HENRY, Circuit Judge.
In October 2001, the Forest Service approved a timber-harvesting project in Utah’s Fishlake National Forest. Utah Environmental Congress (“UEC”), an environmental organization, filed a petition for review, and the district court dismissed the petition and affirmed the project’s authorization. UEC alleges on appeal that the Forest Service (1) did not properly select and monitor the Management Indicator Species (“MIS”) that it used to determine the effects of management activities on other species, and (2) did not consider a reasonable range of alternatives to the project. Exercising jurisdiction under 28 U.S.C. § 1291, we reverse the judgment below and remand to the district court with instructions to vacate the Forest Service’s approval of the project. Intervening circuit precedent requires the agency to collect quantitative population data on actual MIS populations before authorizing a project under the rules the Forest Service uses.
I. BACKGROUND
The Thousand Lakes Community Forestry Initiative Project (“the Project”) is located in Wayne County, Utah, on the 1.5-million-acre Fishlake National Forest. A collaboration of timber representatives, environmentalists, politicians, and federal land managers undertook the Project (1) “to reduce the overall stand densities of the [spruce and aspen] stands that are at the highest risk of [spruce beetle] infestation, while maintaining a forested appearance” and (2) “to provide forest products to resource dependent industries in an economically feasible manner.” Administrative Record vol. I, at (“1 AR”) 000044. The Project encompasses timber harvests on 219 acres, with approximately one-half mile of road re-construction and post-treatment activities to minimize erosion and the use of authorized vehicles. No new road building is involved. The Project would use salvage (removal of unhealthy trees), sanitation (removal of dead trees in excess of resource needs for habitat), and commercial thinning (removal to reduce overall stand densities).
The Forest Service manages the Fish-lake National Forest at two different levels. “At the first level, the Forest Service develops the Forest Plan, [which is] a broad, programmatic document, accompa[1109]*1109nied by an Environmental Impact Statement and public review process conducted in accordance with the National Environmental Policy Act [‘NEPA’].” Colorado Envtl. Coalition v. Dombeck, 185 F.3d 1162, 1167-68 (10th Cir.1999); 16 U.S.C. § 1604. To this end, the Forest Service adopted the Fishlake National Forest Plan (the “Forest Plan”) in 1986 to maintain the Fishlake National Forest. As part of its substantive responsibilities, the Forest Plan must “provide for diversity of plant and animal communities based on the suitability and capability of the specific land area in order to meet overall multiple-use objectives.” 16 U.S.C. § 1604(g)(3)(B).
At the second level, “the Forest Service implements the Forest Plan by approving (with or without modification) or disapproving particular projects,” such as the Thousand Lakes Community Forestry Initiative Project. Dombeck, 185 F.3d at 1168. Individual projects are also subject to NEPA and must comply with the Forest Plan and the National Forest Management Act (“NFMA”). 16 U.S.C. § 1604(i).
In November 1999, the Forest Service began preparing an Environmental Assessment (“EA”) of the Project. The Forest Service examined the Project’s potential impact to wildlife, soils, vegetation, and other resources, and it prepared a Biological Assessment and Biological Evaluation of impacts to sensitive plant and animal species. In May 2001, the Forest Service published its EA for the Project, and in October 2001, the district ranger issued a finding of no significant impact (“FONSI”) and Decision Notice approving the Project. UEC brought an administrative appeal, and the Forest Service issued a Final Decision in February 2002 that affirmed the district ranger. 1 AR 000032.
UEC challenged the Forest Service’s approval of the Project in federal district court. In September 2003, the district court dismissed UEC’s petition for review and affirmed the agency’s Decision Notice and FONSI. UEC now appeals the Forest Service’s approval of the Project on two of the three grounds rejected by the district court. UEC alleges that the Forest Service, (1) did not properly select and monitor certain Management Indicator Species and (2) did not consider a reasonable range of management alternatives.1
II. STANDARD OF REVIEW
We take “an independent review of the agency’s action” and are not bound by the district court’s factual findings or legal conclusions. Olenhouse v. Commodity Credit Corp., 42 F.3d 1560, 1569 n. 16, 1577 n. 27 (10th Cir.1994). We review the Forest Service’s decision under the Administrative Procedures Act and set aside agency action if it is “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” 5 U.S.C. § 706(2)(A). We give deference to an agency’s interpretation, “especially when that interpretation involves questions of scientific methodology.” Dombeck, 185 F.3d at 1170. In addition, “[t]he agency, not the reviewing court, is entrusted with the responsibility of considering the various modes of scientific evaluation and the[1110]*1110ory and choosing the one appropriate for the given circumstances.” City of Bridgeton v. FAA, 212 F.3d 448, 459 (8th Cir.2000) (quotation marks and internal citation omitted), cited with approval in Custer County Action Ass’n v. Garvey, 256 F.3d 1024, 1036 (10th Cir.2001). “[T]he court is not empowered to substitute its judgment for that of the agency.” Citizens to Preserve Overton Park v. Volpe, 401 U.S. 402, 416, 91 S.Ct. 814, 28 L.Ed.2d 136 (1971).
III. APPLICABLE FOREST SERVICE REGULATIONS
Before we proceed to UEC’s separate challenges, we briefly describe the relevant regulations at issue. Forest Service regulations implement NFMA’s requirement that the government address how forest plans provide for plant and animal diversity. 36 C.F.R. § 219.19 (1999). The Department of Agriculture (“Department”) substantially amended the regulations at issue in November 2000. 65 Fed.Reg. 67,-568 (Nov. 9, 2000). The 2000 regulations included a transition provision that initially delayed their application to project decisions until November 2003. See 36 C.F.R. § 219.35(d) (2001). The Department subsequently proposed revisions to the 2000 regulations and extended the transition period for their application to project decisions. See 36 C.F.R. § 219.35(d) (2004); 68 Fed.Reg. 53,294 (Sept. 10, 2003); 67 Fed.Reg. 72,770 (Dec. 6, 2002). Under the transition provisions, from November 9, 2000, until the promulgation of final planning regulations, the Forest Service was directed to “consider the best available science in implementing” a forest plan. 36 C.F.R. § 219.35(a), (d) (2004). When the Forest Service issued its Decision Notice in October 2001 and filed its appellate brief in this court in May 2004, it did not contend that the transition provisions of the 2000 regulations applied to the Project. The Forest Service only considered authorization of the Project under the “1982 rule,” those regulations in place prior to the 2000 amendments. See 36 C.F.R. § 219.19 (1999).
In a Rule 28(j) letter filed one week before oral argument, the Forest Service informed us of the Department’s publication of an interpretative rule in September 2004. 69 Fed.Reg. 58,055 (Sept. 29, 2004); see also Fed. R.App. P. 28(j). The interpretative rule explained that the 2000 regulations rendered the 1982 rule inoperative for project-specific decisions made after November 9, 2000. The interpretative rule clarified that, during the transition period between November 2000 and promulgation of a final rule, the Forest Service should use the “best available science” under § 219.35(a) for project decisions. 69 Fed.Reg. at 58,056. New rules replaced the 2000 planning regulations in January 2005. See 70 Fed.Reg. 1023 (Jan. 5, 2005).
We conclude, however, that the transition provisions of the 2000 regulations do not apply to the Project’s authorization, and we review Forest Service obligations under the 1982 rule instead. First, the Forest Service stated at oral argument that it “elected” to apply the 1982 rule when it approved the Project. See 36 C.F.R. § 219.35(b) (2001); see also Aple’s Br. at 8 n. 2 (“The Forest Service promulgated the [] Forest Plan and decided to undertake the Thousand Lakes project under the 1982 planning rule.”) (emphasis added). While the Forest Service now contends that it was not required to comply with the 1982 rule, it nonetheless exercised its discretion to apply the 1982 rule when issuing its Decision Notice and Final Decision. The agency’s choice of methodology is entitled to deference. See Custer County Action Ass’n, 256 F.3d at 1036.
[1111]*1111Second, in another case involving a Fishlake National Forest project, we treated the 1982 rule, rather than the transition provision of § 219.35(a), as “[t]he regulations in effect at the time of the disputed Forest Service decisions” made in December 2000. Utah Envtl. Cong. v. Bosivorth, 372 F.3d 1219, 1221 n. 1 (10th Cir.2004) (“UEC I ”). While we now have the benefit of the Department’s post hoc explanation of the applicable regulations since we issued UEC I, interpretative rules “do not have the force and effect of law and are not accorded that weight in the adjudicatory process.” Shalala v. Guernsey Mem’l Hosp., 514 U.S. 87, 99, 115 S.Ct. 1232, 131 L.Ed.2d 106 (1995). Because we review the Forest Service’s authorization for an abuse of discretion, it is logical to review that discretion based on the 1982 rule that was, according to UEC I, in place when the discretion was authorized. But see Forest Watch v. United States Forest Serv., 410 F.3d 115, 118 (2d Cir.2005) (citing UEC I for the proposition “that the relevant date for the purpose of determining which rule applies is the date the final agency decision was made,” while concluding that the 2000 transitional provisions govern an August 2002 project decision, despite the Forest Service’s application of the 1982 rule to the project’s approval).
Third, the Forest Plan identifies several MIS and specifies in its “Implementation” chapter that the Forest Service use “Population Trends” for MIS monitoring. 1 AR 000157 (Forest Plan at 11-29 to 11-30, V-6); UEC I, 372 F.3d at 1232-33 (Baldock, J., concurring) (finding it “unnecessary to delve into the arduous task of interpreting [the 1982 rule]” because “[t]he' Fishlake Forest Plan requires the collection of quantitative data to fulfill its MIS monitoring requirements”). Projects must comply with the Forest Plan, see 16 U.S.C. § 1604(i), and the Forest Plan’s diversity provisions reflect the 1982 rule.
Thus, we have multiple grounds to apply the 1982 rule and its attendant reliance on MIS analysis to our review of the Project’s authorization. Code of Federal Regulations citations (unless otherwise noted) refer to the 1999 edition of Part 219, the last published edition before the publication of the 2000 amendments.
IV. ANALYSIS
UEC claims that the Forest Service acted arbitrarily, capriciously, and contrary to law by (1) failing to select and monitor Management Indicator Species and (2) failing to consider a reasonable range of project alternatives.
A. Selection and monitoring of MIS under 36 C.F.R. § 219.19
To assess habitat viability, the Forest Service estimates a project’s effect on certain representative species.
Fish and wildlife habitat shall be managed to maintain viable populations of existing native and desired non-native vertebrate species in the planning area.
(a)(1) In order to estimate the effects of each [management] alternative on fish and wildlife populations, certain vertebrate and/or invertebrate species present in the area shall be identified and selected as management indicator species and the reasons for their selection will be stated. Those species shall be selected because their population changes are believed to indicate the effects of management activities.
36 C.F.R. § 219.19.
MIS are similar to the famed canaries used to monitor air in coal mines. They are a “bellwether” “for other species that have the same special habitat needs or population characteristics,” Inland Empire [1112]*1112Pub. Lands Council v. United States Forest Service, 88 F.3d 754, 762 n. 11 (9th Cir.1996), and serve as “a proxy for determining the effects of management activities on other species.” Forest Guardians v. United States Forest Serv., 180 F.Supp.2d 1273, 1281 (D.N.M.2001). The regulations require that “[p]opulation trends of the management indicator species will be monitored and relationships to habitat changes determined.” 36 C.F.R. § 219.19(a)(6). In addition, “[inventories shall include quantitative data making possible the evaluation of diversity in terms of its prior and present conditions.” Id. § 219.26.
To meet the diversity provisions under § 219.19, the Forest Plan designates two categories of MIS: “ecological indicator” and “high interest” species. The ecological indicators include five groups of species and one individual species: the sage-nester guild, riparian guild, cavity-nester guild, resident trout, macro-invertebrates, and northern goshawk. A “guild” is a group of species that exploits the same class of environmental resources in the same way. The Forest Service selects ecological indicator species according to five criteria: (1) “affinity for the vegetation type,” (2) “life cycle keyed to a vegetation type,” (3) “[sjensitivity to habitat change,” (4) “[r]el-ative ease of monitoring, i.e., easily recognized and adequate numbers,” and (5) being “[sjomewhat representative of other species which use the same vegetation type.” 1 AR 000157 (Forest Plan’s Final Environmental Impact Statement (“FEIS”) at III — 34).
“High interest” MIS include elk, mule deer, Bonneville cutthroat trout, and Rydberg’s milkvetch. The Forest Service chose “high interest” MIS “because of their threatened, endangered, or sensitive status, social or economic importance, or high public interest.” Id. at III — 36. UEC does not challenge, and we need not examine, the Forest Service’s selection of “ecological indicator” and “high interest” guilds for the Project. Rather, UEC challenges the chosen MIS within these groups of species.
Our recent decision in UEC I addressed the Forest Service’s MIS obligations with regard to another project in the Fishlake National Forest, and it resolves two important issues here. First, UEC I makes clear that “the regulations anticipate application of § 219.19 to project level as well as plan level management actions.” 372 F.3d at 1225. As we noted in UEC I, this approach is consistent with other circuits. See Sierra Club v. Martin, 168 F.3d 1, 6 (11th Cir.1999) (recognizing “that the regulations refer to the formulation of Forest Plans rather than to specific projects proposed under already enacted Forest Plan” but that “the planning process does not end with the Forest Service’s approval” and “continuefs] throughout the Plan’s existence”); Inland Empire Pub. Lands Council, 88 F.3d at 760 n. 6 (“Because any district contained within the boundaries of a forest having a plan would be an ‘area ... covered by a ... forest plan,’ it would [] also be a planning area governed by Regulation 219.19.”) (quoting 36 C.F.R. § 219.3). Thus, the Forest Service’s obligations under § 219.19 apply to the Project.
Second, we decided in UEC I that the Forest Service must use “actual, quantitative population data” to meet MIS monitoring obligations under § 219.19. 372 F.3d at 1226. “[T]o effectuate its MIS monitoring duties under the language of its regulations, the Forest Service must gather quantitative data on actual MIS populations that allows it to estimate the effects of any forest management activities on the animal population trends, and determine the relationship between manage[1113]*1113ment activities and population trend changes.” Id. at 1227; see also Martin, 168 F.3d at 6 (examining § 219.19(a)(6) and concluding that “[i]t is implicit that population data must be collected before it can be monitored and its relationship determined”). Prior to UEC I, the Forest Service contended that it need not conduct “head-counts” of MIS in a planning area because it had discretion to assess a project’s effects on MIS using habitat data, population data, or both. See generally Andrew Orlemann, Note, Do the Proposed Forest Service Regulations Protect Biodiversity? An Analysis of the Continuing Viability of “Habitat Viability Analysis, ” 20 J. LAND RESOURCES & Envtl. L. 357, 360-374 (2000) (describing the split among federal circuits as to the adequacy of habitat viability analysis under § 219.19).
With these requirements in mind, we examine UEC’s challenges to the Forest Service’s monitoring of the (1) sage-nester guild, (2) riparian guild, (3) cavity-nester guild, (4) northern goshawk, and (5) Mexican spotted owl.
1. Sage-nester guild: sage grouse as MIS
UEC first maintains that the Forest Service (1) did not determine the presence of sage nesters in the Project area and (2) did not explain why it chose the sage grouse as the guild’s sole representative. The Forest Service concluded, based on a July 2000 field survey and unpublished radio telemetry data, that sage grouse are not present in the Project area. 1 AR 000079, 2 AR 000773, 001247. It also used aerial photos and vegetation mapping to conclude that the Project includes no sagebrush habitat. The EA noted that the Project’s road reconstruction would occur along an existing road through “potential sage nesting habitat.” 1 AR 000079. However, “because sage nesters have not been documented along the road and the additional disturbance would involve only .01% or less of total available sagebrush habitat on Thousand Lakes Mountain, [effects] to sage nesting species [are] expected to be minimal as a result of implementing the Proposed Action.” Id.
Because no sage grouse were documented within the Project area, UEC contends that the Forest Service improperly selected an MIS for the sage-nester guild. The logic of this position finds support in § 219.19. The Forest Service must identify and select “certain vertebrate and/or invertebrate species present in the area ” “to estimate the effects of each alternative on fish and wildlife populations,” and it must also state “the reasons for their selection.” 36 C.F.R. § 219.19(a)(1) (emphasis added). “These [MIS] shall be selected because their population changes are believed to indicate the effects of management activities.” Id.
The plain language of § 219.19 requires that selected MIS have some documentation of a presence within the contemplated project area. First, we recognized in Dombeck that § 219.19 “clearly presupposes the ascertainable presence of a species’ population within a given planning area.” 185 F.3d at 1170. Noting UEC I’s conclusion that § 219.19 applies at the project level, it follows that the regulations’ requirement to select MIS “present in the area” should also apply to the project level. Second, it is axiomatic that choosing MIS actually “present in the [project] area” most appropriately advances the regulations’ goal of diversity and best meets the UEC I directive to monitor MIS population trends and gather inventory data. In Dombeck, the “Forest Service logically did not select the rare and elusive lynx as a Management Indicator Species” because “there simply is no lynx population data available to the Forest Service.” 185 F.3d [1114]*1114at 1169. See also Oliver A. Houck, On the Law of Biodiversity and Ecosystem Management, 81 Minn. L.Rev. 869, 888 (1997) (“Congress intended not only that diversity be ‘considered’ in forest planning, but that it be achieved.”).
This court has not previously addressed the Forest Service’s selection of an MIS with no confirmed or unconfirmed population in a project area. In UEC I, the Forest Service acknowledged that sage grouse are potentially present, referenced “unconfirmed sightings” in the project area, but provided no quantitative data to confirm the sage grouse’s presence. 372 F.3d at 1229. We determined in UEC I that the Forest Service did not comply with § 219.19, without addressing the propriety of the sage grouse’s selection as an MIS. In UEC I, “despite unconfirmed sightings of the sage grouse within the [ ] project area, the record reflected] no attempt by the Forest Service to confirm the presence of this MIS.” Id. at 1230.
Under a plain reading of § 219.19 and UEC I, we conclude that the Forest Service must select an MIS with some evidence that it is “present in the [project] area.” The Forest Service must then collect “actual, quantitative population data,” id. at 1226, to monitor population trends and to determine relationships to habitat changes. See 36 C.F.R. § 219.19(a)(6). It must also confirm, with “good faith efforts,” the presence of the selected MIS within a project area. UEC I, 372 F.3d at 1230. If no MIS representative is “present in the [project] area,” the Forest Service must show good-faith efforts to confirm and explain the absence of selected MIS. It may be that the Forest Service selected an improper guild, or actions previously taken may have had a significant deleterious effect on the chosen MIS. “[W]here impossible, the Forest Service is not required by the applicable statutes and regulations to collect population data.” Id. at 1229.
The Forest Service must select within each guild an appropriate MIS that is present in the project area. Selecting only one or two (or a few) acceptable MIS actually present in a project area cannot satisfy the overall monitoring obligations of § 219.19. See Martin, 168 F.3d at 7 (concluding that the Forest Service violated §§ 219.19 and 219.26 because it “ha[d] no population data for half of the MIS in the Forest and thus [could not] reliably gauge the impact of the timber projects on these species”). Here, the Forest Service selected its ecological indicator MIS according to several criteria, including the “[r]elative ease of monitoring.” 1 AR 000157 (FEIS at III-34). The Forest Plan specifically notes that the sage-nester guild “[ijncludes sage dependent species, [i.e.,] sage thrasher, vesper sparrow, sage sparrow, etc. (on a case by case basis) dependent upon proposed project or management activity.” Id. at III — 34 to III — 35.
Even with its discretion to select an MIS based on the proposed project, the Forest Service erred when it selected a sage-nester MIS with no documentation of any presence within the Project area. First, neither the Forest Service’s 2000 field survey nor its radio telemetry data documented sage grouse in the Project area. Second, a Fishlake National Forest breeding bird survey route and the United States Geological Survey’s bird population trends also provide no quantitative data on sage grouse in or near the Project area. 2 AR 001326-30, 001335-49. Third, while not dispositive, the sage grouse was considered but ultimately rejected as an MIS in the larger Forest Plan. 1 AR 000157 (FEIS at III — 30) (listing sage grouse as a “Species Considered But Not Selected for MIS,” with an estimated population of “Unknown”).
[1115]*1115In sum, the Forest Service provides neither evidence of the sage grouse population nor even a single sighting in the Project area. We note that UEC does not contest the Forest Service’s finding of virtually no sagebrush habitat within the Project. Nonetheless, once the Forest Service determined that the sage-nester guild was “appropriate” to monitor under § 219.19(a)(1), it was obligated to select an MIS representative with some evidence of a presence in the Project area. The Forest Service does not contend that the selected MIS is a “more reclusive species” or that “there is no technically reliable and cost-effective method of counting individual members of the species,” especially given the Project’s relatively small area. Inland Empire Pub. Lands Council, 88 F.3d at 763 n. 12 (approving of an alternative method to population trend analysis under such circumstances). On remand, the Forest Service must select a more appropriate MIS for the sage-nester guild and monitor its population trends.
2. Riparian guild: Southwestern willow flycatcher as MIS
Similar to its contentions regarding the sage grouse are UEC’s arguments that the Forest Service (1) has not collected quantitative population data on the Southwestern willow flycatcher and (2) has not adequately explained why it chose this MIS for the riparian guild. According to the EA, willow flycatchers were observed in the 1990s in thick willow stands eight air miles northwest of the Project area. 1 AR 000069; see also 2 AR 000768 (Project’s Biologist Report). However, the EA concluded that no potential habitat occurs in the Project area because, “[bjased on field surveys, no willow flycatchers had been documented on the Thousand Lakes Mountain.” 1 AR 000069. The Forest Service also references formal flycatcher surveys in the Fishlake National Forest, including areas of the Loa Ranger District, conducted since 1994. 2 AR 001336. However, we cannot assess these field surveys because they are not in the administrative record. See UEC I, 372 F.3d at 1227 (instructing courts to examine the Forest Service’s approval of a project “based on the administrative record”).
Like the sage-nester guild, the Forest Service selected an MIS representative for the riparian guild that, according to the agency, was not “present in the area.” While the Forest Service may choose a guild representative based “upon the proposed project or management activity,” 1 AR 000157 (FEIS at III-34), it erred in selecting the willow flycatcher as the riparian guild’s MIS. No evidence from prior surveys or unconfirmed sightings suggested that the flycatcher was actually present in the Project area. Thus, the Forest Service has not met its MIS obligations under § 219.19 for this guild.
3. Cavity-nester guild: three-toed ivoodpecker and flamulated owl as MIS
The Forest Service selected two species, the three-toed woodpecker and the flamulated owl, to represent the cavitynester guild. The three-toed woodpecker lives in spruce and aspen habitat and relies on spruce beetles as a primary food source; its population fluctuates with prey abundance. 1 AR 000073, 2 AR 001250. The flamulated owl also feeds on spruce beetles. The Project’s primary objective is to reduce “the overall stand densities of the stands that are at the highest risk of infestation” of spruce bark beetles. 1 AR 000044.
Visual and auditory surveys detected woodpeckers on Thousand Lakes Mountain during 2000, but nesting pairs were not located. 1 AR 000151. The EA con-[1116]*1116eluded that “the proposed treatment acres represent approximately 2% of the total available spruce/aspen habitat for the three-toed woodpecker on Thousand Lakes Mountain. Direct and indirect effects to the three-toed woodpecker are therefore not expected to cause a loss of population viability.... ” 1 AR 000075. We also acknowledge the Forest Service’s pronouncement of “stable trends” in the populations of eighteen cavity nesters, including the three-toed woodpecker, based on a breeding bird survey in Wayne County. 2 AR 001348. However, we do not find sufficient quantitative data on the three-toed woodpecker in the administrative record to determine population trends or to forecast the Project’s effects on its population viability. Nor do we conclude that the Forest Service has demonstrated good-faith efforts to confirm the woodpecker’s absence in the Project area. Accordingly, the Forest Service has not met its MIS monitoring responsibilities under § 219.19 with respect to the three-toed woodpecker.
As to flamulated owls, the Forest Service cites a 1991 study and 1992 study that detected the owls in mixed conifers one mile west and 2.5 miles east of the Project area. 1 AR 000075. The EA recognized that the Project’s proposed salvage harvest would remove potential nesting habitat for the owl, and concedes “[wjhether these owls would tolerate selective logging during the breeding season near a nest site is unknown.” 1 AR 000076. The Forest Service contends that the Project’s impact on the owl will be minimal because (1) the Project area’s vegetation is not the owl’s preferred habitat and (2) the Project area constitutes only a small percentage of available habitat on Thousand Lakes Mountain. Based on the administrative record, the Forest Service selected the flamulated owl as an MIS despite having no documentation of any presence in the Project area. The Forest Service therefore erred because it did not satisfy the requirement under § 219.19 to select an MIS “present in the area.”
4. Northern goshawk
UEC next maintains that the Forest Service did not satisfy its MIS monitoring for the northern goshawk. The Forest Service collected monitoring data through (1) four aerial surveys of the Project area in 2000, (2) ground surveys in 2000 and 2001, and (3) additional surveys and a statewide habitat assessment conducted between 1998 and 2000. 2 AR 001234-45, 001279, 001364-65, 001369. UEC I relied in part on the same population data when concluding that the Forest Service’s actions were adequate. 372 F.3d at 1227-28. We agree with UEC I that, based on the administrative record, the Forest Service fulfilled its § 219.19 MIS monitoring with respect to the northern goshawk.
5. Threatened and endangered species: Mexican spotted owl
UEC finally contends that the Forest Service did not properly monitor the Mexican spotted owl, a threatened and endangered species.2 Regulations implementing NFMA’s diversity mandate provide that “[e]ndangered and threatened plant and animal species identified on state and Federal lists for the planning area” should be selected as MIS where appropriate. 36 C.F.R. § 219.19(a)(1). The Forest [1117]*1117Plan mandates that the Forest Service annually monitor by visual reconnaissance any threatened, endangered, and sensitive animals in the Fishlake National Forest to ensure “no decrease attributed to management activities.” 1 AR 000157 (Forest Plan at V-6).
The Forest Service conducted a Thousand Lakes Mountain survey for the Mexican spotted owl between 1991 and 1994, and no breeding pairs were documented at that time. 1 AR 000070. Surveyors detected a single female in June and July 1991. Because only single birds were detected, the Forest Service concluded that “incidental use by non-breeding individuals was occurring.” Id. Due to the “incidental use,” the EA reported that “[e]ffects to the Mexican spotted owl as a result of the Proposed Action [are] expected to be minimal.” Id. at 000071. However, without more recent and complete quantitative data to monitor population trends, the Forest Service has not satisfied § 219.19 as to the Mexican spotted owl.
6. Conclusion
The Forest Service has not met its obligations under § 219.19 to select and monitor MIS with respect to the sage-nester guild, riparian guild, cavity-nester guild, or Mexican spotted owl.
B. Consideration ofNEPA alternatives
UEC next contends that the Forest Service acted arbitrarily when the Project’s EA formally considered only two alternatives: a no-action alternative and the modified proposed action. Under NEPA, federal agencies prepare an EIS when they propose a “major Federal action[ ] significantly affecting the quality of the human environment.” 42 U.S.C. § 4332(2)(c). “Agencies need not prepare a full EIS, however, if they initially prepare the less detailed [EA] and, based on the EA, issue a [FONSI], concluding that the proposed action will not significantly affect the environment.” Pennaco Energy, Inc. v. Dep’t of Interior, 377 F.3d 1147, 1150 (10th Cir.2004) (internal quotation marks omitted); see also 40 C.F.R. §§ 1501.4, 1508.9. “A properly drafted EA must include a discussion of appropriate alternatives to the proposed project.” Davis v. Mineta, 302 F.3d 1104, 1120 (10th Cir.2002).
We review the Forest Service’s actions only to “insure that the agency has taken a ‘hard look’ at environmental consequences.” Kleppe v. Sierra Club, 427 U.S. 390, 410 n. 21, 96 S.Ct. 2718, 49 L.Ed.2d 576 (1976). “In deciding whether the [agency] acted arbitrarily by not considering certain alternatives, we remain mindful that an agency decision concerning which alternatives to consider is necessarily bound by a rule of reason and practicality.” Airport Neighbors Alliance, Inc. v. United States, 90 F.3d 426, 432 (10th Cir.1996). “[A]n agency need not analyze the environmental consequences of alternatives it has in good faith rejected as too remote, speculative, ... impractical or ineffective.” Id. “While it is true that defendants could reject alternatives that did not meet the purpose and need of the project, they could not define the project so narrowly that it foreclosed a reasonable consideration of alternatives.” Davis, 302 F.3d at 1119 (internal quotation marks omitted).
The Project was designed “to improve or maintain habitat conditions in order to promote long-term ecosystem health for the benefit of people, wildlife, and fish.” 1 AR 000136 (Decision Notice and FONSI). The primary objective is to reduce densities of aspen and spruce stands that are at most risk of spruce beetle infestation, while maintaining a forested appearance. Under the secondary objective, the Project [1118]*1118will supply local resource-dependent enterprises with raw materials in an “economically feasible manner.” Id.
UEC argues that the Forest Service violated NEPA by defining the project’s objectives so narrowly that the only possible alternatives were the proposed project or no action. It also contends that the Forest Service could have addressed the spruce beetle infestation in non-commercial ways and considered alternatives that did not revalidate roadless areas. According to the Forest Service, a non-commercial alternative would achieve neither project objective. Given the Project’s dual objectives and the agency’s discretion to chose those objectives, see Airport Neighbors Alliance, 90 F.3d at 432, the Forest Service examined a reasonable range of alternatives and did not act arbitrarily when it considered only the no-action alternative and the modified proposed action.
V. CONCLUSION
We REVERSE the district court’s order affirming authorization of the Project and REMAND to the district court with instructions to VACATE the Forest Service’s approval of the Project.3