Greater Yellowstone Coalition v. Kempthorne

MEMORANDUM OPINION EMMET G. SULLIVAN, District Judge. The instant case represents the latest in a series of challenges to the regulations promulgated by the National Park Service (“NPS”) concerning snowmobile use in the National Parks. The regulations currently at issue propose new restrictions on recreational snowmobiling in Yellowstone and Grand Teton National Parks and the John D. Rockefeller Jr. Memorial Parkway (collectively “the parks”). There are two plaintiffs in this action. The first is the Greater Yellowstone Coalition, a group of conservation organizations that “take an active interest in maintaining the integrity of the National Park System.” This group includes the Sierra Club, the Winter Wildlands Alliance, the Wilderness Society, and the Natural Resources Defense Council (collectively “GYC”). GYC Compl. ¶ 7. The second Plaintiff is the National Parks Conservation Association (“NPCA”), the largest national organization in the United States dedicated to the protection and enhancement of the National Park System. NPCA Compl. ¶ 8. Defendants are the National Park Service, Dirk Kempthorne, in his official capacity as the Secretary of the Interior, Mary Bomar in her official capacity as Director of the National Park Service, and Mike Snyder in his official capacity as Director of the Intermountain Region of the U.S. National Park Service (collectively “NPS”). The new Winter Use Plan (“WUP,” “Rule,” or “Plan”) promulgated by Defendants allows 540 recreational snowmobiles and eighty-three snowcoaches to enter Yellowstone National Park every day. Plaintiffs allege that this number is so high as to render the plan arbitrary and capricious in violation of the Administrative Procedure Act (“APA”) and proeedurally flawed in violation of the National Environmental Policy Act (“NEPA”). Plaintiffs also claim that the plan violates the NPS Organic Act, the Yellowstone Enabling Act, NPS regulations, and two Executive Orders. Specifically, Plaintiffs’ arguments focus on the WUP’s substantive and procedural deficiencies as they relate to the plan’s impacts on the parks’ natural soundscapes, air quality, and wildlife. Agreeing that there are no facts in dispute, Plaintiffs and Defendants have filed cross-motions for summary judgment. The Court held a hearing on the motions on August 27, 2008, and the parties filed short post-hearing briefs. Upon consideration of the motions, the responses and replies thereto, oral argument at the hearing, the post-hearing briefs, the applicable law, and the entire administrative record in this case, the Court GRANTS Plaintiffs’ Motion for Summary Judgment and DENIES Defendants’ Motion. The 2007 Winter Use Plan, the 2007 Final Environmental Impact Statement (“FEIS”), and the 2007 Record of Decision (“ROD”) are hereby vacated and remanded to the agency for proceedings consistent with this opinion. I. Procedural History This Court’s involvement in the ongoing series of cases regarding Yellowstone’s winter management began in 1997 and has continued nearly without pause to the present day. See Fund for Animals v. Norton, 323 F.Supp.2d 7 (D.D.C. 2004)(“FFA II”); Fund for Animals v. Norton, 294 F.Supp.2d 92 (D.D.C. 2003)(“FFA /”); Fund for Animals v. Babbitt, 97-cv-1126 (EGS) (filed May 20, 1997). Over the years, environmental and recreation groups have challenged the Park Service’s restrictions on the use of *187 snowmobiles in the parks, with the more recent controversies growing out of a year 2000 Record of Decision (“2000 ROD”) which found that the use of snowmobiles at present levels so harmed the integrity of the parks’ resources and values that it violated the NPS Organic Act. See Record of Decision, Winter Use Plans for the Yellowstone and Grand Teton National Parks and John D. Rockefeller Jr. Memorial Parkway, 65 Fed.Reg. 80,908, 80,916 (Dec. 22, 2000). In light of this finding, in 2001, NPS published a Final Rule calling for the eventual phase-out of personal snowmobiles in the parks, and instead recommended continued winter access through the use of a snowcoach mass transit system. FFA I, 294 F.Supp.2d at 100. The “phase-out rule,” promulgated by the Clinton administration, was published the day after President George W. Bush took office, and was immediately stayed pending a review of the Rule by the new administration. Id. In response to litigation brought by snowmobile manufacturers and enthusiasts, NPS prepared a Supplemental EIS (“SEIS”) in 2003. The SEIS proposed a dramatic change of course. In place of the planned phase-out, NPS set a new limit of 950 snowmobiles per day in Yellowstone. Id. at 101. Following two lawsuits in this Court and one in the District of Wyoming, NPS put into effect a “Temporary Winter Use Plan” which allowed a daily limit of 720 snowmobiles. Under the temporary plan, all snowmobiles entering the parks were required to meet “best available technology” standards for noise and emissions, and were also required to be accompanied by a commercial guide. This temporary plan was to be in effect for three winter seasons, from 2004 through 2007, and then replaced with a long-term winter use plan in 2007/2008. It is the new long-term plan that is the subject of the instant case. On September 24, 2007, NPS published its Winter Use Plan Final Environmental Impact Statement (“FEIS”). The complete plan was published in a November 20, 2007 Record of Decision (“2007 ROD”). The 2007 ROD claims to address this “Court’s various concerns regarding the winter use 2003 Supplemental EIS” and allows 540 recreational snowmobiles per day, subject to “best available technology” standards, (hereinafter, “BAT”), 100% commercial guiding, and a requirement that all snowmobilers travel in groups of eleven or less. The Rule also requires that all snowcoaches and administrative snowmobiles implement BAT standards by 2011. 2007 ROD at 5. On November 20 and 21, 2007, two lawsuits were filed in this Court by GYC and NPCA, respectively. Both suits allege that the FEIS and 2007 ROD in this case failed to comply with the National Environmental Policy Act (“NEPA”) and the Administrative Procedure Act (“APA”). On December 18, 2007, NCPA amended its complaint to include a challenge to the 2007 Final Rule, which was published on December 13, 2007. In addition to NEPA and the APA, NPCA contends that the 2007 Final Rule violates the National Park Service Organic Act, and governing Executive Orders and NPS Regulations. The GYC Plaintiffs likewise amended their complaint on January 11, 2008 to also challenge the Final Rule bringing similar claims. The cases were consolidated by Order of this Court on March 19, 2008. The Plan at issue was selected as one of seven alternatives analyzed in the 2007 FEIS. The alternatives ranged from a “no action” alternative which would have ended all oversnow vehicle (“OSV”) use in the parks, to an “expanded recreational use” alternative which would have allowed up to 1025 snowmobiles per day. The details of the Winter Use Plan (also known as “Alternative 7” in the FEIS) are as follows. Recreational snowmobiles are limited to 540 per day in Yellowstone and snowcoach- *188 es are limited to eighty-three per day. All snowmobiles must meet Best. Available Technology (“BAT”) standards for emissions and noise. Snowcoaches and administrative snowmobiles (including park staff and concessionaires) must also meet BAT standards by the 2011-2012 winter season. All snowmobiles must be accompanied by commercial guides and must travel in groups of one to eleven. (Note that an individual snowmobiler and his/her paid guide constitute a “group” under this definition). The plan calls for continuing the “Adaptive Management Program” created under the Temporary Rule to determine if certain goals relating to soundscapes, air quality, and the protection of wildlife are being met. The original plan called for Sylvan Pass, which connects the East Entrance of the Park to Cody, Wyoming, to be closed. However, the ROD was amended July 10, 2008 and now allows for the pass to be open subject to “full avalanche forecasting.” Amended 2008 ROD at 3. On December 13, 2007, the state of Wyoming filed a petition for review of agency action in the U.S. District Court for the District of Wyoming. The petition challenged the 2007 FEIS, 2007 ROD, and 2007 Final Rule, alleging that those actions violate NEPA, the APA, the Organic Act, the Yellowstone National Park Act, and the United States Constitution insofar as they (1) impose daily limits on snowmobile access to Yellowstone (2) impose a commercial guide requirement; and (3) impose a new management scheme for Sylvan Pass. Defs.’ Mot. at 9. On January 2, 2008, the Board of County Commissioners of the County of Park filed a nearly identical petition. Id. The two Wyoming cases were consolidated by Order dated February 19, 2008. Id. On February 22, 2008, the International Snowmobile Manufacturers Association, the American Council of Snowmobile Associations, the Blue Ribbon Coalition, and Terri Manning (collectively “ISMA”) filed a motion to intervene as plaintiffs in the consolidated Wyoming cases, challenging the 2007 Final Rule’s reduced limit of 540 snowmobiles per day in Yellowstone and the commercial guide requirement. Id. ISMA’s motion was granted the same day. On April 24, 2008, this Court denied Federal Defendants’ Motion to transfer this case to the District of Wyoming. Greater Yellowstone Coalition, et al. v. Kempthorne, et al., 2008 WL 1862298 (D.D.C. April 24, 2008). On the same day, the Court granted ISMA’s Motion to Intervene insofar as it sought to intervene as a Defendant in this case, but denied its motion to assert cross-claims against the Federal Defendants because ISMA is pursuing identical claims against the Federal Defendants as a plaintiff-intervenor in the Wyoming action. Id. at *7. Accordingly, based on principles of comity and the first-to-file rule, the Court found that it must yield to the Wyoming Court’s jurisdiction over ISMA’s affirmative claims against the NPS. Id. II. Legal Standards 1. Administrative Procedure Act Under the APA, federal agency actions are to be held unlawful and set aside where they are “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” See 5 U.S.C. § 706(2)(A). While this standard does not empower courts to substitute their judgment for that of the agency, it requires “a thorough, probing, in-depth review” of challenged decisions. Citizens to Preserve Overton Park v. Volpe, 401 U.S. 402, 415-16, 91 S.Ct. 814, 28 L.Ed.2d 136 (1971). Accordingly, an administrative action must be vacated where the agency relied on factors which Congress has not intended it to consider, entirely failed to consider an important aspect of the problem, offered an explanation for its *189 decision that runs counter to the evidence before the agency, or is so implausible that it could not be ascribed to a difference in view or the product of agency expertise. Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43, 103 S.Ct. 2856, 77 L.Ed.2d 443 (1983); see also Daingerfield Island Protective Soc’y v. Babbitt, 40 F.3d 442, 446 (D.C.Cir.1995) (deference only to “reasoned, permissible construction^] of ... relevant statute[s]”) (internal quotations omitted). Review of an agency action is more demanding where the challenged decision stems from an administrative about-face. “For [an] agency to reverse its position in the face of a, precedent it has not persuasively distinguished is quintessentially arbitrary and capricious.” La. Pub. Serv. Comm’n v. FERC, 184 F.3d 892, 897 (D.C.Cir.1999). Thus, when reversing itself, “[an] agency is ‘obligated to supply a reasoned analysis for the change beyond that which may be required when an agency does not act in the first instance.’ ” FFA I, 294 F.Supp.2d at 104 (quoting State Farm, 463 U.S. at 41-42, 103 S.Ct. 2856) (emphasis in original). This obligation is all the more pronounced where the agency’s reversal is at odds with a clear statutory mandate governing the agency’s actions. See id. at 105, 108. 2. National Environmental Policy Act NEPA generally requires federal agencies to examine the environmental effects of proposed federal actions and to inform the public of the environmental concerns that were considered in the agency’s decision making. Citizens Against Rails to Trails v. Surface Trans. Bd., 267 F.3d 1144, 1150 (D.C.Cir.2001) (citing Baltimore Gas v. Natural Res. Def. Council, 462 U.S. 87, 97, 103 S.Ct. 2246, 76 L.Ed.2d 437 (1983)). Specifically, NEPA requires agencies to prepare an environmental impact statement (“EIS”) for all proposals for “major Federal actions significantly affecting the quality of the human environment.” Id. (citing 42 U.S.C. § 4332(2)(C)). “Environmental impact statements shall state how alternatives considered in it and decisions based on it will or will not achieve the requirements of [NEPA] and other environmental laws and policies.” 40 C.F.R. § 1502.2(d). An EIS must discuss “[p]ossible conflicts between the proposed action and the objectives of Federal ... land use plans, policies and controls for the area concerned,” id. § 1502.16(c), and “should present the environmental impacts of the proposal and the alternatives in comparative form, thus sharply defining the issues and providing a clear basis for choice among options by the decisionmaker and the public.” Id. § 1502.14. “Such information may cause the agency to modify its proposed action.” See, e.g., Natural Res. Def. Council v. Morton, 458 F.2d 827, 831 (D.C.Cir.1972). NEPA creates no private right of action and therefore challenges to agency compliance with NEPA must be brought pursuant to the APA, 5 U.S.C. § 551 et seq. Karst v. EPA 475 F.3d 1291, 1295 (D.C.Cir.2007). The APA requires “final agency action for which there is no other adequate remedy in a court.” 5 U.S.C. § 704. Accordingly, persons claiming a right to sue under NEPA must identify some final agency action that adversely affects them. In the NEPA context, the “final agency action” required by the APA must also be “major federal action” under NEPA. Karst, 475 F.3d at 1295. In the instant case, the final rule implementing the WUP, the 2007 ROD, and the 2007 FEIS constitute final agency action such that review under NEPA is appropriate. 3. Governing Statutory Mandates In addition to the general requirements of the APA and NEPA, NPS is also bound *190 by specific statutory mandates that define the Service’s mission and impose independent requirements upon the agency. Plaintiffs challenge the WUP as contrary to the National Park Service Organic Act, the Yellowstone Enabling Act, two Executive Orders and the NPS Snowmobile Regulation. The NPS was created in 1916 and charged with the duty to promote and regulate the use of the ... national parks, monuments, and reservations hereinafter specified ... by such means and measures as conform to the fundamental purpose of the said parks, monuments, and reservations, which purpose is to conserve the scenery and the natural and historic objects and the wild life therein and to provide for the enjoyment of the same in such manner and by such means as will leave them unimpaired for the enjoyment of future generations. 16 U.S.C. § 1. Congress supplemented and clarified these provisions through the General Authorities Act in 1970, and again through enactment of the “Redwood Amendment” in 1978. That Act, as amended, reinforced that management of the parks “shall be consistent with the Organic Act” and declared that the “protection, management, and administration of these areas shall be conducted in light of the high public value and integrity of the National Park Service and shall not be exercised in derogation of the values and purposes for which these areas have been established, except as may have been or shall be directly and specifically provided by Congress.” Id. The NPS’s 2006 Management Policies, which interpret the above directives, designate the Organic Act as “the most important statutory directive for the National Park Service.” National Park Service 2006 Management Policies at 1.4.1 (“NPS Policies”), A.R. 120645. 1 The Yellowstone Enabling Act, the federal statute governing the agency’s administration of Yellowstone, requires that the NPS preserve “from injury or spoliation” the “wonders” of the park and insure “their retention in their natural condition.” 16 U.S.C. § 22. The Secretary is also required to “provide against the wanton destruction of the fish and game found within the park, and against their capture or destruction for the purposes of merchandise or profit.” Id. Finally, two Executive Orders specifically address the use of snowmobiles in the Parks. Executive Order 11644, signed by President Nixon in 1972, established procedures for controlling the use of off-road vehicles, specifically including snowmobiles, on public lands. The Executive Order mandated that each agency establish regulations designating specific zones of use for off-road vehicles, and that such chosen areas be located to “minimize harassment of wildlife and significant disruption of wildlife habitats.” Exec. Order *191 No. 11644, 37 Fed.Reg. 2877 (Feb. 8,1972). Executive Order 11989, signed by President Carter in 1977, amended and strengthened the 1972 Order, stating that if an agency head determines that the use of off-road vehicles will cause “considerable adverse effects on the soil, vegetation, wildlife, wildlife habitat or cultural or historic resources of particular areas or trails of the public lands” the agency head shall “Immediately close such areas or trails to off-road vehicles.” Exec. Order No. 11989, 42 Fed.Reg. 26,959 (May 24, 1977). In response to the Executive Orders, NPS promulgated 36 C.F.R. 2.18(c), which prohibits snowmobiles “except where designated and only when their use is consistent with the park’s natural, cultural, scenic and aesthetic values, safety considerations, and park management objectives, and will not disturb wildlife or damage park resources.” 36 C.F.R. § 2.18(c). III. Discussion 1. Statutory Interpretation of Conservation Mandate As an initial matter, both parties agree that the Organic Act imposes a “conservation mandate” upon NPS, and that that mandate is articulated in § 1.4.3 of the 2006 NPS Policies. However, the parties disagree over precisely what the mandate requires and when it is triggered. Section 1.4.3 provides, in its entirety, The fundamental purpose of the national park system, established by the Organic Act and reaffirmed by the General Authorities Act, as amended, begins with a mandate to conserve park resources and values. This mandate is independent of the separate prohibition on impairment and applies all the time with respect to all park resources and values, even when there is no risk that any park resources or values may be impaired. NPS managers must always seek ways to avoid, or to minimize to the greatest extent practicable, adverse impacts on park resources and values. However, the laws do give the Service the management discretion to allow impacts to park resources and values when necessary and appropriate to fulfill the purposes of the park, so long as the impact does not constitute impairment of the affected resources and values. The fundamental purpose of all parks also includes providing for the enjoyment of park resources and values by the people of the United States. The enjoyment that is contemplated by the statute is broad; it is the enjoyment of all the people of the United States and includes enjoyment both by people who visit the parks and by those who appreciate them from afar. It also includes deriving benefit (including scientific knowledge) and inspiration from the parks, as well as other forms of enjoyment and inspiration. Congress, recognizing that the enjoyment by future generations of the national parks can be ensured only if the superb quality of park resources and values is left unimpaired, has provided that when there is a conflict between conserving resources and values and providing for enjoyment of them, conservation is to be predominant. This is how courts have consistently interpreted the Organic Act. NPS Policies, § 1.4.3. Relying on the above passage, NPS argues that the conservation mandate of the Organic Act is only triggered when the impacts from a particular use rise to the level of “unacceptable impacts.” See Hearing Transcript, 95, August 27, 2008 (hereinafter “Transcript”). Defendants paraphrase the mandate as follows: “[i]f unacceptable impacts are found, the Service deems the proposed use of park recourses to be in conflict with their conservation and therefore prohibits the proposed *192 use.” Fed. Defs.’ Supp. Brief at 4. Applying this theory to the instant case, Defendants argue that because NPS has determined that the impacts of snowmobiling are “acceptable,” then there is no “conflict” between conservation and use, and therefore the requirement that conservation predominate is not implicated. See id.; see also Transcript, 95. NPS reasons that the above-referenced “management discretion to allow impacts” encompasses the decision to allow the WUP’s admittedly adverse impacts to the parks’ natural soundscape, air quality, and wildlife, because those impacts do not conflict with conservation. At the hearing, NPS argued that the adverse impacts of snowmobiling are acceptable because the Organic Act allows adverse impacts if they are unavoidable and appropriate. Government counsel offered the following in support of this argument. “A good analogy here is on these battlefield parks where a tree will block a vista and it’s important to restore that vista, that tree being a national resource, will nonetheless be removed, and that is an unavoidable impact, but it is an appropriate impact. That what the fourth sentence [of the Policies] is about, so it doesn’t limit the Park Service’s discretion to allow impacts in the way that the plaintiffs think it does.” Transcript, 97. Plaintiffs strenuously disagree with this characterization of the Organic Act. While recognizing that the NPS has broad discretion to carry out its mission, Plaintiffs contend that the WUP impermissibly permits adverse impacts to park resources merely to provide another form of recreation. NPCA Mot. at 41. Plaintiffs argue that unlike removing a tree to restore a vista, NPS has not explained how snowmobiling is “necessary and appropriate to fulfill the purposes of the park” such that the adverse impacts are acceptable, nor have they explained how the Plan seeks “ways to avoid, or to minimize to the greatest extent practicable, adverse impacts on park resources and values.” NPS Policies, § 1.4.3. Plaintiffs point out that under the temporary use plan in place over the past three years, NPS’s own “adaptive management thresholds” for air quality and soundscape protection have been exceeded on multiple occasions without generating a response from NPS. See GYC Mot. at 21. Plaintiffs argue that because actual daily use under the temporary plan has averaged only between 260-290 snowmobiles, and NPS’s own thresholds for noise and air pollution have been exceeded in spite of the low numbers, allowing up to 540 snowmobiles per day will effectively double the environmental harms seen under the temporary plan. Plaintiffs insist that this result cannot be squared with the Organic Act, regardless of how NPS chooses to define “conflict.” The Court agrees. The Organic Act clearly states, and Defendants concede, that the fundamental purpose of the national park system is to conserve park resources and values. Section 1.4.3 of the NPS Policies, which provides the NPS’s official interpretation of the Organic Act, states that “conservation is to be predominant.” Defendants claim that the Act “establishes the fundamental purposes of conservation and enjoyment, but is silent as to how those two purposes should be analyzed.” Fed. Defs.’ Supp. Brief at 1. While it is true that the Act is “silent as to the specifics of Park management,” Davis v. Latschar, 202 F.3d 359, 365 (D.C.Cir.2000), Defendants’ own official interpretation of the Organic Act explicitly instructs NPS on how to balance conservation and enjoyment. Namely, in the case of a conflict, “conservation is to be predominant.” NPS Policies, § 1.4.3. Moreover, while it is true *193 that “enjoyment” is also a fundamental purpose of the parks, enjoyment is qualified in the Organic Act in a way that conservation is not. The Organic Act charges NPS with the duty to “provide for the enjoyment” of the parks’ resources and values in “such manner and by such means as will leave them unimpaired for the enjoyment of future generations.” 16 U.S.C. § 1. This is not blanket permission to have fun in the parks in any way the NPS sees fit. As Plaintiffs articulated at the hearing, the “enjoyment” referenced in the Organic Act is not enjoyment for its own sake, or even enjoyment of the parks generally, but rather the enjoyment of “the scenery and natural and historic objects and the wild life” in the parks in a manner that will allow future generations to enjoy them as well. Id. Accordingly, while NPS has the discretion to balance the “sometimes conflicting policies of resource conservation and visitor enjoyment in determining what activities should be permitted or prohibited,” S. Utah Wilderness Alliance v. Dabney, 222 F.3d 819, 826 (10th Cir.2000), that discretion is bounded by the terms of the Organic Act itself. NPS cannot circumvent this limitation through con-clusory declarations that certain adverse impacts are acceptable, without explaining why those impacts are necessary and appropriate to fulfill the purposes of the park. See NPS Policies, § 1.4.3. The limits on NPS’s discretion have been recognized by this Circuit. In Daingerfield, this Circuit upheld the NPS’s choice of an interchange design that NPS had concluded “would have the least deleterious effect on the environment.” 40 F.3d at 446. The Court noted that the Organic Act “gives the Park Service broad, but not unlimited discretion in determining what actions are best calculated to protect Park resources.” Id. (emphasis added). While not explicitly holding that NPS is required to choose the least deleterious option, the Circuit did cite with approval to the District Court’s observation that “the only choice left to the Park Service was to approve the least intrusive interchange possible, which it did, or refuse to approve any interchange at all.” Id. at n. 3. Accordingly, at the very least, NPS is required to exercise its discretion in a manner that is “calculated to protect park resources” and genuinely seeks to minimize adverse impacts on park resources and values. See Daingerfield, 40 F.3d at 446; NPS Policies, § 1.4.3. Many courts before this one have interpreted the Organic Act to require as much. See Bicycle Trails of Marin v. Babbitt, 82 F.3d 1445, 1453 (9th Cir.1996) (The “overarching concern” of the Organic Act is “resource protection.”); Edmonds Inst. v. Babbitt, 42 F.Supp.2d 1, 16 (D.D.C.1999)(citing cases interpreting Organic Act “amendments to reflect a renewed insistence on the part of Congress that the national parks be managed in accordance -with the primary purpose of the [Act], namely the conservation of wildlife resources.”); Nat’l Rifle Ass’n of Am. v. Potter, 628 F.Supp. 903, 909 (D.D.C. 1986) (“In the Organic Act, Congress speaks of but a single purpose, namely conservation.”). To hold otherwise now would depart from years of well-reasoned precedent and undermine over 100 years of park management. The record contains many compelling examples of the magnitude of this decision, but none more so than a letter from eleven former National Park Service Directors. AR 120941. Dated March 26, 2007, the letter was written in response to the original “preferred alternative” for the WUP, which would have allowed 720 snowmobiles per day, a proposal the former directors urged “would radically contravene both the letter and spirit of the 2006 Management Policies.” Id. While that number was ultimately reduced to 540, the letter contends *194 that further reducing the limit to zero, “while expanding public access on modern snowcoaches, would further improve the park’s health.” Id. Relying on the same studies found in the FEIS, the former directors argue that increasing snowmobile use over the current average of 250 snowmobiles per day would increase air and noise pollution and “sidestep[] a recent recommendation by Park Service scientists” that traffic should be kept “at or below current levels, not expanded.” Id. The letter concludes, It is our profound hope that in our country’s oldest national park you will insist that your commitment [to the 2006 Management Policies] be upheld — that the traditional conservation emphasis of the national parks will be continued. The current proposal to accommodate increasing snowmobile use in Yellowstone is at odds with these policies. Id. 2. Impact and Non-Impairment Determinations As explained above, the Organic Act prohibits uses which impair park resources and values. NPS has defined impairment as an impact that “would harm the integrity of park resources and values, including the opportunities that otherwise would be present for the enjoyment of those resources or values.” See ROD at 28-29; NPS Policies § 1.4.5. Whether an impact meets this definition depends on the particular resources and values that would be affected; the severity, duration, and timing of the impact; the direct and indirect effects of the impact; and the cumulative effects of the impact in question and other impacts. An impact is more likely to constitute impairment to the extent that it affects a resource or value whose conservation is necessary to fulfill specific purposes identified in the establishing legislation or proclamation of the park; key to the natural or cultural integrity of the park or to opportunities for enjoyment of the park, or identified in the park’s general management plan or other relevant NPS planning documents as being of significance. Id. NPS has also determined than an impact would be “less likely to constitute an impairment if it is an unavoidable result of an action necessary to preserve or restore the integrity of park resources or values and it cannot be further mitigated.” Id. In addition, NPS has interpreted the Organic Act to prohibit uses which cause “unacceptable impacts.” Section 1.4.7.1 of the NPS Policies defines “unacceptable impacts” as “impacts that, individually or cumulatively, would • Be inconsistent with a park’s purposes or values, or • Impede the attainment of a park’s desired future conditions for natural and cultural resources as identified through the park’s planning process, or • Create an unsafe or unhealthful environment for visitors or employees, or • Diminish opportunities for current or future generations to enjoy, learn about, or be inspired by park resources or values, or • Unreasonably interfere with ■ park programs or activities, or an appropriate use, or the atmosphere of peace and tranquility, or the natural soundscape maintained in wilderness and natural, historic, or commemorative locations within the park. NPS Policies § 1.4.7.1; see also 2007 ROD at 29-30; FEIS at 170. To illustrate this definition, the Policies provide a graph that “illustrates the relationship between *195 appropriate use, unacceptable impacts and impairment.” AR 120645 at 14. The 2000 FEIS determined that then-existing conditions constituted “impairment” of the parks’ soundscapes, wildlife, and air quality in violation of the Organic Act. At that time, the average number of snowmobiles entering the park was 795. According to the 2007 ROD, “[t]here is no new evidence contradicting the finding that historically unlimited snowmobile ... use impaired park resources and values.” 2007 ROD at 30. However, the 2007 FEIS concluded that none of the seven alternatives studied would constitute impairment or unacceptable impacts, even the alternative that would have allowed up to 1025 snowmobiles per day. As explained below, the Court finds that NPS fails to articulate why the WUP’s impacts are “acceptable.” NPS simply repeats the above standards in the context of the WUP’s impacts on soundscapes, wildlife, and air quality, but fails to provide any supporting analysis of how the impacts relate to those standards. The ROD makes no effort to explain, for example, why impacts on soundscapes characterized as “major and adverse” do not “unreasonably interfere with the soundscape” and cause an unacceptable impact. Compare ROD at 34 (discussing soundscape impacts) with NPS Policies § 1.4.7.1 (defining unacceptable impacts). Similarly, NPS fails to explain why increasing the amount of benzene and formaldehyde to levels that broach (and sometimes exceed) the minimum risk levels applicable to hazardous waste sites does not “create an unsafe or unhealthful environment for visitors or employees.” NPS Policies § 1.4.71. Defendants claim that the unacceptable impacts determination is one left to the expertise of park managers and argue that Plaintiffs are wrong to rely on the data collected under the temporary plan, in which the thresholds for noise and air quality were exceeded on multiple occasions, as evidence of unacceptable impacts. Defendants state, “GYC provides no basis for its assumption that those adaptive management thresholds coincided with the threshold for unacceptable impacts, nor could it. As the ROD explains, such preliminary thresholds are established to be one of the many guides for possibly taking actions if a problem is perceived.” Fed. Opp’n to GYC Mot. at 15. NPS argues that the thresholds “are designed to indicate when conditions could be moving away from those that are desirable” and they do not set the standard for what constitutes an unacceptable impact. Id. Tellingly, the government fails to identify what does set that standard, referring only to the “graphic” which indicates solely that “unacceptable impacts” are greater than acceptable impacts and less than impairment. Other than this unhelpful image, NPS provides no quantitative standard or qualitative analysis to support its conclusion that the adverse impacts of the WUP are “acceptable.” NPS simply repeats the above definition of unacceptable impacts, but does not state why the impacts of the WUP fail to meet this definition. As illustrated in more detail below, this conclusion, with no supporting analysis or explanation, is quintessentially arbitrary. See Mainella, 459 F.Supp.2d at 103 (rejecting NPS’s conclusory non-impairment determination that “largely parrots the definition of impairment set forth in the NPS Management Polices”). IV. NPS’s Studies, Findings, and Conclusions Concerning Impacts of the WUP As indicated above, Plaintiffs challenge the WUP’s impacts on the natural sounds-capes, wildlife, and air quality in the parks. The Court has conducted “a thorough, probing, in-depth review” of the challenged decision as required by the APA and the Supreme Court. See Citizens to Preserve *196 Overton Park, 401 U.S. at 415-16, 91 S.Ct. 814. Below, the Court will summarize Defendants’ key findings and conclusions and Plaintiffs’ primary challenges to them. A. Impacts on Soundscapes 1. NFS’s Findings and Conclusions The natural soundscape in Yellowstone is explicitly recognized in the 2001 and 2006 NPS Management Policies as a key park resource, and is also identified as such in the FEIS. Accordingly, the above legal mandates protecting park resources and values apply equally to the conservation of the parks’ natural soundscapes. The environment of sound that exists in Yellowstone in the absence of human-caused noise is considered to be the “baseline condition” from which to gauge the impacts of human use. FEIS at 137. NPS conducted soundscape monitoring during the past four winter seasons to evaluate the impacts of OSV use on the natural soundscapes of the parks. According to NPS, the data collected as a result of this effort constitutes “one of the most extensive national park acoustic datasets in existence.” FEIS at 139. The monitoring data focused on the percent of the day between 8:00 a.m. and 4:00 p.m. that OSVs were audible at a particular location (known as “percent time audible”), and the sound level (or loudness) of a particular sound at a particular location. This monitoring took place at five Yellowstone sites that were representative of the various developed areas and travel corridor management zones within the Park: Old Faithful Weather Station; Old Faithful Upper Basin; Spring Creek; West Thumb Geyser Basin; and a point 2.3 miles west of Madison Junction. AR 125050 at 2. An impact was characterized as an “adverse major effect” when a sound of greater than seventy decibels was heard or when any OSV sound was audible more than 75% of the time in a developed area or more than 50% of the time in a travel corridor. Id. at 12. The FEIS included a table of commonly known sounds for a better understanding of the significance of a sound at particular decibel levels. A sound measuring seventy decibels is perceived to be “noisy” and is the equivalent of being in a room with a running vacuum cleaner. FEIS at 140. The average daily use by OSVs at the above sites during the 2005-2006 monitoring season was 256 snowmobiles and thirty-two snowcoaches, approximately half the number of OSVs permitted by the WUP. The 2007 ROD indicates that even at the “lower levels of use experienced over the past three years, oversnow vehicles could be heard more than expected.” 2007 ROD at 20. The ROD indicated that “the 59% time audible along the West Yellowstone to Madison road exceeded the 50% monitoring threshold that was established to help park managers evaluate the effectiveness of winter use management.” Id. at 21. At Old Faithful, the 75% time audible threshold was exceeded 26% of the time. In the Old Faithful Geyser developed area, OSVs were audible about 67% of the day. The monitoring thresholds for sound level were also exceeded more often than expected, as was the backcountry audibility threshold. Id. The backcountry audibility threshold for a “major adverse impact” is 20% time audible, but the average percent time audible for all days analyzed in the backcountry was 26%. Id. Regardless of these frequent exceedances, the ROD concludes that “[although the numbers of snowmobiles permitted to enter Yellowstone each day under this decision are greater ... than those actually experienced during the previous three winters, the impacts of this use are manageable, and the use of snowmobiles at this level, combined with other elements of this decision, is an appropriate use of the park.” Id. The ROD further concludes that “NPS expects to be fully within the *197 soundseapes thresholds identified in the monitoring and adaptive management program.” Id. Though NPS does not explicitly explain why it “expects” to be within all thresholds under the Plan, NPS appears to rely on the requirement that all snowmobiles and snowcoaches will meet BAT standards. However, recreational snowmobiles during the monitoring period (when the excee-dances were recorded) were already required to meet BAT standards, so that factor alone is not a guarantee of future compliance. While it is true that the Plan will require snowcoaches to meet BAT standards for the first time, it is unclear how snowcoach BAT will reduce the levels of percent time audible, when the snowmobiles already meeting BAT standards accounted for 72% of the OSV sound heard during the day. 2007 ROD at 20. Moreover, the ROD relies on the 100% commercial guide requirement to further reduce sound impacts, but that requirement has also been in place throughout the monitoring period, so that element of the Plan does not promise to bring future sounds-cape impacts within the monitoring thresholds either. NPS argues that requiring administrative OSVs to meet BAT standards will also dramatically mitigate noise impacts, because while administrative snowmobiles account for only 12% of all OSVs, they are heard 29% of the time between 8:00 a.m. and 4:00 p.m. Fed. Opp’n to GYC at 20. ■While the Court has no reason to doubt that requiring administrative OSVs to use BAT standards will reduce some noise, precisely by how much is impossible to evaluate on this record. As Plaintiffs point out, NPS has been “converting their own administrative fleet of snowmobiles to four stroke machines” since 2001. See 72 Fed.Reg. at 70,785. NPS has not disclosed how many administrative snowmobiles already employ BAT standards; a figure that would certainly impact NPS’s ability to predict the expected reduction in noise by requiring administrative OSVs to meet BAT standards. In addition to monitoring noise levels in the park, NPS enlisted the U.S. Department of Transportation to conduct acoustical modeling using the Federal Aviation Administration’s (“FAA”) Integrated Noise Model (“INM”), as adapted for use with oversnow vehicles. FEIS at 301. The INM was used to model various vehicle types consistent with the scenarios described in each of the seven examined alternatives, including two- and four-stroke snowmobiles, various models of snowcoach-es, and wheeled vehicles. Fed. Defs.’ Opp’n to GYC at 26-27. NPS contends that “the development of this modeling tool was critical to the Service’s analysis because, unlike historical monitoring data, the INM allowed the Service to compare the potential environmental impacts corresponding to each of the seven alternatives.” Id. Based on the model, NPS predicted for each of the seven alternatives in the FEIS: 1) the total area within the parks that would have any level of OSV audibility, however slight; 2) the percentage of time that OSVs would be audible, as measured both park-wide and at representative points throughout the park; and 3) the intensity (decibel level) of the sound that would be caused by OSVs at various points within the park. 2 Id. On this basis, *198 the FEIS concludes with respect to Alternative 7 (the preferred alternative) that the authorization of up to 540 snowmobiles and eighty-three snowcoaches will cause a moderate, adverse, short-term, and direct impact with respect to Park-wide audibility; a major impact with respect to percent time audible; and a minor, adverse, and short term impact with respect to maximum sound levels. FEIS at 339. Under the modeled conditions, the FEIS concluded that none of the seven alternatives would result in “impairment” under the Organic Act, even the Alternative that modeled sound conditions for 1025 snowmobiles per day. 2. Plaintiffs’ Claims Regarding Soundscape Impacts a. Modeling Data v. Monitoring Data Plaintiffs charge that the FEIS places too much emphasis on the modeling data created by the Volpe report, rather than the monitoring data collected over four seasons by NPS scientists. While acknowledging that a model is appropriate to predict the future impact of the various alternatives, Plaintiffs contend that the conclusions rely too heavily on modeled impacts, to the exclusion of reliable monitoring data, despite criticisms from NPS’s own scientists regarding the problem of “treating the model as literal.” AR 125263. Plaintiffs further contend that the discrepancies between the modeling and monitoring data in the soundscape analysis render the FEIS unlawful under NEPA. See NPCA Reply at 28 (“[T]he Volpe modeling results are not a reasonable basis on which to judge the impact of the Winter Use Plan because those results are totally at odds with the known data derived from the monitoring studies.”). Though an agency’s expert choice of methodology is entitled to deference, see Sierra Club v. DOT, 753 F.2d 120, 128 (D.C.Cir.1985) (“It is clearly within the expertise and discretion of the [agency] to determine proper [noise] testing methods.”), a model must be rejected as arbitrary and capricious “if there is simply no rational relationship between the model and the known behavior of [the items] to which it is applied.” Chemical Mfrs. Ass’n v. EPA 28 F.3d 1259, 1265 (D.C.Cir.1994). The FEIS indicates that the model underestimated the sound level of OSVs at eight out of twelve monitoring sites compared to the field measurements; it never overestimated sound level. FEIS at 302. The model also underestimated percent time audible at seven of the twelve monitoring sites but overestimated audibility at only one site. Id. These discrepancies alone raise doubts about the validity of the government’s impact conclusions, particularly when the justification for the Plan seems to be based almost exclusively on the modeled data. For example, based on the modeled results, the FEIS concludes that “impacts due to maximum sound levels would be minor, adverse and short-term” at the various sites the model represented. Id. at 339. However, the NPS monitoring studies state that sound levels exceeded seventy decibels at many of those same sites, including Old Faithful, along the groomed travel corridor between Madison Junction and the West Yellowstone Entrance, and between West Thumb and Old Faithful. AR 125255 at 2. Sound levels above seven *199 ty decibels are considered an “adverse major effect” by NPS’s own definitions, and yet based on the model, the FEIS concludes that sound level impacts under the Plan will be “minor.” While requiring snowcoaches to meet BAT standards will reduce sound levels, it is unclear how that requirement alone will change the impact level from major to minor, particularly when the monitored conditions represent 250 BAT snowmobiles and the modeled conditions represent 540. There is also record evidence that several of NPS’s own scientists were highly critical of the model, stating that “the model didn’t realistically calculate OSV travel in the area, it just truncated all traffic at a point on the main road. This is why the modeled results are so far from reality.” See AR 125317 (Email from NPS Scientist, Shan Burson); see also AR 116985 (Email from Karen Trevino, NPS scientist, indicating that “given the political reality, we will not be spending much time on this,” referring to ways to improve the soundscape analysis through different modeling techniques). The model also fails to account for temperature “inversions,” which are common in Yellowstone, and cause sound to travel much further than predicted by the model. See FEIS at 302. b. Use of the Park-Wide Metric Obscures Alternatives Plaintiffs also object to the “park-wide metric” employed to compare the percent of total park area in which OSV noise can be heard under the different alternatives. Under this metric, the various alternatives are measured against a “total-park” audibility standard in which an alternative is deemed to have a “major impact” only when it would spread OSV noise across more than 20% of Yellowstone’s more than two million acres. FEIS at 304, Table 4-48. NPS acknowledges that most of the park is inaccessible to those visitors seeking to enjoy the natural soundscape. Although sounds from OSVs are audible within a relatively small portion of the parks’ total acreage, they are concentrated to a large degree around travel corridors and park attractions and affect the areas most accessible by the vast majority of park visitors. Most areas used by winter visitors seeking solitude and quiet are within 2 miles of travel corridors. Remote backcountry areas that are largely free of non-natural sounds are beyond the reach of most visitors.... FEIS at 139. In light of this admission, NPS defends the metric on the grounds that it is obligated to protect soundscapes throughout the entire park, not only in areas where visitors congregate. While this may be true, the Court finds that the park-wide metric serves mostly to dilute the impacts of OSV noise. This is most evident in applying the metric to historic conditions. Based on the model, under historic conditions, OSV noise could be heard in 16.2% of the park, a percentage considered only a “moderate” impact under this metric. NPS entirely fails to justify how conditions previously found to constitute “impairment” can now cause only a “moderate” impact on the park’s soundscape. Rather than defend this bizarre result, NPS argues that the park-wide metric is only one of the three metrics used to determine soundscape impact, and points to its use of percent time audible and sound level metrics as well. However, these metrics do nothing to support Defendants’ ultimate conclusions. For example, in the ROD and the FEIS, NPS repeatedly touts the reduction in “overall audibility” under the Plan from 14.4% of the park to 13.8% of the park. Overall audibility combines the percent of park and percent time audible metrics. This means that OSVs will be *200 heard at some level under the Plan in 13.8% of the park, whereas they can be heard at some level in 14.4% of the park based on the modeling of current conditions. 2007 ROD at 34. Based on the reduction from 14.4% to 13.8% park-wide audibility, NPS concludes that “[t]his decision will be beneficial in Yellowstone and adverse in Grand Teton compared to current use and beneficial in both parks compared to the historical condition.” Id. Putting aside the inherent problems with the comparison to the modeled historical conditions, see n. 2, supra, the Court finds that this conclusion is unsupported by the data. Table 4-49: Yellowstone: Percent of Park Area for 10% Increments of % Time Audible (updated from Volpe Report Table 36) Percent Time Audible Current Historic (8-hours) Alt 1 Alt 2 Alt 3A Alt 4 Alt 5 Alt 6 Alt 7 Condition Condition 0 12.9 10.2 3.5 13.9 13.9 10.4 13.8 14.4 16.2 10 8.6 6.5 2.4 10.2 9.1 5.8 7.6 5.4 13.6 20 6.1 3.9 1.4 8.3 6.5 3.9 5.6 2.5 12.1 30 4.1 2.1 1.0 6.2 4.6 3.1 4.1 1.3 10.4 40 2.8 1.4 0.8 4.5 3.4 2.6 2.8 0.8 9.0 50 2.2 1.0 0.8 3.8 2.4 2.1 1.8 0.6 8.1 60 1.7 0.9 0.7 3.1 1.8 1.6 1.4 0.5 7.4 70 1.3 0.8 0.6 2.5 1.4 1.0 1.2 0.3 6.6 80 1.1 0.8 0.6 1.9 1.1 0.7 1.0 0.2 5.6 90 1.0 0.6 0.5 1.5 0.9 0.6 0.9 0.2 4.6 Upon closer examination of the data relied upon by NPS, it becomes clear that this “reduction” in OSV noise applies only to the areas of the park in which OSVs can be heard the smallest percentage of the time to begin with. Table 4-49 on page 307 of the FEIS purports to show this “beneficial” impact on the soundscape levels compared to current use. The table, as shown above, demonstrates the percentage of the park in which OSVs can be heard at 10% intervals, from 0 up to 90% of the time. FEIS at 307. For example, under current conditions, the model reveals that OSVs can be heard from 0 to 9% of the time in 14.4% of the Park. The WUP reduces that figure to 13.8% of the park, for the 0 to 9% time audible bracket. But in every other audibility category, the percentage of the park in which OSVs can be heard actually increases to two to three times the current condition. For example, at the 50% time audible level, the percentage of the park in which OSVs can be heard under current conditions is 0.6, but under the WUP, that figure will triple to 1.8. The Court inquired about this seeming discrepancy during oral argument on the pending motions and was assured by NPS counsel that the table was correct: the only category in which park-wide audibility decreases under the Plan is the 0 to 9% time audible bracket. Transcript, 61. This means that the Plan will increase all areas of the park in which OSVs are heard more than 10% of the time. At the hearing, NPS continued to argue that the decision was “beneficial with respect to park-wide audibility” but conceded that “it is not beneficial with respect to percent time *201 audible or maximum sound levels.” Transcript, 63. Given that any arguable “benefit” in terms of park-wide audibility pertains only to the 0 to 9% time audible category, the ROD’S statement that the decision is “beneficial in Yellowstone” compared to current use is extremely misleading. Particularly when in the same paragraph, NPS admits that the percent time audible impact will be “major and adverse.” See 2007 ROD at 33-34. c. Road Closures Unequally Modeled In Alternatives 1 and 7, NPS models the road connecting Madison and Norris (also known as the Gibbon Canyon Road) as closed. As a part of a study regarding the impact of road grooming on bison movement, NPS’s scientists recommended studying bison via GPS collaring and cameras for five years to determine if and how their movements are impacted by road grooming. If those studies suggest that closing the Madison-Norris road would provide informative results, then the park may close the road and cease road grooming operations on it. 2007 ROD at 23 (“It is uncertain until the five-year period of data gathering and analysis has finished whether such closure will yield informative data or conclusions.”). However, NPS does not present analysis of the road closure as a mere possibility after five years of study; rather it analyzes the impacts on the soundscape as if the road was already closed. The ROD states that “[potential closure of the Madison to Norris road segment also removes oversnow vehicle operations that would otherwise contribute to sound impacts.” 2007 ROD at 34. “[T]he impacts have been analyzed assuming that the road segment between Madison and Norris would be closed to all motorized oversnow travel.” 2007 ROD at 16. Because this option is not modeled across the other alternatives (with the exception of Alternative 1), and there is only a remote possibility that the closure will happen after five years of study, modeling the road as closed and relying on the consequent “reduction” in soundscape impacts distorts the adverse impacts of the WUP relative to other alternatives. d. Impact Definitions and Non-Impairment Determination The ROD concludes that the WUP’s impact on the natural soundscapes from the overall park perspective will be “moderate, adverse, short-term, and direct.” 2007 ROD at 33. However, this conclusion directly contradicts the “impact definitions” in the FEIS. There, NPS states that if the assessed impact level for one parameter is higher than for another, the overall impact is judged at the higher level. For example, if one alternative is modeled to result in a moderate impact for percent time audible but a major impact for the maximum sound levels present, the overall impact conclusion is for a major impact. FEIS at 304. The ROD found that impacts due to percent time audible will be “major ... adverse, and short-term” while impacts due to sound level will be “minor, adverse, and short-term.” ROD at 34. Accordingly, under NPS’s own definitions, the overall impact to soundscapes for the WUP should have been assessed as “major.” The ROD then states that the decision will be “beneficial in Yellowstone ... compared to current use,” id., and concludes that “[impairment of park resources will not occur; the level of oversnow vehicle sound under the decision will not harm the integrity of park resources and values.” Id. NPS entirely fails to explain why a finding of minor, moderate, and major adverse impacts on soundscapes does not constitute impairment, let alone the lesser threshold of “unacceptable impacts.” The NPS Policies clearly define an unaccepta *202 ble impact as one that would “unreasonably interfere with ... the natural sounds-cape maintained in wilderness and natural, historical, or commemorative locations within the park.” 2007 ROD at 29. Like the Court in Mainella, this Court is equally perplexed as to why any impact characterized as “major and adverse” does not constitute an unacceptable impact, let alone impairment. This is a distinction NPS again fails to explain. See Mainella, 459 F.Supp.2d at 102, n. 24. There is no higher level than “major” on the impact scale. NPS provides no explanation as to how the most adverse impacts can still be considered acceptable. Accordingly, for the reasons stated above, the Court finds that the non-impairment conclusion regarding impacts to the natural soundscape in the parks is fundamentally arbitrary and capricious. B. Impacts on Wildlife 1. NPS’s Findings and Conclusions Yellowstone is rich in wildlife. It is one of the last refuges in the United States for bison and elk and is also home to wolves, coyote, lynx, wolverines, trumpeter swans, bald eagles, and other wildlife. The FEIS acknowledges that wildlife and wildlife habitats are “highly valued park resources.” FEIS at 111. In addition to the protections offered by the Organic Act and Yellowstone Enabling Act, wildlife is also specifically protected by NPS’s snowmobile regulation. See 36 C.F.R. 2.18(c) (Snowmobiles are prohibited “except where designated and only when their use is consistent with the park’s natural, cultural, scenic and aesthetic values, safety considerations, and park management objectives, and will not disturb wildlife or damage park resources.”). The wildlife impact. analyses in the FEIS indicate that the WUP would result in “negligible to moderate adverse effects.” 2007 ROD at 22. The ROD concludes that the same impact range would result under the WUP as would with the other alternatives, including both the “snowcoaeh-only” alternative and the “expanded recreational use” alternative, which would allow up to 1025 snowmobiles per day. Id. According to the 2007 ROD, “monitoring shows that winter use did not contribute to wildlife effects at the population level under a wide range of snowmobile numbers (between 324 and 1400 per day at the West Entrance).” Id. Accordingly, the ROD concludes that “[s]etting the snowmobile limit lower, at 540 per day, ensures that there will continue to be no population level effects, and will also help limit any impacts to individual wildlife.” Id. NPS claims that commercial guiding also helps reduce any such effects, and the adaptive management program will continue to be used to mitigate any unacceptable impacts. Id. In reaching this conclusion, NPS relies upon several studies of wildlife in Yellowstone. See generally FEIS at 113-29. Impacts of each alternative were analyzed on the basis of five major concerns, including: 1) vehicle-caused mortality to individual animals; 2) displacement impacts; 3) behavioral responses of wildlife groups to OSVs and associated human activities; 4) physiological responses of wildlife groups to OSVs and associated human activities; and 5) demographic effects at the population level. Id. at 251. The FEIS acknowledges that increases in winter traffic levels and associated human recreational activity cause increases in each of the above adverse impact categories. Id. at 270. However, the FEIS concludes that mitigating these impacts through limiting the number of visitors, continued wildlife monitoring, and adaptive management will “limit impacts to acceptable levels.” Id. 2. Plaintiffs’ Claims Regarding Impacts on Wildlife Plaintiffs claim that the WUP will unlawfully “disturb” wildlife within the mean *203 ing of 36 C.F.R. § 2.18(c), that the impacts on wildlife caused by the Plan violate the Organic Act, and that the FEIS analysis is inadequate under NEPA. Specifically, Plaintiffs contend that NPS’s own studies indicate that wildlife are disturbed by snowmobile use and that NPS has failed to articulate why it has chosen not to follow the studies’ recommendations that would further limit OSV use. Plaintiffs also allege that the FEIS improperly minimizes the adverse impact of snowmobiles by analyzing only population-level impacts, rather than adverse impacts on individual animals. NPS and Montana State University conducted two studies to examine the nature of wildlife responses to interactions with OSVs. The first, Behavioral Responses of Bison and Elk in Yellowstone to Snowmobiles and Snow Coaches, Borkowski et al. (2006), (hereinafter “Borkowski study”), analyzed more than 6,500 interactions between elk or bison and OSVs over a five year period. The second study, Behavioral Responses of Wildlife to Snowmobiles and Coaches in Yellowstone, White et al. (2006), (hereinafter “White Study”), analyzed over 5,500 records of interaction from the 2003-2006 winter season. 3 Relying primarily on the White study, Plaintiffs claim that NPS arbitrarily disregarded the recommendation of its own scientists regarding the level of acceptable OSV traffic. The White study concluded that regulations restricting the levels and travel routes of OSVs during the last several years were effective at “reducing disturbances to wildlife below a level that would cause measurable fitness effects.” AR 125701 at 20. Accordingly, the White study recommended keeping “OSV traffic levels at or below those observed during the past three years of our study.” Id. The average daily entry seen during this period was 260 to 290 snowmobiles per day. See 2007 ROD at 20. In its briefs, NPS argues that it chose to lower the limit from 720 to 540 snowmobiles per day in response to the White study’s recommendation that use levels be kept at or below use from the 2003-2006 study period. In spite of the higher caps, NPS claims that “actual use levels will likely be similar to the OSV entry figures observed at the West Entrance during the last three years of the White Study.” Fed. Opp’n to GYC at 40 (emphasis in original); see also Fed. Reply to GYC at 22. NPS argues that the 540 limit corresponds to “peak use” and accordingly, there is no reason to expect that average use will increase over the current average of 260-290 entries per day. The Court is troubled by this argument, which persists throughout NPS’s briefs. See Fed. Defs.’ Reply to GYC at 22 (“Federal Defendants already have demonstrated that the lowering of the daily limit from 720 snowmobiles per day to 540 per day was done with the intention of keeping OSV use at roughly the same levels as the last three winters under the temporary rule.”). NPS appears to disavow the Rule it actually promulgated through repeated insistence that actual use will not likely increase. However, there is no evidence in the record to support the assertion that future use levels will remain as low as those seen in the past three winter seasons. The ROD acknowledges that the *204 lower use seen in past three winters resulted in part from confusion over whether the park was even open to snowmobile use under the temporary plan. Moreover, whether or not usage increases is not the test of the lawfulness of the WUP, rather, the correct inquiry is whether the 540 limit will cause unacceptable impacts on park resources and values. NPS’s repeated assertion that “actual use” will likely be lower than the limit does nothing to justify its conclusion that the 540 limit adequately protects wildlife. Plaintiffs also claim the FEIS is inadequate under NEPA because “the wildlife impacts metric is incapable of assessing the alternatives’ conformity with Park Service mandates.” GYC Mot. At 40. Specifically, Plaintiffs complain that the FEIS analyzes only impacts on populations, while NPS is required to prevent disturbance to individual animals, regardless of population-level effects. Plaintiffs contend that the FEIS simply “acknowledges that adverse impacts to individual animals should be minimized” but fails to actually analyze those impacts. Plaintiffs point to the impact definitions as an example of a population-centric analysis. An impact is considered “major” if it is an “action that will noticeably affect a population or individuals of a species; the effect will be measurable and have a substantial and possibly permanent consequence to the population.” FEIS at 257; see also NPCA Mot. at 24. Plaintiffs argue that even if NPS analyzed impacts on individual animals, it appears to only have relied upon population level impacts. See 2007 ROD at 22 (“Monitoring shows that winter use did not contribute to wildlife effects at the population level under a wide range of snowmobile numbers (between 324 and 1400 per day at the West Entrance.”)). The Court agrees. While the Administrative Record contains voluminous studies by wildlife experts, NPS fails again to explain why the expected adverse impacts on wildlife are acceptable, particularly as they relate to individual animals. The ROD notes that “80% of the time bison displayed no visible response to [OSVs], 12.5% of the time they displayed a ‘vigilant response,’ and 7% of the time they ‘actively’ responded, meaning they either walked or ran away or charged the vehicle.” 2007 ROD at 22. Elk displayed a vigilant response 44% of the time and an active response 8% of the time. Id. The FEIS explains that for each wildlife group studied, the “percentage observing a response (either movement or vigilance) was 83.3% for bald eagles, 60.5% for coyote, 52.4% for elk, 42.5 % for swans and only 19.6% for bison.” While the studies show that only 7% of bison displayed an “active” response, neither the FEIS nor the ROD explains why this impact on individual bison is “acceptable” or how this level of disturbance complies with 36 C.F.R. § 2.18(c). The fact that all alternatives studied resulted in the same determination of “negligible to moderate” impacts on wildlife prevents the decision maker from meaningfully evaluating the difference between the alternatives. If impacts on individual animals were in fact considered, it defies logic that zero snowmobiles (Alternative 2) and 1025 snowmobiles, including 250 that are unguided, (Alternative 4) could possibly produce the same impacts on individual animals in the park. See ROD at 22 (“The same impact range would occur with other alternatives.”); Compare FEIS at 261 (finding negligible to moderate impacts under Alternative 2) with FEIS at 264 (finding minor to moderate impacts under Alternative 4). Accordingly, the Court finds that the agency acted arbitrarily when it failed to explain why the impacts on individual animals do not violate 36 C.F.R. § 2.18(c), why a “moderate” impact on wildlife is *205 acceptable, and why the White study’s recommendation for lower use levels was rejected. C. Impacts on Air Quality 1. NPS Findings and Conclusions The FEIS states that “air quality is a key resource in itself as well as a highly prized (and expected) element of the park visitor experience.” FEIS at S-5. Impacts on air quality secondarily have impacts on human health and the quality of visitor experience. Id. at 95. As with soundscapes and wildlife, the 2000 FEIS found that under historic use air quality had been impaired in violation of the Organic Act. Though BAT snowmobiles are far cleaner than snowmobiles historically used in the parks, the ROD acknowledges that neither they, nor snowcoaches, are pollution free. 2007 ROD at 21. According to the ROD, the WUP will have moderate, adverse, long-term, direct, park-wide impacts, which are more adverse than current conditions. Id. at 33. Under the impact definitions, air quality impacts are considered “major” only if “[t]he impact is substantial and noticeable park-wide” whereas an impact is “minor” when “the impact on air quality is measurable, but localized within a relatively small area.” Id. A “moderate impact” is one that is “measurable and perceptible, possibly throughout the parks, but could be reversed and generally localized.” Id. Based on these definitions, the WUP is determined to have a “moderate, long-term, adverse and direct impact” on air quality. FEIS at 236. NPS defines a “long-term” impact as a “change in a resource or its condition that does not return to pre-disturbance levels and for all practical purposes is considered permanent.” FEIS at 168, Table 4-1. Amazingly, none of the alternatives evaluated in the FEIS were determined to impair air quality, even Alternative 4, which was modeled as generating 234% of the carbon monoxide emissions now present under current conditions. Id. at 236. The ROD concludes that “the level of air pollution under the decision will not harm the integrity of park resources and values and will not constitute unacceptable impacts or impairment.” 2007 ROD at 33. NPS reached this conclusion through both monitoring and modeling studies. NPS conducted winter emissions measurements in Yellowstone that involved collection of emissions data from in-use snowcoaches and snowmobiles, as well as conducted winter air quality monitoring in the Old Faithful developed area. Monitoring of meteorological, gaseous, and particulate data also took place at Yellowstone’s West Entrance. FEIS at 94. 2. Plaintiffs’ Claims Regarding Air Quality Impacts Plaintiffs bring several challenges to the analysis of the Plan’s impacts on air quality. Specifically, Plaintiffs allege that 1) the impact definitions inappropriately rely on park-wide impacts; 2) NPS does not justify its use of the federal or state air quality standards; 3) NPS inappropriately compares the WUP to historical conditions and obscures the negative impacts of the WUP compared to current conditions; and 4) the WUP understates adverse impacts on air quality. a. Impact Definitions and the Park-Wide Metric Plaintiffs contend that the FEIS is flawed with respect to air quality because the impact definitions used are inappropriate for measuring impairment or unacceptable impacts within the parks. GYC Mot. at 39. Specifically, Plaintiffs complain that the FEIS improperly distinguishes between “localized air quality impacts” and park-wide impacts. See FEIS at 222, Ta *206 ble 4-31. Plaintiffs claim that, again, use of a park-wide metric obscures the impact of each alternative on those areas where visitors and employees actually congregate during the winter season. NPS argues that it did not consider only park-wide effects because it conducted monitoring in four locations expected to generate the most elevated air quality impacts. Fed. Opp’n to GYC Mot. at 33. NPS also claims that it set adaptive management threshold levels well below both federal and state standards at which actions would be taken to ensure no impairment occurs. Id. at 34; see ROD at 41-42. NPS argues that monitoring would take place at specific locations throughout the park to determine whether the thresholds were reached and therefore, any response would be based on localized effects, not “park-wide.” While this is true, the impact definitions in the FEIS only find a “major impact” if park-wide impacts are felt. According to Table 4-45 in the FEIS, Alternative 4 is the only one that would constitute a “major” adverse impact in terms of emissions, but as with soundscapes, the FEIS concludes that even Alternative 4 would not cause impairment. FEIS at 236. Again, when the highest impact on the scale does not cause impairment, the Court and the public are left to guess what impacts possibly would. Though Alternative 4 was not selected, this non-impairment decision raises considerable questions about NPS’s assessment of impacts across all alternatives. Alternative 4 was predicted to have both “major” and “long-term” adverse effects. FEIS at 235. Given that major impacts are those that are substantial and noticeable park-wide, and long-term impacts are for all practical purposes permanent, it is inconceivable to the Court that these impacts do not constitute impairment. The FEIS provides no explanation for this conclusion and states only that “[ijmpairment of park resources would not occur; the level of air pollution under Alternative 4 would not harm the integrity of park resources and values.” Id. NPS recites (verbatim) the same non-impairment conclusion for Alternative 7, (the WUP), after stating that the WUP will have moderate, long-term, adverse effects on the park’s air quality. Again, NPS has utterly failed to explain how and why these impacts are acceptable. b. Use of Federal and State Air Quality Standards Plaintiffs also challenge NPS’s reliance on the National Ambient Air Quality Standards (NAAQS) and other state regulations that establish permissible amounts of air pollution. Plaintiffs claim that NPS has failed to explain the relationship between these air quality standards and the NPS’s independent duty to conserve the parks by preserving the “best possible” air quality. GYC Mot. at 40 (citing 2006 Mang. Policies § 4.7.1). Relying on Wilderness Soc’y, 434 F.3d at 596-596, NPS responds that the management policies are not binding, but rather only internal agency documents intended for guidance and therefore not judicially enforceable. NPS also contends that the state and federal air quality standards are entirely appropriate baselines for measuring air quality in the parks. As the Court previously noted, the Management Policies are not independently judicially enforceable, however, they are relevant insofar as NPS puts forth the Policies as justification for the decision under review. In this instance, NPS specifically notes that § 4.7.1 of the Policies was reviewed in formulating the WUP. 2007 ROD at 27. That section provides: “The Service will seek to perpetuate the best possible air quality in the parks to (1) preserve natural resources and systems; (2) preserve cul *207 tural resources; and (3) sustain visitor enjoyment, human health, and scenic vistas.” Id. Following this statement, the ROD states that, “[a]t present, with oversnow vehicle use levels similar to what would likely be experienced under this decision (including the air quality on the busiest days, with OSV numbers at or near the maximum that will be allowed under this decision), air quality in the parks is in full compliance with the requirements of the Clean Air Act and NPS Management Policies.” Id. Plaintiffs complain that NPS has equated the “best possible air quality” envisioned in the Policies with compliance with the NAAQS, and failed to explain why these baselines are appropriate for guaranteeing non-impairment. Though current levels of pollution are well below the NAAQS, Plaintiffs point out that the NAAQS were not exceeded even under the historic conditions that were found to constitute “impairment.” See FEIS at 94. Perhaps a more significant omission, however, is the failure by NPS to explain why current use levels are “similar to what would likely be experienced under this decision.” Yet again, NPS attempts to minimize the reality that the WUP allows nearly twice the number of snowmobiles than have entered Yellowstone under the temporary plan. NPS attempts to defend a rule it did not implement by arguing that the limits set forth in the WUP are not likely to be met, or will only be met over a few busy weekends per year. Fed. Defs.’ Opp’n to GYC at 23. This argument serves only to blur the distinctions between alternatives. NPS is obligated under NEPA and the APA to “[Vigorously explore and objectively evaluate all reasonable alternatives,” thereby “sharply defining the issues and providing a clear basis for choice among options by the decision-maker and the public.” 40 C.F.R. § 1502.14. There is simply no support in the record to justify equating current use levels with the expected impacts of the WUP. NPS is required to select a snowmobile limit that is best calculated to protect park resources and values. It is fundamentally arbitrary to set a higher limit and merely hope it will not be reached. c. Comparison to Historical Conditions Plaintiffs object to the use of “historical conditions” as a baseline for evaluating the various alternatives because those conditions no longer exist. While the WUP would be an improvement over the deplorable conditions seen prior to managed use, the WUP will have considerably worse impacts on air quality in the Park than those seen under current conditions. FEIS at 236. NPS responds that the WUP is explicitly compared to both historic conditions and current conditions and that “air quality will still be excellent in the park” despite the increased emissions under the WUP. See Fed Opp’n to GYC at 20. NPS further argues that in addition to the comparison to past and current conditions, the FEIS also considered an objective standard (NAAQS and state ambient air quality standards) to assess the potential impacts of each alternative. Fed. Opp’n to GYC at 33. The comparison to historical conditions is somewhat misleading because it does make every alternative considered appear to be “an improvement,” even the 1025 snowmobile limit in Alternative 4 is listed as “beneficial.” As Plaintiffs point out, NPS’s mandate is not to “improve” air quality over past impairment, but to ensure that current and future conditions will not suffer unacceptable impacts. NPS Policies §§ 1.4.3,1.4.7.1. d. The WUP Understates Adverse Impacts on Air Quality Finally, Plaintiffs also allege that the FEIS understates the likely adverse im *208 pacts on air quality from the WUP. To support this argument, Plaintiffs cite the discrepancies between modeled and monitored conditions and test results indicating elevated levels of benzene and formaldehyde under current conditions. Plaintiffs point to the discrepancies between monitored and modeled concentrations of particulate matter in Table 4-30 as evidence that the NPS model underestimates the actual effect of snowmobiles on air quality. See FEIS at 221 (indicating that monitored conditions reflected almost 2.5 times more particulate matter than predicted under the model). The model also fails to account for temperature “inversions,” which are common in Yellowstone, and dramatically increase pollution levels. NPS acknowledges that there are inaccuracies between the model and the monitoring data, but concludes that the modeled results “are within a reasonable range of possibility” and that no model will predict precisely the amount of emissions produced under any one alternative. FEIS at 220. Rather, NPS contends that “it is the magnitude of differences between alternatives as shown by modeling that is most useful in comparing one alternative to another.” Id. Plaintiffs also claim that the ROD understates the likely adverse impacts of the WUP in terms of benzene and formaldehyde exposure for park employees. The ROD acknowledges that under the temporary plan, “some monitoring indicated a potential problem with benzene and formaldehyde exposure for employees.” 2007 ROD at 21. This is alarming because the WUP allows nearly twice as many snowmobiles as entered the park during the monitoring period. According to Table 3-11 in the FEIS, a 2005 study found that the average benzene level at Kiosk A of Yellowstone’s West Entrance was 0.0035 parts per million (“ppm”) with an average daily entry of 180 snowmobiles. FEIS at 100. In 2006, the average benzene level at Kiosk A was 0.00325 ppm with a daily average of 216 snowmobiles. Id. Both of these levels exceed the chronic-duration minimum risk level (“MRL”) 4 of 0.003 ppm for benzene exposure. The 2006 study also found that the intermediate-duration MRL of 0.006 ppm was exceeded on two occasions. AR 117413 at 16. However, the FEIS misstates this key finding, implying that only the chronic-duration level was exceeded. The FEIS states, [T]he 2006 report confirmed employee exposures to be below all current standards set by regulatory agencies except for two of thirteen benzene samples (mean concentration of 0.0032 ppm). The MRL for chronic-duration (365 days/year) inhalation exposure is .003 ppm for benzene; the intermediate-duration inhalation exposure is .006 ppm and the PEL is 1.0 ppm. While the two benzene samples averaged slightly higher than the MRL, employees would have to be exposed to these levels every day of the year (which they are not) for a concern to be present. Rather, the two samples that were higher than 0.003 ppm were short term samples taken to minimize dilution effects and thereby obtain a better idea of potential worst case exposures. FEIS at 99. A review of the report indicates that the two exceedances referenced were of the intermediate threshold, not the lower chronic threshold. The report *209 provides, “The benzene samples with concentrations of 0.0072 ppm and 0.0086 ppm are above the intermediate-duration inhalation exposure of 0.006 ppm, but below the acute-duration inhalation exposure of 0.009 ppm.” AR 117413 at 16-18. Contrary to the statement in the FEIS, the report indicates that there were actually five instances where the sample was at or above 0.003 ppm. Id. at 18. While the report ultimately concludes that benzene levels are within legal limits and the samples represent “worst case scenarios,” the Court is troubled by the discrepancy between the report and the FEIS, and the appearance of attempting to minimize the recorded exceedances. NPS does not dispute the exceedances, but instead argues that the levels recorded would only constitute a health risk if they were experienced every day of the year, which they are not. FEIS at 99. NPS also acknowledges that four-stroke snowmobiles produce more benzene (and other toxic chemicals) than their older two stroke counterparts. Id. Defendants conclude that “recent benzene exposures are an order of magnitude lower than they were when two-strokes were allowed in the parks” but then acknowledge that the decrease is “possibly attributable to fewer numbers of snowmobiles.” Id. Plaintiffs argue that this conclusion is at odds with the decision to increase the number of snowmobiles allowed compared to those seen under current use levels when the thresholds were exceeded. Plaintiffs also complain that maximum health thresholds for formaldehyde have been approached. Again, NPS does not dispute this finding. See 2007 ROD at 21; FEIS at 100. In response, NPS proposes to “continue air quality and health and safety monitoring.” As Plaintiffs pointed out at the hearing, “these organic standards are used to assess health threats at hazardous waste sites, so for the administration to disregard them in Yellowstone is rather stunning.” Transcript, 48. Furthermore, these thresholds have been incorporated into the adaptive management thresholds in use during the temporary plan. See 2007 ROD at 46. Plaintiffs contend that NPS’s failure to take action in the face of past exceedances makes the promise of future adaptive management action all the more illusory. Transcript, 48. As with soundscapes and wildlife, the Court finds that NPS has failed to articulate why a plan that will admittedly worsen air quality complies with the conservation mandate. Defendants’ analysis of the air quality impacts violates NEPA primarily because Defendants have failed to explain why the NAAQS and the Class I Federal Standards for visibility constitute the adequate baselines for unacceptable impacts and non-impairment. These standards were not violated under the abysmal conditions found to constitute “impairment” in the 2000 FEIS, a conclusion with which the 2007 FEIS agrees. Accordingly, these standards do not provide a meaningful relationship between air quality and non-impairment. Defendants’ attempt to fall back on the much lower “adaptive management thresholds” is unavailing. Defendants themselves have indicated that the thresholds are not enforceable and not indicative of “impairment” or even “unacceptable impacts.” Rather, they are indicators that “conditions could be moving away from those that are desirable” and are only “preliminary in nature.” See AR 126499 at 40; Fed Opp’n to GYC at 15. Accordingly, by Defendants’ own admission, there is no standard for unacceptable impacts or impairment of air quality, which makes the determination that impacts will be “acceptable” fundamentally arbitrary. Y. Conclusion For the reasons stated above, the Court finds that the Winter Use Plan, as codified *210 in the Final Rule and explained in the 2007 ROD, is arbitrary and capricious, unsupported by the record, and contrary to law. In contravention of the Organic Act, the Plan clearly elevates use over conservation of park resources and values and fails to articulate why the Plan’s “major adverse impacts” are “necessary and appropriate to fulfill the purposes of the park.” NPS Policies § 1.4.3. NPS fails to explain how increasing snowmobile usage over current conditions, where adaptive management thresholds are already being exceeded, complies with the conservation mandate of the Organic Act. In violation of the APA, NPS also fails to provide a rational explanation for the source of the 540 snowmobile limit. Furthermore, the FEIS in support of the Plan does not provide the decisionmaker with a clear analysis of the alternatives that NEPA requires. Chief among its failings, the FEIS relies on admittedly inaccurate sound modeling data, employs a park-wide metric that dilutes the Plan’s impacts on soundscapes and air quality, and utterly fails to explain why none of the seven alternatives would constitute “impairment” or unacceptable impacts. According to NPS’s own data, the WUP will increase air pollution, exceed the use levels recommended by NPS biologists to protect wildlife, and cause major adverse impacts to the natural p soundscape in Yellowstone. Despite this, NPS found that the plan’s impacts are wholly “acceptable,” and utterly fails to explain this incongruous conclusion. Put simply, the WUP provides “no rational connection between the facts found and the choice made.” Alpharma, Inc. v. Leavitt, 460 F.3d 1, 6 (D.C.Cir.2006). While the Court will defer to an agency’s exercise of expertise, the “Court will not defer to the agency’s conclusory or unsupported assertions.” McDonnell Douglas Corp. v. U.S. Dep’t of the Air Force, 375 F.3d 1182, 1187 (D.C.Cir.2004). Accordingly, Plaintiffs’ Motion for Summary Judgment is GRANTED and Defendants’ Cross-Motion for Summary Judgment is DENIED. The Winter Use Plan, 2007 ROD, and 2007 FEIS are vacated and remanded to the agency for proceedings consistent with this opinion. An appropriate Order accompanies this Memorandum Opinion. SO ORDERED. Footnotes [1] Defendants concede that' § 1.4 serves as NPS’s official interpretation of the Organic Act and is therefore enforceable against NPS. See Fed. Reply to GYC at 5, n. 5. Relying on Wilderness Society v. Norton, 434 F.3d 584, 596-97 (D.C.Cir.2006), however, Defendants contend that other portions of the NPS Policies, which are not official interpretations of a statute or regulation, are not judicially enforceable. The Court agrees. Accordingly, the Court will focus solely on § 1.4 of the NPS Policies as the official and enforceable interpretation of the Organic Act. However, the Court finds that statements made in other portions of the NPS Policies may be relevant to the same extent that NPS relied upon those statements in making the decisions under review. See 2007 ROD at 27 (Relying on NPS Policies §§ 4.7.1, 4.9, and 8.2). As other courts in this district have done, this Court "will refer to the NPS Policies insofar as the decisions under review have done so.” Sierra Club v. Mainella, 459 F.Supp.2d 76, 79 n. 1 (D.D.C.2006). [2] NPS also compared each of the alternatives to "current conditions” and "historical conditions.” According to the Final Sound Modeling Report, "current conditions” were modeled at an average of 260 snowmobiles per day and twenty-nine snowcoaches per day. AR 125352 at 31. Historical conditions were modeled at 1400 snowmobile entries per day and forty snowcoaches. Id. The Sound Modeling Report indicates that these figures were provided by the NPS as estimates of actual use. However, the 2007 ROD states that the historical average was 795 snowmobiles per *198 day. 2007 ROD at 20. Accordingly, it is unclear why the model based the results for “historic conditions” on a daily average of 1400, almost twice the actual daily average under "historic conditions.” This major discrepancy between the definition of historical conditions in the modeling study and the ROD undermines any arguable usefulness of employing historical conditions as a point of comparison for the evaluation of the alternatives. [3] The winter season in Yellowstone runs from December to March. The White et al. report labels winter seasons according to the year in which the season ended. For example, the 2002-2003 winter season is known as "winter 2003.” Accordingly, the data collected between the 2003 and 2006 winters actually spans four winter seasons. The report specifically indicates that the recommended level of OSV traffic should be kept at or below that observed "during the past three years” of monitoring, which are the years in which the temporary plan was in effect. [4] The MRLS are set by The Agency for Toxic Substances and Disease Registry (ATSDR). “An MRL is an estimate of the daily human exposure to a hazardous substance that is likely to be without appreciable risk of adverse non-cancer health effects over a specified duration of exposure.” AR 117413 at 16. These estimates are used to identify contaminants and potential health effects that may be of concern at hazardous waste sights. Id.