that the proposed action, would have no significant impact on the environment, including particularly the endangered Alabama Beach Mouse. As discussed below, the court finds that plaintiffs are likely to succeed on the merits of their claim that the FWS’s Finding of No Significant Impact (“FONSI”) in the absence of such data was arbitrary and capricious, that the threatened impact-as found by the FWS-to the ABM’s habitat is significant, that such impact is more significant when considered in connection with the cumulative impact of other prior and reasonably foreseeable future losses, and that the issuance of the ITPs without completion of an EIS was thus improper. The court further finds that the plaintiffs are entitled to entry of a preliminary injunction. II. Jurisdiction This court has subject matter jurisdiction over this action pursuant to 28 U.S.C. § 1331. A. Standing The court has previously held, in a closely analogous case brought by Sierra Club against the same governmental defendants over incidental take permits for the Alabama Beach Mouse issued to developers on the Fort Morgan Peninsula, that Sierra Club has standing to 'bring such a claim. Sierra Club v. Babbitt, 15 F.Supp.2d 1274, 1276-77 (S.D.Ala.1998). Sierra Club makes similar allegations in this action. The court adopts its former analysis, and determines that standing is no impediment to the court’s exercise of jurisdiction. B. Reviewability: This action challenges a final agency action, the issuance by the Fish and Wildlife Service of an Incidental Take Permit for the endangered Alabama Beach Mouse without an Environmental Impact Statement, and thus presents a reviewable issue. See Citizens to Preserve Overton Park v. Volpe, 401 U.S. 402, 91 S.Ct. 814, 28 L.Ed.2d 136 (1971)(overruled on other grounds by Califano v. Sanders, 430 U.S. 99, 105, 97 S.Ct. 980, 51 L.Ed.2d 192, (1977)); 5 U.S.C. § 701. III. Standard of Review A. Preliminary Injunction: In deciding the plaintiffs’ Motion for Preliminary Injunction, the court considers whether the evidence supports a finding in plaintiffs’ -favor on four elements: “(1) a substantial likelihood of success on the merits, (2) a substantial threat of irreparable, injury if the injunction were not granted, (3) that the threatened injury to the plaintiff outweighs the harm -an injunction may cause the defendant, and (4)that granting the injunction would not disserve the public interest.” Am. Red Cross v. Palm Beach Blood Bank, Inc., 143 F.3d 1407, 1410 (11th Cir.1998). 1. Review of the Merits In reviewing the preliminary injunction claim, the court considers the likelihood of success on the merits. That analysis requires the court to consider the merits of plaintiffs’ claim under the appropriate legal standard for review of that decision. Of particular importance in this action is NEPA’s requirement that the agency prepare an Environmental Impact Statement (“EIS”) for any “major” federal action which will “significantly” affect the “human environment.” In this action, the parties acknowledge that the action is federal and major; the only issue affecting the decision to prepare an EIS is whether the action’s environmental impact is “significant.” To determine if an action’s threatened impacts are significant, the agency must prepare an Environmental Assessment (“EA”), followed either by a Finding of No Significant Impact (“FON-SI”)'or, if the impact is found to be significant, the creation of an EIS. 42 U.S.C. § 4332(2)(C). In this action, plaintiffs challenge the FWS’s decision that it was unnecessary to prepare an Environmental Impact Statement (“EIS”) for the proposed issuance of the two ITPs at issue. The FWS issued the ITPs without preparation of an EIS because they found the impact insignificant; that determination centered on the effects on the endangered Alabama Beach Mouse and its habitat. Plaintiffs assert that this Finding of No Significant'Impact (“FONSI”) was arbitrary and capricious, in violation of NEPA, the APA and the ESA. In Marsh v. Oregon Natural Resources Council, 490 U.S. 360, 371, 109 S.Ct. 1851, 104 L.Ed.2d 377 (1989), the Supreme Court addressed the scope of review of an agency decision not to prepare an EIS. The case involved preparation of a supplemental EIS, but the Court noted that the standard was identical to that for an EIS in the first instance. Id. at 374, 109 S.Ct. 1851. The Court held that court review of such a decision was subject to the APA’s “arbitrary and capricious” standard. Id. at 375-76, 109 S.Ct. 1851. The Court noted, however, that there is.little difference between this standard and the “reasonableness” standard previously applied in this and other Circuits. Id. at 377 n. 23, 109 S.Ct. 1851 (citing Manasota^88, Inc. v. Thomas, 799 F.2d 687, 692, n. 8 (11th Cir.1986)). This holding expressly overruled prior precedent from several circuit courts, including the Eleventh Circuit, National Wildlife Federation v. Marsh, 721 F.2d • 767, 782 (11th Cir.1983), which had applied the reasonableness standard of review. In reviewing the agency’s action under the arbitrary and capricious standard, the court considers whether the agency acted within the scope of its legal authority, whether it has explained its decision, whether the facts on which it purports to rely have some basis in the record, and whether the agency considered the relevant factors. Overton Park, 401 U.S. at 415-16, 91 S.Ct. 814. This determination is necessarily based on the record presented by the agency as the basis for its decision. The court defers to the expertise of the agency, may not substitute its judgment for that of the agency, must limit its review to the record at the time the agency acted, and may not consider subsequent matters. Overton Park, 401 U.S. at 414, 91 S.Ct. 814. The court “must judge the propriety of [the agency’s determination] solely by the grounds invoked by the agency. If those grounds are inadequate or improper, the court is powerless to affirm the administrative action.” ICC v. Brotherhood of Locomotive Eng’rs, 482 U.S. 270, 290, 107 S.Ct. 2860, 96 L.Ed.2d 222 (1987). Overton Park ... vastly expanded the range of arbitrary and capricious review under § 706(2)(A).... Overton Park added teeth to arbitrary and capricious review: when applying arbitrary and capricious review to an agency decision, a court must “engage in a substantial inquiry ... a probing in-depth review-” 401 U.S. at 415, 91 S.Ct. 814. The Court further observed that to “find arbitrariness, the court must consider whether the decision was based on a consideration of relevant factors and whether there has been a clear error of judgment.” Id. at 416, 91 S.Ct. 814, 401 U.S. 402, 91 S.Ct. 814. While “[t]he court is not empowered to substitute its judgment for that of the agency,” it niust engage in a “searching and careful” inquiry into the facts. Id. Sierra Club v. Peterson, 185 F.3d 349 (5th Cir.1999). 2. NEPA The National Environmental Policy Act (“NEPA”), 42 U.S.C. §§ 4321, et seq., requires all federal agencies to consider the effects of their projects on the human environment. The Supreme Court has described NEPA’s requirements. NEPA does not work by mandating that agencies achieve particular substantive environmental results. Rather, NEPA promotes its sweeping commitment to “prevent or eliminate damage to the environment and the biosphere” by focusing Government and public attention on the environmental effects of proposed agency action. By so focusing agency attention, NEPA ensures that the agency will not act on incomplete information, only to regret its decision after it is too late to correct. Similarly, the broad dissemination of information mandated by NEPA permits the public and other government agencies to react to the effects of a proposed action at a meaningful time. Marsh, 490 U.S. at 371, 109 S.Ct. 1851(in-ternal citations omitted). The Act is a purely procedural statute, and does not impose substantive requirements. In this case, for example, it would not have violated NEPA if the Forest Service, after complying with the Act’s procedural prerequisites, had decided that the benefits to be derived from downhill skiing at Sandy Butte justified the issuance of a special use permit, notwithstanding the loss of 15 percent, 50 percent, or even 100 percent of the mule deer herd. Other statutes may impose substantive environmental obligations on federal agencies, but NEPA merely prohibits uninformed — rather than unwise — agency action. Robertson v. Methow Valley Citizens Council, 490 U.S. 332, 350-51, 109 S.Ct. 1835, 104 L.Ed.2d 351 (1989). Under Eleventh Circuit precedent, there are four criteria to be considered in determining whether an agency’s decision not to prepare an EIS is arbitrary and capricious: First, the agency must have accurately identified the relevant environmental concern. Second, once the agency has identified the problem it must have taken a “hard look” at the problem in preparing the EA. Third, if a finding of no significant, impact is made, the agency must be able to make a convincing case for its finding. Last, if the agency does find an impact of true significance, preparation of an EIS can be avoided only if the agency finds that changes or safeguards in the project sufficiently reduce the impact to a minimum. Hill v. Boy, 144 F.3d 1446 (11th Cir.1998); see also North Buckhead Civic Assn. v. Skinner, 903 F.2d ' 1533 (11th Cir.l990)(adopting arbitrary and capricious review in all NEPA cases reviewing agency decisions). Section 10 of the Endangered Species Act of 1973, (“ESA”), 16 U.S.C. § 1339(a)(1)(B), provides that the FWS may issue an “Incidental Take Permit” under certain conditions to allow a “take” of a threatened or endangered species which would be otherwise illegal under the Endangered Species Act, where the take would be incidental to some other lawful purpose. As part of the permit process, the applicant is to develop and submit a conservation plan for the remainder of the species. 16 U.S.C. § 1539(a)(2)(A); 50 C.F.R. § 17.22(b)(1) and 17.32(b)(1).' The plan must “specif[y] .,. the impact which will likely result from such taking,” and the FWS shall issue an ITP only if it finds several facts, including that “the taking will not appreciably reduce the likelihood of the survival and recovery of the species in the wild.” 16 U.S.C. § 1539(a)(2)(B)(iv). This provision formed the basis of the agency’s decision to issue the ITPs and also underscores its determination of significance under the CEQ regulation. However, as noted above, plaintiffs do not bring a claim for violation of the ESA at this point in the action. The parties do not contest that the ITPs at issue constitute major federal action. The plaintiffs do not challenge any effect other than the impact on the Alabama Beach Mouse and its habitat. However, on that issue the parties are at loggerheads. , As currently presented, the issue is whether plaintiffs are entitled to a preliminary injunction against further construction premised pn the ITPs; as part of that determination, the court considers the likelihood that plaintiffs will ultimately prevail on the merits. , Plaintiffs challenge the FONSI under NEPA as arbitrary and capricious, and assert that the FWS failed to make a convincing case for the finding of no significant impact. They assert that the ITPs should not have issued without preparation of an EIS. Defendant asserts that its finding of no significant impact was properly supported, and alternatively that, even if the action were significant, the Habitat Conservation Plan reduces that effect to insignificance. Hill v. Boy, supra. The, Center for Environmental Quality (“CEQ”) has issued regulations governing the NEPA process. Of particular note is 40 C.F.R. § 1508.27, in which the CEQ identifies ten factors to be considered in determining whether the threatened impact of an action is significant. An environmentally significant action need not involve a threat of extinction to a federally-protected species. Lesser impacts, including impacts on non-listed species, can constitute a significant effect. See e.g. Makua v. Rumsfeld, 163 F.Supp.2d 1202, 1218 (D.Hawai’i 2001) (distinguishing the ESA standard of “jeopardy” from the NEPA standard of significant impact.); Environmental Defense Fund, Inc. v. Massey, 986 F.2d at 536. However, the involvement of a federally protected threatened or endangered species and the possibility of a threatened impact as potentially serious as extinction or extirpation from a traditional habitat are factors weighing in favor of a finding of significance. Another factor of particular relevance in this action is the existence of uncertainty regarding the environmental effect of the proposed action. 40 C.F.R. § 1508.27(b)(5). IV. Facts A Agency Findings at Time of Listing In 1985, the Fish and Wildlife Service published its Final Ruling listing the Alabama Beach Mouse as an endangered species and designating its critical habitat. 50 Fed.Reg. 23872. That document contains a description of the beach mouse subspecies, its situation, and its historical and current habitat. The findings made at that time have not been amended, though the FWS is considering whether to amend the designation of its critical habitat. The Alabama Beach Mouse was one of sixteen recognized subspecies of Peromys-cus polionotus, otherwise known as the old field mouse. The Alabama Beach Mouse was originally found on coastal dunes from Fort Morgan (at the eastern tip of fort Morgan Peninsula) to Alabama Point and on Ono Island, as well; all of those locations are on the Gulf Coast of Baldwin County, Alabama. The Service noted that a 1983 study had determined that the Alabama Beach Mouse survived on “disjunct tracts of the sand dune system from Fort Morgan State Park to the Romar Beach area, but [has] apparently disappeared from most of its original range, including all of Ono Island.” At the time of listing, the Service made extensive findings concerning the perils which the Alabama Beach Mouse had endured and which it continued to face. It had been extirpated from much of its original habitat. It suffered a severe decline in population due to “human and natural alteration of coastal ecosystems.” The Service found that “most suitable habitat has been lost because of residential and commercial development, recreational activity, beach erosion, and vegetational succession. Competition from introduced house mice (Mus musculus) and predation by domestic cats (Felis catus) also seem to be problems. Tropical storm's’ are a constant threat to the remnant, fragmented populations of beach mice.” “Residential and commercial development have already isolated the remaining areas of beach mouse habitat, fragmenting populations. Because of the history of devastating tropical storms, often extirpating beach mice, it is necessary to maintain several suitable areas of habitat, irrespective of ownership, if the beach mice are to have a reasonable chance of survival and recovery.” The agency stated that “[a] substantial decline of beach mouse habitat, through destruction or adverse impact by development, has been noted just since data were collected for the proposed rule of June 7, 1984.” The FWS concluded that “[t]he major threat to beach mouse habitat continues to be human destruction of the coastal sand dune ecosystem for commercial and residential development.” . In designating critical habitat, the FWS relied on a 1983 study which concluded that “the minimum area needed to maintain a population of beach mice is 50 hectares (ha) (124 acres (A)), that preferable habitat size is at least 100-200 ha (247-494' A), and that there should be natural corridors for migration between areas.” The Service determined that “[t]he protection of several separate areas of habitat for each species of beach mouse is essential for the conservation of these animals. Should a. species of beach mouse exist in only one small stretch of suitable habitat, it would be much more vulnerable to extinction through the effects of tropical storms and other deleterious factors.” In designating the critical habitat of the three subspecies of beach mice, the FWS held that Section 4(b)(8) of the [Endangered Species] Act requires, for any proposed or final regulation that designates critical habitat, a brief description and evaluation of those activities (public or private) which may adversely modify such habitat or may be affected by such designation. Activities most likely to adversely modify the critical habitat of the three beach mice are the continued destruction of sand dunes for residential and commercial development. Indiscriminate pedestrian and vehicular use also adversely impacts the sand dunes. 50 Fed.Reg., at 23881 (emphasis added). In denying a request by the Federal Highway Commission and the Florida Department of Transportation to exclude existing rights of way from the designation of critical habitat for the two other beach mice addressed in the final rule, the FWS stated that “[pjrotection of the remaining habitat is essential for the long-term- survival and recovery of the beach mouse.” Id. at 23878 (Emphasis added) B. Relevant Portions of the Administrative Record The proposed development involves two adjacent tracts totaling 196.5 acres of ABM available habitat. The two developments, while separate, are treated together because the developers offer a consolidated Habitat Conservation Plan and sought joint consideration. The projects involve construction of large condominium developments near the Gulf of Mexico on the Fort Morgan Peninsula in Baldwin County, Alabama. The construction is expected to destroy 58.9 acres of beach mouse habitat, through conversion to paved and other impervious features, including buildings, roads, patios, pools, ponds, tennis courts, a club house, shops, a medical facility, and a fire station, as well as and open areas planted in grasses, trees and shrubs. Six 20-story condominiums ware to be built, as well as thirteen single-family units, and a commercial development which will also contain additional housing. Intervenors assert that their project will impact only .35 acres of critical habitat. The Habitat Conservation Plan (“HCP”) submitted by the developers includes a proposal to set aside 105.5 acres of the 196.5 acre tract in perpetuity for the protection of the Alabama Beach Mouse, and also provides for certain other mitigating steps. One notable aspect of the mitigation is the protection of land on which an ITP had previously been granted. The developers have purchased the 17 acre tract on which the FWS had previously granted án ITP for a development known as “French Car-ribean Plantation Resort.” That proposed development would have led to the destruction of approximately 3.7 acres of habitat. The current HCP would protect that tract from development. By setting aside this 17-acre tract as part of the beach mouse habitat “preserve”, the developers have in effect rescinded the prior ITP and reduced the cumulative effects of prior ITPs. The Intervenors also tout their plan to grade the dunes in front of their buildings, rather than erect a retaining wall. While it is true that the slope of sand will avoid the creation of another physical barrier to Beach Mouse movement from the shore to locations farther inland, it has not been established that the mice will use that slope for that or any other purpose. The developers plan to re-plant this area with native vegetation, and making it possible that the ABMs- will use this are for foraging; nonetheless, ‘the slope appears unsuitable for habitation. Despite the mitigation measures, however, the record clearly indicates that the projects will cause the loss of 58.9 acres of "ABM habitat. Of that, 34.5 acres are considered “optimal habitat.” The agency designated a 711-acre “action area,” upon which it focuses in addressing significance. The FWS calculates the relative loss of habitat as a result of these projects as follows: 8.3 percent of the total available ABM habitat in the action area. 20 percent of the total optimal ABM habitat in the action area. 5.6 percent of the total available ABM habitat range-wide. 13.3 percent of the total optimal ABM habitat range-wide. The projects will also destroy 68 percent of the escarpment on the project site: 1935 linear feet of escarpment out of a total of 2844 linear feet on the property will be permanently affected. - The HCP is intended to reduce the adverse effects- to the beach mouse in the 196.5 acre tract as a whole. It does not purport to “undo” the destruction of that portion of ABM habitat. Some less drastic effects will also be felt in the remaining portions of the tract and in the remainder of the project area, as well, but the court will presume for purposes of this opinion that these other effects are to be successfully mihimized under the plan to the point of insignificance. Since the species was listed in 1985, it has continued to lose habitat. The Biological Opinion contains a table providing habitat loss figures for nine (9) projects within the 771 acre action area. See Record, Tab 8, Table 12. That table shows that 337 acres of habitat were lost to large-scale development in the action area between 1985 and 1996. Further, in 2001, the developers of Kiva Dunes announced a proposal to build high-rise condominiums where the plat (and ITP application) provided for single-family residences. Development of single-family structures on individual lots has also contributed to the loss of habitat. V. Discussion Plaintiffs’ two most persuasive challenges to the Service’s Finding of No Significant Impact, and thus to the resulting decision to issue the ITPs without preparation of an Environmental Impact Statement, are as follows: 1) that the planned destruction of such a large portion of the ABM’s remaining habitat, particularly optimal habitat, is so clearly significant that the FWS’s FONSI was “a clear error of judgment,” ■.Sierra Club v. Peterson, 185 F.3d 349 (5th Cir.l999)(citing Citizens to Preserve Overton Park v. Volpe, 401 U.S. 402, 414, 91 S.Ct. 814, 28 L.Ed.2d 136 (1971)), especially where the affected species is endangered because of prior losses of habitat and has suffered large losses of habitat since listing; and that the agency’s failure to explain and support its counter-intuitive conclusion of insignificance renders the FONSI arbitrary and capricious under the third element of the test identified in Hill v. Boy, 144 F.3d 1446 (11th Cir.1998) (“if a finding of no significant impact is made, the agency must be able to make a convincing case for its finding.”); see also Over-ton Park, supra (in reviewing agency action under arbitrary and capricious standard, court to consider whether the agency has explained its decision); and 2) that the agency’s admitted uncertainty concerning such central issues as the effect of destruction of optimal habitat on the Alabama beach mouse renders the finding of no significant impact arbitrary and capricious. The court finds that plaintiffs have demonstrated a substantial likelihood of success on the merits on each of these issues. The court also notes but does not decide various other subsidiary questions raised in this case. A. A Clearly Significant Impact: It seems almost beyond question that the destruction of 20% of an endangered species’ “optimal habitat” (defined to include all habitat needed by the species), on top of post-listing losses of optimal habitat of approximately 20%, for a species which was rendered endangered by prelisting losses of habitat, will have an impact on that species which is “significant” under any reasonable definition of that term. The FWS reached the opposite conclusion, and does not adequately explain this apparently unsupportable determination. The factors listed in the CEQ regulation do not lead to an easily quantifiable result. Despite the fact that the FONSI lists and addresses each of the ten factors identified by the CEQ in its regulation defining “significantly,” the FWS does not clearly state how it weighed or balanced those factors or how it reached its determination of no significant impact under that framework. Nor, despite the inclusion of large amounts of data, findings and analysis, has the agency answered how such a large absolute and relative habitat loss can be said not to be significant. The FWS stated that the loss of an acre in one location would have a different impact than an acre from another location. One might therefore expect the Service to claim that the large loss of habitat would not significantly affect the ABM because the land was not used by the species, or was not important to the species. Instead, trapping data indicates that the species uses the land which is to be destroyed, and a large part of that affected land consists of “optimal habitat.” Defendant does not offer data concerning the population living on or using the 58.9 acres, because it does not have such information. An agency could potentially conclude that a large loss of habitat might not significantly affect the beach mouse if the species did not react to the loss of habitat. In this case, however, the beach mouse was found to be' endangered precisely because of population losses due to the loss of its traditional habitat. Further, habitat destruction has continued since the species was listed; the species lost 24% of its optimal habitat range-wide and 20.6% of its optimal habitat in the action area since it was listed in 1985. Almost all of that habitat loss has been due to development, most of which has been built pursuant to ITPs previously granted by the FWS. An agency could perhaps argue that it does not deem it necessary to prepare an EIS because it has previously prepared Environmental Impact Statements for one or more of the prior ITPs, either those in the action area or at least one outside that area. However, that is not the case. Though the Service has issued numerous ITPs for the beach mouse, it has never prepared an EIS. An agency could argue that the loss of habitat in the action area would only affect the local colony, and would not significantly affect the species as a whole. First, local extirpation can be a significant impact. Second, the beach mouse must exist in several discrete, but connected, colonies so that local extirpation from storms, succession of the food-producing plants, or other environmental factors does not affect all of the colonies at once, and also for other reasons such as the need for breeding between the separate populations. Fragmentation of the available habitat is another habitat-related problem for the ABM. Otherwise useful habitat which is cut off from other occupied habitat is not useful to the ABM if the area is too'small or insufficiently diverse to support a colony. Even if the area is large enough, the colony can have too little genetic variety in the absence of recruitment from outside the colony. Also, if the population is extirpated from the area, there is no natural means of re-population. Barriers to movement between the shore and interior dunes and scrub also adversely affects the species. When primary and secondary dunes are inundated by storms, the mice need to be able to seek shelter temporarily in interior habitat. Further, the mice need to be able to reach interior habitat for foraging and habitat, particularly when population pressure increases, either due to increases in population or decreases in available habitat or in food supply. The agency acknowledges that it does not know what constitutes barriers to movement along the shoreline (east to west) or between the shore and interior habitat (north to south) The agency further acknowledges that the planned buildings will act as a barrier to north-south movement. Additionally, the agency, though unsure what effect the loss of escarpment will have, notes that the damming effect of the buildings coupled with the loss of escarpment will deflect storm surges toward adjacent interior habitat, making flooding more likely. Tab 7(EA) section 4.8.2, p. 62. The agency has not addressed the effect of this additional flooding of interior habitat on the beach mice, particularly on their access to interi- or habitat and their use of interior habitat as a refuge from storms. The FWS has not clearly addressed how much contiguous habitat will remain.- The HOP calls for preservation of 105.5 acres. Intervenors add several corridors for passage of the ABM between contiguous tracts of habitat, and assert that their projects will leave the ABM with approximately 140 acres of habitat. This gerrymandering is a questionable analytical approach. Further, the FWS does not have reliable data on the minimum habitat requirements for a viable beach mouse colony. An early study concluded that the ABM needed at least 50 hectares (124 acres) but should preferably have two to four times that much habitat available. Intervenors note that their calculation of gerrymandered habitat slightly exceeds the smallest rough estimate of minimum habitat, and claim this as proof that the loss of habitat will not have a significant impact on the species. This argument in specious. The use of figures rounded to the nearest half-hundred hectares indicates that the estimate was not intended as a precise measure of minimum habitat requirements. It is certainly not precise enough to warrant Intervenors’ conclusion. Even if the estimate were intended to be precise and took into account different types of habitat needed, the argument that the project will have little impact on the species because Intervenors’ plan will leave slightly more habitat than the lowest estimate of minimum habitat requirements is not supportable; significant impacts could include adverse effects not just on survival but on recovery and maximum supportable population size. The Service notes that the action area “currently has a relatively low population”: which it “believes” to be due to the species’ failure to recover from losses due to prior storms and other factors which locally reduced the food supply. (Tab 8(BO) section VI, pp. 91-92) The loss of 58.9 acres of habitat which includes food sources for the ABM, and additional losses of individuals, would be likely to have a greater-than-usual impact on the local population and its recovery because of current conditions.in that area. Further, the Service lacks .a reliable estimate of the minimum population requirements for long-term viability and recovery. Without this information, any conclusion regarding the health of this colony despite the killing of mice and the loss of burrows and forage by construction is flawed. Further, there was evidence presented to the FWS that the ABMs had lost so much habitat that they were at or near the point at which they could not survive or recover, particularly if subjected to storms and other environmental risks, or if they suffered additional habitat losses. This court had previously remanded ITPs on the basis of lack of necessary data, including the lack of an estimate of minimum population necessary for survival and recovery, and the lack of estimates of range-wide population and distribution within the range. Defendant has focused on this prior decision in the instant action. However, the agency does not offer that information. Instead, the agency’s analysis focuses on losses of habitat instead of population as the basis of its determination. The agency claims that this habitat-based analysis can better be translated (by some means not made clear) into a more reliable analysis of the effects of the loss of habitat on the species. However, the agency lacks a reliable estimate of the minimum habitat requirements of the species with which to compare any loss of habitat. Because the remaining habitat left by the projects at issue is near- or below the agency’s only, general estimate for minimum -habitat, the court finds that a determination of insignificance is. unsupportable in the absence of a reasonably reliable estimate of minimum habitat requirements to analyze the significance of the impact of such loss on the ABM’s viability and recovery in this area. Under these circumstances,, the court cannot accept defendant’s largely implicit conclusion that the effects of a cumulative loss of occupied and optimal habitat of this magnitude on the beach mouse is so insignificant that preparation of an EIS is not required. Despite mountains of data, this crucial, counter-intuitive conclusion goes largely unexplained. The court thus finds that plaintiffs have a substantial likelihood of demonstrating that the record does not adequately support the conclusion that losses of 20 percent of optimal habitat in the action area, on top of prior losses of approximately 20.6 percent of optimal habitat in the action area, will not cause a significant effect on the Alabama beach mouse. 1. Mitigation: The defendants and intervenors assert that, even if the projects are found to threaten a significant impact, that impact is mitigated to insignificance by the proposed Habitat Conservation Plan. Based upon the record presented, the court cannot agree. At this stage, it appears that plaintiffs are likely to prevail on this argument, as well. The CEQ regulation defining significance states that “[a] significant effect may exist even if the Federal agency believes that on balance the effect will be beneficial.” 40 C.F.R. § 1508.27. This provision appears to be directly applicable to this aspect of the instant action. Despite all of the beneficial aspects of the HCP, particularly as to the 105.5 acres to be protected, the destruction of 20 percent of the remaining optimal habitat in the action area can nonetheless be a significant environmental impact regardless of what beneficial measures may also exist. However, the FWS ignores this provision, citing instead provisions which allow the agency to find that mitigation measures will reduce an otherwise significant environmental effect to- insignificance. See Hill v. Boy, supra (“if the agency does find an impact of true significance, preparation of an EIS can be avoided only if the agency finds that changes or safeguards in the project sufficiently reduce the impact to a minimum.”). Construing this language in the context of the CEQ regulation cited above, the court finds them to be consistent, and that the CEQ’s provision is applicable to the instant ITPs. The mitigation measures, admittedly beneficial and admittedly minimizing many of the lesser adverse effects of this action, do nothing to lessen the irreducible destruction of habitat found by the agency. The remaining elements of the HCP relate to mitigating lesser effects on the habitat; they likewise do not alter the fact that habitat will be lost through more direct and concrete effects. Throughout this opinion, however, the court has focused on the loss of habitat, including the loss of optimal habitat, directly and permanently resulting from the construction of structures on the ABM habitat. Even if the court were to presume that the HCP would completely reduce any effect on that Í05.5 acres portion of the 196.5 acre tract at issue to insignificance, the HCP will not “undo” that destruction, and will not create new habitat to replace that lost through this construction; The court’s analysis based on the agency’s findings of losses of habitat is unaffected by the mitigation measures. B. Uncertainty: The reason for this lack of explanation becomes clearer. The agency admits at several places in its EA that it lacks information on the effect loss of optimal habitat will have on the species. (Tab7 (EA) Sections 4.4.4 (p.51); 4.8.1 (p. 61); 5.1.1 (p. 64)) In the absence of such fundamental information, it would seem that any alleged “finding” that the project will not significantly affect the species is the purest sophistry. The agency also has noted other areas' of uncertainty; the reliability of still other of the “facts” the agency relied on are called into question based on the assumptions made. On November 8, 1999, the FWS announced its “90-day finding” on a petition filed by Sierra Club and others on February 2, 1999. The petition sought- an amendment of the critical habitat designation for the ABM and the two other subspecies of beach mouse which were .listed with the ABM in 1985; the stated reason for the petition was the assertion that the current designated habitat was inadequate and coastal development had destroyed and would continue to destroy part of their habitat. The FWS found that “[available information and data indicate that secondary and scrub dune habitat may be essential to the survival and recovery of all three subspecies”. (64 Fed.Reg. 63004 (November 19, 299.9)). Thus, the Service indicated almost two years ago that it would probably amend the critical habitat designation to include additional secondary and scrub dune habitat; however, because that process has not been completed, the instant ITPs are being granted on the basis of the old designation. The ITPs allow the destruction of habitat types which are “essential to the survival and recovery” of the ABM. The significance of this destruction to the ABM thus may be understated in the instant analysis due to the agency’s delay in addressing this issue, and the protections granted to critical habitat are withheld from the optimal habitat destroyed under the current proposal. Private development of single-family structures is another area of uncertainty. The EA states that there are approximately 180 undeveloped lots in the action area, averaging approximately 0.5 acres each, and more than 600 lots outside the action area but within ABM habitat on Fort Morgan Peninsula. Tab 7, section 5.1.2. The agency estimated that up to 40 new single family residence are constructed on Fort Morgan Peninsula annually, and that each single-family residence-if constructed on pilings as is eommon-destroys approximately 0.25 acres of habitat. Tab 7, section 3.10, pp. 31-32. The agency estimates that annual construction within the action area would be.only 4 new residences per year, for an anticipated annual loss of habitat of 2 acres within the action area. Id. section 5.1.2. The factual basis for this estimate does not appear in the record. Such construction is not presently regulated, but the FWS asserts that it “anticipates” that such construction will become regulated (at an undisclosed time “in the future”) as a result of “a county-wide HCP or other section 10 compliance methodology.” Id. The record, however, does not offer a factual basis for this conclusion, and thus for the optimistic assumptions that land undisturbed under an agency-issued ITP will remain undisturbed. Even if it is not the most probable outcome, it is at least foreseeable that additional losses of habitat due to unregulated construction of single-family residences could accelerate and cause substantial future losses of habitat in the action area. However, as noted at the beginning of this section, the central area of uncertainty centers on the effects of the loss of habitat-particularly optimal habitat-on the species. In the BO, the FWS included population data in an appendix. The agency notes that such population data was not the primary basis for their FONSI analysis. The Service decided instead to use a habitat-based assessment to determine the effect of the development on the beach mice. The FWS stated that “[t]he detailed discussion of ABM population data is included in Appendix B — Point 6. This information is placed in the Appendix because the widely fluctuating characteristics of the ABM population makes it impossible to identify the impact of a development to the populations directly. Therefore, we have determined to do the assessment as a result of acres of habitat that is used by the ABM and impacted by the development.” TAB 8, BO, C.l. p. 28 (emphasis added). Because of dramatic seasonal variations in ABM population, the Service analyzed trapping figures by month to estimate the population. The FWS found that data taken strictly from the grids did not include a population estimate for the entire Perdue Unit. of the ABM’s habitat and that the estimates were statistically insufficient to allow a range-wide estimate. Thus, the Service instead estimated ABM population in the action area by aerial photography. From the photographs, it identified similar habitat types and estimated the amount of such habitats. From that information, the FWS estimated a population of between 646 and 5,013 beach mice. The court also notes with concern the wide range of the population estimate given; it appears that the Service’s estimates of population from habitat types (and vice versa) involve a substantial rate of error, at least when applied to a large area. This error rate leads to substantial uncertainty in using these estimates, and may significantly limit their reliability and usefulness. Further, the record does not indicate that the stated error rate takes into account the issues as to which the agency admits a lack of knowledge of the effects of the proposed changes. The court further notes its concern that the apparently simplistic habitat model may gloss over the agency’s lack of information concerning the minimum requirements of the species, and may ignore what little is known of the minimum habitat requirements for individual ABM colonies. In order for the Service to use a habitat model to “identify the impact.of a development to the populations [in]directly,” the agency clearly must have a basic understanding of the effects of losses of habitat on the species. The agency found that the project would destroy 58.9 acres of ABM habitat, including 34.5 acres of optimal habitat. Optimal habitat for the species is identified as “those sections of coastal strand that contain a primary/secondary dune, escarpment and adjacent scrub. These habitats areas make up, based on current scientific information, all requirements necessary for ABM recovery and survival.” Tab 8(BO) section I, p. 4.' Part of the optimal habitat is the federally-designated critical habitat, which is defined for the ABM as that portion of the relevant section of shoreline extending 500 feet inland from the mean high tide line of the Gulf of Mexico. Because it is closest to the shore, the critical habitat consists primarily of primary and secondary dunes. The project will destroy approximately 34.5 acres of optimal habitat. According to the FWS, this constitutes 20 percent of the optimal habitat in the action area (172.2 acres), and 13.3 percent of the optimal habitat remaining in the entire range (260 acres). The FWS candidly acknowledges that it does not know what the effect of the loss of optimal habitat will have on the Alabama beach mouse. Tab 7(EA) 4.4.4; Tab 7(EA) 5.1.1. Despite this lack of information, the FWS nonetheless claims to be able to determine that the loss will not have a significant effect on the ABM; that finding avoids the additional investigation required in preparation ,of an EIS which might provide the needed information. See also 40 C.F.R. § 1502.22 (CEQ regulation providing for means of addressing lack of certainty in context of EIS). The record does not explain how the FWS can overcome the lack of such information and reach a valid finding of insignificance. In the absence of information on what effect the loss of this key habitat type will have on the beach mouse, and particularly in view of the size of the loss, the court finds that plaintiffs have a substantial likelihood of demonstrating that the agency’s FONSI is inadequately supported by the record, and is thus arbitrary and capricious. Though 34.5 acres of optimal habitat are to be destroyed, only approximately 0.35 acres of critical habitat will be affected; the majority of optimal habitat to be lost is escarpment and adjacent scrub. The proposed construction will destroy 1935 linear feet (68%) of the total 2844 linear feet of escarpment in the project area. The FWS acknowledges that it does not know what effect the loss of over 1935 linear feet of escarpment will have on the landscape or what effects might arise from wind-borne sand. The FWS also , notes that the dével-opment will result in fragmentation of the ABM-occupied interior scrub habitat along 1935 feet of escarpment and would affect adjacent scrub areas to the north, and that it will act as a barrier to movement of the ABM. One effect it could predict is that “Moss of the escarpment and the damming effect of the buddings would deflect storm surges toward adjacent interior habitat, making flooding more likely.” Tab 7(EA), section 4.8.2. The agency does not separately address the effect of such flooding on the ABM, particularly on its access to the inland habitat (including the escarpment and adjacent scrub) or use of such areas as refuge from inundation of critical habitat during storm events. In the absence of such information, the court finds that plaintiffs have a substantial likelihood of success on the merits. Additionally, the court questions the sufficiency of the discussion .of the combined effects of all of the barriers to ABM movement created by the project, but does not here decide that issue. Though mentioned, the parties do not discuss the plaintiffs challenges to the use of the habitat model in place of the population-based analysis previously criticized by this court as incomplete. Sierra Club v. Babbitt, supra. Nonetheless, the court notes some concerns. The Service’s estimate of population in the action area was made by extrapolation from a catalogue of the types of habitat in the action area. The effect on that population is estimated based on a similar cataloguing of the types and amounts of habitat lost compared to the baseline. Though the FWS asserted that the population data was not completely divorced from their analysis, the record does not clearly explain their method of determining population from a determination of the amounts and types of habitat affected, particularly in the absence of such basic information as that noted above. 1. Uncertainty and the Need for an EIS: It would seem self-apparent that uncertainty concerning the effects of important aspects of the proposed action on the ABM would preclude a “finding” that the effects of the proposed action on the ABM would not be significant, and thus finding that the additional study which goes with preparation of an Environmental Impact Statement to be unnecessary. The defendants and intervenors assert otherwise, however, claiming that the Service was required to issue the permits unless they found that the action would jeopardize the species, and was required to make that determination on the basis of the best information available. NEPA was designed to prevent uninformed action. “NEPA requires each agency to undertake research needed adequately to expose environmental harms.” Save Our Ecosystems v. Clark, 747 F.2d 1240, 1248 (9th Cir.1984) (citations omitted). Defendant’s argument in this case would turn NEPA on its head, making ignorance into a powerful factor in favor of immediate action where the agency lacks sufficient data to conclusively show not only that proposed action would harm an endangered species, but that the harm would prove to be “significant.” The agency cites regulations which require the agency to issue the permits unless their findings preclude such action, and language limiting the agency’s consideration to the “best scientific and commercial data available.” Title 50, section 17.22(b)(2) of the Code of Federal Regulations, established by the FWS, addresses the ITP application procedure under the Endangered Species Act. It identifies the information which must be considered by the Director, and provides that he or she “shall issue the permit if he or she finds” various elements, including that “the taking will not appreciably reduce the likelihood of the survival and recovery of the species in the wild.” Similarly, 50 C.F.R. § 18.21(b) provides that “the Director shall issue the appropriate permit unless ... (4) The authorization requested potentially threatens a wildlife or plant population .... ” These sections address the Director’s duties under the ESA, rather than under NEPA. The ESA’s “jeopardy” analysis is distinct from the “significant impact” standard of NEPA. See Makua v. Rumsfeld, 168 F.Supp.2d at 1218. Neither of these regulations addresses NEPA’s preliminary determination: whether NEPA requires the agency to prepare an EIS in: order to adequately assess these jeopardy issues under the ESA. As already noted, NEPA’s requirement of an EIS does not apply if the agency determines that the action does not reach the threshold level of a significant environmental impact. NEPA’s purpose is to avoid decisions made without adequate information. Uncertainty is a factor in determining the significance of an action. 40 C.F.R. § 1508.27. 40 C.F.R, § 1502.22 establishes a procedure for reaching a decision where the agency may lack sufficient information: it allows the agency to weigh the importance of the missing information against any barriers and costs to gathering such information. Because this provision is only available in the context of an EIS, it appears that uncertainty over the effects of an action would generally require the preparation of an EIS. Cf. Robertson v. Methow Valley Citizens Council, 490 U.S. 332, 354, 109 S.Ct. 1835, 104 L.Ed.2d 351 (1989)(Though uncertainty existed over data relevant to environmental decision, agency decision not flawed by failure to conduct worst case analysis where agency followed new provisions of § 1502.22). The Service is not required to find a proposed project insignificant in the absence of readily available information to the contrary; rather, it is required to create an EIS for any project which may significantly affect the environment. Under NEPA, it cannot use the lack of existing information as a basis for acting without preparing an EIS. If information sufficient to reach the ultimate determination is not currently available, then the agency shall conduct an EIS, and may weigh whether or not to gather such additional information. C. Conclusion on the Likelihood of Success Based upon the record, the court finds that plaintiffs have demonstrated a substantial likelihood that they will' succeed in demonstrating that the FONSI issued by the FWS is arbitrary and capricious. 1)The court finds that it was arbitrary and capricious to hold that the loss of 20% of optimal habitat in the action area, on top of previous losses of 20.6% of optimal habitat in the action area since the species was listed, would have no significant impact on the species, despite the fact that the species was found to be endangered because of prior losses of habitat. Such a conclusion is at odds with the absolute, cumulative and relative losses of habitat, and is “so implausible that it could not be ascribed to a difference in view or the product of agency expertise.” Motor Vehicle Mfr. Ass’n. v. Stae Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43, 103 S.Ct. 2856, 77 L.Ed.2d 443 (1983). 2) The court finds that the FONSI was arbitrary and capricious on the additional basis that the record does not explain how such a substantial cumulative loss of habitat would not have a significant impact on the ABM. Such an implausible conclusion from the facts found was not explained on the basis of agency expertise or any other fact or analysis. Hill v. Boy, 144 F.3d 1446 (11th Cir.1998). 3) The court further finds that the FONSI was arbitrary and capricious because the agency lacked necessary information upon which to base its finding, including: a) the minimum habitat requirements of the species and/or of a colony of the species; b) the effect of the loss of optimal habitat on the species and/or a colony of the species; c) the effect of the loss of such a substantial portion of escarpment on the terrain and thus on the species. D. Other' Requirements for Preliminary Injunction The court - has found that, on several theories, the plaintiffs have demonstrated a substantial likelihood that they will prevail on the merits. The court thus considers the remaining factors for issuance of a preliminary injunction. 1. Presumptions The parties disagree over the correct standard to be applied in addressing these factors in a NEPA case. With respect to the second element of the preliminary injunction standard, plaintiffs cite Loggerhead Turtle v. County Council of Volusia, Fla, 896 F.Supp. 1170, 1180 (M.D.Fla.1995), for the proposition that, in cases involving threatened harm to an endangered species, there is an irrebuttable presumption that the threatened harm is irreparable. Plaintiffs further assert that the balancing of harm in the third and fourth elements has been definitively determined by Congress. See Sierra Club v. Marsh, 816 F.2d 1376, 1383 (9th Cir.1987); TVA v. Hill, 437 U.S. 153, 98 S.Ct. 2279, 57 L.Ed.2d 117 (1978). The Intervenors disagree, and claim that the general balancing of harm has not been displaced in suits under NEPA and the APA, as in the instant action, as opposed to suits brought pursuant to the ESA. In TVA v. Hill, the Supreme Court discussed the ESA’s statement of policy: One would be hard pressed to find a statutory provision whose terms were any plainer than those-in §§ 7 of the Endangered Species Act. Its very words affirmatively command all federal agencies “to insure that actions authorized, funded, or carried out by them do not jeopardize the continued existence” of an endangered species or “result in the destruction or modification of habitat of such species ....” 16 U.S.C. §§ 1536 (1976 ed.). (Emphasis added.) This language admits of no exception. TVA v. Hill, 437 U.S. at 173, 98 S.Ct. 2279. The Ninth Circuit has held that this legislative command effectively strips federal courts of discretion in determining the balance of hardships between the parties. Sierra Club v. Marsh, 816 F.2d at 1383 (citing TVA v. Hill, at 173, 98 S.Ct. 2279). In Amoco Production Co. v. Village of Cambell, Alaska, 480 U.S. 531, 107 S.Ct. 1396, 94 L.Ed.2d 542 (1987), the Supreme Court held that the statute at issue therein — the Alaskan National Interest Lands Conservation Act (“ANILCA”), 16 U.S.C. § 3120 — did not contain a clear Congressional statement of intent to displace the traditional equitable power of the courts, or the traditional balancing of equities. It further held that injunctive relief in an action brought under ANILCA was not governed by the legislatively mandated presumptions of the ESA merely because the action involved environmental considerations. Th