LUCERO, Circuit Judge.
At issue is the extent of the United States Forest Service’s obligations under its regulations to monitor management indicator species before approving forest management activities. Finding the Forest Service’s monitoring efforts sufficient, the district court affirmed the Record of Decision authorizing the Monroe Mountain Ecosystem Restoration Project. Exercising jurisdiction pursuant to 28 U.S.C. § 1291, we REVERSE and REMAND.
I
Monroe Mountain, part of the Fishlake National Forest approximately 140 airline miles south of Salt Lake City in Central Utah, is the center of the Monroe Mountain Ecosystem Restoration Project (the “Monroe Project”). The Monroe Project area covers approximately 50,000 acres, consisting of forest (primarily fir occasionally mixed with aspen) and non-forest (grass and sagebrush) land.
As part of the National Forest System, the Fishlake National Forest is maintained under a Forest Management Plan (the “Fishlake Forest Plan”), pursuant to the National Forest Management Act (“NFMA”), 16 U.S.C. § 1604; the Fish-lake Forest Plan was approved in 1986. Forest Management Plans are to be revised every ten to fifteen years and can be amended. 16 U.S.C. § 1604(f). To accomplish NFMA’s broad land management responsibilities, the Forest Service has promulgated regulations, including: further clarification of general forest planning procedures, 36 C.F.R. § 219.10; minimum forest management requirements, 36 C.F.R. § 219.27; requirements to ensure diversity of plant and animal species, 36 C.F.R. § 219.26; and habitat requirements for native and desirable non-native species, 36 C.F.R. § 219.19.
To execute specific Forest Management Plans, the Forest Service proposes individual projects. See, e.g., Ohio Forestry Assoc., Inc. v. Sierra Club, 523 U.S. 726, 729-730, 118 S.Ct. 1665, 140 L.Ed.2d 921 (1998) (describing the steps involved in proposing and adopting a site-specific action under the Wayne National Forest’s Management Plan); Colorado Environmental Coalition v. Dombeck, 185 F.3d 1162, 1168 (10th Cir.1999) (“The Forest Service implements the Forest Plan by approving ... or disapproving particular projects_”). Individual projects must comply with the NFMA’s enacting regulations. 16 U.S.C. § 1604(i). Specifically, any projects undertaken by the Forest Service in the Fishlake National Forest must be consistent with the Fishlake Forest Plan. 16 U.S.C. § 1604(i), 36 C.F.R. § 219.10(e) (1999).1
In 1997, a Forest Service Interdisciplinary Team considered portions of Monroe Mountain for “immediate ecosystem restoration treatment.” (Administrative Rec[1222]*1222ord Vol. 2 (“2 AR”) Doc. 485 at 1-2.) After assessing conditions in the relevant area, the Monroe Mountain Ecosystem Restoration Project (the “Monroe Project”) was developed. The Monroe Project’s Final Environmental Impact Statement (“FEIS”), which analyzes the potential effects of several management alternatives, states that the primary purpose of the Monroe Project “is to move the forest and grassland ecosystems within the project area toward properly functioning condition of ecological structure and function.” (2 AR Doc. 485 at 1-3.) To that end, six distinct objectives were developed: (1) restore aspen ecosystems, (2) improve watershed conditions, (3) reduce the risk of large wildfires, (4) reduce the potential of epidemic spruce beetle infestation in spruce/fir stands, (5) restore the grass/forb ecosystem to improve wildlife and livestock habitat, and (6) provide timber products for the local economy. To effectuate these objectives, five management alternatives were analyzed to determine the potential effects the Monroe Project might have on the area, as required by the National Environmental Policy Act (“NEPA”), 42 U.S.C. § 4331, et seq.
Proposing several actions to meet the specific objectives of the Monroe Project, the selected alternative includes timber harvesting, prescribed burns, and sagebrush treatments to Monroe Project land outside of Inventoried Roadless Areas (“IRAs”).2 Action within IRAs was limited by the plan to prescribed burning. The Forest Service’s Record of Decision found that the selected alternative met the six objectives of the Project by: (1) converting over 2,500 acres of mixed conifer/aspen stands to aspen stands, thus restoring an abundance of aspen to the forest; (2) increasing riparian resiliency by improving understory vegetation and decreasing sediment production; (3) treating 2,645 acres of mixed conifer/aspen stands to reduce fuel, thereby decreasing the risk of intense wildfires; (4) reducing the density of 248 acres of spruce/fir stands, thereby decreasing the risk of epidemic spruce beetle infestation; (5) converting over 1,300 acres of sagebrush into grass/forb areas, creating more habitat for livestock and wildlife; and (6) producing approximately 8.7 million board feet of timber, thus supporting the local economy. Slightly under 5,000 of the 50,000 Monroe Project area acres were subject to treatment under the selected alternative.
Following the Forest Supervisor’s authorization of the Monroe Project, Utah Environmental Congress (“UEC”) filed suit in district court, alleging the Forest Service’s authorization of the Monroe Project violated NEPA, the NFMA, the Administrative Procedure Act (“APA”), and various Forest Service regulations. In March 2003, the district court found in favor of the Forest Service on all claims. UEC appeals, raising two primary allegations of error: (1) that the Forest Service’s failure to monitor Management Indicator Species (“MIS”) renders its authorization of the Monroe Project violative of the NFMA, the NFMA’s implementing regulations, and the Fishlake Forest Plan; and (2) that the Forest Service’s “revalidation” of roadless areas during the Monroe Project development, which resulted in some previously labeled roadless areas no longer being categorized roadless, was illegal be[1223]*1223cause such an action is not authorized at an individual project level.
II
A
We first address the Forest Service’s Management Indicator Species (“MIS”) monitoring.3 Judicial review of agency action under 5 U.S.C. §§ 701-706 directs reviewing courts to 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). Although our review of an agency’s decision is deferential, we will reverse if the decision is not supported by substantial evidence. Hoyl v. Babbitt, 129 F.3d 1377, 1382 (10th Cir.1997) (citations omitted); Federal Power Comm’n v. Transcontinental Gas Pipe Line Corp., 423 U.S. 326, 331, 96 S.Ct. 579, 46 L.Ed.2d 533 (1976) (recognizing that if an agency’s decision is not supported by the administrative record made, it must be vacated and the matter returned to the agency for further consideration). “Because we are obligated to undertake an independent review of the agency’s action in this appeal, we are not bound by the District Court’s recitation of the facts below.” Olenhouse v. Commodity Credit Corp., 42 F.3d 1560, 1569 n. 16 (10th Cir.1994) (citation omitted).
When reviewing an agency’s interpretation of its own regulations, we defer to the agency’s view, unless it is “unreasonable, plainly erroneous, or inconsistent with the regulation’s plain meaning.” Bar MK Ranches v. Yuetter, 994 F.2d 735, 738 (10th Cir.1993) (citation omitted). In the context of technical or scientific matters particularly, judicial deference to agency decisions is appropriate, Wyoming v. United States, 279 F.3d 1214, 1240 (10th Cir.2002), and choices of suitable scientific methodology are entrusted to the agency, Custer County Action Assn. v. Garvey, 256 F.3d 1024,1036 (10th Cir.2001).
In the instant case, UEC argues that the district court erred when it found that “[f]or each of the five MIS at issue [in this case], the Forest Service has collected appropriate quantitative population data or has a valid reason for not collecting such data.” Utah Environmental Congress v. Bosworth, et al., No. 2:01-CV-00316PGC, slip op. at 18 (D.Utah Mar. 27, 2003) (“Dist.Ct.Order”). On appeal UEC maintains that the Forest Service did not (1) collect appropriate quantitative data and (2) improperly relied on habitat trends rather than actual species population trends. As a result, UEC argues, the Forest Service’s decision to authorize the Monroe Project was arbitrary and capricious.
The Fishlake Management Plan subdivided the Fishlake National Forest into nineteen separate Management Areas, each with its own management practices. The Monroe Project covers six different Management Areas within the Fishlake National Forest; however, ninety-two percent of the Monroe Project area is within Management Area 4B. Area 4B is designated “Habitat for Management Indicator Species,” which indicates that priority should be given to managing the lands in accordance with the habitat needs of MIS. (1 AR Doc. 460 at S-9.) To this end, the Forest Service is required to “estimate the effects of each [management] [1224]*1224alternative on fish and wildlife” species that are selected “because their population changes are believed to indicate the effects of management activities.” 36 C.F.R. § 219.19(a)(1). The species thus selected as MIS should represent endangered or threatened plants and animals; species with special habitat requirements which may be affected significantly by planned activities; commonly hunted, trapped, or fished species; special interest species; and plant or animal species selected because their population changes are believed to indicate the effects of management activities on other species. Id. Because of their representative capacity, MIS have been variously described as “a bellwether-a class representative,” Inland Empire Pub. Lands Council v. United States Forest Serv., 88 F.3d 754, 762 n. 11 (9th Cir.1996), “a management shortcut,” Forest Guardians v. United States Forest Serv., 180 F.Supp.2d 1273, 1281 (D.N.M.2001), and “a proxy for determining the effects of management activities on other species,” Dombeck, 185 F.3d at 1170. As the Ninth Circuit has observed, “[t]he use of management indicator species is intended to allow the [Forest] Service to thoroughly evaluate the effects of the [forest management] alternatives on fish and wildlife populations by using a ‘class representative,’ without having to evaluate each species individually.” Inland Empire, 88 F.3d at 762.
Two groups of indicator species are utilized in the Fishlake Management Plan; one group serves as “ecological indicators,” (2 AR Doc. 544 at 11-27), and the other “represent[s] species of high interest.” (Id.) Goshawk, cavity nesters, riparian dependent guild, sage nesters, macroinver-tebrates, and resident trout, each selected for their sensitivity to habitat change and representative characteristics, were chosen as ecological indicator MIS in the Fishlake Forest Management Plan. They were specifically selected to ensure that species with special habitat needs are adequately considered in forest management activities. Elk, mule deer, and Bonneville cutthroat trout were selected as high interest species MIS. These high interest species were selected because of their threatened, endangered, or sensitive status, or because of high public interest in the species.
UEC argues that the Forest Service authorized the Monroe Project without the relevant MIS population data — here, the effect of management activities on several MIS — in violation of the APA, the NFMA, the NFMA’s implementing regulations, and the Fishlake Forest Plan. In support of this argument, UEC claims that the MIS population data was deficient because the Forest Service used habitat trend modeling as a proxy for actual population surveys, rather than actual quantitative data, which they argue is required by 36 C.F.R. § 219.19. While conceding that § 219.19 applies to the development of Forest Management Plans, as opposed to specific project level actions, UEC insists that because Forest Service obligations under the Fishlake Forest Plan continue as long as the Plan is in existence, the Forest Service is required to evaluate planning alternatives under § 219.19 prior to authorizing specific projects, like the Monroe Project.
Accordingly, as a preliminary matter, we must determine whether § 219.19 applies to the authorization of the Monroe Project. Consistent with UEC’s argument, we have previously applied the § 219.19 requirements to project level actions. For example, in Dombeck, after noting that the Forest Service implements the Forest Plan through individual projects and that these projects must be con[1225]*1225sistent with the Forest Plan, we applied § 219.19 to a project level action. 185 F.3d at 1168. This application is in accord with the analysis of other circuits. See, e.g., Sierra Club v. Martin, 168 F.3d 1, 6 (11th Cir.1999) (recognizing that although § 219.19 applies to the “formulation of Forest Plans rather than to specific projects proposed,” the duties of the Forest Service “continue throughout the Plan’s existence”); Inland Empire, 88 F.3d at 760 n. 6 (9th Cir.1996) (reasoning that “[bjecause any [area] 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). Several district court opinions in this circuit have also applied § 219.19 to project-level decisions, see Utah Environmental Congress v. Zieroth, 190 F.Supp.2d 1265, 1270 n. 1 (D.Utah 2002); Forest Guardians, 180 F.Supp.2d at 1279-1280. Moreover, our analysis comports with the NFMA, which requires that resource management activities be consistent with the Forest Plan. 16 U.S.C. § 1604®. Similarly, Forest Service regulations direct that activities “affecting [forest] lands ... shall be based on the plan,” 36 C.F.R. § 219.10, and at a minimum, all management practices should “[p]rovide that habitat for species chosen under § 219.19 is maintained and improved to the degree consistent with multiple-use objectives established in the plan,” 36 C.F.R. § 219.27. Because § 219.27 contemplates the application of § 219.19 to management activities, and as the Forest Service implements Forest Management Plan activities by approving specific projects, we are convinced that read as a whole, the regulations anticipate application of § 219.19 to project level as well as plan level management actions. Accordingly, we apply § 219.19 to the authorization of the Monroe Project.
Having so concluded, we consider UEC’s argument that 36 C.F.R. § 219.19 requires the Forest Service to gather quantitative population data for MIS in its management planning activities. The Forest Service counters that the regulation does not require the use of quantitative population data in the monitoring of MIS. Relying on the plain language of the regulation, the Forest Service claims that § 219.19 requires only that it monitor population trends of MIS, leaving the method of monitoring — whether habitat information, actual population data, or some combination of the two — to the discretion of the Forest Service. Although it recognizes that our recent decision in Dombeck, 185 F.3d 1162, did not address MIS specifically, the Forest Service nevertheless argues that Dom-beck supports using habitat information to satisfy § 219 requirements in general. We decline to read Dombeck so broadly.
Dombeck addressed a challenge to the Forest Service’s decision to allow the Vail ski resort to expand. 185 F.3d at 1165-1167. Colorado Environmental Coalition, among others, appealed, arguing that the Forest Service failed to properly analyze the effects of the proposed expansion on the Canada lynx populations in the area. Id. at 1165. The Coalition argued that the Forest Service was required to collect actual population data on the lynx before moving forward with the expansion; the Forest Service maintained that habitat data could be substituted in cases where actual population data was unavailable and where actual data would not enhance the Forest Service’s analysis because the proposed action would not result in species loss. Id. at 1168. We were careful to note in Dombeck that “the population inventory requirements of § 219.19 that apply to Management Indicator Species are irrele[1226]*1226vant to the issue before us” because the Canada lynx was not an MIS. Id. at 1170. Therefore, our narrow holding was that “the regulatory language does not require the Forest Service, under the circumstances of this case, to collect actual lynx population data.” Id.
Moreover, our reasoning in Dombeck rested on the observation that the “regulatory language clearly presupposes the ascertainable presence of a species’ population within a given planning area.” Id. at 1169. In the ease of the Vail expansion project, “despite good faith efforts to confirm the presence of lynx, no one has seen an actual lynx in the project area in over twenty-five years.... ” Id. Because we concluded that Congress could not have intended to require the Forest Service to collect quantitative population data “where no population exists,” id. at 1170, the Forest Service’s use of habitat analysis information to examine the effects of the Vail expansion with an eye toward preserving the potential for a viable lynx population in the area in the future was all that was required under the regulations. Id. Additionally, we noted our agreement with the Ninth Circuit that the Forest Service should use actual population data when such data is available. Id. at 1170 n. 9 (citing Inland Empire, 88 F.3d at 761 n. 8). We are therefore unconvinced that Dombeck supports the Forest Service’s interpretation that habitat information alone is sufficient to fulfill its obligations under § 219.
In keeping with the reasoning of the Eleventh Circuit and the district courts of this circuit, we conclude that § 219.19 requires the Forest Service to use actual, quantitative population data to effectuate its MIS monitoring obligations. Section 219.19 mandates that as part of forest planning, “[f]ish and wildlife habitat shall be managed to maintain viable populations of existing native and desired nonnative vertebrate species.” Further, forest management “[planning alternatives shall be stated and evaluated in terms of both amount and quality of habitat and of animal population trends of the management indicator species,” § 219.19(a)(2); similarly, “[pjopulation trends of the management indicator species will be monitored and relationships to habitat changes determined,” § 219.19(a)(6). Plainly the regulations require that the Forest Service monitor population trends of the MIS in order to evaluate the effects of forest management activities on the MIS and the viability of desired fish and wildlife populations in the forest more generally.
Our reading of the requirements of § 219.19 is strengthened by § 219.26, which provides that to ensure diversity of plant and animals in forest planning, inventories which “include quantitative data making possible the evaluation of diversity in terms of its prior and present condition” shall be taken. We agree with the Eleventh Circuit in Martin that these two sections of regulation 219 “are harmonious when read together.” Martin, 168 F.3d at 7. Because
MIS are proxies used to measure the effects of management strategies on Forest diversity ... [and because § ] 219.26 requires the Forest Service to use quantitative inventory data to assess the Forest Plan’s effects on diversity .... then, taken together, the two regulations require the Forest Service to gather quantitative data on MIS and use it to measure the impact of habitat changes on the Forest’s diversity.
Id. Similarly, the court in Forest Guardians reasoned that the language of § 219.19 required the Forest Service “to acquire and analyze hard population data of its selected management indicator species” before approving a timber sale, be[1227]*1227cause these regulations clearly preclude reliance “solely on habitat trend data as a proxy for population data or to extrapolate population trends.” 180 F.Supp.2d at 1281. Likewise, we agree that a reading of § 219.19 as requiring only habitat analysis is “inconsistent with the regulation’s plain meaning,” Yuetter, 994 F.2d at 738. Accordingly, we conclude that in order to effectuate its MIS monitoring duties under the language of its regulations, the Forest Service must gather quantitative data on actual MIS populations4 that allows it to estimate the effects of any forest management activities on the animal population trends,5 and determine the relationship between management activities and population trend changes.
B
Having determined that § 219.19 applies to project level activities and requires that the Forest Service gather quantitative data on specific MIS populations, we turn to UEC’s specific objections to the Forest Service’s approval of the Monroe Project. In particular, UEC challenges the Forest Service’s monitoring of the goshawk, Bonneville cutthroat trout, southwestern willow flycatcher, sage grouse and sage nesters, and cavity nesters. Although we are not bound by the district court’s findings of fact regarding the Forest Service’s monitoring, see Olenhouse, 42 F.3d at 1569 n. 16, we will set aside the Forest Service’s approval of the Monroe Project only if based on the administrative record, the decision is arbitrary, capricious, not in accordance with the law, or not supported by substantial evidence, Hoyl, 129 F.3d at 1382 (citations omitted). “Substantial evidence” has been variously characterized as evidence sufficient to justify a court’s refusal to direct a verdict in a jury trial, or “such relevant evidence a reasonable person would deem adequate to support the ultimate conclusion.” Id. at 1383 (quotations omitted).
With respect to the northern goshawk, UEC maintains that although the Forest Service has gathered population data in the Monroe Project area, the Forest Service has not gathered forest-wide data. Because the Forest Service lacks a forest-wide context in which to place its project population data, UEC argues, it cannot measure the impact of any change in northern goshawk habitat in the project area on the diversity of the entire Fishlake Forest. See Martin, 168 F.3d at 7 (holding that NFMA regulations “require the Forest Service to gather quantitative data on MIS and use it to measure the impact of habitat changes on the Forest’s diversity”). Conversely, the agency maintains that it documented eleven goshawk territories in the Fishlake Forest, and therefore that completion of a forest-wide survey is [1228]*1228unnecessary and not required by the regulations.
The Forest Service conducted the following surveys: (1) helicopter surveys of the nesting and foraging habitat for the goshawk on Monroe Mountain in 1999 and 2000; (2) goshawk surveys, relating to active nests and territories, two years prior to any proposed timber sale in the Fish-lake Forest; and (3) a goshawk survey in the 1994 Burnt Flat Analysis. In addition, the Forest Service mapped documented goshawk territories on Monroe Mountain and compiled goshawk flight data as well as numerous completed goshawk monitoring forms dating from 1999 to 2000. Finally, the Forest Service relied upon information and data received from an inter-agency Goshawk Conservation Strategy group which concluded that “the current goshawk population is viable in the State of Utah.” (5 AR at 2301.)
In finding that the NFMA does not necessarily require context data, the district court reasoned that because the NFMA requires that the Forest Service maintain viable populations of MIS, if the Forest Service is able to determine viability on the basis of the data gathered in the Monroe Project area, then it is not required to gather forest-wide data. We conclude that forest-wide data is not required if the Forest Service can determine the viability of the MIS at issue without a forest-wide survey. We also conclude, based on the data in the administrative record, that the Forest Service’s monitoring of the goshawk fulfilled its obligations under § 219.19.
As to the Bonneville cutthroat trout, the Forest Service relied on a Biological Opinion which indicates that the trout were “cursor[ily] check[ed]” (5 AR at 2221) in 1999 after being reintroduced to Manning Creek in 1996. UEC argues that this Biological Opinion data is insufficient for two reasons: first, because the check was cursory and second, because § 219.19(b)(6) requires “population trends” of the MIS be monitored and trends cannot be determined based upon an isolated data point. The district court rejected these arguments, reasoning that § 219.19’s monitoring requirements cannot be read to require the Forest Service to sample the Manning Creek Bonneville cutthroat trout population because the sampling may harm the reintroduction of the trout in that stream.
The Forest Service also contends that quantitative data on the Bonneville cutthroat trout had been collected annually at Manning Meadow Reservoir (also in the Monroe Project area) because the Reservoir operates as Bonneville cutthroat trout brood stock habitat for Utah’s sportfishing stock program. Six years following the reintroduction of 714 Bonneville cutthroat trout to the Reservoir, the sportfishing stock program collected over 100,000 eggs from the brood stock. Manning Creek is stocked from eggs taken from Manning Reservoir. Even though this data is somewhat limited, on balance, we agree with the district court that the Forest Service has collected adequate, appropriate quantitative population data in regards to the Bonneville cutthroat trout in support of its decision to approve the Monroe Project.
Turning to the southwest or southwestern willow flycatcher, a small bird on the Endangered Species list located within the proposed project area, the Monroe Project Biological Assessment concludes that the project would create “no direct adverse effects on southwest willow flycatcher or their habitat.” (5 AR at 2288.) However, no quantitative data on the southwestern willow flycatcher, either within the Project area, or within the Fishlake National For[1229]*1229est, appears in the Biological Assessment. Although the Forest Service states, in the FEIS, that “[ajnnual monitoring for presence, abundance, and nesting [of the Southwestern willow flycatcher] are [sic] conducted by the Forest Service and the Utah Division of Wildlife Resources,” (2 AR Doc. 485 at III — 64), we were unable to locate quantitative data from either the Forest Service or the Utah Division of Wildlife Resources in the record.
Despite the absence of actual data in the record, the effects analysis in the FEIS concludes that the current management of the Monroe Project area could harm the southwestern willow flycatcher’s habitat. While annual monitoring of the southwestern willow flycatcher’s presence, abundance, and nesting is sufficient to meet the Forest Service’s obligations under the Fishlake Forest Plan and § 219.19, because we find no such data in the record, we are unable to conclude that the Forest Service fulfilled its obligations under § 219.19 in monitoring this MIS.
Our review of the record regarding sage grouse and sage nesters leads to a similar conclusion. In its brief, the Forest Service acknowledges that “[s]age grouse are potentially present,” “unconfirmed sightings” of sage grouse have been made, and habitat suitable for the sage grouse is present in the Monroe Project area. (Appellees’ Br. at 23.) Because, however, the Monroe Project is anticipated to “alter only two or three percent of the sagebrush habitat,” the Forest Service concluded that the Monroe Project “will not threaten the viability of the sage nesters.” (Appellees’ Br. at 48-49.)
As to its monitoring obligations, the Forest Service directs our attention to the Burnt Flat Analysis, which included an analysis of a survey of the sage nesters (the guild to which the sage grouse belong). The Burnt Flat Analysis, conducted in 1994, surveyed thirty-seven different species of birds, yet reviewing the Burnt Flat Analysis included in the record, we discovered no quantitative data regarding either the sage grouse, or the sage nester guild specifically.
As a final matter, notably, the district court determined that because there “have been no known sightings of ... Sage Nes-ters on Monroe Mountain,” the Forest Service could not collect hard population data, and relying in part on Dombeck, the court refused to interpret the language of the NFMA to require the impossible. (Dist. Ct. Order at 19.) Although we agree with the district court that where impossible, the Forest Service is not required by the applicable statutes and regulations to collect population data, we part with the district court’s further reasoning that when faced with “difficult-to-count” species, the Forest Service may reasonably disregard its obligation to obtain quantitative data and choose to allocate its resources elsewhere. (Id.) This extension of the reasoning of Dombeck is troublesome in this context, where the Fishlake Forest Plan specifically states that the MIS selected for the plan were chosen, in part, due to “[r]elative ease of monitoring.” (2 AR Doc. 544a at 11-27.) Moreover, because MIS, if they are to serve their purpose as a class representative in the Fishlake Forest Plan, should generally not be difficult-to-count species, and because the Forest Service did not argue to us that the sage grouse and sage nesters are reclusive and hard to track, we decline to follow the reasoning in Inland Empire, 88 F.3d 754, 763 n. 12 (finding that the MIS at issue was a “reclusive species” and there was no “technically reliable and cost-effective method of counting individual members of the species”).
Of course, not all MIS designated in the Fishlake Forest Plan are necessarily [1230]*1230present within the Monroe Project area, and we do not require the Forest Service to attempt to track species “where no population exists,” Dombeck, 185 F.3d at 1170. We do, however, conclude that “good faith efforts to confirm,” Dombeck, 185 F.3d at 1169, the absence or presence of an MIS species by the Forest Service is necessary in order to fulfill its MIS monitoring obligations under § 219.19. Here, despite unconfirmed sightings of the sage grouse within the Monroe Mountain project area, the record reflects no attempt by the Forest Service to confirm the presence of this MIS. Accordingly, we find that the Forest Service did not comply with § 219.19 with respect to the sage grouse and sage nester guild MIS within the Monroe Project area.
Finally, we address the Forest Service’s monitoring of the cavity nesters. The district court stated that there “have been no known sightings of Cavity Nesters ... on Monroe Mountain.” (Dist. Ct. Op. at 19.) The Forest Service, however, does not assert on appeal that there have been no sightings of cavity nesters, nor does it direct this court to record reference where we might locate evidence to support the district court’s conclusion. Rather, to support its monitoring of the cavity nesters, the Forest Service relies upon the aforementioned Burnt Flat Analysis from 1994 and additional studies of the three-toed woodpecker, a member of the Cavity Nes-ter guild, as evidence of its monitoring of this MIS guild. Reviewing the three-toed woodpecker information in the record, we arrive at a brief discussion in the Biological Evaluation on Sensitive Species of the effects of the Monroe Project on the woodpecker that observes prey abundance, soil, water, and air quality changes resulting from the project could indirectly effect the three-toed woodpecker. For example, the Biological Evaluation speculates that tree harvesting could result in woodpecker relocation, possibly causing a compression of the woodpeckers in one area. Also included is a map indicating the “Cavity, Feed & Nest Tree Locations” of the three-toed woodpecker in the Fishlake National Forest as of the summer of 2000. (5 AR at 2234.) The Biological Evaluation concludes that individual woodpeckers and woodpecker territory may be affected by the Monroe Project, however those effects will not lead to a loss of population viability for the three-toed woodpecker, and that the Monroe Project should result in a increased availability of snag trees, which is cavity nester habitat.
Our review of the record reveals no quantitative data for either the cavity nester guild or the three-toed woodpecker species; the woodpecker map is apparently based on actual data collected, and the FEIS observes that “nest building activity” for the three-toed woodpecker in the Monroe Project area has been observed. (2 AR Doc. 485 at III-68.) While the Forest Service contends that “woodpecker survey work ... [and] a graduate study on three-toed woodpecker distribution” have been completed, the record citation is to a letter asserting this information has been collected, not to actual data. (5 AR at 2205.) It appears that hard quantitative data — at least sufficient to create the woodpecker map — was collected on the three-toed woodpecker; nevertheless, quantitative data sufficient to determine population trends of the three-toed woodpecker was absent from the record before us. Accordingly, we determine that the Forest Service’s monitoring of the cavity nester guild, through its monitoring and evaluation of the three-toed woodpecker, was insufficient to comply with the requirements of § 219.19.
Ill
We now turn to UEC’s claim that the Forest Service illegally “revalidated” road-[1231]*1231less areas during the development of the Monroe Project. In 1979, the Roadless Area Review and Evaluation II (RARE II) resulted in a nationwide inventory of road-less areas potentially suitable for wilderness designation. See Block, 690 F.2d at 758. The Forest Service concedes that four areas of the Fishlake National Forest were designated IRAs in the RARE II inventory. Forest Service regulation § 219.17(a) provides that, “[ujnless otherwise provided by law, roadless areas within the National Forest System shall be evaluated and considered for recommendation as potential wilderness areas during the forest planning process....” 36 C.F.R. § 219.17(a).
In conjunction with the Monroe Project, the Forest Service changed the boundaries of certain IRAs within the project area. UEC argues that the Forest Service’s approval of the Monroe Project was unlawful because it depended upon an “illegal and improper ‘revalidation’ of the boundaries of the roadless areas in the Monroe Mountain project area.” (Appellant’s Br. at 16.) According to UEC, the roadless area inventory may be modified only at the forest planning level, and thus the Forest Service’s approval of the Monroe Project, which UEC argues was based on the revalidation of the IRA boundaries, was arbitrary and capricious. Assuming that UEC is correct that the Forest Service may only redraw IRA boundaries at a forest plan level, nonetheless, UEC has failed to identify any applicable limitation on the Forest Service’s discretion to undertake the proposed Monroe Project management activities within the IRAs regardless of the Forest Service’s revalidation. See infra Part I (outlining the proposed management activities).
Nothing in § 219.17 restricts the Forest Service’s discretion in managing roadless areas. Under § 219.17, moreover, areas which have already “been designated as wilderness or for nonwilderness uses by law” need not be evaluated as potential wilderness areas. 36 C.F.R. § 219.17(a)(l)(i). In this case, the subject roadless areas were designated for multiple use by Congress in the Utah Wilderness Act of 1984, Pub.L. No. 98-428, § 201(a)(2), 98 Stat. 1657 (1984), which was passed after completion of RARE II. The stated purposes of the Act were to “designate certain national forest system lands in the State of Utah for inclusion in the National Wilderness Preservation System” and “to release other forest lands for multiple use management.” Utah Wilderness Act of 1984. Accordingly, the Act designated twelve specific IRAs in Utah as components of the National Wilderness Preservation System. Id. at § 102. None of the Fishlake National Forest’s IRAs were so designated. By default, the Act provides that all areas covered by RARE II not designated as wilderness areas by the Act “shall be managed for multiple use in accordance with land management plans.Id. § 201(b)(3).
Although UEC directs our attention to the Draft Inter-Mountain Region Planning Desk Guide, the Guide merely contains guidelines for implementing § 219.17 and the Forest Service Handbook. There is no indication in the Desk Guide itself or elsewhere that the Desk Guide restricts the Forest Service’s discretion in managing IRAs. In addition, UEC relies on the Forest Service Handbook, § 1909.12, Chapter 7 to support its position. However, that chapter merely “describes the process for identifying and evaluating potential wilderness in the National Forest System,” id. at chap. 7, and sets forth criteria for determining whether “[r]oadless areas qualify for placement on the inventory of potential wilderness,” id. at 7.11. Nothing in § 1909.12 restricts the Forest Service’s discretion in managing IRAs.
[1232]*1232Finally, in its reply brief, UEC directs our attention to the so-called “Roadless Rule,” 36 C.F.R. §§ 294.10-14 (2003). Section 294.12 prohibits, with certain exceptions, road construction in certain inventoried roadless areas. Section 294.13 prohibits, with certain exceptions, timber cutting or removal in inventoried roadless areas. We need not decide whether the Monroe Project runs afoul of these new regulations, or if they are even valid, because they became effective in January 2001 and do not apply retroactively. See 36 C.F.R. § 294.14(c) (stating new rules do “not revoke, suspend, or modify any project or activity decision made prior to January 12, 2001”). If anything, the fact that these rules were propounded suggests that the Forest Service previously was not restricted in its management of roadless areas. See 36 C.F.R. § 294.10 (stating purpose of rules “is to provide, within the context of multiple-use management, lasting protection for inventoried roadless areas within the National Forest System”). In sum, UEC has failed to identify any applicable limitation on the Forest Service’s discretion to implement the proposed Monroe Project within the pre-revalidation IRAs, and thus we cannot conclude that the Forest Service’s authorization of the Monroe Project was dependent upon the revalidation. Accordingly, we reject this claim of error.
IV
Because we conclude that the Forest Service has not complied with its duties under Forest Service regulations to monitor several of the relevant management indicator species, we conclude that the Forest Service’s authorization of the Monroe Project was arbitrary and capricious. We REVERSE the district court’s order affirming the Forest Service Record of Decision authorizing the Monroe Project and REMAND for further proceedings in compliance with this ruling.