Plaintiff Utah Environmental Congress (UEC) brought this action alleging that defendants, representatives of the United States Forest Service (Forest Service), violated federal law by authorizing six separate projects in four national forests in the State of Utah. The district court entered judgment in favor of defendants. UEC now appeals. We exercise jurisdiction pursuant to 28 U.S.C. § 1291, affirm the district court's decision affirming authorization of three of those projects, reverse the district court's decision affirming authorization of the remaining three projects, and remand to the district court with directions to vacate the Forest Service's approval of those latter three projects.

The Forest Service, an agency within the United States Department of Agriculture, manages the National Forest System under several federal statutes and accompanying regulations. We begin our review of this appeal, as we recently did in a similar case filed by plaintiff UEC, by outlining the specific statutory and regulatory framework applicable to the issues raised by UEC. Utah Envt'l Cong. v. Bosworth, 443 F.3d 732, 735-36 (10th Cir.2006) (UEC III).

The National Forest Management Act of 1976 (NFMA) requires the Forest Service to "develop a land and resource management plan, commonly known as a forest plan, for each unit of the National Forest System." UEC III, 443 F.3d at 736 (citing 16 U.S.C. § 1604(a), (e), (g)(3)(B)). Each forest plan "envisions the forest will be used for multiple purposes, including `outdoor recreation, range, timber, watershed, wildlife and fish, and wilderness.'" Id. at 737 (quoting 16 U.S.C. § 1604(e)(1)). "At the same time, the forest plan provides for `diversity of plant and animal communities based on the suitability and capability of the specific land area.'" Id. (quoting 16 U.S.C. § 1604(g)(3)(B)). In turn, the Forest Service implements each forest plan "by approving or disapproving particular projects," each of which "must comply with the applicable forest plan." Id.

"Before a forest plan may be created, NFMA `explicitly requires the Secretary of Agriculture to issue regulations that set out the process for the development and revision of land management plans for units of the National Forest System, and regulations that establish management planning standards and guidelines....'" Id. (quoting 47 Fed.Reg. 43,026, 43,037 (Sept. 30, 1982)). "Of particular concern in this case are planning rules the Forest Service adopted in 1979 and revised in 1982, codified at 36 C.F.R. § 219 (1982), which govern Forest Service management at both the programmatic and project levels." Id. "In November 2000, the Forest Service substantially amended these regulations, known as the 1982 planning rules, replacing them with the 2000 planning rules, codified at 36 C.F.R. § 219 (2001)." Id.

As part of the 2000 planning rules, the Forest Service added a new regulation, 36 C.F.R. § 219.35, entitled "Transition," that attempted to outline how responsible officials within the Forest Service were to "transition" from the 1982 rules to the 2000 rules. Section 219.35 indicated that the forest plan for each unit of the National Forest System would be revised pursuant to a schedule established by the Chief of the Forest Service. 36 C.F.R. § 219.35(g) (2001). Consistent with that goal, § 219.35(a) announced an indefinite "transition period," "begin[ning] on November 9, 2000 and end[ing] upon the completion of the revision process .. . for each unit of the National Forest system." Importantly, for purposes of this case, § 219.35(b) created a grace period during which responsible officials within the Forest Service could, at their discretion, revise forest plans utilizing either the 1982 planning rules or the 2000 planning rules.

With respect to the implementation of forest plans, § 219.35(a) provided that, "[d]uring the transition period," responsible officials would be required to "consider the best available science in implementing and, if appropriate, amending" the then-existing forest plans (i.e., forest plans that had yet to be revised). At the same time, however, § 219.35(d) also stated: "Site-specific decisions made by the responsible official 3 years from November 9, 2000 and afterward must be in conformance with the provisions of this subpart." Thus, responsible officials were left to resolve the tension between § 219.35(a)'s mandate that they "consider the best available science in implementing" existing forest plans during the transition period, and § 219.35(d)'s provision that site-specific decisions did not have to conform to "this subpart," including, presumably, § 219.35(a), until November 9, 2003 and thereafter.

In May 2002, the Forest Service publicly noted that "field personnel" had raised concerns regarding their ability to comply with the 2000 rules in terms of making site-specific decisions. 68 FR 53295. At that time, the Forest Service indicated "that it expected to address these concerns by removing the requirement or extending the original transition date for site-specific projects." Id. After publishing a proposed revision on December 6, 2002, id., the Forest Service issued an interim final rule on September 10, 2003, amending § 219.35(d) to read: "The date by which site-specific decisions made by the responsible official must be in conformance with the provisions of this subpart is extended from November 9, 2003, until the Department promulgates [its] final planning regulations...." In doing so, the Forest Service acknowledged that "[t]here [wa]s a lack of clarity about how projects [we]re to be compliant with the rule," and that "[t]his uncertainty and lack of clarity [could have] pose[d] an unreasonable analysis burden on field units when planning for site-specific project decisions." 68 FR 53295. Unfortunately, the Forest Service's extension of subsection (d)'s grace period did nothing to eliminate the tension between subsections (a) and (d). Indeed, as exemplified by some of the projects at issue in this appeal, the September 10, 2003 interim final rule could have been interpreted by responsible officials as an indication that they were not obligated to apply the best available science standard to site-specific decisions, including project approvals, until the final regulations were promulgated.

On September 29, 2004, the Forest Service issued a final rule "to clarify the intent of the transition section of the planning regulations," i.e., § 219.35, "regarding the consideration and use of the best available science to inform project decision making that implement[ed] a [pre-2000] land management plan . . . ." 69 FR 58055. In doing so, the Forest Service acknowledged that "[c]onsiderable uncertainty ha[d] arisen regarding ... the transition provisions," id., and that "[t]his uncertainty ha[d] affected the ability of the Forest Service to utilize fully the provisions of § 219.35 paragraph (a) to consider the best science available in ... project decision making." Id. at 58056. The final rule ultimately added an appendix (Appendix B) to the end of § 219.35 adding clarifying language. Id. at 58057.

"These transition provisions remained on the books until January 2005 when the new rules were finally implemented." UEC III, 443 F.3d at 737 (citing 36 C.F.R. §§ 219.1 to .16 (2005), and 70 Fed.Reg. 1,023 (Jan. 5, 2005)). "The 2005 rules retained the best available science standard, requiring the `Responsible Official [to] take into account the best available science' by `document[ing] how the best available science was taken into account in the planning process,' and evaluating and disclosing substantial uncertainties and risks in that science." Id. (quoting 36 C.F.R. § 219.11(a) (2005) (emphasis added)).

3) National Environmental Policy Act (NEPA)

"The National Environmental Policy Act (`NEPA') requires federal agencies such as the Forest Service to analyze environmental consequences before initiating actions that potentially affect the environment." Id. at 735-36. "In conducting this analysis, the Forest Service must prepare one of the following: (1) an environmental impact statement [EIS], (2) an environmental assessment [EA], or (3) a categorical exclusion." Id. at 736.

Of these three, "[a]n environmental impact statement involves the most rigorous analysis, and is required if a proposed action will `significantly affect[ ] the quality of the human environment.'" Id. (quoting 42 U.S.C. § 4332(2)(C) and 40 C.F.R. § 1502.4). An environmental assessment is "considerably less detailed," and may be prepared "[i]f an agency is uncertain whether the proposed action will significantly affect the environment...." Id. at 736 (citing 40 C.F.R. § 1508.9). "An environmental assessment provides `sufficient evidence and analysis' to determine whether a proposed project will create a significant effect on the environment." Id. (quoting 40 C.F.R. § 1508.9). "If so, the agency must then develop an environmental impact statement; if not, the environmental assessment results in a `Finding of No Significant Impact,' and no further agency action is required." Id. (quoting 40 C.F.R. § 1508.9).

Occasionally, a proposed action will "fall[ ] within a categorical exclusion, i.e., those actions predetermined not to `individually or cumulatively have a significant effect on the human environment.'" Id. (quoting 40 C.F.R. § 1508.4). In such circumstances, the agency "is not required to prepare either an environmental assessment or an environmental impact statement." Id.

At issue in this case are six separate projects encompassed within four different National Forests in Utah. The projects, categorized by the National Forest of which they are a part, are discussed in detail below.

The Wasatch-Cache National Forest (WCNF) encompasses two of the six projects at issue: the Bear Hodges II Timber Sale and the East Fork Fire Salvage projects. The Forest Plan for the WCNF, which was revised in 2003, states, in pertinent part, that management indicator species (MIS) "are required under 36 CFR 219.19," App. at 759, and "are used to assess the effects of a management activity on wildlife." Id. at 760. For purposes of managing the WCNF, the Forest Plan designates beaver, snowshoe hare, and cutthroat trout as MIS.1 Id. at 756-57, 760. The Forest Plan indicates that these MIS are to be monitored in the following manner:

• Beaver—"Beaver populations across the Forest" are to be monitored. Id. at 756. These populations are to be measured by the "[n]umber of ... active dams." Id. The measurement and reporting frequency is: "1-2 4th order HUC's [hydrologic units] per year after baseline establishment." Id.

• Snowshoe hare—The Forest Plan requires "[s]nowshoe hare presence and population index" to be determined. Id. This is to be performed by way of "[p]ellet counts along transects." Id. The measurement and reporting frequency is: "First 2 yrs [e]stablish baseline. Annual update. Evaluated at 5-year intervals." Id.

• Cutthroat trout—The Forest Plan requires the Forest Service to obtain "[c]utthroat trout population estimates." Id. at 757. These estimates are to be made by way of the "Fish Condition index." Id. The measurement and reporting frequency is: "One 4th order HUC per year." Id.

a) The Bear Hodges II Timber Sale. This project involves a harvest of approximately 3.1 million board feet of timber from twelve separate units comprising approximately 701 acres in and adjacent to the T.W. Daniels Experimental Forest, located in the Bear Hodges area of the Logan Ranger District, all of which are part of the WCNF. Id. at 793-94, 807. The project, which involves both permanent and temporary road construction in the project area, is intended to: "1) provide research and demonstration opportunities on the Utah State University T.W. Daniel Experimental Forest through testing and demonstrating different silvicultural methods, 2) maintain the mature character of the spruce-fir forest in the near future while beginning to develop more age class diversity, and 3) provide a merchantable product for the timber industry ...." Id. at 793.

The Forest Service first prepared an EA for this project, followed by an EIS. Id. at 793-94, 815. The project's Record of Decision (ROD) issued on May 14, 2004, and concluded, in pertinent part, that the project was consistent with the Forest Plan. Id. at 806, 816. The EIS concluded that, with respect to the three relevant MIS: (1) there are no wetland or riparian areas, and thus no beaver habitat, within the project area, id. at 800, 802, 1128; (2) snowshoe hare habitat exists within the project area, but the project would not significantly impact snowshoe hare populations or their habitat because project treatments were designed to maintain spruce-fir habit used by the species, id. at 804, 1128; and (3) Bonneville cutthroat trout are not present in the project area, id. at 803, 805.

b) The East Fork Fire Salvage project. The Forest Service planned this project in response to a fire that burned approximately 14,200 acres of the Evanston Ranger District in the WCNF in late June and most of July 2002. Id. at 2034. The project includes the "harvest of fire killed trees," id., in twenty units totaling 781 acres. Id. at 2035. The project also includes the "maintenance and improvement of drainage on about 20 miles of existing roads" in the project area. Id. The purpose of the project "is to provide timber for commercial harvest and to capture economic value of timber killed by the . . . [f]ire consistent with goals for watershed health, sustainable ecosystems, biodiversity and viability and scenic/recreation opportunities." Id.

An EIS was prepared for this project, and the project ROD issued on June 14, 2004. The EIS concluded, with respect to the MIS: (1) beaver activity within the project area was stable or increasing, and the project would have no effect on beaver habitat because no salvage activity is proposed within riparian areas, id. at 2040, 2043; (2) there was no snowshoe hare population within the project area because the 2002 fire that precipitated the project destroyed the hares' habitat, and the salvage sale would not have any adverse effect on the reestablishment of that habitat, id. at 2042-43, 2061; and (3) population trends for Bonneville cutthroat trout within the project area were flat, and the project included mitigation measures designed to reduce the potential for sediment from buffer strips on two salvage units from negatively impacting fish-bearing streams, id. at 2045, 2047.

The Manti-La Sal National Forest (MLNF) covers two projects at issue in this appeal: the SITLA and the South Manti Timber Salvage projects. The Forest Plan for the MLNF was approved on November 5, 1986. Id. at 590, 2085. Although the Forest Plan does not expressly cite to the 1982 planning rules (i.e., 36 C.F.R. § 219.19), it lists various MIS, including macroinvertebrates.2 Id. at 79, 2087. Macroinvertebrates, which are found in the forest in stream fisheries, lakes, reservoirs, tunnels and canals, are "ecological indicator species in aquatic habitats and the ability of that habitat to support fisheries." Id. at 80. They are monitored and reported every five years using Biotic Condition Index (BCI) and the Habitat Condition Index (HCI). Id. at 79, 82, 89, 91, 104. BCI "is a tool for assessing overall aquatic-ecosystem health." Id. at 2069. "The Forest Plan Standard [for BCI] is 75 or greater." Id. at 96, 2069. However, the Forest Plan "provides no direction about where to sample or the desired density of sampling sites across the Forest," and "[t]he majority of stream systems on the Forest have not been sampled or monitored." Id. at 97. And, with respect to those streams that have been sampled, it appears that the resulting data is so sparse that it is (or at least was) "not really possible to calculate a statistically valid trend with three to four data points." Id. at 100.

In an undated Forest Monitoring Report contained in the administrative record, it is noted that freshwater macroinvertebrate communities "go through cyclical highs and lows relating to natural conditions," and, "[d]ue to the natural year to year variability of these" conditions, "several years of data are required to establish adequately the range of variation in community structure and productivity and to distinguish between natural variability and that due possibly to human-caused land uses or activities." AA, Vol. 12 at SM0537a (internal quotation marks omitted). The report further notes that "[t]he Forest Plan provides no direction about where to sample or the desired density of sampling sites across the Forest," and that "[t]he majority of the stream systems on the Forest have not been sampled or monitored." Id. at SM05378c.

a) The SITLA project. The State of Utah School and Institutional Trust Lands Administration (SITLA) owns certain "in-holdings" located on East Mountain in the Manti-La Sal NF, approximately 15 miles northwest of Huntington, Utah.App. at 2072, 2075. SITLA applied to the Forest Service in order to obtain "permanent road access to the[se] in-holdings" in order "to facilitate the harvest of up to five million board feet of timber on two sections of [its] land, and the construction of one exploratory natural gas well on a separate section of [its] land[ ]."3 Id. The project "incidentally would provide access to an additional proposed exploratory natural gas well on National Forest lands by Prima Oil and Gas Company." Id. The project "involves road construction and reconstruction on National Forest lands in and outside the boundaries of the East Mountain Inventoried Roadless Area (IRA) and aggregate pit development within the IRA boundaries." Id.

An EIS was prepared for the project. Id. at 2077-78, 2101. The ROD issued on July 12, 2004, id. at 2091, and concluded, in pertinent part, that the project "would not affect drainages with aquatic macroinvertebrate habitat," and therefore "there would be no affect [sic] to aquatic macroinvertebrates or their habitat." Id. at 2087.

UEC filed an administrative appeal challenging the ROD. The Forest Service denied that administrative appeal on September 29, 2004, AA, Vol. 46 at S003298, concluding in pertinent part that they "ha[d] complied with the monitoring provisions of the NFMA." Id. at S003299.

b) The South Manti Timber Salvage project. The South Manti Timber Salvage project "was initiated in response to epidemic spruce beetle ... activity across the South Manti landscape." App. at 578. "The project area includes approximately 24,597 acres of National Forest lands within the southern portion of the Wasatch Plateau on the Ferron-Price and Sanpete Ranger Districts of the" MLNF. Id. Because "[a]n abundance of dead trees can predispose an area to the occurrence of a large, intense wildland fire," id. at 580, the purpose of the project is to reduce the potential for such wildfires by salvage harvesting approximately 19 to 25 million board feet "of dead and dying timber ...." Id. at 581. The project also includes "reforestation across a total range of 970 to 1,251 acres ...." Id.

An EIS was prepared for the project and completed in May 2000. Id. at 578. It noted, in pertinent part, that macroinvertebrate "[m]onitoring stations [we]re located at the Forest boundary on Ferron, Muddy, and Twelvemile Creeks," all of which were in the project area. AA, Vol. 16 at SM09053. Although the EIS noted that "[w]ater quality in Twelvemile Creek appear[ed] to have improved steadily since ... landslide and flooding events [in] 1983 and 1984," it stated that "[d]ata results for Ferron Creek and Muddy Creek [we]re so variable that there [wa]s no apparent trend." Id. The EIS further noted that the "[p]rimary effects of concern when assessing timber treatment projects are increases in sediment to streams, which . . . degrades habitat diversity of aquatic macro-invertebrates." Id. at SM09109. The EIS concluded that, with respect to three of the four alternatives being considered, "[h]arvest activity in the watersheds would increase sediment yield, and affect aquatic habitat," thereby resulting in "[d]eclines in filter-feeding macroinvertebrate species ...." Id. at SM09111.

The ROD was issued by the Forest Supervisor on June 16, 2004. App. at 577, 593. The ROD notes, in pertinent part, that "[t]emporary increases in sedimentation will be expected from temporary road construction, reconstruction, maintenance, and decommissioning." Id. at 583. "Over the long-term," however, "road reconstruction, maintenance, and decommissioning associated with the" project "will result in reductions in sediment." Id. Further, the ROD notes that "[n]o harvesting or mechanical entry ... will be permitted within 100 feet of each side of perennial streams, seeps, lakes, reservoirs, or wetlands." Id. Although the ROD briefly discusses the effect of the project on some of the MIS listed in the Forest Plan, it does not mention macroinvertebrates. Instead, in a separate section, the ROD simply states that "[t]he diversity index DAT ... will be maintained at or above 11, the standing crop at or above 1 and the biotic condition index at or above 75," and that if these indicators "fall below set levels then evaluation of cause of sediment source would be done and corrective measures taken as soon as possible." Id. at 605.

UEC appealed from the ROD, claiming in part that the Forest Service "failed to meet the mandate of the [NFMA] by failing to adequately maintain monitoring data . . . ." Id. at 612. In responding to this appeal, the Forest Service noted that "[t]he 1982 [NFMA] regulations (36 CFR 219) provide[d] the specific [monitoring] requirements." Id. The Forest Service further noted that the Forest Plan "d[id] not indicate any specific project level monitoring requirements," and that, "[a]s with goals and objectives, no single project can provide monitoring information that meets all of the requirements for the Forest Plan." Id. Thus, the Forest Service emphasized, "[t]he direction to monitor ... applies to the National Forest, not individual projects." Id.

UEC also appealed the ROD claiming that the Forest Service "failed to meet its mandate under NFMA to collect required quantitative population trend data and determine relationships between management activities or habitat changes and population trend changes for ... macroinvertebrates ...." Id. at 613. In response, the Forest Service noted:

The Forest Plan Monitoring Report (1987-1991) states that stream conditions across the Forest generally appear to be improving based on macroinvertebrate sampling data and the Biotic Condition Index....

Macroinvertebrate inventories from 1990-2000 are in the project record and were used to make the determination that "for all streams and reservoir habitats, macroinvertebrate indices would not be expected to fall below Forest Management Plan standards" (Final Environmental Impact Statement (FEIS), p. 4-21; Aquatic Macroinvertebrate Monitoring Report; 5/18/2000).

Id.

c) The Uinta National Forest

The Uinta National Forest (UNF) covers the White River Salvage Sale project. The Forest Plan for the UNF, which was revised in 2003, lists the Colorado River cutthroat trout (CRCT) as one of several MIS. Id. at 622, 728. The Forest Plan indicates that, for the CRCT, population estimates must be conducted annually on at least 33% of the streams in the UNF (but reported only every five years). Id. The Forest Plan further provides that, for the CRCT, habitat conditions must be monitored and reported every five years. Id. As of the time the Forest Plan was amended in 2003, "[s]elected streams ha[d] been inventoried" for CRCT "and/or monitored in conjunction with" the Utah Department of Wildlife Resources, and a "[f]orest-wide monitoring protocol [was being] developed." Id.

The White River Salvage Sale project, located within the Left Fork of the White River drainage in the Spanish Fork Ranger District, "encompass[es] [the sale of] approximately 300 acres [worth] of mixed conifer stands." Id. at 720. The purpose of the project is two-fold: (1) "to reduce ... the amount of tree mortality and deformity caused by the Douglas-fir beetle and dwarf mistletoe," and (2) "to restore a more natural variability of age classes and vegetative composition in the forested stands within [the] treatment area." Id. at 725. "Temporary roads and skid trails would be utilized to access stands, and would be rehabilitated following completion of harvest activities." Id. at 720. In addition, portions of two National Forest System Roads in the project area "will be maintained to bring the[m] closer to Objective Maintenance Levels to address erosion and run-off concerns." Id. at 726.

On December 12, 2003, a report regarding "Cutthroat Trout Population Trends" in the UNF was published. Id. at 653. That report stated that "[p]opulation trend data using indices of overall condition" for CRCT "indicate[d] a slight increase ... in the overall condition of [CRCT] ... during the period between 1975 and 2002," but "no observable change" "during the time period between 1991 and 2002...." Id. This data was subsequently incorporated into a February 6, 2004 report entitled "Cutthroat Trout Population Monitoring— Uinta National Forest FY 2003." Id. at 654. This 2004 report indicated that during fiscal year 2003, UNF staff "conducted population inventory and/or monitoring surveys on ... eight streams containing" CRCT, which "equate[d] to ... 53 percent of the [CRCT] populations on the Forest." Id. at 655.

An EA was prepared for the project and published in March 2004. Id. at 723. It noted, in pertinent part, that "[p]opulation trend data" in the project area "indicate[d] that the overall condition" of CRCT "was lower in 1998 than in 1995," even though forest-wide trend data "show[ed] no observable change in the overall condition" of CRCT populations. Id. at 731. The EA further noted that "[t]he existing high sediment loads, stream channel instability, and impacted riparian areas may currently be restricting the rate of recovery of" CRCT populations in the project area. Id. Although the EA noted that "existing roads [we]re major contributors of sediment" in the project area, it stated "it [wa]s not anticipated that [the] timber harvesting [associated with the project] w[ould] increase risks to the [CRCT] within the [project area] beyond those currently present under existing environmental conditions." Id. at 736. The EA also stated that the project was "not likely to []either result in a trend toward federal listing of the species [ ]or affect their viability." Id.

On March 4, 2004, a Decision Notice and Finding of No Significant Impact (FONSI) was issued by the Forest Service for the project. Id. at 737. The FONSI stated, in pertinent part, that "[a]fter considering the environmental effects described in the EA," the Forest Service "ha[d] determined that [the project] w[ould] not have a significant effect on the quality of the human environment considering the context and intensity of impacts," and that, in turn, an EIS "w[ould] not be prepared" for the project. Id. at 745.

UEC filed an administrative appeal from the Decision Notice, asserting in part that the Forest Service "ha[d] failed to be consistent with the [NFMA] and the 2003 Forest Plan because they ha[d] not gathered quantitative [MIS] trend or determined its relationship to habitat changes." App. at 751. The Forest Service, in response, began by noting that "[s]urveys for [CRCT] occurred in the White River System in 1995, 1998, and 2003," and that "physical habitat data was also collected." Id. The Forest Service further noted that "[t]he monitoring direction in the [Forest Plan] for aquatic species [wa]s not project level specific, but [wa]s used in an adaptive management approach to validate assumptions made during the Forest planning process and as an information source for adaptive management." Id. The Forest Service also noted that the revised Forest Plan "was signed April 2003 and monitoring to satisfy plan requirements [wa]s expected to begin in 2004," and thus "[i]t [wa]s premature to determine consistency with Forest Plan monitoring requirements since they [we]re in their first year and ha[d] not yet been initiated." Id. Thus, the Forest Service concluded, "[t]he District Ranger [who made the ROD] had the benefit of existing data from surveys in 1995, 1998, and 2003 to make a reasonable analysis and decision." Id.

d) The Dixie National Forest

The Dixie National Forest (DNF) covers the Dark Valley Vegetation Management project. The Forest Plan for the DNF lists as MIS Bonneville cutthroat trout (BCT) and other types of trout. Id. at 61. Monitoring of these MIS is to be conducted and reported annually. The Forest Plan also provides, however, that "[i]f fish population data is not available for a particular water body, the macroinvertebrate biotic condition index (BCI) will be used to assess fish habitat capability." Id.

The Dark Valley Vegetation Management project encompasses the commercial harvest of Douglas fir, spruce and ponderosa pine trees on approximately 237 acres located in the Loa/Teasdale Ranger District of the DNF. Id. at 70. The purpose of the project "is to reduce vegetation densities in order to protect the area from insect infestations and potential stand replacing wildfires." Id. The project area "has a build-up of fuels and some insect and disease outbreaks that include mistletoe, spruce beetle, Douglas-fir beetle, and root rot." Id. "The proposed treatments are needed to reduce stand densities, reduce fuel loads, and protect forest health." Id. "To access areas needing vegetation management, about½ mile or less of temporary road construction will be required." Id. at 74.

A Decision Memo approving the project issued on October 7, 2004. Id. at 73. In analyzing the project, the Forest Service concluded that potential habitat and existing populations of BCT or other trout varieties did not exist in the project area. Id. at 71-72. According to the parties, the project was categorically excluded under NEPA. Aplt. Br. at 53.

UEC initiated this action on October 27, 2004. On February 15, 2005, UEC filed an amended complaint alleging that the Forest Service's approval of the six projects at issue was contrary to NFMA, its implementing regulations, and the APA. On July 6, 2005, the district court issued a memorandum decision rejecting all of UEC's claims and entering judgment in favor of defendants. UEC has since filed a timely notice of appeal.

"Because neither NEPA nor NFMA provide a private right of action, this court reviews the Forest Service's approval" of the projects at issue "as ... final agency action[s] under the APA." UEC III, 443 F.3d at 739. Under that standard, we "will not overturn the agency's decision unless it is `arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.'" Id. (quoting 5 U.S.C. § 706(2)(A)). An agency's decision will be deemed "arbitrary and capricious `if the agency ... entirely failed to consider an important aspect of the problem, offered an explanation for its decision that runs counter to the evidence before the agency, or is so implausible that it could not be ascribed to a difference in view or the product of agency expertise.'" Id. (quoting Motor Vehicle Mfrs. Ass'n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43, 103 S. Ct. 2856, 77 L. Ed. 2d 443 (1983)). Likewise, an agency's decision will be deemed arbitrary and capricious if the agency failed to base its decision on "consideration of the relevant factors," or if "there has been a clear error of judgment" on the agency's part. Id. Does the 1982 rule apply to the Forest Service's approval of the six projects?

In its first issue on appeal, UEC argues that the district court committed reversible error by applying the 2000 planning rules, and their "best available science" standard, to each of the six projects at issue, rather than the 1982 planning rules and their focus on MIS. We review the district court's decision on this issue de novo. UEC III, 443 F.3d at 739. In doing so, however, we must "grant[] `substantial deference' to the agency's interpretation of its own regulations." Id. "We may reject the agency's interpretation only when it is `unreasonable, plainly erroneous, or inconsistent with the regulation's plain meaning.'" Id. (quoting Bar MK Ranches v. Yuetter, 994 F.2d 735, 738 (10th Cir.1993)).

Defendants argue that, "under the 2000 rule's transition provisions, the 1982 rule did not apply to site-specific decisions, such as the six projects" at issue. Aplee. Br. at 26. Instead, defendants argue, "the Forest Service was required to consider `best available science' during the transition period that began in November 2000," and "[t]he remainder of the 2000 rule did not apply until the expiration of the transition period." Id. at 27. According to defendants, "[t]he transition period ended in January 2005, with the repeal of the 2000 rule and the promulgation of the 2005 rule." Id. Defendants also assert that "UEC's argument that the 1982 rule is controlling is ... contrary to the Forest Service's [September 29,] 2004 interpretative rule explaining the 2000 rule." Id. That interpretative rule, defendants assert, "unequivocally stated that, during the transition period, the 1982 rule did not apply to site-specific decisions." Id.

We accept in part and reject in part defendants' arguments. Three of the six projects at issue were intended to implement forest plans that pre-dated the 2000 planning rules: the SITLA and South Manti Timber Salvage projects, both of which implemented the Manti-La Sal National Forest Plan, and the Dark Valley Vegetation Management project, which implemented the Dixie National Forest Plan. Neither of those two forest plans "explicitly reference[d] or adopt[ed] § 219.19 of the 1982 [planning] rules, concerning the selection and monitoring of management indicator species." UEC III, 443 F.3d at 748. Thus, "[u]nder the transition provision of the 2000 regulations, the Forest Service was required to consider the `best available science' when implementing [the three] site-specific projects within [those two] forest plan[s]." The Ecology Ctr., Inc. v. United States Forest Serv., 451 F.3d 1183, 1190 (10th Cir.2006) (citing 36 C.F.R. § 219.35(a) (2001)).

The remaining three projects, i.e., the Bear Hodges II project, the East Fork Fire Salvage project, and the White River Salvage Sale project, present a different situation. These three projects were intended to implement two forest plans that were revised in 2003 in accordance with the 2000 planning rules, i.e., the Wasatch-Cache National Forest Plan and the Uinta National Forest Plan. By its plain language, § 219.35(a) was inapplicable to those projects. That is, § 219.35(a) mandated application of the best available science standard to the implementation of "current plan[s]," which we construe, in the context of the transition provision as a whole, as encompassing only plans that pre-dated the 2000 planning rules.4 Thus, we look to see what standards were incorporated in the two revised forest plans. As expressly authorized by § 219.35(b), both of the revised forest plans incorporated the old MIS standards, rather than the best available science standard. Indeed, the Wasatch-Cache National Forest Plan expressly referenced the 1982 planning rules. Accordingly, it was the MIS standards, rather than the best available science standard, that was applicable to these three projects. To the extent that defendants contend to the contrary, we conclude their position is inconsistent with the plain language of § 219.35 and our existing precedent.5 See UEC III, 443 F.3d at 748 (indicating that if a forest plan explicitly references or adopts § 219.19 of the 1982 planning rules, a project implementing that plan has to comply with the MIS monitoring requirements set forth in the 1982 planning rules, notwithstanding the fact that the project at issue was approved after the September 29, 2004 final rule).

Did the district court err in considering post-decisional data?

In its second issue on appeal, UEC argues that the district court erred in considering, rather than striking, documents in the administrative record "that contained information about particular MIS in the Wasatch-Cache National Forest that post-date the dates when the projects on this Forest were approved." Aplt. Br. at 21. According to UEC, this includes "[t]he bulk of the Wasatch-Cache's snowshoe hare data," as well as "[p]ost-decisional data for Bonneville cutthroat trout...." Id. UEC argues that these documents should have been stricken for purposes of review because they "were not before the decision-maker at the time of approval of these projects...." Id. at 22.

We agree with UEC. As discussed in greater detail below, the 1982 planning rules required the Forest Service, prior to approving a particular forest project, to gather quantitative population data on the various MIS listed in the particular forest plan at issue, determine population trends for those MIS, and determine the likely effect of the proposed project on those MIS. Obviously, post-decisional data regarding the population numbers and/or trends of a particular MIS is irrelevant to whether the Forest Service properly fulfilled these obligations prior to approving a particular project. Thus, the district court should not have taken any post-decisional data into account in determining whether defendants violated federal law in approving the projects at issue.

Sufficiency of the Forest Service's MIS monitoring

In its third issue on appeal, UEC argues that the Forest Service lacked adequate MIS population trend data at the time it approved each of the projects at issue. At the outset, we summarily reject UEC's argument with respect to the SITLA, South Manti, and Dark Valley projects. As we have already concluded, those three projects were, pursuant to § 219.35(a), subject to the best available science standard. Thus, UEC's specific challenges to those projects are inapposite.

Turning to UEC's challenges to the remaining three projects, we begin by first reviewing the requirements imposed on the Forest Service under the 1982 planning rules. Section 219.19 of the 1982 planning rules provided, in pertinent part, as follows:

Fish and wildlife habitat shall be managed to maintain viable populations of existing native and desired non-native vertebrate species in the planning area. For planning purposes, a viable population shall be regarded as one which has the estimated numbers and distribution of reproductive individuals to insure its continued existence is well distributed in the planning area. In order to insure that viable populations will be maintained, habitat must be provided to support, at least, a minimum number of reproductive individuals and that habitat must be well distributed so that those individuals can interact with others in the planning area.

(a) * * *

(1) In order to estimate the effects of each 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.... These species shall be selected because their population changes are believed to indicate the effects of management activities.

* * *

(2) 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.

* * *

(6) Population trends of the management indicator species will be monitored and relationships to habitat changes determined. This monitoring will be done in cooperation with State fish and wildlife agencies, to the extent practicable.

* * *

36 C.F.R. § 219.19 (1999).

In Utah Environmental Congress v. Bosworth, 372 F.3d 1219, 1224-25 (10th Cir.2004) (UEC I), we held that the provisions of § 219.19, although primarily applicable to the development of Forest Plans, also apply to the Forest Service's authorization of individual projects implementing a Forest Plan. Further, we held in UEC I "that § 219.19 requires the Forest Service to use actual, quantitative population data to effectuate its MIS monitoring obligations." Id. at 1226. More specifically, we concluded

that in order to effectuate its MIS monitoring duties under the language of [the 1982 planning rules], 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 management activities and population trend changes.

Id. at 1227.

UEC contends that, with respect to the two projects within the Wasatch-Cache National Forest (Bear Hodges II and East Fork Fire Salvage), defendants failed to gather adequate data regarding three MIS species: snowshoe hare, beaver, and Bonneville cutthroat trout. UEC's specific arguments regarding each of these species will be addressed below.

a) Snowshoe hare

UEC complains that there was insufficient data collected regarding snowshoe hare because (a) the Forest Service itself acknowledged that survey data it collected in 1973 and 1978 was not sufficient to determine population trends, (b) data cited in the Forest Plan Monitoring Report is primarily from an area distant to the Bear Hodges II project area, and (c) although the Monitoring Report references data taken from one transect by a Utah Department of Wildlife Resources employee, the Report itself acknowledges that data from only one transect is insufficient to draw conclusions on population trends and/or habitat relationships. Aplt. Br. at 26.

Defendants first respond by arguing that UEC has overlooked their specific findings "respecting the lack of impact of the [two] projects on snowshoe hare populations and habitat." Aplee. Br. at 37. As for UEC's specific complaints, defendants argue that, "because the data [wa]s to be evaluated at 5-year intervals under the revised ... forest plan, and given that the revised forest plan issued in 2003, the first evaluation of snowshoe hare population data ... was not yet due when the Bear Hodges II EIS issued in 2004." Id. Further, defendants argue that "even though [they] did not yet have the benefit of the first 5-year population trend analysis contemplated by the forest plan when the Bear Hodges II and East Fork RODs issued in 2004," the available data "indicated that snowshoe hare populations ha[d] been stable across the [forest] over a substantial swath of time." Id. at 38. Defendants also point to post-decisional data that they say supports the conclusions they reached at the time of the RODs. Id. at 38-39.

We find defendants' arguments, other than their reliance on post-decisional data, persuasive. Turning first to the Bear Hodges II project, we conclude the Forest Service effectively complied with the Forest Plan's requirement to determine "[s]nowshoe hare presence" because, in the ROD, it noted that snowshoe habitat existed in the project area, and in turn assumed that a snowshoe population also existed. It is true that the Forest Service did not have an estimate of the precise snowshoe hare population in the project area (or, for that matter, in the Forest as a whole), at least in part because the Forest Plan was revised only a year prior to the project approval, and the 2004 annual survey results for snowshoe hare had not yet been received.6 However, UEC has failed to establish how the presence of such data could have altered the Forest Service's decision regarding the impact of the Bear Hodges II project. As previously noted, the Forest Service concluded that the project would not significantly impact any snowshoe hare populations or habitat in the project area because the project treatments were designed to maintain sprucefir habitat used by the species. The fact that the Forest Service was unaware of the precise numbers of snowshoe hare in the project area does not, in our view, undermine this conclusion—nor, for that matter, does UEC expressly suggest that it does.

With respect to the East Fork project, UEC's arguments overlook the fact that the 2002 fire that precipitated the project effectively destroyed all available snowshoe hare habitat and, in turn, left no snowshoe hare population in the project area. In light of these facts, defendants reasonably concluded that the project would not have a negative impact on snowshoe hare population or habitat in the project area. Moreover, any failure on the part of defendants to adequately monitor snowshoe hare populations in other portions of the Forest is irrelevant in light of these findings and conclusions.

b) Beaver

UEC argues that there was insufficient data regarding beaver because (a) the Monitoring Report indicates that beaver populations were estimated per square mile based on potential habitat, and such extrapolation is not permitted under the Plan or the 1982 planning rules, (b) in approving the Bear Hodges II project the Forest Service improperly relied on the analysis of aerial photos (UEC argues that photo interpretation of habitat or population trends is subjective and inexact), and (c) data from one mountain range within a forest cannot be used to determine the effects of management activities in a separate mountain range within the same forest (as UEC alleges occurred with respect to approval of the East Fork Fire Salvage project).

We conclude that UEC's claims lack merit. To begin with, the Forest Plan indicates that beaver populations are to be measured by the number of active dams, and is silent with respect to the precise method of measurement. In a subsequent Final Environmental Impact Statement, the Forest Service concluded that monitoring of beaver populations could "be done from the ground or the air...." App. at 770. Although UEC objects to the Forest Service measuring beaver populations from the air, it has not persuasively established why this method is deficient, and we conclude we must defer to the defendants on this technical issue. UEC I, 372 F.3d at 1223 (noting that "choices of suitable scientific methodology are entrusted to the agency"). Thus, we conclude that defendants' use of aerial photos was proper under the Plan. Further, because defendants specifically monitored by air the beaver populations in the two project areas (concluding with respect to the Bear Hodges II project that there was no beaver habitat and thus no beaver population, and concluding with respect to the East Fork Fire Salvage project that the beaver population in the project area was either stable or increasing), and in turn concluded based on that monitoring that the two projects would not negatively impact beaver populations or habitat (conclusions which UEC does not specifically challenge), it is irrelevant whether or not defendants adequately monitored beaver populations in other, non-project areas of the Forest. See UEC I, 372 F.3d at 1228 ("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.").

c) Bonneville cutthroat trout

Lastly, UEC complains that, with respect to Bonneville cutthroat trout, (a) there was insufficient quantitative population trend data to predict trends for this species (UEC notes that most of the streams in the WCNF have only been sampled once or twice in the past twenty years), and (b) the most recent data regarding the Bear Hodges II project area is from 1999, even though the Forest Plan requires such data to be collected annually.

We find UEC's arguments unpersuasive. In UEC I, we held that the Forest Service must make good faith efforts to determine "the absence or presence of an MIS species ... in order to fulfill its MIS monitoring obligations under § 219.19." 372 F.3d at 1230. If the Forest Service has made this good faith effort to determine whether or not an MIS is present, however, it is not otherwise required "to attempt to track species where no population [thereof] exists . . . ." Id. (internal quotation marks omitted). That is precisely the situation presented here. Defendants assert, and UEC does not dispute, that no BCT reside in the Bear Hodges II project area, and that the project area in fact contains no habitat suitable for BCT. Accordingly, we conclude it was unnecessary, and indeed impossible, for defendants to have produced quantitative population data for the BCT in the project area. See Utah Envt'l Cong. v. Bosworth, 439 F.3d 1184, 1192 (10th Cir.2006) (UEC II) (reaching similar conclusion where the Forest Service concluded there was no sagebrush habitat within project area); UEC III, 443 F.3d at 751 (concluding Forest Service did not have to collect project-level data on BCT "because no fish-bearing streams and lakes exist[ed] within the cumulative effects area"). In turn, UEC has not offered any persuasive reasons why defendants were wrong in concluding that the Bear Hodges II project would have no effect on the BCT.

UEC contends that, with respect to the Uinta National Forest and the White River Salvage Sale project within it, defendants failed to collect adequate population trend data for the Colorado River Cutthroat Trout in the Left Fork of the White River (the project area is in the Left Fork White River watershed). Although UEC acknowledges that the Forest Service and the Utah Department of Wildlife Resources collected annual data on CRCT between 1991 and 2002, UEC argues that the two agencies "used various survey protocols in different locations," thus making it impossible to precisely determine population trends as required by the Forest Plan. Id. at 50. Further, although UEC admits that defendants collected population data for CRCT in 2003, this did not include any data for the Left Fork of the White River. Because it is undisputed there is no migration between CRCT populations within the Forest (from stream to stream), UEC argues that the 2003 data cannot serve to satisfy the requirements of the Forest Plan. Nor, UEC asserts, can the habitat trend data cited by the district court satisfy the requirements of the Forest Plan and the NFMA.

Defendants argue, in response, that the data cited by UEC does, in fact, have "scientific utility." Aplee. Br. at 49. In particular, defendants assert that the available data "indicate[s] that during the time period between 1991 and 2002 the overall condition of cutthroat trout within these sites show[ed] no observable change." Id. at 49-50. As for their alleged failure to obtain data for the Left Fork of the White River in 2003, defendants argue that they did, in fact, attempt to obtain such data but "the stream channel was dry when monitored that particular year." Id. at 50. Defendants also dispute UEC's assertion that the Forest Plan requires each of the four "conservation populations" of CRCT to be separately monitored. In fact, defendants argue, the Plan "does not identify" these four "conservation populations" "as separate MIS populations." Id. at 51. Lastly, defendants deny that they used habitat data from the Left Fork of the White River as a proxy to satisfy the requirements of the Forest Plan and the NFMA. Instead, defendants assert, this habitat data was used to satisfy the Plan's requirement "for measuring habitat conditions every five years, as well as considering 33% of sample streams annually to determine CRCT population estimates." Id. at 52.

We find defendants' arguments persuasive. Although in UEC I we interpreted § 219.19 as requiring quantitative population data, we otherwise imposed no specific requirements on the type of data that must be collected. In this regard, defendants' conclusion that the available population data was scientifically useful is neither arbitrary nor capricious. See UEC III, 443 F.3d at 739 (noting that "[d]eference to the agency is especially strong where the challenged decisions involve technical or scientific matters within the agency's area of expertise."). As for UEC's assertion that defendants failed to gather quantitative data for the Left Fork of the White River in 2003, defendants have persuasively demonstrated that they made a good faith attempt to collect such data but that it was impossible to do so because the stream channel was dry that particular year. Given the defendants' good faith effort, we therefore conclude, consistent with our decision in UEC I, that no more was required of the Forest Service.

Defendants' failure to apply the best available science standard

Although we have concluded that the Forest Service was bound to apply the best available science standard in approving the SITLA, South Manti, and Dark Valley projects, it is obvious from the record on appeal that the Forest Service failed to do so. In particular, it is beyond dispute that the Forest Service relied solely on MIS-based analysis in approving these three projects. Further, we are not persuaded, as the district court apparently was, that the Forest Service's failure was harmless, i.e., that the Forest Service's reliance on other available data effectively satisfied the best available science requirements. See The Ecology Ctr., 451 F.3d at 1193-95 (concluding, under similar circumstances, that Forest Service's failure to apply best available science standard was not harmless). We therefore conclude, consistent with our decision in The Ecology Center, that it is necessary to vacate the Forest Service's approval of these three projects and remand so that the Forest Service "can have the first chance to apply [the proper] standards." Id. at 1194; see SEC v. Chenery Corp., 332 U.S. 194, 196, 67 S. Ct. 1575, 91 L. Ed. 1995 (1947) (noting that the propriety of an agency determination must be judged "solely by the grounds invoked by the agency"); Forest Watch v. U.S. Forest Serv., 410 F.3d 115, 119 (2d Cir.2005) (vacating Forest Service's approval of project under similar circumstances).

In reaching this conclusion, we have given careful consideration to the procedural history of this case. In the district court, UEC challenged the SITLA, South Manti, and Dark Valley projects on the grounds that the Forest Service failed to collect the requisite MIS data. The Forest Service, in response, made no mention of the 2000 planning rules or the best available science standard. Instead, the Forest Service argued that it had gathered sufficient quantitative data to satisfy the 1982 planning rules.

The district court, in its memorandum opinion and decision rejecting UEC's claims on the merits, was the first to mention the 2000 planning rules and their best available science standard. In particular, the district court concluded that the 2000 planning rules required the Forest Service to consider the best available science standard in implementing the projects at issue. In turn, the district court, in upholding these three project decisions, concluded that the Forest Service had effectively complied with the best available science standard by considering only MIS-based data. That is, the district court affirmed the three project approvals based on its own after-the-fact application of the best available science standard.

UEC, in challenging the district court's ruling on appeal, correctly points out that the Forest Service itself never applied the 2000 planning rules in approving the projects at issue. For example, in its opening brief, UEC notes that the district court incorrectly held "that the 2000 NFMA regulations were applied in the approval for each project," Aplt. Br. at 13, and that "[e]vidence that the Forests `elected' to apply the 1982 regulations (and never elected to apply the 2000 NFMA regulations) can be found throughout the [administrative] record." Id. at 14. Similarly, in its appellate reply brief, UEC notes that the administrative record "is silent in regards to the new `best available science' standard because it was not applied or adhered to in the planning or approval of these" projects. Aplt. Reply Br. at 7. In turn, UEC argues: "Had the [Forest Service]'s responses to UEC's administrative appeal points been premised on the `best available science' standard UEC would have had the option to challenge these projects under this rule. But this was not the case." Id. Although the Forest Service on appeal asks us to adopt the same approach as the district court and conclude that the analysis actually engaged in by the Forest Service effectively satisfies the best available science standard, UEC takes this position to task, correctly noting that "[a] reviewing court may not supply the basis for the agency's decision that the agency itself has not given." Id. at 8; see Chenery, 332 U.S. at 196, 67 S. Ct. 1575 (holding that a reviewing "court is powerless to affirm the administrative action by substituting what it considers to be a more adequate or proper basis"); The Ecology Center, 451 F.3d at 1195 (concluding that a reviewing court cannot properly affirm an administrative decision when the issuing agency clearly failed to apply the proper standard); Forest Watch, 410 F.3d at 119 (same).

The procedural posture of this case thus distinguishes it from UEC III. Unlike UEC III, UEC in this case clearly argued, and after reviewing the administrative record we agree, that the Forest Service acted arbitrarily and capriciously by failing to apply the applicable regulations. Thus, consistent with the Supreme Court's decision in Chenery, we may not affirm the project approvals on a "basis ... [not] ... set forth [in the record] with such clarity as to be understandable." 332 U.S. at 196, 67 S. Ct. 1575. Accordingly, the only solution is to vacate the project approvals and remand so that the Forest Service, in the first instance, may apply the proper standard.

We AFFIRM the district court's order affirming authorization of the Bear Hodges II Timber Sale, East Fork Fire Salvage, and White River Salvage Sale projects, REVERSE the district court's order affirming authorization of the SITLA, South Manti and Dark Valley projects, and REMAND to the district court with directions to VACATE the Forest Service's approval of the SITLA, South Manti and Dark Valley projects.

McCONNELL, J., concurring in part and dissenting in part.

The majority reverses the district court and overturns an agency decision on the basis of an argument never made by the plaintiff in this case. I therefore respectfully dissent from the portions of the decision reversing and remanding the SITLA, South Manti, and Dark Valley projects.

This case involves six unrelated forest projects in four national forests in the State of Utah. All were challenged by Utah Environmental Congress (UEC) on the ground that the Forest Service had failed to collect adequate data regarding Management Indicator Species (MIS), as required by the 1982 planning regulations and the forest plans. The Forest Service defended on two primary grounds: that the 1982 planning regulations have been superseded by a new requirement that forest managers employ the "best available science" for the purpose of forest management decisions, and, in the alternative, that the projects satisfied the requirements of the 1982 planning regulations. The district court upheld all six projects on the first theory, holding that the "best available science" standard—rather than the 1982 planning regulations—applied. At no point has plaintiff UEC argued that the projects violated the "best available science" standard. Both in district court and on appeal, UEC's sole argument has been that the forest managers were required to comply with the 1982 planning regulations, and did not.

In a careful analysis of the regulatory landscape, the majority concludes that the "best available science" standard applies to the SITLA, South Manti, and Dark Valley projects, but that the forest plans applicable to the Bear Hodges II, East Fork, and White River projects incorporate specific MIS data collection requirements. I agree with this analysis. The majority also concludes that Bear Hodges II, East Fork, and White River project data collection satisfies those requirements. I also agree with those conclusions.1 Having rejected UEC's argument that the MIS data requirements of the 1982 planning regulations apply to the SITLA, South Manti, and Dark Valley projects, and given that UEC has not argued that these projects failed to comply with the "best available science" standard, that should be the end of the matter.

Remarkably, however, the majority reverses the district court and overturns these projects on the basis of an argument never made by the plaintiff: "Although we have concluded that the Forest Service was bound to apply the best available science standard in approving the SITLA, South Manti, and Dark Valley projects, it is obvious from the record on appeal that the Forest Service failed to do so." Maj. Op. at 1287. The majority then proceeds to apply harmless error analysis—not plain error analysis, and not waiver analysis—to the issue, and concludes (without further discussion) that the error was not harmless.

This violates well-established principles of appellate review. Harmless error analysis is reserved for cases in which the appellant properly objects to an error at trial, the appellate court finds there was error, and the appellee contends that the decision should nonetheless be affirmed. See, e.g. Polys v. Trans-Colorado Airlines, Inc., 941 F.2d 1404, 1409-10 (10th Cir. 1991); United States v. Rivera, 900 F.2d 1462, 1469-70 (10th Cir.1990). Here, UEC never objected to the error, if it was error. Thus, at most, we should be applying plain error review. Id. Moreover, in its opening brief to this Court, UEC did not challenge the district court's conclusion that the disputed projects satisfied the "best available science" standard. UEC argued only that the district court erred in concluding that the "best available science" standard was applicable. UEC's "Statement of Issues" reads as follows:

1. Did the district court properly apply the correct regulations under the National Forest Management Act?

2. Did the district court properly allow the inclusion of post-decisional population information for various Management Indicator Species (MIS) into the record and can this post-decisional information be used to support the Forest Service's decision?

3. Did the Forests have adequate quantitative population trend data for Management Indicator Species (MIS) pursuant to the Forest Plans and 1982 regulations, under the National Forest Management Act (NFMA), that were used by the various Forest Service decisionmakers when they made their original decisions approving these projects?

Appellant's Br. at 7. Conspicuously absent from the Statement of Issues is any argument in the alternative: that even if the district court was correct to apply the "best available science" standard, it nonetheless erred in its application of that standard to this record. Nor can any such argument be teased out of UEC's opening brief.

It follows that the issue was waived. Dubbs v. Head Start, Inc., 336 F.3d 1194, 1202 n. 4 (10th Cir.2003); Perry v. Woodward, 199 F.3d 1126, 1141 n. 13 (10th Cir.1999). To be sure, "when manifest injustice would otherwise result," this Court can reach issues that were not specifically raised on appeal. Sussman v. Patterson, 108 F.3d 1206, 1210 (10th Cir. 1997) (internal quotation marks omitted). But I can find no precedent justifying reversal of significant agency action, affirmed by the district court, on the basis of a challenge the plaintiff did not make in district court and did not make in this Court. UEC is a sophisticated litigant, specializing in precisely this sort of case. If it chose not to bring a particular legal challenge, we should not comb the record and bring the challenge sua sponte.

The majority takes the district court to task for being "the first to mention the 2000 planning rules and their best available science standard" and for "affirm[ing] the three project approvals based on its own after-the-fact application of the best available science standard." Maj. Op. at 1287-88. The majority thus characterizes the district court's approach as a violation of SEC v. Chenery Corp., 332 U.S. 194, 196, 67 S. Ct. 1575, 91 L. Ed. 1995 (1947). I offer no opinion on whether that characterization is correct, because UEC did not make any such argument (at least not in its opening brief). Indeed, in its opening brief UEC does not cite Chenery Corp. or any of its progeny. UEC crafted its appeal as an argument that the Forest Service remains bound to follow the MIS monitoring standards of the 1982 regulations— not as an argument that the district court affirmed on grounds not supported by the record.

To be sure, the majority quotes two sentences from UEC's reply brief that can be read to support such an argument. See Maj. Op. at 1288, quoting Aplt. Reply Br. at 7. But "[t]his court does not ordinarily review issues raised for the first time in a reply brief." Stump v. Gates, 211 F.3d 527, 533 (10th Cir.2000).2 The majority's sole quotation from the opening brief on this point, see Maj. Op. at 1287, quoting Aplt. Br. at 13-14, is not an argument that the district court erred in its manner of application of the "best available science standard"; it is an argument that, because the Forest Service "elected" to employ the 1982 regulations in the project decision, those regulations are legally binding.

The majority's decision directly conflicts with a recent decision by this Court. In UEC III, this Court rejected the plaintiff's contention that a particular forest project was subject to the 1982 MIS data-collection rules and instead concluded that the "best available science" standard was applicable. 443 F.3d at 748. As here, UEC in that case never argued that the "best available science" standard was misapplied. Indeed, UEC framed its arguments identically to the arguments in the case before us: that the 1982 regulations controlled, and that the 2000 regulations "could not have replaced [them]." UEC III, Appellant's Br. at 46. Accordingly, the Court in UEC III held that: "Since UEC does not argue that the Forest Service failed to consider the best available science when it implemented the Seven Mile Project, we find no error in the district court's order affirming the agency's decision." Id. at 749. That decision should control our disposition of this case.

The majority's attempts to distinguish this case from UEC III are unconvincing; the arguments in the two cases' opening briefs are, on this issue, substantially identical. The majority notes that, in our case, UEC's opening brief states that "[e]vidence that the Forests `elected' to apply the 1982 regulations (and never elected to apply the 2000 NFMA regulations) can be found throughout the [administrative] record." Maj. Op. at 1287 (quoting Appellant's Br. at 14). In UEC III, the appellants argued that "the district court's reliance on the transitional 2000 rule conflicts with statements in the project record that state that the 1982 regulations were the applicable regulations." UEC III, Appellant's Br. at 47. In UEC III, this Court held that such language could not properly be understood as an argument that the "best available science" standard applies, and I see no reason to come to a different conclusion here.

Nor could Appellants have had any doubt that they were expected to make such an argument explicitly if they wished it to be considered. After UEC III, Appellants were on notice that a direct attack on the application of the "best available science" standard would be required, if that standard was found to be controlling. UEC submitted briefs in this case on June 1, 2006, two months after the UEC III opinion was issued. Their failure to argue, at that point, that the "best available science" standard was incorrectly applied appears to be a tactical litigation choice, and we do not typically reward such choices by making the appellant's argument for him.

The majority suggests that another recent decision, Ecology Center, Inc. v. United States Forest Serv., 451 F.3d 1183 (10th Cir.2006), points the other way. See Maj. Op. at 1286-87. I cannot agree. To be sure, the Court in Ecology Center remanded a proposed forest project to the agency for application of the "best available science" standard. Ecology Center, 451 F.3d at 1195. But the opinion in that case makes no reference to any failure on the part of the plaintiff to raise the issue, and thus cannot be treated as precedent for disregarding our traditional waiver rules.3 UEC III explicitly addresses the plaintiff's failure to raise the issue, and thus provides the precedent we must follow.

Moreover, Ecology Center differs from this case in a fundamental respect. In Ecology Center, the Court reversed the district court's application of the 1982 planning regulations to the disputed project. The question, then, was whether that error was harmless. In this case, by contrast, we have affirmed the district court's decision to apply the "best available science" standard to these three projects, and no one has raised any claim that the district court erred in its application of that standard. We need not ask whether the district court's decision was harmless, because it was not error.

And finally, assuming arguendo that these two precedents directly conflict, we are obliged to follow the ruling of the earlier panel—in this case, UEC III. Hiller v. Oklahoma ex rel. Used Motor Vehicle and Parts Comm'n, 327 F.3d 1247, 1251 (10th Cir.2003) ("To the extent that [two 10th Circuit panel decisions] are in conflict, ... we are obligated to follow the earlier panel decision over the later one."); see also McMellon v. United States, 387 F.3d 329, 333 (4th Cir.2004) (noting eight circuits that give precedence to the earlier of two conflicting panel decisions).

The majority states that "it is obvious from the record on appeal" that the Forest Service failed to apply the best available science standard. Maj. Op. at 1287. To me, this statement illustrates why we should not go venturing off in pursuit of arguments not made by the parties. The record in this case comprises 46 volumes, with many thousands of pages of text and exhibits. It contains a great deal of scientific detail. I do not know how the agency's analysis of the various issues stacks up against the "best available science" standard. But neither does the majority. For all we know, the MIS data analyzed in connection with these projects are the best available science. UEC does not argue otherwise. It argues only that the data are deficient under the 1982 planning guidelines, which is a different question. If UEC had argued that the decisions in question were deficient under the "best available science" standard, the Forest Service would have been able to respond, and the district court would have been able to make appropriate findings. That is why we have appellate rules: to ensure that both sides have notice of the disputed issues and an opportunity to weigh in.