GINSBURG, Senior Circuit Judge:

The Secretary of the Interior appeals the district court's grant of summary judgment to the Friends of Blackwater et al. The district court held the Fish and Wildlife Service, an agency in the Department of the Interior, violated the Endangered Species Act by removing the West Virginia Northern Flying Squirrel from the list of endangered species when several criteria in the agency's Recovery Plan for the species had not been satisfied. We hold the district court erred by interpreting the Recovery Plan as binding the Secretary in his delisting decision. Because we also reject the Friends' alternative arguments that the Service's action was arbitrary, capricious, and contrary to law, we reverse the judgment of the district court.

I. Background

The West Virginia Northern Flying Squirrel (Glaucomys sabrinus fuscus) is one of 25 distinct subspecies of the Northern Flying Squirrel. It is a "small, nocturnal, gliding mammal[]" with a "long, broad, flattened tail ..., prominent eyes, and dense, silky fur" that lives in West Virginia and Virginia. U.S. FISH AND WILDLIFE SERVICE, APPALACHIAN NORTHERN FLYING SQUIRRELS RECOVERY PLAN 1-3 (Sept. 24, 1990). Despite its name, the flying squirrel cannot fly; but the patagia, or folds of skin, that stretch between its arms and legs allow it to glide for a distance when it leaps from a tree branch. Historically, its habitat consisted of the spruce-fir and northern hardwood forests of the southern Appalachian Mountains. Id. at 2, 6. In 1985, when scientists had documented only ten living squirrels, the Fish and Wildlife Service concluded it was endangered* and suggested that, although the squirrels' population "may have been declining since the Pleistocene, ... [t]heir decline ha[d] probably been accelerated through clearing of forests and other disturbances by people." 50 Fed.Reg. 26,999, 26,999 (July 1, 1985).

As required by § 4(f) of the Endangered Species Act, 16 U.S.C. § 1533(f), the Service created a recovery plan for the "conservation and survival" of the squirrel,* enumerating the following "criteria which, when met, would result in a determination... that the species be removed from the list" of endangered species, id. § 1533(f)(1)(B)(ii):

Recovery Plan at 18.

In 2002, the Service hired a biologist to investigate the possibility of removing the squirrel from the list of endangered species, and the next year began to draft its five-year review of the squirrel's status. In the review, published in 2006, the Service concluded the Recovery Plan, which had been created in 1990, "d[id] not have up to date recovery criteria," and the squirrel did "not meet the definition of endangered or threatened" because it "persist[ed] throughout its historic range." U.S. FISH AND WILDLIFE SERVICE, WEST VIRGINIA NORTHERN FLYING SQUIRREL 5-YEAR REVIEW: SUMMARY AND EVALUATION 5, 20 (April 2006). Whereas only ten squirrels had been sighted at the time of the original listing in 1985, by 2006 scientists had captured 1,063 individual squirrels at 107 sites, id. at 7, which suggested to the Secretary the population was robust, see U.S. FISH AND WILDLIFE SERVICE, ANALYSIS OF RECOVERY PLAN CRITERIA FOR THE WEST VIRGINIA NORTHERN FLYING SQUIRREL 3 (Dec. 18, 2007).

Later in 2006 the Service proposed to remove the squirrel from the list of endangered species. See 71 Fed.Reg. 75,924 (Dec. 19, 2006). The agency explained the squirrel no longer faced any of the threats listed in § 4(a)(1) of the Act so as to warrant its continued designation as either endangered or threatened. Id. at 75,925-29. With regard to the 1990 Recovery Plan, it said that because the "recovery criteria do not specifically address the five threat factors used for ... delisting a species," the plan "does not provide an explicit reference point for determining the appropriate legal status of" the squirrel. Id. at 75,925. In any event, such plans "are not regulatory documents and are instead intended to provide guidance to the Service, States, and other partners on methods of minimizing threats to listed species and on criteria that may be used to determine when recovery is achieved." Id. at 75,924-25. The Service emphasized its view that delisting the squirrel was appropriate because, among other things, "long-term nest box monitoring data provide[d] strong evidence of [its] continued presence throughout its range," id. at 75,926, and "habitat trends [were] moving in a positive direction in terms of forest regeneration and conservation," id. at 75,927.

Various scientists and conservation groups filed comments criticizing the Service's use of "persistence," which it defined as "continuing captures of [a species or subspecies] over multiple generations at previously documented sites throughout the historical range," 73 Fed.Reg. 50,226, 50,227 (Aug. 26, 2008) ("Delisting Rule"), to gauge the squirrel's recovery; the measure could not provide estimates of population levels or trends and, they pointed out, persistence so defined could not rule out the possibility the squirrel's population was declining.

In its final rule delisting the squirrel the Service responded to these comments as follows: The data showing persistence across 80 percent of the squirrel's historic range were simply "not indicative of a declining population." Id. at 50,227. Data for the remaining 20 percent need not indicate a lack of persistence because the squirrels are "elusive and hard to capture." Id.

The Friends of Blackwater filed a complaint in the district court claiming (1) promulgation of the Delisting Rule violated the Endangered Species Act by ignoring the objective, measurable criteria in the Recovery Plan and (2) the Rule itself was arbitrary and capricious because it was not based upon the best available science. The district court entered summary judgment for the plaintiff, Friends of Blackwater v. Salazar, 772 F.Supp.2d 232 (D.D.C. 2011), on the ground the Service was bound by the criteria in the Recovery Plan and its decision to delist the squirrel without following those criteria therefore constituted a revision to that plan, made without going through notice and comment rulemaking as required by the Act, id. at 241-42. In a footnote, the court also directed the agency on remand to modify its analysis of the statutory factors relevant to delisting "to the extent the agency's decision [to delist] was based on an analysis that did not separately assess the adequacy of existing regulatory mechanisms," as required by § 4(a)(1)(D) of the Act. Id. at 245 n. 17. The district court vacated the Delisting Rule, id. at 245, and the Service appealed to this court.

II. Analysis

In a case like the present one, "where the district court was reviewing an agency rulemaking under the Administrative Procedure Act ... we review the administrative record directly." Troy Corp. v. Browner, 120 F.3d 277, 281 (D.C.Cir. 1997) (internal quotation marks and citation omitted). We review the Secretary's interpretation of the statute under the familiar two-step framework from Chevron U.S.A. Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984). At Step One, the court asks "if the statute unambiguously forecloses the agency's interpretation," Nat'l Cable & Telecomms. Ass'n v. FCC, 567 F.3d 659, 663 (D.C.Cir.2009); if it does not, then at Step Two "we defer to the administering agency's interpretation as long as it reflects `a permissible construction of the statute,'" Sherley v. Sebelius, 644 F.3d 388, 393 (D.C.Cir.2011) (quoting Chevron, 467 U.S. at 843, 104 S.Ct. 2778).

A. The Legal Effect of the Recovery Plan

The Friends claim the statutory requirement that for each endangered species the Service draft a recovery plan with "objective, measurable criteria" unambiguously means those criteria must be met before a species may be delisted. In response, the Service argues the criteria in the Recovery Plan, unlike the factors in § 4(a)(1) of the Act, are not binding upon the agency in deciding whether a species is no longer endangered and therefore should be delisted.

To resolve this dispute, we "begin[] with the words of the statute." Pharm. Research & Mfrs. of Am. v. Thompson, 251 F.3d 219, 224 (D.C.Cir. 2001). Section 4(a)(1) of the Act provides the Secretary "shall" consider the five statutory factors when determining whether a species is endangered, and § 4(c) makes clear that a decision to delist "shall be made in accordance" with the same five factors. 16 U.S.C. § 1533(a), (c). Although § 4(f) states the Secretary "shall develop and implement" a recovery plan and "shall ... incorporate in [the recovery] plan ... objective, measurable criteria," id. § 1533(f), the Act does not similarly say the Secretary "shall" consult those criteria in making a delisting decision. Rather, § 4(f)(1)(B)(ii) states simply that the criteria in the recovery plan should be those "which, when met, would result in a determination, in accordance with the provisions of this section, that the species be removed from the list." This provision is ambiguous in a relevant respect: It can be read, as the Friends suggest, to place a binding constraint upon the Secretary's delisting analysis, but it can also be read as the Secretary suggests, based in part upon the absence of the word "shall," to indicate the "objective, measureable criteria" are predictive of the Service's delisting analysis rather than controlling that analysis. See Russello v. United States, 464 U.S. 16, 23, 104 S.Ct. 296, 78 L.Ed.2d 17 (1983) ("Where Congress includes particular language in one section of a statute but omits it in another section of the same Act, it is generally presumed that Congress acts intentionally and purposely in the disparate inclusion or exclusion" (internal quotation marks and alteration omitted)). On the Secretary's reading, the criteria would serve as proxies, tailored to what is known about the particular species, standing in for the statutory factors of § 4(a)(1) that ultimately control the Secretary's delisting decisions for all species.* The ambiguity is magnified because § 4(f) of the Endangered Species Act qualifies the Secretary's duty to incorporate "objective, measurable criteria" in "developing and implementing recovery plans" with the phrase "to the maximum extent practicable." 16 U.S.C. § 1533(f). Cf. Oceana, Inc. v. Locke, 670 F.3d 1238, 1242-43 (D.C.Cir.2011) (reading statutory provision as mandatory where, in contrast to a neighboring provision, duty imposed was not modified by phrase "to the extent practicable"); Biodiversity Legal Found. v. Babbitt, 146 F.3d 1249, 1253-54 (10th Cir.1998) (phrase "to the maximum extent practicable" in § 4(b)(3)(A) of Endangered Species Act indicates non-mandatory character of provision).

Other "traditional tools of statutory construction," Chevron, 467 U.S. at 843 n. 9, 104 S.Ct. 2778, do not reveal any more clearly the intent of the Congress on this question. The Friends argue the legislative history indicates the criteria in the recovery plan must be binding because the Congress added the call for "objective, measureable criteria" specifically in order to "improve the development, implementation and review of plans for the recovery of listed species." S.Rep. No. 100-240 (1987), reprinted in 1988 U.S.C.C.A.N. 2700, 2700. The Friends also argue the structure of the Act confirms their view because the Secretary's interpretation would render the requirement of "objective, measurable criteria" meaningless.

These arguments from legislative history and structure come down to the single claim that interpreting the Recovery Plan as non-binding would render § 4(f) of the Act a nullity. That is not correct. With an exception not relevant here, § 4(f) obliges the Secretary to "develop and implement plans" for the recovery of any species designated as endangered. 16 U.S.C. § 1533(f)(1). If the Secretary wants to change the plan, then he first must let the public comment. Id. § 1533(f)(4). It does not follow, however, that with each criterion he includes in a recovery plan the Secretary places a further obligation upon the Service. A plan is a statement of intention, not a contract. If the plan is overtaken by events, then there is no need to change the plan; it may simply be irrelevant. If someone said he would see me in Cleveland while on his way to Chicago and would let me know before changing his plan, it would hardly be sensible to say he must "revise" his plan before he can tell me that he no longer needs to make the trip.

Nor is there anything unusual about a statute that requires an agency to publish a non-binding document. See Norton v. S. Utah Wilderness Alliance, 542 U.S. 55, 69, 72, 124 S.Ct. 2373, 159 L.Ed.2d 137 (2004) (statute required BLM to promulgate land use plan, but plan itself was "designed to guide" BLM, not to be legally enforceable). Contrary to the Friends' argument, the Secretary's interpretation of the plan as non-binding does not render meaningless the Secretary's statutory obligations to create and to implement a recovery plan and to use notice and comment in order to revise such a plan. On the contrary, a recovery plan, even if not binding, so long as the species is endangered provides "objective, measurable criteria" by which to evaluate the Service's progress toward its goal of conserving the species.

It is a short hop from here to conclude under Step Two of Chevron that the Secretary's interpretation is a "permissible" one. The Service fairly analogizes a recovery plan to a map or a set of directions that provides objective and measurable steps to guide a traveler to his destination. Cf. Fund for Animals, Inc. v. Rice, 85 F.3d 535, 547 (11th Cir.1996) (holding "recovery plans are for guidance purposes only"). Although a map may help a traveler chart his course, it is the sign at the end of the road, here the five statutory factors indicating recovery, and not a mark on the map that tells him his journey is over. Moreover, as with a map, it is possible to reach one's destination — recovery of the species — by a pathway neither contemplated by the traveler setting out nor indicated on the map.

B. The Measure of "Persistence"

The Friends of Blackwater contend in the alternative the Service (1), by using data on the species' "persistence" rather than data on its population and population trends, violated the statutory requirement that it use the "best ... data available," 16 U.S.C. § 1533(b)(1)(A), and (2), by failing adequately to explain its departure from the population-based criterion in the Recovery Plan, rendered its decision arbitrary and capricious within the meaning of the Administrative Procedure Act, 5 U.S.C. § 706(2)(A), see Motor Vehicle Mfrs. Ass'n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 42, 103 S.Ct. 2856, 77 L.Ed.2d 443 (1983).*

The Friends' first argument runs afoul of Southwest Center for Biological Diversity v. Babbitt, in which we explained that under the "best ... data available" standard, "the Secretary has no obligation to conduct independent studies"; the Service is entitled to rely upon the best data available to it, which in that case were existing scientific estimates of the species' population, rather than conducting its own population count in order to determine whether a species is endangered. 215 F.3d 58, 60-61 (D.C.Cir.2000). In this case, where it is undisputed the "best ... data available" related to the persistence of the species, the Service was entitled to rely upon those data in its delisting analysis, just as it was entitled to list the squirrel in the first place even though no estimate of the squirrel's population was then available. See id. (requirement of best data available "merely prohibits the Secretary from disregarding available scientific evidence that is in some way better than the evidence he relies on" (internal quotation marks and citation omitted)). Moreover, the Service reasonably determined the data on persistence, which showed the squirrel "persist[ed] throughout its historic range," 5-Year Review at 20, were "not indicative of a declining population," 73 Fed.Reg. at 50,227; hence it could reasonably find the species' survival was no longer threatened by loss of habitat.

This would end the matter were it not for the Service's statement in the Recovery Plan that it would look to estimates of population trends. See Recovery Plan at 18 (saying it could delist the squirrel when "squirrel populations are stable or expanding" in at least 80 percent of certain designated areas). Although, as we explained above, that plan was not binding upon the agency, the Friends maintain the Service had an obligation adequately to account for any departures from the guidelines described in the plan, citing Motor Vehicle Mfrs. Ass'n, 463 U.S. at 42, 103 S.Ct. 2856 ("an agency changing its course ... [must] supply a reasoned analysis"). Whether an agency must account for a departure from a prior non-binding statement of intent is not entirely clear. Compare Sitka Sound Seafoods, Inc. v. NLRB, 206 F.3d 1175, 1182 (D.C.Cir.2000) ("Manual does not bind the Board ... [and so] the relevant question is whether, quite apart from the Manual, the Board acted unreasonably"), with Edison Elec. Institute v. EPA, 391 F.3d 1267, 1269 & n. 3 (D.C.Cir.2004) ("[report] is not strictly binding upon EPA and any deviation from the Report is not per se arbitrary and capricious"; "real question is whether EPA adequately accounted for any departures" from factors described in report). We need not resolve this question today because the Service adequately explained that population data were not available whereas data on persistence were. Still, the Friends contend the Service's stated reason for not itself estimating the population, viz., the cost and difficulty of doing so, fails because the Service knew of the difficulty of estimating the squirrel population when it adopted the population criterion in the Recovery Plan. See Recovery Plan at 11 (noting the squirrels were "extremely difficult to collect and study").

Though not illogical, neither does the Friends' argument show the Service was arbitrary and capricious. The agency did realize when it put the population-based criterion into the Recovery Plan in 1990 that the squirrels were difficult to monitor. After more than 15 years of gathering more data and capturing more squirrels, however, the Service could reasonably conclude, and the Friends do not dispute, that "[a]n adequate monitoring scheme to estimate population numbers across a representative sample of the entire range of the [squirrel] would require many thousands of nest boxes and traps," Analysis of Recovery Plan Criteria at 1. The Friends have not shown the Service's judgment that a project of that magnitude was simply too difficult and too costly for the agency to undertake was arbitrary and capricious. Therefore, we conclude the Service has met any burden it may have to account for its departure from the criterion it contemplated when it developed the Recovery Plan in 1990.

C. Inadequacy of Regulatory Mechanisms

Finally, the Friends argue the Service failed to conduct an independent analysis of the fourth statutory factor, "the inadequacy of existing regulatory mechanisms," 16 U.S.C. § 1533(a)(1)(D), which factor they claim must be analyzed without regard to whether there are any threats arising under the other provisions of § 4(a)(1). The Service did consider the adequacy of existing regulatory mechanisms, but it did not do so in isolation. On the contrary, having considered all the other types of threats listed in § 4(a)(1) and found no existing conditions such as disease or destruction of habitat threatened the subspecies, the Service could reasonably, indeed readily, conclude the squirrel did not require additional regulatory protection. See Delisting Rule, 73 Fed.Reg. at 50,237.

Under the Friends' approach to § 4(a)(1)(D), the Service would have to assess the adequacy of regulatory mechanisms without regard to its analysis of the threats listed in clauses A, B, C, and E of the same section. This contention is most peculiar. If the adequacy or "inadequacy of existing regulat[ion]" is to be judged without considering the level, or even the existence, of any threat the regulation is designed to meet, then it would follow that the Service could never delist a species unless some regulatory mechanism was in place to protect it — whether needed or not. Moreover, because the Service is to apply the same factors to listing as to delisting decisions, 16 U.S.C. § 1533(c), it would follow that every species (except pests, see id. § 1532(6)) must either (1) be protected by regulations of some sort or (2) be classified as endangered or threatened. Absent compelling evidence, we will not attribute to the Congress the intent to create such an absurd overabundance of regulation — and a further abundance of, for example, white-tailed deer, which have long since moved into metropolitan areas in search of better forage and fewer predators, including hunters. See Steeve D. Côté et al., Ecological Impacts of Deer Overabundance, 35 ANN. REV. ECOLOGY EVOLUTION & SYSTEMATICS 113, 116 (2004); Robert K. Swihart et al., Ecology of Urban and Suburban White-Tailed Deer, in URBAN DEER: A MANAGEABLE RESOURCE? 35, 35, 42 (Jay B. McAninch ed.1993). By considering the adequacy or inadequacy of regulations in light of other threats to the species, the Secretary's interpretation of § 4(a)(1)(D) is certainly reasonable, and the Friends therefore have failed to demonstrate the Service violated the Act by acting upon that interpretation.

III. Conclusion

We hold the Secretary reasonably interpreted the Endangered Species Act as not requiring that the criteria in a recovery plan be satisfied before a species may be delisted pursuant to the factors in the Act itself. Because the Secretary's determination the West Virginia Northern Flying Squirrel was no longer endangered was neither arbitrary and capricious nor in violation of the Act, the judgment of the district court is

Reversed.

Appendix: Notes on the Dissent

Our dissenting colleague labors at length to prove "shall" indicates an action is mandatory and "to implement" means to give practical effect, see Dissent at 441-42, two points we nowhere dispute. Nor do we doubt § 4(f)(1) of the Act imposes mandatory obligations upon the Secretary. The Secretary shall, for example, develop a recovery plan for an endangered species, 16 U.S.C. § 1533(f)(1), which plan shall include "objective, measurable criteria," id. § 1533(f)(1)(B)(ii). The Secretary, moreover, must implement the plan. Id. § 1533(f)(1). That is, as long as a species is listed as endangered, the agency is obligated to work toward the goals set in its recovery plan. None of this, however, implies the objective, measurable criteria in the plan limit the agency when it is deciding whether to delist a species.

The foregoing interpretation of § 4(f)(1) does not render the notice and comment requirement of § 4(f)(4) "superfluous." Cf. Dissent at 445. If the Service believes the goals in the recovery plan for a species are outdated but the species is still endangered, then the Secretary must either continue to pursue those goals or, more sensibly, update the plan, which requires notice and comment.

The dissent's claim (at 442) that our interpretation "erases `(f)(1)' from `(f)(1)(B)(ii)'" ignores our reading of those provisions (at 443) as together indicating the Secretary "shall ... incorporate in [the recovery] plan ... objective, measurable criteria." Accordingly, we agree with the dissent to the extent that subsection (B)(ii) imposes upon the Secretary a mandatory duty to incorporate criteria in a recovery plan but, the dissent's insistence to the contrary notwithstanding, that understanding alone does not clarify how such criteria relate to the Secretary's delisting decision. Our dissenting colleague correctly identifies (at 442) the "future conditional tense" in § 4(f)(1)(B)(ii), but misstates the logical relation in that statement, and thereby mistakenly reads the provision as unambiguous. As we note (at 444 n.*), the phrase "which, when met, would result" most plainly expresses a sufficient, not a necessary, condition: It says what must happen when the criteria are met, but is silent — and therefore ambiguous — with respect to what may or must happen when the criteria are not met. Although one could read the word "only" into the statute so that it states a necessary condition ("which, [only] when met, would result"), one can as well — as the Secretary's interpretation suggests — read the conditional "would" as referring to the agency's likely delisting analysis pursuant to the factors prescribed in § 4(a)(1). See Knight v. Comm'r, 552 U.S. 181, 192, 128 S.Ct. 782, 169 L.Ed.2d 652 (2008) ("In the context of making ... a prediction, ... the word `would' is best read as expressing concepts such as ... probability" (internal quotation marks, alteration, and citation omitted)). Hence the ambiguity.

The "context and structure of the statute," Dissent at 442, only underscore this ambiguity and therefore support our deferring under Chevron to the Secretary's interpretation. Section 4(c) of the Act, which describes the Secretary's delisting analysis, provides: "Each determination [that a species be delisted] shall be made in accordance with the provisions of subsections (a) and (b) of this section." 16 U.S.C. § 1533(c). Section 4(c), however, makes no mention of subsection (f) or its requirement of a recovery plan. When the Congress amended the Act to add the requirement of "objective, measurable criteria," see Pub.L. No. 100-478, 102 Stat. 2306 (1988), it could have, but did not, revise § 4(c) to require a delisting determination also be made in accordance with the recovery plan criteria adopted pursuant to § 4(f).*

The dissent suggests (at 442-43) the legal effect of the qualifying phrase "to the maximum extent practicable," 16 U.S.C. § 1533(f)(1), insofar as it applies to the Secretary's duty to "incorporate in each plan ... objective, measurable criteria," id. § 1533(f)(1)(B)(ii), runs out once the Secretary has included such criteria in a plan. The statute, however, applies this qualification to the Secretary's actions both "in developing and [in] implementing recovery plans." Id. § 1533(f)(1) (emphasis added). The dissent says "implementing" can have no reference to subsection (B) of § 4(f)(1) because "incorporat[ing]" something in a plan occurs solely in the course of "developing" the plan, but we are not so quick to abandon the plain text of the statute. Although "implementing" has a more obvious connection to subsection (A), with respect to subsection (B) it at least tells us the Secretary must incorporate criteria that are practicable to implement. The dissent's interpretation would have the Secretary measure the practicability of incorporating criteria in a recovery plan without respect to the practicability of their implementation, but that cannot be correct. If the practicability of incorporating criteria is to be determined without a view to the practicability of their implementation, then any imaginable criterion may be incorporated so long as the agency has the wit to place the requisite words upon a page. (Why not when the cow jumps over the moon? Or when Birnam Wood be come to Dunsinane?) Following a more reasonable interpretation, it would be "impracticable" for the Secretary to adopt criteria that by their nature could never be met and hence would preclude delisting a species so long as those criteria remain in effect. Similarly, if the Secretary foresees that adopting certain criteria would unduly restrict his delisting analysis, then he may decide it is practicable only to adopt criteria that guide but do not constrain that analysis.*

Our dissenting colleague next offers (at 444) her "flight plan" analogy in an effort to show certain administrative plans may not be "discarded" even if "overtaken by events," but the analogy in fact supports our interpretation. As the dissent notes, a portion of the regulation regarding flight plans allows a pilot to deviate from a flight plan if "an emergency exists," 14 C.F.R. § 91.123(a), which the pilot may declare in his discretion, id. §§ 91.123(c), 91.3(b), but the dissent misses the significance of this exception. An emergency negates the need for a fixed flight plan much as the recovery of a species negates the need for a plan designed to bring about that recovery; in either event, the formal revision of such a plan would not be useful and therefore, unsurprisingly, is not required by law.

The dissent addresses (at 449-50) a facially plausible "logical outgrowth" argument that appears nowhere in the Friends' brief, was not raised in the district court, and therefore is not properly before us. See United States v. Southerland, 486 F.3d 1355, 1360 (D.C.Cir.2007) ("argument ... raised for the first time at oral argument... is forfeited"); Benoit v. Dep't of Agric., 608 F.3d 17, 21 (D.C.Cir.2010) (argument not raised in district court is forfeited). As part of their argument that the statute unambiguously requires the Secretary either to meet the criteria in the recovery plan or to modify them through notice and comment prior to delisting, the Friends did argue the notice and comment process the Secretary used to delist the Squirrel did not constitute a revision of the recovery plan. See Br. of Appellees at 35-40. Because we rejected the Friends' must-meet-or-modify premise, however, there is no need to address that dependent argument. In the Friends' "alternative argument that the Secretary violated the Administrative Procedure Act," Dissent at 448, they alleged the delisting process was arbitrary and capricious and not based upon the best data available, see Br. of Appellees at 45-51, not that the final rule was not a logical outgrowth of the proposed rule.

With respect to the Friends' argument that the Act precludes the Secretary from relying upon data concerning persistence, the dissent suggests (at 450-61) the Secretary must have data on the population of a species before he may decide to delist it. What § 4(b)(1)(A) and § 4(c) of the statute require, however, is that the Secretary use the "best ... data available" when, respectively, listing or delisting a species. 16 U.S.C. § 1533(b)(1)(A), (c). Population data were not available when the Secretary listed the squirrel as endangered. Nor were such data available when he delisted the squirrel. To require the Secretary before acting to obtain such data as are not then "available" is clearly foreclosed by the statute. See id.; Am. Wildlands v. Kempthorne, 530 F.3d 991, 1001 (D.C.Cir.2008) ("in the absence of available evidence, Congress does not require the agency to conduct its own studies").

The dissent compounds the error by claiming (at 451) data on persistence do not "answer the relevant question," and asserts upon this basis the Secretary relied upon "no data." Evidence is relevant to a particular question of fact if "it has any tendency to make [that] fact more or less probable." Fed.R.Evid. 401(a). The question at issue here is whether the squirrel is an "endangered species," which the Act defines as "any species which is in danger of extinction throughout all or a significant portion of its range...." 16 U.S.C. § 1532(6). Because extinction is less likely where there is widespread persistence than where there is not, persistence is relevant to a determination whether a species is endangered; and if the only data available concern persistence, then they are quite clearly the "best" data available. To be sure, data on persistence would also be relevant to the question of a species' "survival," see Dissent at 450, 451-52, a term with a meaning distinct from "recovery," see 50 C.F.R. § 402.02 (defining "[r]ecovery" as "improvement in the status of listed species to the point at which listing is no longer appropriate under the criteria set out in section 4(a)(1) of the Act"), but that is neither here nor there. Evidence may be relevant to two distinct legal questions, and therefore its relevance to a question not at issue (survival), does not imply or even suggest its irrelevance to the question that is at issue (recovery).

ROGERS, Circuit Judge, dissenting:

Because Congress "has directly spoken to the precise question at issue," Chevron USA Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 842, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984), the court's job is done. Instead, the court defers to the Secretary's interpretation, contrary to the plain text of the Endangered Species Act ("ESA"), 16 U.S.C. §§ 1531-1544, that the West Virginia Northern Flying Squirrel ("Squirrel"), an endangered species, loses all protections even though the recovery criteria in its recovery plan have not been met and those criteria are revised, while the Squirrel was listed as endangered, without required notice and prior consideration of public comments. But even assuming, as the court concludes, the ESA is ambiguous, the Secretary was arbitrary and capricious in delisting the Squirrel based in material part on an analysis revising the recovery plan criteria that was not publically noticed until the final delisting rule, and then only on the basis of available scientific and commercial evidence showing the Squirrel persists (i.e., is not yet extinct) as distinct from recovered so as no longer to require ESA's protections. Accordingly, I respectfully dissent.

I.

"As in all statutory construction cases," the court must "begin with the language of the statute." Barnhart v. Sigmon Coal Co., Inc., 534 U.S. 438, 450, 122 S.Ct. 941, 151 L.Ed.2d 908 (2002). "[C]ourts must presume that a legislature says in a statute what it means and means in a statute what it says there. When the words of a statute are unambiguous, then, this first canon is also the last: judicial inquiry is complete." Id. at 461-62, 122 S.Ct. 941 (quoting Connecticut Nat. Bank v. Germain, 503 U.S. 249, 253-54, 112 S.Ct. 1146, 117 L.Ed.2d 391 (1992) (internal quotation marks and citation omitted)). Congress's requirements in the ESA for delisting an endangered species for which the Secretary of the Interior has developed a recovery plan are unambiguous with respect to when that species is eligible for delisting and to the procedure for revising announced recovery plan criteria.

Section 4(f) provides:

16 U.S.C. § 1533(f) (emphases added). These substantive and procedural requirements reflect Congress's finding that "various species of ... wildlife," such as the Squirrel, "are of esthetic, ecological, educational, historical, recreational, and scientific value to the Nation and its people," id. § 1531(a)(3), and adoption of a policy of conservation, id. § 1531(c), which is defined as use of the "methods and procedures which are necessary to bring any endangered species or threatened species to the point at which the measures provided pursuant to [the ESA] are no longer necessary," id. § 1532(3). Applying the traditional, well-settled standards for statutory interpretation,1 it is difficult to imagine how Congress could have spoken more clearly and directly when it strengthened the ESA in 1988, see Pub.L. No. 100-478, 102 Stat. 2306, by mandating both development and implementation, prior to delisting, of recovery plans that include "objective, measurable criteria," 16 U.S.C. § 1533(f)(1)(B)(ii), and the procedures for their amendment, id. § 1533(f)(4). See S.Rep. No. 100-240, at 4 (1987), reprinted in 1988 U.S.C.C.A.N. 2700, 2703 (noting that "far too many recovery plans for listed species have not been implemented ... [and] recovery plans have failed to include consistently criteria, time frames and estimated costs for recovery").

The plain text of section 4(f) answers the questions of whether recovery plans are discretionary, what they must contain, what process must be followed for their adoption and revision, and whether recovery plan criteria must be met before delisting procedures are initiated. First, section 4 provides that the Secretary "shall" implement recovery plans. When Congress uses the word "shall," it intends to communicate a mandatory action. See Gonzalez v. Thaler, ___ U.S. ___, 132 S.Ct. 641, 651, 181 L.Ed.2d 619 (2012). "It is fixed law that words of statutes or regulations must be given their ordinary, contemporary, common meaning. It is also fixed usage that `shall' means something on the order of `must' or `will.'" FTC v. Tarriff, 584 F.3d 1088, 1090 (D.C.Cir.2009) (internal quotation marks and citations omitted). Looking to the usual understanding of the words used by Congress, to "implement" is "to give practical effect to and ensure of actual fulfillment by concrete measures." MERRIAM WEBSTER'S COLLEGIATE DICTIONARY 583 (10th ed.1993). Thus, section 4(f)(1) "is not at all ambiguous, but instead is exquisitely clear," Ctr. for Biological Diversity v. Norton, 254 F.3d 833, 837 (9th Cir.2001) (interpreting ESA section 4(b), 16 U.S.C. § 1533(b)(3)(A)), in requiring the Secretary to ensure the actual fulfillment of species' recovery plans prior to delisting. Where, as here, the Secretary seeks to delist a species whose initial recovery plan criteria have not been met, section 4(f)(4)'s procedures for revising the recovery plan must be followed. This is the only reading of section 4 of the ESA that does not render the mandatory requirements of subsections (f)(1) and (f)(4) superfluous. See Corley v. United States, 556 U.S. 303, 314, 129 S.Ct. 1558, 173 L.Ed.2d 443 (2009).

Eschewing the plain text, the court finds ambiguity for three reasons. First, the court notes that the word "shall" does not appear in section 4(f)(1)(B)(ii) with respect to whether the "objective, measurable criteria" to be included in the recovery plan control delisting. See Op. at 433. Second, this purported ambiguity is "magnified," the court states, because the Secretary's obligation to include such criteria in a recovery plan is qualified by the phase "to the maximum extent practicable." See id. at 433. Third, the court asserts that other "`traditional tools of statutory construction,' Chevron, 467 U.S. at 843 n. 9, 104 S.Ct. 2778, do not reveal any more clearly the intent of Congress on this question." Id. These reasons do not withstand examination.

The first reason erases "(f)(1)" from "(f)(1)(B)(ii)" and ignores English grammar. The court blinds itself to the introductory provision, which provides that the Secretary "shall develop and implement" recovery plans, which "shall ... incorporate" the criteria in (B)(ii), 16 U.S.C. §§ 1533(f)(1) & (f)(1)(B). Subsection (B)(ii) cannot exist dissected from its introductory text in (f)(1); (B)(ii) is not even a complete sentence without (f)(1). At some point Congress surely is permitted to avoid being duplicative (triplicative?). Likewise, the grammatical structure of subsection (f)(1)(B)(ii) is in a simple future conditional tense. Plans that shall contain "objective, measurable criteria which, when met, would result in a determination... that the species be removed from the list," id. § (f)(1)(B)(ii) (emphasis added), shall be implemented. The condition — the time at which the recovery criteria are met — is followed by the consequence — a determination to delist the species.2 Adding a third "shall" to this sentence does not change its plain meaning, nor would it make sense given the simple condition-consequence structure of the sentence. The court agrees the Secretary has mandatory duties, Op. at 436-37, but then disregards the import of its agreement finding ambiguity when there is none. As a further example, the court suggests the possibility that the phrase "would result" could mean "a sufficient, not a necessary" condition. Id. at 437. But "the sort of ambiguity giving rise to Chevron deference is a creature not of definitional possibilities, but of statutory context." New York v. EPA, 443 F.3d 880, 884 (D.C.Cir.2006), and a court "must not `confine [itself] to examining a particular statutory provision in isolation.'" Am. Bankers Ass'n v. Nat. Credit Union Admin., 271 F.3d 262, 267 (D.C.Cir.2001). It hardly would make sense for Congress to mandate implementation, and formal procedures for revision, of criteria in recovery plans that "would result" in "the species be[ing] removed from the list" if Congress intended that other unadopted recovery criteria could suffice, and need not be formally adopted pursuant to the procedures Congress mandated, to determine delisting. The context and structure of the statute are clear that the recovery criteria must be "met" or revised prior to delisting. And, upon "exhausting the traditional tools of statutory construction, including examining the statute's legislative history," id., that history "definitively resolves the debate," id., the court creates over the word "would" for Congress stated that the 1988 amendments added "[t]he requirement that plans contain objective, measurable criteria for removal of a species from the Act's lists." S.Rep. No. 100-240, at 9-10 (emphasis added); see infra n. 4.

Next, the court attempts to find ambiguity in the phrase "to the maximum extent practicable." Section 4(f)(1) provides that "[t]he Secretary, in developing and implementing recovery plans, shall, to the maximum extent practicable ... incorporate in each plan ... objective, measurable criteria." 16 U.S.C. § 1533(f)(1)(B)(ii) (emphasis added). Any potential ambiguity evaporates where, as here, the Secretary has incorporated objective criteria in the Squirrel's recovery plan.3 It obviously was practicable to do so here. Having done so, the Secretary is not free to ignore Congress's mandates to "implement" plans and to delist only "when" the objective criteria are "met," particularly given his determination, in developing the recovery plan, that the plan "will ... promote the conservation," 16 U.S.C. § 1533(f)(1), of the Squirrel. The court protests ambiguity remains because the word "implementing" means "the Secretary must incorporate criteria that are practicable to implement." The court explains, otherwise "any imaginable criterion may be incorporated so long as the agency has the wit to place the requisite words upon a page.... [I]t would be `impracticable' for the Secretary to adopt criteria that by their nature could never be met and hence would preclude delisting a species so long as those criteria remain in effect." Op. at 438. This is word play, not statutory analysis. As Congress crafted the ESA, the Secretary's chosen criteria, once "incorporate[d] in each plan," remain "incorporate[d] in each plan," however "practicable" their adoption might once have been, or their implementation might later become, and Congress provided a remedy for the latter possibility — revision pursuant to public notice-and-comment. See 16 U.S.C. § 1533(f)(4). The recovery criteria, therefore, plainly cannot "preclude delisting a species" because they can be changed, with notice-and-comment, if the Secretary determines "they could never be met" or no longer accurately measure recovery. The court either overlooks section 4(f)(4)'s revision process or drains it of purpose. Furthermore, even if the Secretary could "decide it is practicable only to adopt criteria that guide but do not constrain [the delisting] analysis," Op. at 438, that is not what happened here. The Squirrel's recovery plan states that "[r]ecovery plans delineate reasonable actions believed to be required to recover and/or protect listed species," and notes that recovery plans are subject to amendment — which is consistent with section 4(f)(4). Recovery Plan, Executive Summary (emphasis added). Indeed, the Secretary specified that the first three criteria were required for down-listing the Squirrel to "threatened" status, while the fourth, in combination with the first three, were required for delisting. Id. The court's search for ambiguity here is in vain.

The third reason the court finds ambiguity, which it collapses into a denial that its deference to the Secretary's interpretation of sections 4(a) and (f) renders (f) a nullity, see Op. at 433-34, overrides Congress's repeated use of "shall" in identifying the Secretary's obligations and allows the Secretary to end run section 4(f)'s requirements. Rather than confront Congress's plain mandatory text, or the supportive legislative history,4 the court turns to a travel planning analogy that distorts what happened here. The court posits:

Op. at 434. Here, the Secretary in fact went to Chicago — he declared the Squirrel recovered and delisted it. He just avoided Cleveland altogether (i.e., several of the recovery criteria), and stopped in Detroit instead (i.e., the covertly revised criteria), without telling anyone, despite saying he "would let me know before changing his plan" (i.e., comply with section 4(f)(4)). The court's analogy begs the key question of how it is to be determined that the stop in Cleveland (i.e., meeting the recovery criteria) no longer needs to occur — as Congress directed, or as the Secretary would prefer? A better analogy, grounded in administrative law, is of an airline pilot who determines mid-flight, due to changed circumstances (e.g., turbulence), that the approved flight plan should be revised. Although "overtaken by events," Op. at 434, under the regulations, in the absence of an emergency, the flight plan may not be discarded by the pilot — instead the pilot must follow the revision process set forth by regulations. See 14 C.F.R. § 91.123 ("[N]o pilot in command may deviate from [a] clearance unless an amended clearance is obtained...."). Congress provided no comparable "emergency" exception in the ESA whereby the Secretary may disregard the recovery plan criteria if he decides, insulated from public input, that the species has in fact recovered despite not satisfying the official plan criteria for recovery. Instead, Congress specified in the ESA the process for revising recovery plans.

Furthermore, the court altogether ignores that the Secretary, acting through the Fish and Wildlife Service ("FWS"), revised the recovery plan, while the Squirrel was still listed as endangered, without following the notice and comment procedures required by ESA section 4(f)(4). In December 2007, nearly a year before promulgating the Final Rule Removing the Squirrel from the Endangered Species List, 73 Fed.Reg. 50,226 (Aug. 26, 2008) ("Final Rule"), the FWS revised two of the four criteria in the Squirrel's recovery plan. In an unpublished, publically-unavailable analysis, the FWS concluded that the second and fourth recovery criteria had been met, while the "intent" of the first and third criteria had been met under revised criteria. See U.S. FISH & WILDLIFE SERVICE, ANALYSIS OF RECOVERY PLAN CRITERIA FOR THE WEST VIRGINIA NORTHERN FLYING SQUIRREL 13 (Dec. 18, 2007) ("2007 Analysis"). Specifically, the FWS revised the first criterion from one measuring Squirrel populations in five geographic areas to one measuring "persistence," that is, whether the Squirrel was present (without regard to quantity) or absent, in 3-5 year intervals, in different areas.5 See id. at 2. The FWS revised the third criterion to eliminate its core provision that the geographic recovery area ("GRA") be managed in perpetuity.6 It is undisputed the FWS made these revisions without providing notice or opportunity for comment, and that the 2007 Analysis was publically mentioned for the first time in the Final Rule, 73 Fed.Reg. at 50,227. Even were the court correct that the purported ambiguity of section 4 permits the Secretary's interpretation of recovery plans as discretionary and mere guidance (contrary to the ESA's plain text) and that the Secretary did not need to revise the recovery plan if its criteria could not be met, see Op. at 434, where the Secretary does revise the plan, the court has no explanation for why the requirements of section 4(f)(4) can be ignored, other than to invoke its inapt travel plan analogy.

The circularity of the court's reasoning demonstrates how its reading renders section 4(f) superfluous: According to the court: "[A]s long as a species is listed as endangered, the agency is obligated to work toward the goals set in its recovery plan," but the "criteria in the plan [do not] limit the agency when it is deciding whether to delist a species." Op. at 437. If a species is delisted on the basis of recovery, without regard to whether the recovery plan criteria have been "met," then there is nothing left to section 4(f)(4)'s mandatory requirement that revisions to the criteria by which recovery is evaluated be subject to notice and comment prior to their adoption. Under the court's interpretation, the FWS, on the Secretary's behalf, can dispense with the revision plan criteria by not labeling its changes a "revision," and proceed to delist a protected species pursuant to section 4(a) without regard to the requirements of sections 4(f)(1) and (4). That the FWS would never have an incentive to follow the revision process of section 4(f)(4), because recovery plans could be ignored without consequence, is aptly demonstrated by the facts here: while the Squirrel was still listed as endangered, the FWS covertly changed its recovery plan criteria without following the requirements of section 4(f)(4).

Required procedures are a vital part of the protections afforded by the ESA, in which Congress employed mandatory language regarding the Secretary's obligations. See, e.g., 16 U.S.C. § 1533(a)(3)(A) & (B) (setting procedures for designating and revising critical habitat); id. § (b)(3)(A)-(D) (setting procedures and time period for responding to petitions); id. § (b)(5) (setting procedures and time period for notice-and-comment on listing); id. § (b)(6) (setting time period for publishing final listing or delisting rule); id. § (c)(1) (setting requirements for what endangered and threatened lists must contain); id. § (f)(4) & (5) (setting procedures for revising recovery plans and considering public comments). Whether or not the court views Congress's chosen process as unnecessary, see Op. at 434, "[the court's] job is to interpret the methods that Congress chose to further its goals, not to devise methods of our own." Consolidated Rail Corp. v. United States, 896 F.2d 574, 579 (D.C.Cir.1990). If, as the court asserts, this dissent "labors at length" about Congress's use of the word "shall," see Op. at 436, the "labor[ing]" has been for naught as the court chooses to acknowledge its force selectively, in fact only with respect to section 4(a)(1), see id. Furthermore, the court's revision to section 4(f) does not make much sense for it has the decision to delist driving what the recovery criteria are rather than the recovery criteria driving the decision to delist. Consistent with Congress's choice of words and purpose to strengthen the ESA in the 1988 amendments, the notice-and-comment process for the recovery plan should sensibly precede consideration of delisting. See 16 U.S.C. § 1533(f)(4) ("The Secretary shall consider all information presented during the public comment period prior to approval of the plan.") (emphasis added).

Contrary to the Secretary's suggestion in this court, the requirements added by Congress in 1988 to strengthen the ESA's protections, see S. REP. NO. 100-240, at 8-9, are not a "make-work exercise" or mere "hoop-jumping," Appellant's Br. at 43. Instead, consistent with its ESA findings and policy of conservation, Congress determined to "require deliberation" when the existence of precious species have been found endangered or threatened. Congress instructed in plain terms that only upon subjecting proposed revisions to recovery plan criteria to the rigor of public comment would the Secretary (or the FWS) be in a position properly to assess proposed revisions and undertake to consider, upon applying the revised criteria, whether they are met and the protected species should be delisted pursuant to sections 4(a) and (b). After all, the purpose of notice-and-comment procedures is "to ensure that affected parties have an opportunity to participate in and influence agency decision making at an early stage, when the agency is more likely to give real consideration to alternative ideas." See State of N.J., Dept. of Envtl. Protection v. EPA, 626 F.2d 1038, 1049 (D.C.Cir.1980) (internal quotation marks and citation omitted) (emphasis added). That the Secretary (or the FWS) may find these requirements inconvenient or view section 4(f)(4) as a "make-work" exercise is irrelevant, for "[w]hen a statute commands an agency without qualification to carry out a particular program in a particular way, the agency's duty is clear; if it believes the statute untoward in some respect, then `it should take its concerns to Congress,' for `[i]n the meantime it must obey [the statute] as written.'" Oceana, Inc. v. Locke, 670 F.3d 1238, 1243 (D.C.Cir.2011) (quoting Natural Res. Def. Council v. EPA, 643 F.3d 311, 323 (D.C.Cir.2011)) (second and third alterations in original).

The court's reliance on Norton v. Southern Utah Wilderness Alliance, 542 U.S. 55, 72, 124 S.Ct. 2373, 159 L.Ed.2d 137 (2004), is misplaced, see Op. at 434. In determining that Bureau of Land Management ("BLM") land use plans were not binding documents, the Supreme Court relied on a statutory provision granting the Secretary leeway in implementing plans: "Title 43 U.S.C. § 1712(e) provides that `[t]he Secretary may issue management decisions to implement land use plans' — the decisions, that is, are distinct from the plan itself." Norton, 542 U.S. at 69-70, 124 S.Ct. 2373 (alteration in original). BLM regulations likewise provided that land use plans were "not a final implementation decision on actions which require further specific plans, process steps, or decisions." Id. at 70, 124 S.Ct. 2373 (quoting 43 C.F.R. § 1601.0-5(k) (2003)). By contrast, the ESA includes no provision granting the Secretary leeway in issuing "management decisions" about implementing recovery plans, and unlike land use plans, which may lack specificity and process steps, Congress mandated that recovery plans contain "objective, measurable criteria" to be "met." 16 U.S.C. § 1533(f)(1)(B)(ii). The court's remaining citation is to out-of-circuit precedent providing no statutory analysis and relying on a case decided before the 1988 ESA amendments. See Op. at 434 (citing Fund for Animals v. Rice, 85 F.3d 535, 547 (11th Cir.1996) (citing Strickland v. Morton, 519 F.2d 467, 469 (9th Cir.1975))).

Viewing the ESA as a whole, see Dole v. United Steelworkers of Am., 494 U.S. 26, 42-43, 110 S.Ct. 929, 108 L.Ed.2d 23 (1990), — and consistent with "one of the most basic interpretative canons, that `[a] statute should be construed so that effect is given to all its provisions, so that no part will be inoperative or superfluous, void or insignificant,'" Corley, 556 U.S. at 314, 129 S.Ct. 1558 (internal citation omitted) (alteration in original) — there is only one statutory reading that gives full effect to all of section 4's provisions. For this reason, section 4(c)(2), which provides that determinations to remove a species from the list of endangered or threatened species be "made in accordance with the provisions of subsections (a) and (b)," 16 U.S.C. § 1533(c)(2), must be read in light of Congress's 1988 amendment to section 4(f) to strengthen protections for species. "[W]hen" the existing or properly revised recovery criteria have been "met," delisting is to occur pursuant to section 4(a) and (b). Where a species' recovery plan criteria have not been met, the species remains listed. Where circumstances change, the recovery plan criteria may be revised in the manner prescribed by section 4(f)(4). Because section 4(f)(4) mandates notice be given prior to approval of a plan revision, and that public comments be considered before approval of a plan revision, section 4 is likewise unambiguous that the notice-and-comment period for plan revisions may not run concurrently with the notice-and-comment period for the proposed delisting rule.

A statute that permits only one intelligible outcome is not ambiguous, and the court thus errs in deferring to the Secretary's contrary interpretation, see Final Rule, 73 Fed.Reg. at 50,226, that recovery plans are discretionary and mere guidance documents whose recovery criteria do not inform delisting decisions. Because "the statutory language is unambiguous and the statutory scheme is coherent and consistent," Barnhart, 534 U.S. at 450, 122 S.Ct. 941 (internal citations omitted), the court's inquiry is at an end. For these reasons, I would affirm the judgment of the district court that the Secretary could not delist the Squirrel without either satisfying, or soliciting and considering public comments on the revisions to, the delisting criteria in the Squirrel's recovery plan.

II.

Although the Secretary's statutory challenge is properly resolved under Chevron step one, as there is no ambiguity for the Secretary to interpret, the court errs as well in rejecting appellees' alternative argument that the Secretary violated the Administrative Procedure Act ("APA"). A delisting, no less than a "listing determination[,] is subject to review under the APA and must be set aside if `arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.'" Am. Wildlands v. Kempthorne, 530 F.3d 991, 997 (D.C.Cir.2008) (quoting 5 U.S.C. § 706(2)(A)).

A.

"Under APA notice and comment requirements, `[a]mong the information that must be revealed for public evaluation are the "technical studies and data" upon which the agency relies [in its rulemaking].'" Am. Radio Relay League, Inc. v. FCC, 524 F.3d 227, 236 (D.C.Cir.2008) (quoting Chamber of Commerce v. SEC, 443 F.3d 890, 899 (D.C.Cir.2006)) (alterations in original). "More particularly, `[d]isclosure of staff reports allows the parties to focus on the information relied on by the agency and to point out where that information is erroneous or where the agency may be drawing improper conclusions from it.'" Id. (quoting Nat'l Ass'n of Regulatory Util. Comm'rs v. FCC, 737 F.2d 1095, 1121 (D.C.Cir.1984)) (alteration and emphasis in original). "It is not consonant with the purpose of a rule-making proceeding to promulgate rules on the basis of ... data that, [in] critical degree, is known only to the agency." Portland Cement Ass'n v. Ruckelshaus, 486 F.2d 375, 393 (D.C.Cir.1973), superseded by statute on other grounds, Am. Trucking Ass'ns, Inc. v. EPA, 175 F.3d 1027 (D.C.Cir.1999); see also Idaho Farm Bureau Fed'n v. Babbitt, 58 F.3d 1392, 1403 (9th Cir.1995).

1. In the Final Rule, the FWS relied heavily on the 2007 Analysis of the Squirrel's recovery plan as the basis for eliminating the protections of the ESA for the species. See Final Rule, 73 Fed.Reg. at 50,227. It also relied on the 2006 5-Year Review.7 See id. Neither the 2007 Analysis nor the 2006 5-Year Review were published in the Federal Register, see id.; indeed, the 2007 Analysis was created after the close of the public comment period on the proposed rule, see id., and counsel for the Secretary was unable during oral argument to indicate where a member of the public could gain access to the 2007 Analysis prior to, or even after, promulgation of the Final Rule. See Oral Arg., at 20:44-22:08.

Nor did the notice of proposed rulemaking to delist the Squirrel provide adequate substitute notice of the recovery plan revisions set forth in the 2007 Analysis. Instead, although it generally outlined the FWS's reasons for concluding recovery of the Squirrel had occurred, see Proposed Rule to Remove Squirrel From List of Endangered Species, 71 Fed.Reg. 75,924 (Dec. 19, 2006) ("NPRM"), the public had no opportunity to comment on the revisions of criteria one and three in the Squirrel's recovery plan, including the definition of "persistence" in the 2007 Analysis and in the Final Rule. Neither, therefore, was there an opportunity for the FWS to consider, as Congress required, "all information presented during the public comment period [on the recovery criteria revisions]," 16 U.S.C. § 1533(f)(4), because no such comment period occurred. Even if the FWS was generally aware, through comments submitted in response to the NPRM, of criticisms of rejection of initial recovery plan criteria, "`knowing' is not, in any event, the same as actually considering the problems raised by" commenters. Gerber v. Norton, 294 F.3d 173, 183 (D.C.Cir.2002). The jettisoned procedures of section 4(f)(4) were designed to facilitate such consideration.

2. Furthermore, the NPRM gave no indication that the FWS intended to abandon "population" as the relevant standard in assessing the Squirrel's recovery. "Given the strictures of notice-and-comment rulemaking, an agency's proposed rule and its final rule may differ only insofar as the latter is a `logical outgrowth' of the former." Envtl. Integrity Project v. EPA, 425 F.3d 992, 996 (D.C.Cir.2005). "A rule is deemed a logical outgrowth if interested parties `should have anticipated' that the change was possible, and thus reasonably should have filed their comments on the subject during the notice-and-comment period." Northeast Md. Waste Disposal Auth. v. EPA, 358 F.3d 936, 952 (quoting City of Waukesha v. EPA, 320 F.3d 228, 245 (D.C.Cir.2003)).

The NPRM solicited comments on four topics, three of which explicitly sought input on Squirrel population:

71 Fed.Reg. at 75,924 (emphases added). The NPRM sought no comments on the use of "persistence," rather than population, as the relevant standard, see Ass'n of Private Sector Colls. & Univs. v. Duncan, 681 F.3d 427, 461 (D.C.Cir.2012), and only vaguely referenced "presence" and "persistence," see NPRM, 71 Fed.Reg. at 75,926 (citing "strong evidence of the [Squirrel's] continued presence throughout its range"); id. (citing ability of Squirrel to "adjust its foraging and denning behavior... to persist in and around ... forest patches"); id. at 75,928 (citing Virginia laws as "ensur[ing] the [Squirrel's] persistence in Virginia into the foreseeable future"); id. at 75,929 (noting that southern flying squirrel "does not appear to be affecting population persistence of the [Squirrel]"); id. at 75,929-930 (citing surveys showing Squirrel "persistent at multiple locations for multiple generations," stating that "protected habitat should allow for persistence of viable populations" and concluding that "available information shows that the [Squirrel] is persisting throughout its historic range."). The FWS nowhere indicated, contrary to the request for comments on population size, that "population" would be replaced as the recovery standard with "persistence" (or how "persistence" was to be defined). And two of the vague references to "persistence" fall in the same sentences in which NPRM mentions "population." See id. at 75,929-930.

Given the vague references to "persistence" and the explicit requests to comment on population size, interested parties were not reasonably apprised that they should submit comments on the use of "persistence," rather than population, as the standard. The text of the NPRM provided no basis for anticipating that the FWS "considers persistence to be the best indicator of successfully reproducing populations for [the Squirrel]," Final Rule, 73 Fed.Reg. at 50,227. Yet the FWS's reliance on "persistence," and how it was to be defined, were critical shifts in the standard for the Squirrel's recovery that presented the occasion for notice to the public so comments could address whether the FWS's definition complied with the ESA's stated purpose of "conservation" and the FWS could consider those comments before amending the recovery plan, much less completely delisting the Squirrel. The fact that some commenters criticized the lack of population data in the NPRM and the lack of a definition of "persistence," see Final Rule, 73 Fed.Reg. at 50,227, cannot eliminate the FWS's obligation to provide notice of its intent to substitute "persistence" for a "population" standard and of the definition of "persistence" and how it related to ESA's policy of conservation. See Fertilizer Inst. v. EPA, 935 F.2d 1303, 1312 (D.C.Cir.1991). Having directed parties to focus on population size in their comments, the FWS may not "use the rulemaking process to pull a surprise switcheroo." Envtl. Integrity Project, 425 F.3d at 996.

The court nowhere addresses these notice problems, despite the fact that this argument appears explicitly (and repeatedly) in appellees' brief, see, e.g., Appellees' Br. 36-40, with pin cite citations to and quotations from this circuit's logical outgrowth doctrine cases. The argument is not, as the court concludes, see Op. at 438, forfeited.

B.

The court's approval of the Secretary's reliance on the Squirrel's "persistence" as the standard for delisting, see Op. at 434-36, is also contrary to the repeated, unambiguous distinction in the ESA between conservation of a species and its mere survival, id. § 1532(3).

The FWS defined "persistence" as "continuing captures of [the Squirrel] over multiple generations at previously documented sites throughout its historical range," Final Rule, 73 Fed.Reg. at 50,227; see also 2007 Analysis at 2. Stating that "analysis ... shows no evidence of localized extirpation since the [Squirrel] was listed" and that "[t]he [Squirrel] persists in or near all of the historical areas where it was originally known at the time of listing," Final Rule, 73 Fed.Reg. at 50,229, the FWS determined that the Squirrel is not extinct and some (although unclear how many) continue to survive after multiple generations.

1. ESA section 4(b) requires that "[t]he Secretary shall make determinations ... solely on the basis of the best scientific and commercial data available." 16 U.S.C. § 1533(b)(1)(A); see id. § 1533(c)(2). The court concludes the Secretary properly relied on available data on "persistence," see Op. at 434-36, citing Southwest Center for Biological Diversity v. Babbitt, 215 F.3d 58, 60-61 (D.C.Cir.2000), which held that "the Secretary has no obligation to conduct independent studies." But Southwest Center concerned a dispute over whether population estimates supported a decision to list a species, not over a shift in the relevant standard for determining whether to delist a species. In that case, the court approved the Secretary's reliance on estimates of species' population where an actual count was unavailable, noting that the Secretary was not alleged to have "acted on the basis of no data." Id. at 61. By contrast, here the Secretary delisted the Squirrel on the basis of "no data," id. (emphasis added), regarding population. Instead, the Secretary shifted the standard from population numbers, see NPRM, 71 Fed.Reg. at 75,924; Recovery Plan at 18, to the Squirrel's mere presence/persistence, Final Rule, 73 Fed.Reg. at 50,227, an entirely different concept. As appellees point out, see Appellees' Br. at 49, "population" is a measure of quantity; "persistence" is a measure mere survival, or existence, of the species.

Indeed, one of the main reasons stated in the NPRM for delisting was the FWS's conclusion that there had been "an increase in the number of individual squirrels." NPRM, 71 Fed.Reg. at 75,924. In response to comments, however, the FWS acknowledged that "use of the phrase `increase in number if [sic] individual [Squirrels]' was not accurate, as [the FWS] ha[s] not estimated the size of the [Squirrel] population." See Final Rule, 73 Fed.Reg. at 50,230. If there is no data available to answer the relevant question, section 4(b)(1)(A) does not permit the Secretary to answer another question that does have supporting data. The court's reasoning presumes that the Secretary may begin with a conclusion (to delist) and then rely on some data remotely related to the species even if in answer to a question untethered to the ESA's primary goal of recovery and conservation, to satisfy the "best... data available" standard. The "best... data available" standard cannot be used as an excuse to avoid implementing the recovery plan criteria. If the Secretary (or the FWS) concludes the available data suggests recovery but is insufficient to satisfy the recovery plan criteria, then the recovery plan must be revised in the manner prescribed by section 4(f)(4). This is the process Congress mandated, and it ensures that the criteria for recovery, and the data by which they are measured, are the best available.

2. Even assuming section 4(b)(1)(A) permitted the Secretary to change the standard used to measure a listed species' recovery, the plain text of the ESA precludes the Secretary's choice of "persistence." The ESA defines "conservation" as "the use of all methods and procedures which are necessary to bring any endangered species or threatened species to the point at which the measures provided pursuant to [the ESA] are no longer necessary." 16 U.S.C. § 1532(3). It requires post-delisting monitoring for "all species which have recovered to the point at which the measures provided pursuant to [the ESA] are no longer necessary." Id. § 1533(g)(1). As other circuits have recognized, Congress unambiguously required more than simply a species' continued survival in determining whether it is to be protected under the ESA. In considering the Secretary's regulations implementing the critical habitat provision, 16 U.S.C. § 1536(a)(2), the Ninth Circuit concluded that "the ESA was enacted not merely to forestall the extinction of species (i.e., promote a species survival), but to allow a species to recover to the point where it may be delisted." Gifford Pinchot Task Force v. U.S. Fish & Wildlife Serv., 378 F.3d 1059, 1070 (9th Cir.2004). "The purpose[] of [the ESA] ... [is] to provide a program for the conservation of [] endangered species and threatened species....," 16 U.S.C. § 1531(b); consequently "survival" and "recovery" were distinct goals of the ESA, see Gifford, 378 F.3d at 1070. The Fifth Circuit reached the same conclusion in Sierra Club v. Fish & Wildlife Service, 245 F.3d 434, 441-42 (5th Cir. 2001), observing that "`[c]onservation' is a much broader concept than mere survival," id. at 441; cf. New Mexico Cattle Growers Ass'n v. Fish & Wildlife Serv., 248 F.3d 1277, 1283 n. 2 (10th Cir.2001).

Congress repeatedly referred in the ESA to "survival" as a separate status than "conservation" or "recovery." See 16 U.S.C. §§ 1533(f)(1) & (f)(1)(B)(i) (mandating recovery plans for the "conservation and survival" of species); id. §§ 1535(c)(1) & (c)(2) (authorizing cooperative agreements with State agencies that have "an adequate and active program for conservation of endangered and threatened species," id. §§ 1535(c)(1) & (c)(2), and are also authorized to "conduct investigations to determine the status and requirements for survival," id. §§ 1535(c)(1)(C) & (c)(2)(C)); id. §§ 1535(d)(1)(B) & (E) (authorizing financial assistance to states based on the state's capacity to "proceed with a conservation program" and the "urgency to initiate a program to restore and protect [a species] ... in terms of survival of the species"); § 1539(a)(2)(B)(iv) (permitting "taking" of species where, among other things, it "will not appreciably reduce the likelihood of the survival and recovery of the species"). From the statutory text, "it is clear that Congress intended that conservation and survival be two different (though complementary) goals of the ESA." Gifford, 378 F.3d at 1070. The Secretary's regulation on delisting, which provides that "[a] species may be delisted on the basis of recovery only if the best scientific and commercial data available indicate that it is no longer endangered or threatened," 50 C.F.R. § 424.11(d)(2) (1984) (promulgated prior to the 1988 ESA Amendments), must be viewed in light of Congress's distinction between "survival" and "recovery." Consequently, the Secretary's decision to delist the Squirrel on the basis of its "persistence" — that is, its bare survival — is a statutorily insufficient basis for delisting.8

In sum, contrary to Congress's plain text, the court jettisons the protections in the ESA for endangered and threatened species and leaves the Secretary (and the FWS) more insulated and less informed than Congress contemplated in strengthening the ESA in 1988. The court's approval of the FWS's covert revisions to the Squirrel's recovery plan, surprise introduction of a new recovery standard in the Final Rule, and adoption of a delisting standard unambiguously foreclosed by the ESA leaves little of the species' protections Congress provided in the ESA, much less of APA requirements.

Accordingly, I respectfully dissent.