FINDINGS OF FACT AND CONCLUSIONS OF LAW REGARDING PROBABILITY OF SUCCESS AND IRREPARABLE INJURY; REQUEST FOR SUPPLEMENTAL BRIEFING CONCERNING REMEDY; ORDER DENYING DEFENDANTS’ OBJECTION TO EXTRA-RECORD MATERIAL

INTRODUCTION

On motion for preliminary injunction, this order finds that plaintiffs have shown a probability that the National Park Service violated its own regulations requiring notice and opportunity for public comment before implementing a closure of certain park lands, finds that plaintiffs will suffer irreparable injury, and requests further briefing as to a provisional remedy.

STATEMENT

1. The Closure Regulation of the National Park Service

The National Park Service’s regulations require notice-and-comment rulemaking procedures before a closure of a park area that is of a “highly controversial nature” or that will result in “a significant alteration in the public use pattern of the park area”:

36 C.F.R. 1.5(b). Plaintiffs contend that the National Park Service violated this regulation by closing a portion of Fort Funston without first publishing the proposed closure in the Federal Register and allowing comment.

2. Fort Funston

Fort Funston is a multi-use recreational area on the coastal bluffs in southwest San Francisco overseen by the Golden Gate National Recreation Area (GGNRA), a unit of the National Park Service. Fort Fun-ston is one of several separate public lands in the Bay Area that are superintended by the GGNRA. Others are Fort Mason, Fort Baker, the Presidio, Lands End, Alcatraz Island, Muir Woods, and Fort Mi-ley. All told, the GGNRA encompasses approximately 76,500 acres of land and water. Fort Funston itself encompasses approximately 222 acres. Fort Funston became a part of the GGNRA in 1974, when San Francisco transferred its ownership and control to the United States. In the GGNRA’s statement of purpose, Congress acknowledged both maintaining recreational open space within an urban area and preserving that area from uses that would destroy its natural character:

16 U.S.C. 460bb. Although the National Park Service generally requires that pets be on-leash in national parks, the Park Service allows dog owners to walk their . dogs off-leash at Fort Funston.

3. Plaintiff Dog Walkers

Perhaps because they may run off-leash on voice command, dogs and their owners frequent Fort Funston, although they are not its only visitors. Ft. Funston Dog Walkers was formed in 1996 so that the members could get to know one another and organize regular park clean-ups. The group has approximately 600 members. The group meets once a month on a Saturday morning. Clean-up supplies are provided to members who attend. Annual dues are ten dollars. SFDOG was formed in 1976 as a consolidated voice for dog owners in the San Francisco Bay Area. It has approximately 650 members. Both organizations are plaintiffs. Also serving as plaintiffs are four individuals. Plaintiffs Linda McKay and Lindsay Kefauver are members of Ft. Funston Dog Walkers and frequently walk their dogs in Fort Fun-ston. Ms. McKay is one of the organizers of the Fort Funston Dog Walkers. Plaintiff Florence Sarrett belongs to SFDOG and has frequently visited Fort Funston for more than thirty years. Plaintiff Marion Cardinall is a frequent visitor to Fort Funston.

4. The Bank Swallows

The closure at issue is intended to protect the bank swallow, a species listed as threatened by the California Fish and Game Commission. The Fort Funston colony of bank swallows winters in South America and nests in the cliffs at Fort Funston from March or April until August. They burrow small holes horizontally into eroding cliffs, just large enough for a nest. The Fort Funston colony is one of two active colonies remaining on the west coast. Their nesting site used to be at the north end of the Fort Funston cliffs. In 1997 and 1998, however, the colony moved somewhat southward after the El Nino and La Nina winters eroded much of those cliffs. Their new nesting area is still in Fort Funston but is farther south; therein lie the seéds of the controversy.

5. The 1995 Closure

To protect the original nesting area of the bank swallows and for safety reasons, the National Park Service closed two sites in Fort Funston in 1995. One of the two sites covered several acres at the most northern bluffs between the beach and the Coastal Trial. The purpose of this closure was to protect the bank-swallow nesting colonies, located then on the sheer, vertical faces of the bluffs. This closure was unpopular with plaintiffs.

At the time of the' closures, Brian O’Neill, the General Superintendent of the GGNRA, represented to Richard Avanzi-no, president of the San Francisco Society for the Prevention of Cruelty for Animals, that the northern bank-swallow closure would not be expanded southward:

(A.R. US06438-39, ltr. from Gen. Superintendent O’Neill to R. Avanzino, Mar. 15, 1995). At a large meeting in 1995, Ranger Jim Milestone represented to approximately 250 members of the plaintiff organizations that there would be no further closures (Undated Declaration of Linda McKay filed Apr. 3, 2000, ¶ 14).

6.The 2000 Closure

Contrary to its statements that the bank-swallow closure would not be expanded, the National Park Service decided in 1999 to do just that. This decision was provoked by the southward move of the bank-swallow nesting. The new closure, the one now at issue, encompasses another ten acres south of the 1995 bluff closure. Fort Funston, as stated, comprises approximately 222 acres along the Pacific coast. The Coastal Trail runs roughly parallel to the beach on top of the bluffs in the northern part of the park. (See map appended hereto.) It is the main artery for hikers, joggers and dog walkers. The northern terminus of the Coastal Trail runs through the 1995 closure and curves westward and downward to the beach. About half of the new closure is “seasonal,” meaning it is closed only while the bank swallows are in residence, and about half is permanent, meaning year-round. The new permanent closure borders and extends south of the 1995 permanent closure segment. The seasonal closure borders and extends south of the new permanent closure. The new closures are, as stated, about ten acres.

The new closures comprise land between the Coastal Trail and the cliffs. This land includes undulating terrain and dunes on which dogs run, children play and adults hike. Children enjoy sliding down a steep dune called Joey’s Hill. The high points offer excellent elevated scenic views of the surf. Within the closure is the middle of only three gaps providing access to the beach, the steep cliffs otherwise making the descent hazardous. This middle access is particularly important at high tide when the entire beach is awash, providing an avenue to safety midway between the other two gaps.

Plaintiffs oppose the new closure for several reasons. First, they see it as a violation of the representation that there would be no further closures. Second, they contend that the area being closed is especially pleasant in the views and activities it sustains. Third, they are concerned because the closure voids the middle of three draws to and from the beach.

7. The Decision to Implement the 2000 Closure

In 1998, the GGNRA started organizing to close off the bluff area south of the 1995 permanent closure. On February 2, 1999, the National Park Service generated a project statement. It recommended activities such as nest monitoring and educating the visiting public about the bank swallows and their nests (A.R.US00262-66). Importantly, the project statement also recommended restrictions on park use, including erecting a fence to protect the swallows:

(Ibid.). In addition the statement recommended revegetation along commonly used paths in the off-trail area:

(Ibid.). These items show that the National Park Service was contemplating fencing and revegetation as early as February 1999. Funds were received through the Public Land Corps grant program to implement the first phase of the bank-swallow habitat protection project in 1999. The funds were applied to purchasing fencing material. Ranger Sharon Farrell, a National Park Service plant ecologist, was the contact person for the ■ project. Ranger Farrell was a central actor and proponent of the closure.

The National Park Service did not immediately inform the public of its progressing plan. On July 31, 1999, Ranger Mary A. Petrilli, an interpretive ranger at Fort Funston, sent an email to Ranger Chris Powell, a National Park Service public relations staff member, stating that the Bank Swallow Site Extension had been approved through Project Review and that funding had been secured for the new fence line. The email was copied to six other National Park Service staff mem bers. It warned all recipients to be “discreet” with the information in the email and cautioned that “we do not want this to blow up in our faces”:

(A.R. US06293, email from M. Petrilli to C. Powell, July 31, 1999) (emphasis added). In September 1999, the National Park Service was hoping to put up the fences in October before the rainy season (A.R. US06225, email from S. Farrell to T. Fort-mann, et al., Sept. 1,1999).

On October 27, 1999, Ranger Farrell wrote the following email to Ron Schlorff of the California Department of Fish and Game soliciting official DFG input that recreation at Fort Funston should be curtailed to protect the bank swallows:

(A.R. US06939, email from S. Farrell to R. Schlorff, Oct. 27, 1999). A few days later, on November 1, 1999, the park staff briefed the San Francisco Committee of the GGNRA Commission about the need to protect the bank-swallow habitat through closure of a portion of the bluffs (A.R.US04641-44).

8. Reaction of the Dog Walkers

Although the National Park Service had obtained funding for the closure in the summer of 1999 and had originally intended to go forward with the project in the fall of 1999, it was not until December 3, 1999, that the National Park Service informed the Ft. Funston Dog Walkers of their plan, albeit in an indirect way. The topic came up during a so-called walk-through. Walk-through is the term used by the dog walkers and the National Park Service to refer to on-site discussions about park issues. The dog-owner representatives did not know that the walk-through would include a discussion of closures. Instead, they expected to discuss the usual issues such as a dog water fountain on the Coastal Trail, boxes for plastic bags to pick up after dogs, and the locations of several trash cans. During the walk-through, Ranger Chris Powell volunteered that there was a possibility that a further closure south of the 1995 closure might be made for bank-swallow protection and revegetation (A.R.US06200).

On December 21, 1999, Ranger Roger Scott, a public relations staff member, emailed Ms. McKay a set of minutes purportedly summarizing the walk-through. As stated, Ms. McKay is an organizer of the Ft. Funston Dog Walkers. The minutes stated that “Chris Powell began discussion with the group the very likely possibility that a six-acre cliff area might be restricted from access because of a Bank Swallow relocation and ongoing revegetation project” (A.R.US06200). This statement was more definite than had been the casual and hypothetical conversation during the walk-through. Ms. McKay was surprised by the minutes. She had not understood that the National Park Service was planning to close six acres of property between the beach and the Coastal Trail. She so informed Ranger Kevin Turner, the head ranger and an interpretive specialist, who then sent the following email to a fellow ranger:

(A.R. US06197, email from R. Weideman to M. Scott, C. Powell, R. Scott, and Y. Ruan, Dec. 28, 1999, forwarding email from K. Turner to R. Weideman, Dec. 27, 1999). The day after her meeting with Ranger Turner, Ms. McKay wrote the following email to Rangers Fortmann and Scott regarding problems with the minutes from the December 3 walk-through:

(A.R. USÓ6197, email from L. McKay to T. Fortmann and R. Scott, Dec. 27, 1999) (emphasis added). Later, the actual size of the closure proved to be almost ten acres, not five or six acres as the minutes suggested.

The walk-through minutes mentioned that closure was to be on the agenda at the January Advisory Committee meeting; however, the dog walkers contend that they did not get notice of the meeting until January 14, 2000, four days before the meeting (A.R. US060200, Dec. 3 Committee Meeting Minutes). When the dog walkers did get a copy of the agenda, the closure was not listed as a public-discussion item. On January 15, 2000, Ms. McKay sent an email entitled “Fort Fun-ston Crisis” to members .of Ft. Funston Dogwalkers, informing them of the meeting and the bank-swallow closure issue. The email called on members to attend the meeting: “We need bodies there, even though we may not be able to comment— but have comments ready if we can!” (email from L. McKay to Ft. Funston Dog Walkers, Jan. 15, 2000).

The Advisory Committee Meeting occurred on January 18, 2000 (A.R.US06150). Ranger Farrell, a National Park Service plant biologist, and Ranger Hatch, a National Park Service wildlife specialist, gave a slide presentation to illustrate the four objectives of the project: to protect the threatened bank-swallow colony at Fort Funston; to stabilize eroding inland dunes and increase biological diversity by restoring the coastal dune habitat; to increase public safety by reducing visitor exposure to imminent threats; and to protect geologic and historic coastal features from human-induced erosion (A.R.US06151). The meeting was open to the public. Dog walkers, as well as other members of the public, attended and were allowed to speak. Of the thirteen members of the public who spoke, seven were dog owners (ibid.).

During this period, as before, officials were very aware of the possibility that the dog walkers would wish to be heard and sought to preclude or minimize such input. In this vein, Assistant Superintendent Mary Gibson Scott wrote the following in an email on January 24 with the subject line “Ft. Funston dogwalkers and attachment”:

(A.R. US06268, email from M. Scott to D. Mannel, J. Scheumann, Y. Rúan, and T. Thomas and copied to R. Scott, T. Fort-mann, M. Bartling, dated Jan. 24, 2000) (emphasis added). The email was in response to an email from Ranger Roger Scott summarizing a recent meeting with the Ft. Funston Dog Walkers.

Once the dog walkers knew of the planned closure, controversy loomed large. The National Park Service knew it was contentious. Significantly, Ranger Farrell wrote the following to Janet Gomes of the San Francisco Conservation Corps:

(A.R. US04208-09, email from S. Farrell to J. Gomes, Jan. 26, 2000, copied to M. Pe-trilli) (emphasis added).

9. The Demand For Public Input and Rebuff

By early February, the dog owners requested a written record of the public involvement in the National Park Service’s decision to close more of the coastal bluffs. These requests triggered a search within the National Park Service for documentation of any public hearings, as reflected by the following series of emails between Michael Feinstein, Ranger Farrell, and Ranger Fortmann:

(A.R. US06261-62, email from M. Fein-stein to D. Hatch and S. Farrell, Feb. 10, 2000). In response to Michael Feinstein’s email, Ranger Farrell replied that she had not referred to public meetings in. 1996:

(A.R. US06261, email from S. Farrell to M. Feinstein, Feb. 14, 2000).

After Michael Feinstein reviewed Ranger Farrell’s email, he forwarded it to Ranger Fortmann, expressing doubt that any public hearing ever took place:

(Ibid., email from M. Feinstein to T. Fort-mann, Feb. 14, 2000). Ranger Fortmann then forwarded the foregoing emails to Ranger Roger Scott, asking what Ms. McKay had requested:

(Ibid., email from T. Fortmann to R. Scott, Feb. 14, 2000). Ranger Roger Scott then wrote an email to Assistant Superintendent Mary Gibson Scott, Ranger Fort-mann, and Michael Feinstein, concluding that there had been no formal notice or hearing before the 1995 closure. Ranger Roger Scott recommended that the National Park Service take the position that no notice or hearing had been necessary as follows:

(A.R. US06260, email from R. Scott to M. Scott, T. Fortmann, and M. Feinstein, Feb. 15, 2000) (emphasis added). According to Ranger Roger Scott’s email, outreach staff would be on hand one week prior to commencement of the fences and all through the fence building. The outreach was not for the purpose of receiving input on the closure itself. That was a fait accompli The outreach was a public-relations campaign to sell the closure and to create the appearance that the National Park Service wanted the public’s input. General Superintendent Brian O’Neill and Assistant Superintendent Mary Gibson Scott met with representatives of the dog-walking community on February 17, 2000, to discuss their concerns. Fencing began a few days later, as planned, on February 22, 2000.

On February 25, 2000, Lydia Boesch, counsel for plaintiffs, faxed a letter to Ranger Fortmann stating that the recreational users of Fort Funston were considering applying for a temporary restraining order to prevent the closure. The letter requested the results of all public use surveys completed in the preceding three years at Fort Funston. Ranger Roger Scott is quoted in a March 1, 2000, newspaper article as saying “[w]e knew this was not going to be a popular thing” (A.R. US06113, Marianne Constantinou, “Dog Owners Snarl At Fences,” S.F. Exam., Mar. 1, 2000 at A-1, A-8). Ranger Roger Scott admitted in his deposition that he made this statement.

On March 3, 2000, after plaintiffs expressly threatened this lawsuit, Assistant Superintendent Mary Gibson Scott wrote a memorandum on which the subject line read “Justification for Closure of Bank Swallow Habitat and Stabilization/Revege-tation of Eroding Dunes — Fort Funston” (A.R. US06615-23, mem., Mar. 3, 2000, signed by M. Scott, initialed by B. O’Neill). The memorandum began by citing to the requirements of Chapter 36, Sections 1.5(c) and 1.7 of the Code of Federal Regulations: “[t]his memorandum fulfills the requirement of 36 CFR section 1.5, closures and public use limits for the above referenced actions (A.R.US06615). It also documents the public notice required under 36 CFR section 1.7 in support of the closure” (ibid.). The memorandum “serve[d] as a determination that closure of the subject area [was] necessary for the protection of natural resources and public safety, and implementation of management policies, and no less restrictive measures would suffice” (id. at US06617). The National Park Service never prepared any memorandum analyzing whether the closure would be controversial or highly controversial or whether the closure would significantly alter the public-use pattern of Fort Funston.

10. Procedural History

The fences were built between February 22 and mid-March, 2000. On March 13, 2000, plaintiffs simultaneously filed a complaint for injunctive relief and applied for a temporary restraining order preventing the National Park Service from closing off the designated permanent and seasonal closure areas. On March 14, 2000, the Court held a hearing on the motion for a temporary restraining order. The Court granted plaintiffs’ application in limited part, allowing the permanent closure to remain closed, but enjoining the National Park Service from closing the seasonal area barring an emergency under 36 C.F.R. 1.5(b). On April 11, the Court visited Fort Funston, accompanied by the parties and their counsel. On April 12/ the National Park Service reported to the Court that the bank swallows had returned. The National Park Service declared an emergency and closed off the seasonal area. On April 14, the Court held a hearing on plaintiffs' motion for a preliminary injunction, which this order now addresses.

ANALYSIS

“To obtain a preliminary injunction, the moving party must show either (1) a combination of probable success on-the merits and the possibility of irreparable injury, or (2) that serious questions are raised and the. balance of hardships tips sharply in its favor.” Rucker v. Davis, 203 F.3d 627, 634 (9th Cir.2000) (internal citation and quotation marks omitted). “These are not two tests, but rather the. opposite ends of a single continuum in which the required showing of harm varies inversely with the required showing of meritoriousness.” Republic of the Philippines v. Marcos, 862 F.2d 1355, 1362 (9th Cir.1988) (internal citation and quotation marks omitted).

A strong showing of entitlement to a preliminary injunction is required where the moving party seeks to enjoin governmental action taken in the public interest pursuant to a statutory or regulatory scheme. In such cases, the moving party • must establish both irreparable injury and a probability of success on the. merits. NAACP, Inc. v. Town of East Haven, 70 F.3d 219, 223 (2d Cir.1995). But the lower preliminary injunction standard may be applied where there are public interest concerns on both sides. Time Warner Cable of New York City v. Bloomberg L.P., 118 F.3d 917, 923 (2d Cir.1997). As seen below, the,outcome is the same under either the higher or lower standard in this case.

1. Likelihood of Success on the Merits and/or Serious Questions

A. Standard of Review

Under the Administrative Procedure Act, a reviewing court must hold unlawful and set aside an agency action found to be “without observance of procedure required by law.” 5 U.S.C. 706. The issue in this case is whether defendants violated the National Park Service regulation requiring publication in the Federal Register, absent an emergency, for a “highly controversial” closure or a closure that substantially alters public use patterns. As stated, the relevant regulation provides as follows:

36 C.F.R. 1.5(b). The National Park Service did not publish the 2000 Fort Funston closure as rulemaking in the Federal Register. Thus, if the closure was either “highly controversial” or involved a “significant alteration,” the regulation was violated.

The government argues that the National Park Service made an implicit decision, as evidenced by the lack of a published proposed rule, that the closure was neither of a “highly controversial nature” nor a “significant alteration in the public use pattern.” That implicit decision, the government argues, is entitled to deference. According to the government, plaintiffs can prevail only if this implicit decision was arbitrary and capricious based on the administrative record. The Court disagrees.

While an agency’s interpretation of its own regulation is normally given substantial deference by courts, post-hoc rationalizations are entitled to little or no deference. Motor Vehicle Mfrs. Ass’n. of U.S., Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 50 , 103 S.Ct. 2856 , 77 L.Ed.2d 443 (1983) (“It is well-established that an agency’s action must be upheld, if at all, on the basis articulated by the agency itself.”). .In this case, there was no contemporaneous interpretation whether the closure was “highly controversial” or a “substantial alteration of the public use pattern” before the National Park Service made its decision to put up the fences. Before the fencing began and before plaintiffs sent a letter about seeking an injunction, the National Park Service did not prepare a written determination explaining that the closure was neither “highly controversial” nor a “significant alteration in the public use pattern.”

In addition, the administrative record is replete with evidence that the National Park Service was aware the closure would be highly controversial. As far back as 1995, the National Park Service knew that such closures were strongly opposed by the dog walkers. In 1999, Mary Petrilli cautioned fellow rangers “we do not want this to blow up in our faces” and “[everyone copied on this should be very discreet with this information, PLEASE” (A.R. US06293, email from M. Petrilli to C. Powell, July 31, 1999). Assistant Superintendent Mary Gibson Scott indicated that any meetings with the dog walkers should be kept small, so as not to facilitate their further organization by providing them a forum (A.R. US06268, email from M. Scott to D. Mannel, J. Scheumann, Y. Rúan, and T. Thomas and copied to R. Scott, T. Fort-mann, M. Bartling, dated Jan. 24, 2000). In organizing the fence-building, Ranger Farrell referred to the closure as “perhaps one of the more contentious” projects, predicted that the closure would be “much to some of the dog walkers dismay,” characterized the closure as being “in the top 10% of the park’s most visible projects,” and proposed to pay the balance in full if the crew finished earlier than the contract called for because “the project is that political” (A.R. US04208-09, email from S. Farrell to J. Gomes, Jan. 26, 2000, copied to M. Petrilli). After the fences started to go up, Roger Scott admitted to a reporter “[w]e knew this was not going to be a popular thing” (A.R. US06113, Marianne Constantinou, “Dog Owners Snarl At Fences,” S.F. Exam., Mar. 1, 2000 at A-l, A-8) (quoting Roger Scott). Ranger Scott admitted under oath that he made this statement. The National Park Service did not solicit the views of the dog walkers during the decision-making process. Instead, the National Park Service solicited the views of only select groups in favor of the closure, like the California Department of Fish and Wildlife.

Now, after the fact, after litigation has commenced, Assistant Superintendent Mary Gibson Scott has submitted a declaration stating that she determined at the time of the justification memorandum that the closure would be neither highly controversial nor a significant alteration in the public use of the land. This declaration is, of course, outside the administrative record. Regardless, the declaration is of limited value. First, it is not contemporaneous, and likely not as reliable as a document drafted at the time the decision was made. Second, even at the time Assistant Superintendent Scott made the Section 1.5(b) decisions regarding “highly controversial” and “significant alteration,” plaintiffs had already threatened litigation, so the undocumented analysis was made in contemplation of litigation. Third, Assistant Superintendent Scott’s declaration regarding her interactions with the dog walkers is at odds with the contemporaneous evidence. In her declaration, she states that she believed the dog walkers had no on-going concerns about the closure:

(3d Scott Decl., ¶ 12). This statement is at odds with several parts of the record. First, less than a month before, Assistant Superintendent Scott instructed National Park Service staff not to encourage communication with the dog walkers:

(A.R. US06268, email from M. Scott to D. Mannel, J. Scheumann, Y. Rúan, and T. Thomas and copied to R. Scott, T. Fort-mann, M. Bartling, dated Jan. 24, 2000). In addition, between the February 17 meeting and Assistant Superintendent Scott’s March 3 justification memorandum, the dog walkers were actively-opposing the closure. On February 25, counsel for plaintiffs faxed a letter to Ranger Fortmahn stating that the recreational users of Fort Funston were considering applying for a temporary restraining order to prevent the closure. On March 1, a front-page story in the San Francisco Examiner quoted angry comments from dog walkers and noted that the dog walkers were going to seek an injunction. In that same article, Ranger Roger Scott admitted that “[w]e knew this was not going to be a popular thing,” a statement he acknowledged under oath.

The letters appended to Assistant Superintendent Mary Gibson Scott’s declaration as Exhibits A and B, while pleasant in tone, did not suggest that the dog walkers were satisfied with the closure. In Exhibit A, Lindsay Kefauver thanked Assistant Superintendent Scott for meeting with the dog walkers and expressed, hope that the National Park Service and the dog walkers would continue to talk about preserving off-leash privileges at Fort Funston:

(3d Scott Deck, ¶ 12, Exh. A, ltr. from L. Kefauver to M. Scott, Feb. 22, 2000). In Exhibit B, Anne Farrow expressed concern that the Ft. Funston Dog Walkers be a part of decisions concerning the park:

(3d Scott Deck, ¶ 12, Exh. B, ltr. from A. Farrow to T. Fortmann, Feb. 22, 2000). Both letters reminded that there would be a Ft. Funston Dog Walkers meeting on February 29, and requested that a GGNRA representative be present.

The National Park Service regulations do not require a contemporaneous written determination on “highly controversial” or “substantial alteration in the public use pattern.” See 36 C.F.R. 1.5(b). The Court does not herein mean to suggest that such a requirement exists. Where, however, the administrative record shows a heated controversy and a change in the public-use pattern, an after-the-fact declaration outside the administrative record claiming an official thought about (but did not document) whether publication was required is not persuasive, particularly when it appears from the declaration that the decision was made after the threat of litigation.

B. Evidence Outside the Administrative Record

Defendants have objected to plaintiffs’ submission of evidence outside the administrative record. While defendants concede that such evidence may be taken into consideration in connection with plaintiffs’ claim of irreparable harm, they argue that such evidence should not be considered in the Court’s evaluation of plaintiffs’ likelihood of success on the merits. Defendants are correct that a court’s review of an agency’s decision is usually restricted to the evidence in the administrative record. 5 U.S.C. 701-706. In this context, the rule is problematic, given the National Park Service’s studied solicitation of one-sided input and its “discreet” avoidance of the dog walkers. Defendants would have the Court determine, for example, whether the closure was “highly controversial” based on a deck stacked against the dog walkers. That having been said, the fact remains that most of the evidence relied on herein is, in fact, contained in the administrative record and the Court would reach the same result based on the administrative record alone. That record shows the lengths to which the closure architects went in suppressing input. Because the administrative record is skewed, however, the Court has considered extrinsic evidence.

C. Highly Controversial Nature

Only one decision even refers to the “highly controversial nature” phrase in Section 1.5(b). In Henderson v. Stanton, 76 F.Supp.2d 10 (D.D.C.1999), the plaintiffs challenged a National Park Service regulation barring sales of message-bearing T-shirts on federal parkland in National Capital Region. One of the plaintiffs’ arguments was that the definition of “sales” was subject to rulemaking because it was “of a highly controversial nature.” Id. at 16 n. 3. The court rejected this argument on the ground that the definition of “sales” had appeared in the preamble to a final rule that had already been published for notice and comment. Ibid. The court held that it need not be submitted again and did not construe “highly controversial.”

Important guidance comes from the history of the regulation itself. The “highly controversial nature” standard was not included in the regulation as originally proposed. 47 Fed.Reg. 11612. It was inserted in the final rule to address the concerns of commenters that park superintendents could be “arbitrary and excessive in implementing closures.” 48 Fed.Reg. 30254. The commentary to the final publication of 36 C.F.R. 1.5, however, emphasized that public comment and notice were not meant to apply to closures made to achieve routine resource management objectives:

48 Fed.Reg. 30,252, 30,261-62 (June 30, 1983). The present case does not involve construction or facility maintenance. Only a small portion of the closed land is being rehabilitated. The closure is aimed at preserving the viability of the park ecosystem insofar as bank swallows and vegetation are concerned, but the closure cannot be characterized as a “routine practice.”

The' phrase “highly controversial” arises under regulations under the National Environmental Protection Act (NEPA), codified at 42 U.S.C. 4332. Under the NEPA, the government must perform an “intensity analysis” in determining whether an environmental impact statement is required. One criterion for an intensity analysis is the “degree to which the effects on the quality of the human environment are likely to be highly controversial.” 40 C.F.R. 1508.27(b)(4). In this context, the term “highly controversial” refers to instances in which “a substantial dispute exists as to its size, nature, or effect of the major federal action rather than to the existence of opposition to a use.” Foundation for N. Am. Wild Sheep v. United States Dept. of Agric., 681 F.2d 1172 , 1182 (9th Cir.1982) (internal quotation and emphasis omitted).

Opposition to a project does not necessarily require an environmental impact statement. See, e.g., West Houston Air Comm. v. F.A.A., 784 F.2d 702, 705 (5th Cir.1986). “The existence of a disagreement as to whether an EIS should be commissioned is not by itself grounds for a court to require an EIS.” Roanoke River Basin Assoc. v. Hudson, 940 F.2d 58, 64 (4th Cir.1991), cert. denied, 502 U.S. 1092 , 112 S.Ct. 1164 , 117 L.Ed.2d 411 (1992). “Otherwise, opposition, and not the reasoned analysis set forth in an environmental assessment, would determine whether an environmental impact statement would have to be prepared. The outcome would be governed by a ‘heckler’s veto.’ ” North Carolina v. Federal Aviation Admin., 957 F.2d 1125 , 1133-1134 (4th Cir.1992) (citations omitted).

A substantial difference exists, however, between the NEPA and the regulation at issue. Under the NEPA, the ultimate issue is whether the agency action will have a significant enough impact that the scientific and engineering study necessary for an EIS should be required. Under the regulation at issue, the ultimate issue is whether the public should be allowed their say before a closure. In the latter context, the more controversial a proposal in the classic sense of strongly-divided public opinion, the more appropriate is an opportunity for public input, so that the decision-maker has the benefit of all views and advice. The input may not stop a project, but it may revise and improve it. Here, for example, plaintiffs claim to support protection of the cliffs and swallows but question the need to close areas well away from the cliffs, wish to submit suggestions for re-routing the fences, and want to keep the middle access to the beach open.

If the National Park Service had any internal bulletin or field guidance interpreting “highly controversial nature,” the Court would defer to it. See Overnight Motor Transp. Co. v. Missel, 316 U.S. 572, 580-581 , 62 S.Ct. 1216 , 86 L.Ed. 1682 (1942) (“While the interpretative bulletins are not issued as regulations under statutory authority, they do carry persuasiveness as an expression of the view of those 'experienced in the administration of the Act and acting with the advice of a staff specializing in its interpretation and application.”). The National Park Service, however, has provided only a post-hoc interpretation (outside the administrative record) generated after plaintiffs threatened to sue. The Court is left to define “highly controversial.”

To begin with, the Court agrees with the government that newspaper coverage of the closure, after the decision was made, is not proper evidence that the closure was “highly controversial.” Whether the closure was “highly controversial” must be determined based on the evidence available to the National Park Service at the time the decision was made. After-the-fact press coverage could not have been considered by the agency and is sub ject to manipulation. Nevertheless, the Court finds that the closure was of a “highly controversial nature” based on the following.

First, the administrative record is replete with recognition that the closures had been and would be contentious. The dog walkers were known to be a significant and longstanding user of Fort Funston. The previous Fort Funston closure had been controversial, so much so that in 1995, the National Park Service promised the dog walkers that no further bank-swallow closure would be made. Specifically, Brian O’Neill, the General Superintendent of the GGNRA wrote:

(A.R. US06438-39, ltr. from Gen. Superintendent O’Neill to R. Avanzino, Mar. 15, 1995) (emphasis added). There was pervasive recognition within the National Park Service that the new closure would be controversial, evidenced by the following remarks:

These excerpts show an intent on the part of the National Park Service to railroad through the closure, to maintain secrecy, to unleash the fencing with lightening speed, and to establish a fait accompli

Second, contrary to the government’s argument the closure affected a significant and strategically located parcel. The ten-acre closure should be viewed in terms of the 220 acres of Fort. Funston, not the larger GGNRA. Otherwise, a closure within a multi-site park could never be considered of a “highly controversial nature.” The closure must be considered in conjunction with the previous closure of the same type; otherwise, piecemeal closures could never be challenged even though large in overall scope. This parcel is especially significant'. The closure encompasses beach-front land with important beach access and large ’ sand dunes on which people exercise and play with their dogs. Put differently, plaintiffs are able to articulate objectively plausible issues about the closure that explain why the closure is legitimately highly controversial. It is a project whose details'might well be modified by public input even if it goes forward.

D. Substantial Alteration in Public Use Pattern

Two decisions have discussed the term “significant alteration in the public use pattern.” In Mausolf v. Babbitt, 125 F.3d 661 (8th Cir.1997), the snowmobilers sought to enjoin the National Park Service from enforcing restrictions on snowmobiling in Voyageurs National Park. The plaintiffs argued, inter alia, that the restriction should have been published in the Federal Register under Section 1.5(b) because it would result in “a significant alteration in the public use pattern.” The Eighth Circuit rejected the argument. Snowmobiling in Voyageurs National Park was already generally prohibited absent special regulations permitting the activity. Id. at 669 n. 10. The court reasoned that although the National Park Service had neglected to enforce the provision and had allowed snowmobiling to continue despite regulations to the contrary, the challenged closure order restricting snowmobiling did not amount to a “significant alteration” in what was previously an unlawful-public use of the park. Ibid.

In Spiegel v. Babbit, 855 F.Supp. 402 (D.D.C.1994), the court held that the National Park Service’s decision to limit mooring hours in Georgetown Waterfront Park did not create a “significant alteration in the public use pattern.” Without lengthy analysis, the court concluded that plaintiffs’ arguments to the contrary were not sufficient to outweigh the deference a court normally grants to an agency decision. In Spiegel , the National Park Service had a written a letter to the plaintiff explaining that its limited docking restriction did not constitute a significant alteration of public use warranting publication in the Federal Register. Id. at 404 . The District of Columbia Circuit affirmed this portion of the district court’s order. Spiegel v. Babbitt, 56 F.3d 1531 , No. 94-5184, 1995 WL 364555, *1 (D.C.Cir. May 31, 1995).

Here, however, walking dogs off-leash in Fort Funston was expressly permitted (A.R. US01472, GGNRA, 36 C.F.R. 1, Compendium Amend., signed on July 8, 1996, by Brian O’Neill, Gen. Superintendent). Thus, Fort Funston had the reverse scheme from Mausolf , which required. special regulations to allow snowmobiling. The National Park Service’s closure of the coastal bluff area was not simply enforcement of existing law. Likewise, this case differs significantly from Spiegel. There the court’s decision rested solely on deference to the National Park Service, which had memorialized its decision that its restriction did not significantly alter the public use pattern in a letter. Here, as discussed above, deference to the National Park Service is not appropriate.

Comments to the “significant alteration” language in Section 1.5(b) provide:

47 Fed.Reg. 11599-11600.

The record shows that the closure restricts the last large bluff area of Fort Funston. The closure eliminates the central one of only three access points to the beach. This elimination also poses a potential safety hazard for people on the beach during high tide. Hikers trying to go between the northern and southern access trails may get trapped in the- surf. The obvious and logical escape route is the middle draw. That, however, would be barred by the closure. Hikers might be tempted to climb over a rock outcrop at Panama Point, an unsafe endeavor, for most.

The record contains no statistical analysis of use patterns or surveys of users on this issue. Plaintiffs have shown some evidence outside the administrative record regarding use, but much of this'is hearsay. In any event, anecdotal evidence is of limited value. But, again, the National Park Service prepared no contemporaneous memorandum explaining that the closure would not significantly alter the public-use pattern of Fort Funston. The Court conducted a view of the site. Based on the topography and location of the closure, the Court finds that plaintiffs have raised, at a minimum, serious questions regarding whether the National Park Service’s closure will create a “significant alteration in the public use pattern.”

E. Conclusion Re Likelihood of Success

Without doubt, the National Park Service is authorized to manage Fort Funston so as to preserve its natural resources, including closures to protect the bank swallows. The extent to which all or any part of Fort Funston should be closed for that worthwhile .purpose is committed to the sound discretion of the National Park Service. A court could set aside such a substantive judgment only if it were arbitrary and capricious so long as all procedural requirements were followed. In this case, a procedural rule was violated, or at least a likelihood of such a violation has been shown.

2. Irreparable Injury and/or Balance of Hardships

“Plaintiff must demonstrate potential harm which cannot be redressed by a legal or equitable remedy following trial. The preliminary injunction must be the only way of protecting the plaintiff from such, harm.” Campbell Soup Co. v. ConAgra, Inc., 977 F.2d 86, 91 (3d Cir.1992). The deprivation of a source of personal satisfaction and tremendous joy can constitute an irreparable injury. See Chalk v. United States Dist. Court, 840 F.2d 701 , 709 (9th Cir.1988) (possibility of irreparable injury established where school board sought to remove teacher diagnosed with AIDS from teaching hearing-impaired children but continued to-pay his salary).

In Galusha v. New York State Department of Environmental Conservation, the court found that the plaintiffs, people with disabilities requiring motorized vehicles for mobility, would be irreparably harmed by enforcement of regulations prohibiting any non-emergency use of motorized vehicles within Adirondack Park. 27 F.Supp.2d 117 , 122 (N.D.N.Y.1998). The court reasoned that every day the plaintiffs missed in the park constituted irreparable harm because no amount of money could compensate for the loss. The court further found that “[plaintiffs’ access to a naturally ever-changing environment is impermissibly limited” and “[a]bsent preliminary relief, they will suffer an injury that is present, actual, and not calculable.” Ibid. This Court finds the Galusha reasoning sound.

Like the Galusha plaintiffs, plaintiffs in this case have shown the possibility of irreparable harm. Members of the plaintiff groups and the individual plaintiffs used the closure area frequently— some walked their dogs there twice daily. Plaintiffs do not seek money damages in this action. Rather, they seek continued access to recreation that improves the quality of their lives. The harm they face is substantial and irreparable.

Moreover, the balance of hardships favors plaintiffs. Where there is a strong probability of success on the merits, as here on the issue of “highly controversial,” the moving party need only demonstrate that he or she will suffer a degree of hardship that outweighs the hardship facing the opposing party. See Topanga Press, Inc. v. City of Los Angeles, 989 F.2d 1524, 1528 (9th Cir.1993). Defendants would not be hamstrung by an injunction pending a trial on the merits. Defendants need only publish the closure as rulemak-ing, take written comments, and then make a decision. Moreover, as seen by recent events, the National Park Service has the authority to make any closure pursuant to an emergency. When the bank swallows returned, defendants did so. As for the National Park Service’s planned revegetation efforts, defendants have offered no evidence of irreparable harm if the closure does not take effect immediately. Indeed, only one acre of the closure is scheduled to be replanted this year.

CONCLUSION

Plaintiffs have established a probability of success on the merits as well as irreparable injury. In light of the government’s request that the Court remand the matter to the agency rather than issue an injunction, the Court invites all parties to submit briefs on the issue of remedy. Please address what steps are required for notice and comment under Section 1.5. Simultaneous briefs (up to ten pages) should be filed no later than May 4, 2000. Simultaneous replies (up to five pages) should be filed by May 9, 2000. All service should be by hand or by fax. Because the National Park Service has declared an emergency upon the recent arrival of the bank swallows, no injunction will become effective until August upon their annual departure or until further proceedings and determination as to the validity of the emergency declaration or its scope.

IT IS SO ORDERED.

[[Image here]]

. Publication in the Federal Register is a procedure that allows the public to scrutinize a proposed rule. Publication is followed by a period during which the public may file written comments on the proposed rule. The comments become part of the administrative record. The purpose is to gather and record the views of all interested parties. After the comment period ends, the agency is not obligated to follow any particular comment, but the agency’s decision with regard to the proposed rule many not be arbitrary and capricious in light of the' administrative record, of which the comments are a part. . Section (a)(2)(iii) of the GGNRA compendium amendment to 36 C.F.R. 1 provides as follows: The following areas, described below and depicted on maps included as attachments 1-11, are designated as voice control areas where obedient pets, under supervision, may be allowed off leash. Fort Funston and Fort Funston Beach: Beach area south of Sloat Boulevard to San Mateo County line. Pets must be leashed within the trail system of the Bank Swallow Habitat Area and pets are prohibited insides the Battery Davis Hillside Closure. (A.R. US01472, GGNRA, 36 C.F.R. 1, Compendium Amend., signed on July 8, 1996, by Brian O’Neill. Gen. Superintendent). . "Threatened” is a less serious listing than "endangered.” The bank swallow is not federally listed as either "threatened” pr "endangered.” . A.R. means administrative record. This abbreviation will be followed by a bates-number beginning with US, indicating that the document was produced by the government as part of the administrative record.'' .The appended map has been modified to show the approximate boundaries of the new permanent and seasonal closures. It is based on the map appended to the GGNRA compendium amendment to 36 C.F.R. 1, located in the administrative record at US01474. . As discussed in the "highly controversial" analysis, the government's argument regarding the gross size of the closure is not persuasive. The government argues that the closure encompasses less than three percent of Fort Funston and a much smaller percentage of the GGNRA. The government ignores, however, the previous closure, the relative use and distinctive features of the closed property. It is the quality of the closed land that may support plaintiffs’ claim that the closure will prompt "a significant alteration in the public use pattern.”