MEMORANDUM DECISION AND ORDER

WINMILL, Chief Judge.

INTRODUCTION

The Court has before it plaintiffs’ motion for preliminary injunction. The Court heard oral argument on March 29, 2001, and the motion is at issue. The Court finds that the motion should be granted, and explains its reasoning below.

LITIGATION BACKGROUND

Plaintiffs (collectively referred to as ICL) seek to halt a Government program that identifies and kills predators of the sage grouse. Concerned about declining sage grouse populations in southern Idaho, the Idaho Department of Fish and Game (IDFG) approached the Wildlife Service (WS) in 1999 about protecting the sage grouse from predators. In the spring of 1999, the WS and the IDFG conducted a joint artificial nest survey showing a high level of predation by ravens and mammals. This survey did not involve the killing of any the predators.

On the basis of that survey, the WS issued in August, 1999, a “Supplemental EA and Monitoring Report for Predator Management” (referred to as SEA) along with an accompanying FONSI and Decision. The 1999 SEA supplemented a 1996 EA. The 1996 EA discussed predator control generally, but specifically discussed the protection of livestock rather than sage grouse. The 1999 SEA was designed to address the declining sage grouse population. Observing that predation was an important factor in that decline, the 1999 SEA proposed a survey using artificial nests to test the extent of the predation problem. While the 1999 SEA did discuss generally the use of poisoned eggs to reduce avian predators, the SEA was focused on a proposal to determine the extent of predation, not a proposal to kill predators. Thus, the 1999 SEA did not discuss the scope and effect of a program to kill predators. The 1999 SEA concludes with a Final Decision that “predator damage management” should be used to “protect sage grouse.”

Consequently, in the spring of 2000, WS conducted another survey, but this time killed and trapped predators in one area and compared predation results with another area where no killing or trapping took place. For one month, in the Curlew Valley area, the WS placed artificial nests with poison eggs to kill ravens, and trapped foxes, coyotes, badgers, and skunks. In a different area — a control *1100 area — the artificial nests were placed without predator controls. A comparison of the test results showed that predator controls significantly reduced sage grouse nest destruction.

Based on these results, IDFG and WS proposed to expand the program further. As described by the State Director for WS, Mark Collinge, the “proposed research trial” would take place in six designated areas in southern Idaho, each approximately 75 to 100 square miles. In three of those areas, monitoring of the sage grouse population would take place but no predation controls would be instituted. In the other three areas, predation controls would be instituted if early monitoring shows that predation is occurring. Under the proposal, poison eggs would be used to kill avian predators, while “leghold traps, snares, denning, calling and shooting, and aerial hunting might be employed” to control target predators such as coyotes, red foxes, black bears, mountain lions, bobcats, raccoons, badgers, striped skunks, ravens, and magpies. See Declaration of Collinge at ¶ 32.

It is this expanded proposal that ICL seeks to enjoin. ICL asserts that (1) predation control is not an effective means of protecting the sage grouse; (2) the earlier EAs are insufficient under NEPA, and (3) an Environmental Impact Statement (EIS) is required before the project may commence.

ANALYSIS

1. Ripeness

The Government claims this action is not ripe because no killing will occur unless test results from the three areas show that predation is a problem. Because those test results will not be known until sometime after the program has begun, the Government argues that it is premature to stop the program before that point.

“A claim is not ripe for adjudication if it rests upon contingent future events that may not occur as anticipated, or indeed may not occur at all.” Texas v. United States, 523 U.S. 296, 118 S.Ct. 1257, 140 L.Ed.2d 406 (1998). Two prior surveys by WS showed that predation was a problem. In this third survey, the killing and trapping of target predators will begin if the early results of the survey show that predation is a problem. Given the results of the two prior surveys, it is highly likely that the early results will show predation problems, and that the killing and trapping will start very shortly after the survey begins.

Under these circumstances, predator control is only a “contingent future event” in a highly technical sense — in common sense terms, it is inevitable. If ICL must wait until that point to bring its challenge, the Court will be forced to rule on an emergency basis, without the opportunity for a reasoned presentation and analysis. For these reasons, the Court finds that this matter is ripe.

2. Injunctive Relief

Plaintiffs are entitled to an injunction if they demonstrate that they are likely to succeed on the merits of them claims and may suffer irreparable injury, or that serious questions exist on the merits and the balance of hardships tips in their favor. See Self-Realization Fellowship Church v. Ananda, 59 F.3d 902, 913 (9th Cir.1995). The two tests are not separate but represent a sliding scale in which the required probability of success on the merits decreases as the degree of harm increases. Id. “Under any formulation of the test, the plaintiff must demonstrate that there exists a significant threat of irreparable injury.” Oakland Tribune, Inc. v. Chronicle Publishing Co., 762 F.2d 1374, 1376 (9th Cir.1985).

*1101 The harm in this case does not appear to favor either side. The Government has made a colorable claim that the sage grouse is threatened by predators. ICL has made a credible claim that the predator control proposal would endanger other animals. Under these circumstances, ICL cannot show that the harm tips in its favor, and thus must make the highest showing of likelihood that it will suffer irreparable harm.

ICL’s first burden is to show that NEPA applies to the predator control proposal. Given the number of predators being targeted (including coyote, red,fox, black bear, mountain lion, bobcats, raccoons, badgers, striped skunks, and ravens) and the geographical area involved (three 100-square-mile areas where predators will be killed or trapped), it is at least likely that WS’s predator control proposal is a major federal action triggering NEPA scrutiny. See 42 U.S.C. § 4332(2)(C); 40 C.F.R. § 1508.18. 1 If a major federal action is involved, and an agency’s regulations do not categorically require the preparation of an EIS, as here, then the agency must first prepare an (EA) to determine whether the action will have a significant effect on the environment. See 40 C.F.R. § 1501.4.

It is therefore likely that WS must prepare an EA to study its predator control proposal. WS points to two EAs, the 1996 EA and 1999 SEA. However, neither EA examined the specific proposal at issue here, involving the killing and trapping of certain predators in the three 100-square-mile areas. Neither the 1996 EA nor the 1999 SEA discuss, with regard to the three locations, (1) the specific avian predators that might be targeted, (2) any avian killing devices other than eggs poisoned with avicide DRC-1339, (3) any mammalian predators that might be targeted, (4) the types of killing or trapping devices that would be used on mammalian predators, (5) the specific location of any of the artificial nests or predator controls, or (6) the number of animals that might be killed by predation control efforts.

Furthermore, the 1999 Supplemental EA is now close to two years old, and the 1996 EA close to five years old. The age of these EAs and their lack of discussion of the specific proposal at issue here cause the Court concern. ICL has at least a likelihood of success on its claim that WS must prepare a new EA on the specific predator control program at issue here.

WS argues, however, that its predator control program is categorically excluded from NEPA. Pursuant to Council of Environmental Quality (CEQ) regulations, each agency is required to identify categories of actions which do not individually or cumulatively have a significant effect on the human environment. See 40 C.F.R. §§ 1607.3(b)(2)(H) & 1508.4. These actions are classified as “categorical exclusions” for which neither an EA nor an EIS is required. This Court must give an agency’s interpretation of the meaning of its own categorical exclusion substantial deference. See Alaska Center for the Environment v. U.S. Forest Service, 189 F.3d 851 (9th Cir.1999).

Here WS points to two categorical exclusions. First, WS regulations exclude research activities, which are defined as follows:

(i) Activities that are carried out in laboratories, facilities, or other areas designed to eliminate the potential for harmful environmental effects — internal or external- — and to provide for lawful waste disposal.

*1102 (ii) Examples of this category of actions include: (A) the development and/or production (including formulation, repackaging, movement, and distribution) of previously approved and/or licensed program materials, devices, reagents, and biologies; (B) Research, testing and development of animal repellents; and' (C) Development and production of sterile insects.

7 C.F.R. § 372.5(c)(2).

The predation control proposal is not to be carried out in a laboratory or facility, although it might be in an “area designed to eliminate the potential for harmful environmental effects.” Perhaps the 100-square-mile regions were chosen to eliminate the harmful environmental effects of killing and trapping predators. In that sense, the proposal falls within the regulation’s language. However, WS is not entitled to that presumption because the record does not support it — WS points to nothing in the record showing why the WS picked those three regions. This failure is part of the larger problem that WS has no contemporaneous documentation explaining its decision to proceed with the proposal at issue here. See Edmonds Institute v. Babbitt, 42 F.Supp.2d 1, 19 (D.D.C.1999) (denying a claim of categorical exclusion where the agency “provided no evidence whatsoever of such a determination [that the exclusion applied] being made before the [agency action] was finalized.”).

In addition, the examples listed in subsection (ii) do not come close to describing a 100-square-mile region, so there is some question whether this proposal fits into the regulation’s definition of “research activities.” Thus, even giving substantial deference to WS’s own interpretation of its categorical exclusion regulation, the Court must find that ICL is likely to succeed on its claim that this predation control proposal is not a “research activity” categorically excluded from NEPA’s reach by 7 C.F.R. § 372.5(c)(2).

WS asserts a second categorical exclusion covering “routine measures.” See 7 C.F.R. § 372.5(c)(1). The routine measures listed in the regulation include “seizures, quarantines, removals, sanitiging, inoculations, control, and monitoring.” Killing is not included. It appears, however, that the listed routine measures may be euphemisms for killing, as the regulation allows the use of “potentially hazardous devices,” presumably to kill predators, if four criteria are met:

(A) The use is localized or contained in areas where humans are not likely to be exposed, and is limited in terms of quantity, i.e., individualized dosages and remedies; (B) the use will not cause contaminants to enter water bodies, including water bodies; (C) the use does not adversely affect any federally protected species or critical habitat; and (D) the use does not cause bioaccoumulation.

WS has not pointed to anything in the record showing that it considered these four criteria in connection with the predator control proposal at issue. WS argues that it had no duty “to prepare any additional paperwork, other than the administrative record that clearly shows WS finds individual wildlife damage management programs are categorically excluded, such as the sage grouse program.” Response Brief of WS at 11. WS did state in its 1996 EA that “[n]ormally ... individual wildlife damage management actions are categorically excluded.” But this general statement says nothing about whether the particular predation control proposal at issue here falls within the specific terms of a categorical exclusion established by regulation. The regulation on “routine measures” clearly requires four findings. WS points to nothing in the record showing that those four findings were made as to *1103 the particular proposal at issue here. Once again, the lack of an agency record makes WS’s argument unpersuasive. See Edmonds, 42 F.Supp.2d at 19 (D.D.C.1999).

While WS cites to CEQ regulations discouraging additional paperwork, these regulations do not absolve agencies of their duty to record decisions reviewable by courts: “An agency satisfies NEPA if it applies its categorical exclusions and determines that neither an EA nor an EIS is required, so long as the application of the exclusions to the facts of the particular action is not arbitrary and capricious.” Bicycle Trails Council of Marin v. Babbitt, 82 F.3d 1445, 1456, n. 5 (9th Cir.1996) (emphasis added). This Court cannot determine whether the agency’s application of a categorical exclusion to a particular project is arbitrary and capricious unless there is something for the Court to review. While WS submits the Declaration of Mark Collinge to provide the agency’s rationale for proceeding with this proposal, “the Court needs access to the information actually before the agency, not the recollections of a participant.” IWP v. Hahn, Civ. No. 97-0519-S-BLW at p. 8 (March 31, 1999); see also, Animal Defense Council v. Hodel, 840 F.2d 1432, 1436-38 (9th Cir.1988), modified, 867 F.2d 1244 (9th Cir.1989). 2

For these reasons, the Court finds that ICL is likely to succeed in arguing (1) that no categorical exclusion applies to this case, and (2) that WS was required to prepare a new EA on the predator control proposal at issue here. The Court will therefore issue the injunctive relief requested by ICL.

ORDER

In accordance with the Memorandum Decision set out above,

NOW THEREFORE IT IS HEREBY ORDERED, that the motion for preliminary injunction (docket no. 3, part 2) is hereby GRANTED, and that the predator control proposal discussed in the Declaration of Mark Collinge that was previously due to commence on March 15, 2001, as stated in that Declaration, is hereby ENJOINED, and that this preliminary injunction is binding on all those set forth in Federal Rule of Civil Procedure 65(d).

IT IS FURTHER ORDERED, that plaintiffs post a bond of $1,000 with the Clerk of the Court pursuant to Federal Rule of Civil Procedure 65(c).

IT IS FURTHER ORDERED, that if the parties desire further hearings on this matter, evidentiary or otherwise, that they work out an agreed-upon time between themselves and with the Court’s Clerk La-Donna Garcia (208-334-9021). This preliminary injunction shall remain in effect until such time as the Court can hear the parties.

IT IS FURTHER ORDERED, that the motion for temporary restraining order (docket no. 3, part 1) is hereby DEEMED MOOT due to the granting of the motion for preliminary injunction.

IT IS FURTHER ORDERED, that the motions to exceed page limit (docket nos. 4 & 9) are hereby GRANTED, and the Clerk of the Court is hereby directed to file the briefing associated with these motions.

. The same analysis would not apply to the first two surveys, which were more limited in geographic scope, and most likely would not constitute major federal actions under NEPA.

. It is also important that Collinge's Declaration does not refer specifically to any of the categorical exclusion regulations. Thus, even if the Declaration could be considered as an explanation of the agency’s rationale concerning categorical exclusions, it sheds no light on that issue.