GRABER, Circuit Judge.
Plaintiffs, The Wilderness Society and the Alaska Center for the Environment, challenge a decision by Defendant United States Fish and Wildlife Service (the Service) to permit a sockeye salmon enhancement project (the Project) at Tustumena Lake. Tustumena Lake is located in Alaska, within a designated wilderness area in the Kenai National Wildlife Refuge. Plaintiffs argue that the Project violates the Wilderness Act, 16 U.S.C. §§ 1131-1136, because it contravenes that Act’s requirement to preserve the “natural condition” and “wilderness character” of the area, and because it constitutes an impermissible “commercial enterprise” within a wilderness area. Plaintiffs also allege that the Project violates the National Wildlife Refuge System Administration Act of 1966, 16 U.S.C. §§ 668dd-668ee (the Refuge Act), because it is not “compatible” with the purposes of the Refuge Act.
The district court denied Plaintiffs’ motion for summary judgment and entered judgment in favor of Defendant. We have jurisdiction pursuant to 28 U.S.C. § 1331. We affirm.
BACKGROUND
A. Legal Framework
In 1941, President Franklin D. Roosevelt issued an Executive Order that designated 2 million acres on the Kenai Peninsula, including Tustumena Lake, as the Kenai National Moose Range. As the name suggests, the original purpose of the Range was to protect the breeding grounds of the giant Kenai Moose. Exec. Order No. 8979, 6 Fed. Reg. 6471 (Dec. 16, 1941).
In 1964, Congress passed the Wilderness Act, which was intended to designate certain federal lands as “wilderness” areas and to require the “preservation and protection” of those lands “in their natural condition.” 16 U.S.C. § 1131(a). In 1966, Congress passed the Refuge Act. Congress stated several reasons for enacting the Refuge Act; among them was “the purpose of consolidating the authorities relating to the various categories of areas that are administered ... for the conservation of fish and wildlife.” 16 U.S.C. § 668dd(a)(l). All areas that had been designated for wildlife protection were consolidated into the “National Wildlife Refuge System.” Id.
In 1980, Congress passed the Alaska National Interest Lands Conservation Act (ANILCA), Pub. L. No. 96^87 (Dec. 2, 1980), 94 Stat. 2371 (codified as 16 U.S.C. §§ 3101-3233 and scattered sections of the U.S.C.). ANILCA controls the management of refuge lands in Alaska, superseding any conflicting provisions of the Refuge Act. Pub. L. No. 105-57, § 9(b) (Oct. 9, 1997), 111 Stat. 1260. Congress enacted ANILCA intending, among other things, to protect and manage fish and wildlife in a manner that permits those Alaska resi[917]*917dents who are “engaged in a subsistence way of life to continue to do so.” 16 U.S.C. § 3101(c). ANILCA added nearly a quarter of a million acres to the Kenai National Moose Range and renamed it the Kenai National Wildlife Refuge (the Refuge). ANILCA also designated 1.35 million acres of the Refuge as the Kenai Wilderness. ANILCA §§ 302(4)(a), 702(7); 16 U.S.C. § 1132 (notes).
B. The Project
Tustumena Lake is located in the Kenai Wilderness portion of the Refuge. Tus-tumena Lake is the fifth largest freshwater lake in Alaska, measuring about 40 kilometers by 8 kilometers. The lake is fed both by clear-water streams and by glacial creeks originating in the Harding Icefield. The lake’s outlet is the Kasilof River, which in turn drains into the Cook Inlet and hence into the Gulf of Alaska. The Kasilof River watershed supports several species of fish, including sockeye salmon.
For at least one hundred years, intense commercial fishing has occurred in the Cook Inlet. Before Alaska became a state, salmon were fished using stationary traps, which were placed to intercept the salmon as they returned to their birthplaces to spawn. The traps were left open for up to five days at a time. Today, commercial fishing fleets still intercept and harvest salmon, as the salmon return to Tustume-na Lake (and elsewhere) from the ocean via the Kasilof River and other creeks and streams.
Beginning in 1974, the Alaska Department of Fish and Game (the Alaska Department) collected sockeye salmon eggs annually at Tustumena Lake as part of a state-funded research project. The eggs were incubated at the Crooked Creek Hatchery,1 and the resulting salmon fry were later released. Beginning in 1975, the fry were returned to tributaries of Tustumena Lake, as well as to other sites. Between 1976 and 1987, the number of fry released annually at Tustumena Lake varied between 400,000 and 17 million. Each year after 1987 the number of fry released there has remained at about 6 million.2 Until 1980, the Project was conducted without a special use permit, or other kind of permit, from the Service.
In 1980, after ANILCA designated the Refuge and the Kenai Wilderness, the Service’s Refuge Manager notified the Alaska Department that special use permits would be required from then on for projects in the wilderness area. The federal and state agencies negotiated an agreement to continue allowing the Tustumena Lake salmon enhancement program as a research project, subject to annual approval.
In January 1985, as required by ANIL-CA, the Service issued a Final Comprehensive Conservation Plan, Environmental Impact Statement, and Wilderness Review (Final Plan) for the Refuge. Kenai National Wildlife Refuge Final Comprehensive Conservation Plan (Jan. 1985); see ANILCA § 304(g). Approved after an opportunity for public notice and comment, the Final Plan identified as a “significant problem” the off-Refuge commercial and recreational harvest of adult salmon:
[T]his heavy commercial harvest significantly reduces the number of salmon entering rivers to spawn decreasing sport-fishing opportunities both on and off the refuge. Furthermore, the re-[918]*918dueed run sizes can affect other species and aquatic habitat quality in several ways. Fewer living adult salmon are available to predators such as brown bear and eagles. Fewer carcasses are available to scavengers and decomposing organisms that recycle nutrients and maintain the fertility of aquatic habitats. Also, because fewer eggs are laid and fewer fry are hatched, the food supply of a variety of fish, birds, and mammals that feed on salmon fry is reduced.
In addition, commercial fishery management greatly tempers the fluctuations in run sizes that would occur naturally due to variations in spawning and hatching success and survival in freshwater nursery areas and the ocean. It is not clear whether this impact is significant. All of these impacts warrant increased study and assessment of long term effects.
Final Plan at 10-11.
In part because of the need for more information, the preferred alternative (Alternative C) called for the continuation of ongoing fish studies. Alternative C also noted that
[a] variety of fishery enhancement efforts would be undertaken although natural processes would continue to dominate fish production on most areas of the refuge. Management efforts would focus on the refuge’s most popular fisheries. Most current fishery management and research projects, including ... the experimental stocking of sockeye fry in Tustumena and Hidden lakes, ... could continue. Abundance and diversity of salmon ... probably would remain near current levels.
Id. at 128 (emphasis added). The Refuge would continue to supply eggs, including salmon eggs from Tustumena Lake, for state fish hatcheries. The overall effort of the fisheries management would be “to achieve a balance of active management, wise use, and conservation of fishery resources,” with active management benefiting certain species in certain waters. Id.
Additionally, the Final Plan said the following about maintaining “[n]atural diversity of wildlife populations”: “This purpose will be achieved by insuring that fish ... populations and habitats in refuge wilderness and minimal management areas will be regulated or appear to be regulated primarily through natural processes.” Id. at 95 (emphasis added). Specifically with respect to fish habitats, the Final Plan sought to protect these locations “to ensure the production, on a continuing basis, of the recreational and commercially important fish populations.” Id. at 96. Importantly:
Populations of species of high public interest will continue to receive management emphasis in the most highly accessible locations. Populations of species such as ... salmon, will be maintained at relatively constant, high levels. The substantial recreational and commercial benefits ( ... fishing ... and commercial fishing) associated with these species will continue.
... The refuge will continue to serve as a donor of fish and wildlife stocks (i.e., fish eggs for hatcheries).
Id. (emphasis added).
In 1989, the Alaska Department and the Service agreed that a decision would be made within four years either to discontinue the Tustumena Lake research project or to elevate it to an operational “commercial enhancement project.” To that end, a report by the Alaska Department, released in 1992, requested that the Project become operational instead of experimental.
In 1993, the Alaska Department contracted with the Cook Inlet • Aquaculture Association (CIAA) to staff and operate the Crooked Creek Hatchery and its various programs. The CIAA is a private, nonprofit corporation “organized ... for [919]*919the purpose of engaging in salmon enhancement work throughout the Cook Inlet Region.” Its mission is as follows:
The [CIAA] is a non-profit regional association which exists to: (1) protect self-perpetuating salmon stocks and the habitat, upon which they depend; (2) rehabilitate self-perpetuating salmon stocks; (3) rehabilitate salmon habitat[;] and (4) maximize the value of the Cook Inlet (Area H) common property salmon resource by applying science and enhancement technology to supplement the value attained from protection and habitat rehabilitation of self-perpetuating salmon stocks.
The CIAA relies on funding from two sources: (1) a voluntary, self-imposed “2 percent tax” on the value of the annual salmon harvest by Cbok Inlet commercial fishers, and (2) monies recovered from its own harvesting of “supplementally produced fish.”
In May 1994, the Regional Director of the Service contacted the Alaska Department about the Project’s future. Acknowledging that the Project had moved beyond the experimental stage and had become a fishery management measure, the Regional Director determined that a decision had to be made regarding the Project’s future. The director listed several environmental concerns that required careful evaluation. Staff members from the Service, the Alaska Department, and the CIAA agreed that the CIAA would prepare a draft Environmental Assessment (EA) for the Project.
In June 1995, the Service produced a Fishery Management Plan for the Refuge. That Plan discussed the Project in detail, identifying the controversy surrounding an enhancement activity within a wilderness area. Without reaching a recommendation, the Fishery Plan stated' that the “[operational status of the [Project] will be determined through the National Environmental Policy (NEPA) process.” Fishery Management Plan, Kenai National Wildlife Refuge at 29 (June 1995).
In late 1995, the CIAA submitted a draft EA to the Service for review and comment. The draft considered five action alternatives, ranging from eliminating the Project to increasing it threefold, and recommended that the Project continue at the same level, with an annual stocking of about 6 million fry.
The Service reviewed and circulated comments on CIAA’s draft EA. The Alaska Department also reviewed the draft. In June 1997, the Service and the CIAA jointly released a draft EA. During a 45-day public review period, both supporters and opponents voiced their opinions about the Project’s future. Plaintiff Wilderness Society commented and urged the Service to eliminate the Project altogether.
Several developments occurred in August 1997:
• A final EA was released.
• At the same time, the Service issued a “Mitigated Finding of No Significant Impact,” concluding that mitigation measures contained in the special use permit would minimize risks associated with the Project, so that preparation of an Environmental Impact Statement was not required.
• The Refuge Manager released a “Wilderness Act Consistency Review” posing, and answering in the negative, two questions: (1) whether the Project was inconsistent with the Wilderness Act’s requirement to maintain the natural condition of the wilderness, and (2) whether the Project constituted a commercial activity precluded by regulation, 50 C.F.R. § 35.5. Relying on an opinion prepared by the Interior Department’s Solicitor’s Office, the Refuge Manager found the Project consistent with relevant environmental statutes, including both the [920]*920Wilderness Act and the Refuge Act. The Solicitor’s opinion concluded that, “[ujnder present law, the Fish and Wildlife Service has administrative discretion sufficient to grant Cook Inlet Aquaculture Association a ‘special use permit’ for operation of a compatible enhancement project within the Kenai Wilderness.” The opinion explained that the Wilderness Act was a legislative compromise that did not mandate pure preservationism. With respect specifically to wilderness areas in Alaska, the opinion noted Congress’ intention to allow activities enhancing fish populations.
• Finally, the Service also issued a “Compatibility Determination” acknowledging that “the [PJroject cannot ... be considered as supporting Refuge purposes,” but concluding that the Project still was not “incompatible with” Refuge purposes.
Following the issuance of those documents, in August 1997 the Service issued a Special Use Permit to the CIAA for the Project. Under the permit, Project activities occur twice a year. Each summer, a temporary camp is constructed at the mouth of Bear Creek, which flows into Tustumena Lake. Over a period of several weeks in July and August, a temporary weir captures approximately 10,000 fish. The fish yield about 10 million sockeye salmon eggs. Those eggs incubate at the Trail Lakes Salmon Hatchery, outside the wilderness area. The following spring, stocking is accomplished by the release of about 6 million fry to the mouth of Bear Creek, where they remain for at least three hours to imprint the characteristics of the creek on the fry. The permit eon-tains several pages of special conditions and required mitigation measures designed to address concerns about the effect of hatchery fry on the genetic makeup of wild salmon.
In 1998, Plaintiffs filed a complaint seeking to enjoin the Project and seeking a declaration that the Service’s decision to permit the Project violated the Wilderness Act, ANILCA, the Refuge Act, NEPA, and the Administrative Procedure Act (APA). Plaintiffs moved for summary judgment on their Wilderness Act and Refuge Act claims, which are the only claims that are before us on appeal. Specifically, plaintiffs argued that the Project (1) violates the Wilderness Act, which requires preservation of the “wilderness character” and “natural conditions” of the Kenai Wilderness, 16 U.S.C. § 1131(a), (c), and prohibits “commercial enterprise” in the wilderness area, 16 U.S.C. § 1133(c); and (2) violates the Refuge Act’s requirement of “compatibility]” with Refuge purposes, 16 U.S.C. § 668dd(d)(l)(A). The district court disagreed with Plaintiffs’ theories under both statutes and denied their motion for summary judgment. After considering a supplemental record, the court again denied Plaintiffs’ motion. The parties stipulated to voluntary dismissal of the claims as to which Plaintiffs had not sought summary judgment. The district court dismissed those claims without prejudice and entered a final judgment in favor of the Service.3
This timely appeal followed.
STANDARDS OF REVIEW
Our review of a district court’s order granting or denying summary judgment is [921]*921de novo. Bianchi v. Walker, 163 F.3d 564, 569 (9th Cir.1998).
The APA, 5 U.S.C. § 706, governs judicial review of agency action. We may set aside an agency’s action only if it is “ ‘arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.’ ” United States v. Snoring Relief Labs Inc., 210 F.3d 1081, 1085 (9th Cir.2000) (quoting 5 U.S.C. 706(2)(A)).
ANALYSIS
A. Level of Deference Owed to the Agency’s Action
As an initial matter, the parties disagree about how much deference we owe to the Service’s decision to permit the Project. Plaintiffs challenge the Service’s interpretation of its own statutory mandate. Under United States v. Mead Corp., 533 U.S. 218, 221, 121 S.Ct. 2164, 150 L.Ed.2d 292 (2001), we must first assess whether the permit is the type of agency decision that Congress intended to “carry the force of law.” If so, we owe Chevron4 deference to the agency’s interpretation.
[2] In Mead, the Court concluded that a letter ruling issued by the Customs Service, interpreting a regulation defining the appropriate category for a certain type of goods, did not warrant Chevron deference. The Court explained that deference lies on a continuum based on the nature of the agency’s action. At one end of that continuum, Chevron deference applies when Congress has delegated to the agency authority “generally to make rules carrying the force of law, and ... the agency interpretation claiming deference was promulgated in the exercise of that authority.” Mead, 533 U.S. at 226-27, 121 S.Ct. 2164. For example, the highest deference applies to actions that are the result of “express congressional authorizations to engage in the process of rulemaking or adjudication that produces regulations or rulings for which deference is claimed.” Id. at 229, 121 S.Ct. 2164. At the other end of the continuum, an agency interpretation “advanced for the first time in a litigation brief’ is due almost no deference at all. Id. at 228, 121 S.Ct. 2164 (citing Bowen v. Georgetown Univ. Hosp., 488 U.S. 204, 212-13, 109 S.Ct. 468, 102 L.Ed.2d 493 (1988)).
Congressional delegation of authority “may be shown in a variety of ways.” Id. at 227, 121 S.Ct. 2164. “Congress contemplates administrative action with the effect of law when it provides for a relatively formal administrative procedure tending to foster the fairness and deliberation that should underlie a pronouncement of such force.” Id. at 230, 121 S.Ct. 2164. Athough notice-and-comment rulemaking is a good indicator that Chevron deference is warranted, the absence of notice-and-comment rule-making does not foreclose the application of Chevron deference. Id. at 230-31, 121 S.Ct. 2164; Barnhart v. Walton, 535 U.S. 212, 122 S.Ct. 1265, 1271-72, 152 L.Ed.2d 330 (2002).
As the Court itself recognized, “the limit of Chevron deference is not marked by a hard-edged rule.” Mead, 533 U.S. at 237 n. 18, 121 S.Ct. 2164. After Mead, we are certain of only two things about the continuum of deference owed to agency decisions: Chevron provides an example of when Chevron deference applies, and Mead provides an example of when it does not. Id. In those “other, perhaps harder, cases” that do not clearly track either Chevron or Mead, we must “make reasoned choices between the two examples, the way courts have always done.” Id.
[922]*922Reasoning through the criteria for determining the level of deference applicable to agency action, we conclude that Chevron deference applies to the Service’s decision to permit the Project. First, Congress unquestionably delegated to the Service the authority to manage the Kenai Wilderness, including the authority to issue regulations permitting compatible uses within the Refuge. 16 U.S.C. §§ 668dd(d)(l)(A), 1133(b), 3124. Second, the permit was issued after the public had an opportunity to comment on the EA prepared for the Project. Third, the decision to permit the Project is consistent with the Service’s Final Plan for the Refuge, which itself is undoubtedly owed Chevron deference. The Final Plan is analogous to a rule; Congress expressly mandated its preparation, and the Service had to comply with notice-and-comment rulemaking procedures in preparing the Plan. ANILCA § 304(g)(1). Finally, the permit was issued after NEPA procedures were followed. Although the permit is not a rule, NEPA provides a “relatively formal administrative procedure tending to foster ... fairness and deliberation.” Mead, 533 U.S. at 230, 121 S.Ct. 2164.
Even if we were convinced that the permit is more analogous to the letter in Mead, the Service’s decision still would merit respect.5 “[A]n agency’s interpretation may merit some deference whatever its form,” id. at 235, 121 S.Ct. 2164, depending on a number of factors that imbue the interpretation with the “ ‘power to persuade,’ ” id. at 228, 121 S.Ct. 2164 (quoting Skidmore v. Swift & Co., 323 U.S. 134, 140, 65 S.Ct. 161, 89 L.Ed. 124 (1944)). Under Skidmore, the Service’s decision still would warrant respect because of its persuasiveness. The Service has considered the issue thoroughly. Its reasoning is not unsound. The Service had adequate information to permit the Project to operate at its current level, consistent with its operation since 1974. Even if we were to conclude that the Service’s decision was “beyond the Chevron pale,” we nevertheless would respect the decision because it has the “power to persuade” as defined in Skidmore. Mead, 533 U.S. at 234-35, 121 S.Ct. 2164; see Pronsolino v. Nastri, 291 F.3d 1123, 1135 (9th Cir.2002) (concluding that “[i]n the end, though, it does not much matter in this case whether we review the EPA’s position through the Chevron or Skidmore/Mead prism. Under both the more and less rigorous versions of the judicial review standard, the Agency’s position is ... more than sufficiently supported by the statutory materials.”).
B. Chevron Analysis
Having held that Chevron establishes the appropriate level of deference, we turn to the familiar two-step analysis that the Supreme Court prescribes: (1) If the statute is unambiguous and the intent of Congress is clear, we — and the Service — must give effect to that unambiguous intent; but (2) if the statute is ambiguous, Congress implicitly has left a gap for the Service to fill. Chevron, 467 U.S. at 842, 104 S.Ct. 2778. “In such a case, a court may not substitute its own construction of a statutory provision for a reasonable interpretation made by the administrator of an agency.” Id. at 844, 104 S.Ct. 2778.
[923]*923For the reasons explained below, we conclude that the statutes in question are materially ambiguous. Congress has not expressly prohibited fishery enhancement within wilderness areas. To the contrary, in some circumstances enhancement projects are expressly allowed. See Forest Guardians v. Animal & Plant Health Inspection Serv., 309 F.3d 1141, 1142 (9th Cir.2002) (per curiam) (noting that predator control is not expressly prohibited in determining that the Forest Service may control mountain lion populations within a wilderness). The Service has interpreted its statutory mandate in a manner that permits the Project. See 16 U.S.C. § 668dd(d)(l)(A) (delegating authority to the Service to permit compatible uses). Its interpretation is a reasonable one in view of the statutory ambiguities. That conclusion ends our inquiry. See Chevron, 467 U.S. at 843 n. 11, 104 S.Ct. 2778 (explaining that we “need not conclude that the agency construction was the only one it permissibly could have adopted to uphold the construction, or even the reading [we] would have reached if the question initially had arisen in a judicial proceeding”).
C. Wilderness Act: “Wilderness Character” and “Natural Conditions”
The Wilderness Act requires the Service to “preserv[e] the wilderness character” of the Refuge. 16 U.S.C. § 1133(b). In that statute, “wilderness” means
an area where the earth and its community of life are untrammeled by man, where man himself is a visitor who does not remain. An area of wilderness is further defined to mean in this chapter an area of undeveloped Federal land retaining its primeval character and influence, without permanent improvements or human habitation, which is protected and managed so as to presewe its natural conditions and which (1) generally appears to have been affected primarily by the forces of nature, with the imprint of man’s work substantially unnoticeable; (2) has outstanding opportunities for solitude or a primitive and unconfined type of recreation; (3) has at least five thousand acres of land or is of sufficient size as to make practicable its preservation and use in an unimpaired condition; and (4) may also contain ecological, geological, or other features of scientific, educational, scenic, or historical value.
16 U.S.C. § 1131(c) (emphasis added).
The foregoing definition is ambiguous in two important ways. First, when read as a whole, it is clear that a “wilderness” is not absolutely off limits to all human interference. By implication some human activities are to be allowed. A “visitor” can come if the visitor does not “remain”; how long is a “visit” and for what purpose may a “visit” take place? Only “permanent” improvements and human habitation are forbidden; what kinds of “impermanent” improvements and human habitation may occur? The area “generally” must be affected “primarily” with the forces of nature and with human imprints “substantially” unnoticeable; what “secondary” forces may come into play, and how “insubstantial” must human imprints be?
Second, it is not obvious how an agency must protect and manage an area “so as to preserve its natural conditions.” In general, the term “natural” means wild, formed by nature, and not artificially made. See Black’s Law Dictionary 1026 (6th ed. 1990). However, there are two plausible inferences to be drawn from the quoted phrase in this context. See FDA v. Brown & Williamson Tobacco Corp., 529 U.S. 120, 132, 120 S.Ct. 1291, 146 L.Ed.2d 121 (2000) (“The meaning — or ambiguity — of certain words or phrases may only become evident when placed in context.”). On the one hand, to preserve the “natural condi[924]*924tions” of the Refuge could mean protecting against the introduction of artificial propagation programs, like the Project, that alter the natural ecological processes within the Refuge. On the other hand, to preserve the “natural conditions” of the Refuge could mean preserving the natural ecological processes as they would exist in their wild state, in the absence of artificial disturbance from outside the wilderness area.
The difference between those two interpretations is profound. If, for example, hunting outside the wilderness area threatens a particular animal with extinction, and that animal used to be plentiful in the wilderness area, under the former interpretation no project to reintroduce the animal into the wilderness area can occur if it involves any artificial process (e.g., trapping and artificial insemination). By contrast, under the latter interpretation interventions can occur if they restore the “natural” balance of the wilderness area as it would exist without the external human forces that have altered it. Both interpretations are plausible and permissible.
Indeed, the definition’s use of the phrase “protected and managed” highlights this ambiguity. “Management” suggests affirmative steps taken to maintain wilderness character, while “protection” suggests a more hands-off approach. Compare Webster’s Third New Int’l Dictionary 1372 (unabridged ed. 1993) (defining “manage” as “to adjust the ecological factors to best meet the needs and ensure the survival of (a wild animal) usu. by controlling predators and hunting and by providing shelter or supplementary food supplies”), with id. at 1822 (defining “protect” as “to cover or shield from that which would injure, destroy, or detrimentally affect”). If “natural conditions” may be preserved only through a program of strict nonintervention, what is the purpose of the word “managed” in the definition? See City of Los Angeles v. United States Dep’t of Commerce, 307 F.3d 859, 870 (9th Cir.2002) (stating that it is “ ‘a cardinal principle of statutory construction that a statute ought, upon the whole, to be so construed that, if it can be prevented, no clause, sentence, or word shall be superfluous, void, or insignificant’ ” (quoting TRW Inc. v. Andrews, 534 U.S. 19, 122 S.Ct. 441, 448, 151 L.Ed.2d 339 (2001))). If strict nonintervention was Congress’ intent, the word “protect” would have sufficed.
A reasonable interpretation of this ambiguity is that a “wilderness” does not exist in a vacuum. Human activities outside the wilderness continue, with effects that most certainly are felt within the wilderness area. While the wilderness must be “protected” so that its natural processes dominate, it also must be “managed” so that human activities from outside the area do not interfere unduly. See Daniel Rohlf & Douglas L. Honnold, Managing the Balances of Nature: The Legal Framework of Wilderness Management, 15 Ecology L.Q. 249, 259 (1988) (discussing the statutory direction to wilderness managers and noting that, “[significantly, Congress phrased this preservation mandate affirmatively, suggesting that wilderness managers may be obligated to take affirmative actions to preserve or even restore wilderness character” (emphasis added) (footnote omitted)).
In the face of these ambiguities, the Service permissibly interpreted 16 U.S.C. § 1133(b). When mitigation measures are undertaken, the Project is not contrary to the agency’s charged mandate to “pre-serv[e] the wilderness character” and the “natural conditions” of the Refuge. The temporary camp and weir are minor intrusions that reasonably can be understood to be “substantially unnoticeable.” 16 U.S.C. § 1131(c). In fact, the Final Plan expressly recognizes that fly-in tent camps could [925]*925exist in the area. Final Plan at 92. The Project reasonably can be seen as a legitimate measure taken to restore fish runs to their “natural conditions,” that is, to make the runs as plentiful as they would be in the absence of interference from outside the Refuge. The statute is reasonably susceptible to the agency’s interpretation, which is all that we require. See Brandt-Erichsen v. United States Dep’t of Interior, 999 F.2d 1376, 1381 (9th Cir.1993) (affirming an agency decision when the agency’s interpretation of a statutory term was “reasonable”).
We are particularly persuaded by the fact that the Final Plan for the Refuge supports the Service’s decision to permit the Project. The Project addresses one of the most significant problems identified by the Final Plan — commercial fishing outside the Refuge. The Final Plan determines that a reduced salmon population within the Refuge, caused by commercial fishing outside the Refuge, endangers the viability of other species and the aquatic fertility of waters within the Refuge. Brown bears, bald eagles, and other wildlife rely on the salmon runs for their very existence. After more than 25 years of the Project’s enhancements, the ecosystem may have come to depend on the “supplemental” fish that the Project provides.
Under the preferred management alternative, the Final Plan contemplates that the Project would continue. As we already have explained, the Final Plan was adopted after a period of notice and comment. Congress explicitly mandated its preparation. As a result of these formal procedures, the Final Plan is entitled to the highest form of deference. See Mead, 533 U.S. at 229, 121 S.Ct. 2164.
The 1995 Fishery Management Plan identified concerns, both from the public and from within the Service, about the long-term effects of the Tustumena Lake stocking program. Nevertheless, the Plan recognized that “[ejnhancement programs can be an effective management tool in increasing the numbers of fish available to commercial fishermen and in increasing sport fishing opportunities.” Fishery Management Plan at 53. Moreover, the Fishery Plan concluded 'that the controversy would be resolved through NEPA’s procedures, which are not at issue in this appeal.
Permitting the Project is consistent with the Final Plan. More importantly, it is consistent with the statutory mandate in the Wilderness Act to preserve the area’s “natural conditions.” Although Plaintiffs would prefer an alternative reading of the statute, the Service’s chosen reading is permissible. Therefore, we must uphold the decision of the Secretary.
D. Refuge Act
1. The Project’s “Compatibility” With the Major Purposes of the Refuge
The Refuge Act tells the Service to permit only those uses within the Refuge that “are compatible with the major purposes” for which the area was established. 16 U.S.C. § 668dd(d)(l)(A). Congress has clearly delegated to the Secretary the authority to determine whether a use is “compatible” with those purposes. Id. (“The Secretary is authorized, under such regulations as he may prescribe, to ... permit the use of any area within the System for any purpose ... whenever he determines that such uses are compatible with the major purposes for which such areas were established.” (emphasis added)). The definition of “compatible use” similarly confers broad discretion: “ ‘compatible use’ means a wildlife-dependent recreational use or any other use of a refuge that, in the sound professional judgment of the Director, will not materially interfere with or detract from the ... purposes [926]*926of the refuge.” 16' U.S.C. § 668ee(l) (emphasis added).6
Plaintiffs argue that the Project is incompatible with the first stated purpose of the Refuge: to “conserve fish and wildlife populations and habitats in their natural diversity.” ANILCA § 303(4)(B)(i).7 The Service defines “natural diversity” to mean the “number and relative abundance of native species which would occur without human interference.” Final Plan at 174.
Again, the ambiguity in the text is apparent. Is the purpose of the Refuge to conserve the number and relative abundance of native species within the Refuge as they would be without human interference from outside the Refuge? If so, then a restoration program may be “compatible.” Or, is the purpose of the Refuge to leave the Refuge wholly untouched so that, whatever a species’ “natural diversity” is or becomes, no human interference may occur? Aso, does “diversity” refer to the number of species within a Refuge, or to genetic diversity within a single species?
Perhaps more important to our analysis is the fact that the statute refers not to a single purpose, but instead refers in the plural to “the major purposes” for which the refuge is established. 16 U.S.C. § 668dd(d)(l)(A). When the entire Refuge Act is considered, it becomes apparent that one of the other major purposes may be to maintain the size of a population of fish even if that result requires artificial aquaculture activities in a refuge. The Refuge Act contains the following relevant provisions:
• One purpose of the Act is the “conservation of fish ..., including species that are threatened with extinction.” 16 U.S.C. § 668dd(a)(l).
• The mission of the [National Wildlife Refuge] System is “to administer a national network of lands and waters for the conservation, management, and where appropriate, restoration of the fish, wildlife, and plant resources and their habitats.... ” Id. § 668dd(a)(2) (emphasis added).
• The System must be administered not only to “provide for the conservation of fish,” but also to “ensure that the biological integrity, diversity, and environmental health of the System are maintained.” Id. § 668dd(a)(4).
[927]*927The Service’s decision that the Project is “compatible” with the purposes of the Refuge is entitled to deference. Considering the ambiguities within the list of statutory purposes and within the underlying definition of “natural diversity,” the Service’s conclusion that the Project could continue is permissible. Further, a use need not support Refuge purposes in order to be compatible, as the definition clearly provides. In order to be “compatible,” a use simply must not “materially interfere” with stated Refuge purposes. The Service concluded that this use did not so interfere. Because that is a reasonable conclusion, we must affirm the decision of the Service.
2. ANILCA and Fishery Enhancement
The Service’s decision is further supported by special provisions that apply specifically to Alaskan refuges. Because we are dealing with a refuge established in Alaska, the Refuge Act must be interpreted consistently with ANILCA. Public Law No. 105-57, § 9 (Oct. 9, 1997), 111 Stat. 1260, provides that nothing in the Refuge Act affects certain aspects of Alaskan lands. “If any conflict arises between any provision of this Act and any provision of [ANILCA], then the provision in [AN-ILCA] shall prevail.” Id. § 9(b); 16 U.S.C. § 668dd (notes); see also 16 U.S.C. § 668dd(e) (providing that refuge conservation planning programs apply only to non-Alaskan refuge lands).
Three provisions of ANILCA expressly support the Service’s decision to permit the Project:
First, ANILCA specifically contemplates fishery enhancement projects as a part of refuge management. Section 304(e) states: “Where compatible with the purposes of the refuge unit, the Secretary may permit, subject to reasonable regulations and in accord with sound fisheries management principles, scientifically acceptable means of maintaining, enhancing, and rehabilitating fish stock.” ANILCA § 304(e). Athough § 304(e) presumes compatibility, it nonetheless suggests that fishery enhancement is not anathema to the management of Aaskan refuges. The purpose on which Plaintiffs’ argument depends, “to conserve fish ... in their natural diversity,” is one of the stated purposes for every refuge in Aaska. ANILCA § 303(4)(B)(i); see id. §§ 302, 303. Section 304(e) suggests, at least, that fishery enhancement is not incompatible with the purposes of every refuge. If the stated purpose “to conserve ... fish in their natural diversity” precluded any enhancement project, then ANILCA § 304(e), which expressly allows those very projects, would be meaningless.
Second, Title 16 U.S.C. § 3203 governs wilderness management in Aaska. There, Congress “recogni[zes] ... the unique conditions in Aaska.” 16 U.S.C. § 3203(a). In § 3203(b), Congress establishes “the goal of restoring and maintaining fish production in the State of Alaska to optimum sustained yield levels ... in a manner which adequately assures protection, preservation, enhancement, and rehabilitation of the ivildemess resource.” (Emphasis added.) To that end, in certain areas fishery enhancement activities, including fish weirs, fish ladders, fish hatcheries, egg planting, and the like, expressly are authorized. Id.
To be sure, the areas in which Congress explicitly authorizes aquaculture do not include the Refuge. If the legislation were that clear, the answer to this litigation would be easy, but it is not. However, the legislative history of ANILCA supports the Service’s interpretation that aquaculture was not meant to be restricted to the areas specifically mentioned in § 3203(b). See N.W. Forest Res. Council v. Pilchuck Audubon Soc’y, 97 F.3d 1161, 1168 (9th [928]*928Cir.1996) (explaining that legislative history is relevant when the statute is unclear). Referring specifically to the section on aquaculture, the Senate Report states:
In considering wilderness designation in Alaska, both the House-passed bill and the committee amendment adopt several special provisions relating to wilderness management in Alaska. Of particular interest to the committee is the future of fish enhancement and aquaculture activities in the state. The committee adopted language making it very clear that various fisheries enhancement activities could be permitted by the appropriate secretary within wilderness or wilderness study areas, subject only to reasonable regulation.
S.Rep. No. 96-413, at 308 (1980), reprinted in 1980 U.S.C.C.A.N. 5070, 5252.
Third, ANILCA contains provisions protecting recreational and subsistence uses. Title 16 U.S.C. § 3101(b), states expressly the intent of Congress respecting the lands and waters in Alaska that contain significant natural landscapes, wildlife, and other values:
It is the intent of Congress in this Act ... to provide for the maintenance of sound populations of, and habitat for, wildlife species of inestimable value to the citizens of Alaska and the Nation ... and to preserve wilderness resource values and related recreational opportunities including but not limited to ... fishing ....
(Emphasis added.) Additionally, 16 U.S.C. § 3123 requires the Secretary of the Interior to report on the status of fish populations on public lands that are subject to Alaskans’ subsistence uses. Under 16 U.S.C. § 3120(d), public lands may be managed or disposed of for “any” of the “uses or purposes authorized by this Act or other law.”
3. The Reasonableness of the Service’s Decision
The foregoing provisions of the Refuge Act and ANILCA, when read together, establish that
• Congress intends for “compatibility” determinations to be made by the Service, using “sound professional judgment”;
• “Compatible uses” are uses that do not materially interfere with the purposes of a refuge;
• Congress views Alaska as unique and intends Alaska-specific laws to trump more general laws in some instances;
• Congress contemplates that management of both refuges and wilderness areas in Alaska can include fishery enhancement;
• Congress has as one of its goals the restoration of fish production in Alaska “to optimum sustained yield levels,” including “enhancement” and “rehabilitation” activities;
• Congress intends to ensure that fish populations in Alaska remain adequate to sustain at least subsistence uses and also some recreational uses; and
• Congress intends more generally to prevent the extinction of species of fish and, where “appropriate,” to permit restoration of fish resources.
A particular project, such as the one at issue in this case, may advance some of Congress’ goals while thwarting others. It is precisely that kind of balancing that is entrusted to the Service. The Service interpreted the many inconsistent purposes and permissibly concluded that the Project is “compatible with the major purposes” for which this Alaskan wilderness area was established. That is, in the “sound professional judgment” of the Service, the Project does “not materially interfere with or [929]*929detract from” the purposes of the Refuge. 16 U.S.C. § 668ee(l). We are obliged to defer to the agency’s judgment. See Chevron, 467 U.S. at 845, 104 S.Ct. 2778 (stating that deference is owed to a choice that represents a “reasonable accommodation of conflicting policies that were committed to the agency’s care by ... statute” (citation and internal quotation marks omitted)).
E. Wilderness Act: Commercial Enterprise
Finally, Plaintiffs argue that the Project is a commercial enterprise because the fish produced in the enhancement project are commercially harvested when they leave the Refuge. Proceeds from this commercial fishing fund the CIAA. Therefore, Plaintiffs argue, the Project is a “commercial enterprise.”
With certain exceptions that do not apply here, “there shall be no commercial enterprise and no permanent road within any wilderness area designated by this chapter.” 16 U.S.C. § 1133(c) (emphasis added). There are two important ambiguities in the deceptively simple phrase emphasized above.
The first pertains to the word “within.” Does the statute preclude only commercial activities actually occurring inside the wilderness area (e.g., a restaurant in the Refuge)? Or, does the statute also preclude indirect commercial benefits, that is, noncommercial activities “within” a wilderness area when a primary beneficiary of those noncommercial activities is a commercial enterprise outside the wilderness area?
For example, imagine that the wilderness area protects a certain small furry animal. Within the Refuge itself, volunteers who act entirely out of a desire to enjoy the beauty of that animal hike in the wilderness area in places and at times designed to protect the species from its natural predators. However, the main beneficiaries of the animals’ survival and proliferation within the wilderness area are commercial trappers who take the animals when they leave the confines of the wilderness area. Doubtless, the trappers are engaging in a commercial enterprise. But is the trappers’ commercial enterprise “within” the wilderness area? Does the fact that commercial entities benefit from the preservation of the animal within the refuge convert the hikers’ noncommercial preservation activities into a commercial enterprise?
The more natural interpretation is the former one; the “commercial activity” must occur “within” the Refuge. However, Plaintiffs’ proposed alternative reading also is reasonable. Again, we are confronted with an ambiguity that it is up to the Service to reconcile.
The second ambiguity relates to the term “commercial enterprise.” Did Congress intend a mere facial look at the enterprise that is conducting the activity? If so, then the Project is not prohibited by § 1133(c), because it is operated by a nonprofit entity that benefits noncommercial users (e.g., subsistence and recreational fishers) as well as commercial users of the fish resource. Indeed, the Project formerly was operated by the State of Alaska, which plainly is a government and not itself a commercial enterprise. Or, as Plaintiffs posit,, did Congress intend a functional look at the enterprise to find out who provides funding and who the major beneficiaries may be?
Again, Plaintiffs’ interpretation is permissible, but so is the equally or even more persuasive alternative construction that the agency chose. The fact that the nonprofit CIAA is funded by commercial activities that occur outside the Refuge does not necessarily render the Project a “commercial enterprise.” Most importantly, the Service has concluded that the [930]*930enterprise is not commercial, and its interpretation is permissible. In the circumstances, we owe Chevron deference to the Service.
CONCLUSION
The Project is, without a doubt, controversial. But Congress has authorized the Service to make exactly the kind of policy decision that it made here. As long as the CIAA complies with a long list of mitigating conditions, the Service has determined that the Project is permissible.
Our standard of review dictates the outcome. Plaintiffs permissibly interpret the statutes in question and reach what many people may believe to be the better result from a policy perspective. Nonetheless, the Service also permissibly interpreted the relevant acts of Congress to authorize this fish enhancement Project. In the circumstances we may not overturn the Service’s decision.
AFFIRMED.