OPINION

In their suit filed pursuant to the Administrative Procedures Act (APA), 5 U.S.C. § 706 , the League of Wilderness Defenders — -Blue Mountains Biodiversity Project and Cascadia Wildlands Project (collectively, LOWD) sought declaratory and injunctive relief to halt the Deep Creek Vegetation Management Project (the Project), which called for the selective logging of 12.8 million board feet of timber in the Ochoco National Forest. LOWD claims in its suit that the United States Forest Service (Forest Service) failed to comply with the National Environmental Policy Act (NEPA), 42 U.S.C. § 4231 et seq., and the National Forest Management Act (NFMA), 16 U.S.C. § 1600 et seq., in developing and implementing the Project. The district court denied LOWD’s motion for summary judgment and granted the Forest Service’s cross-motion for summary judgment. Because the Final Supplemental Environmental Impact Statement (FSEIS) may not tier to a non-NEPA watershed analysis to consider adequately the aggregate cumulative effects of past timber sales, we reverse the district court’s grant of summary judgment in favor of the Forest Service, and we remand this case so the Forest Service can reissue its NEPA documentation to include the omitted information regarding past timber sales contained in the watershed analysis.

FACTUAL AND PROCEDURAL BACKGROUND

A. 1999 Deep Creek Watershed Analysis

The Forest Service manages the Ochoco National Forest under the Ochoco Nation al Forest Land and Resource Management Plan (Ochoco LRMP). In 1993, the Regional Forester directed the Paulina Ranger District, among others, to conduct a watershed analysis that identified “processes and functions within the Deep Creek watershed that are key to maintaining sustainable and resilient terrestrial and aquatic ecosystems.” To carry out these instructions, the Forest Service created an interdisciplinary team of employees that eventually documented its findings and conclusions in the August 1999 Deep Creek Watershed Analysis (Watershed Analysis).

The Watershed Analysis collectively considered past actions in the Deep Creek watershed and the results of those actions to determine existing conditions and trends. Based on this cumulative evaluation, the Watershed Analysis determined that many of the tree stands were overly dense and susceptible to high-intensity wildfires and forest diseases; that some types of tree stands were over-represented as compared to historical conditions, while other types of stands were under-represented; and that riparian habitats along streams exhibited similar problems. The Watershed Analysis accordingly concluded that the Forest Service should undertake a blend of management activities to alleviate and improve these unsatisfactory conditions.

B. The Project

In October 1999, based on the results of the Watershed Analysis, the Forest Service initiated the NEPA documentation process for the Project. In April 2001, the Forest Service published a draft Environmental Impact Statement (EIS) and, after receiving public comments and making adjustments, issued a final EIS in September 2001. The Forest Service also issued a corresponding Record of Decision (ROD).

Shortly thereafter, LOWD filed an administrative appeal of the Project on behalf of itself and a number of environmental organizations. In response, the Forest Service withdrew the original ROD to perform additional analyses on the Project’s effects, and dismissed the appeal as moot.

In July 2002, the Forest Service issued a draft supplemental EIS for the Project. After receiving and considering public comments, the Forest Service issued a FSEIS in January 2004. According to the FSEIS, the stated purpose and need for the Project are to: (1) “move the landscape-level diversity of vegetation and associated wildlife habitat closer to the [historic range of variability (HRV) ] ... in terms of species composition and structure,” given that “forest vegetation is outside the [HRV] for 57% of the [watershed]”; (2) “increase the amount of single strata late and old structure (LOS) stands” and move “the overall abundance of LOS closer to the [HRV]”; (3) “reduce the forest’s susceptibility to moderate and high severity fires” by lowering fuel levels, reducing stand densities, “increasing the relative abundance of fire tolerant species, and re-introducing fire into the watershed”; (4) “reduce the [forest’s] susceptibility ... to insects, diseases, and wildfires by reducing their stocking levels”; (5) “enhance vegetative conditions in the aspen, riparian, upland shrub, and meadow communities” that have gradually declined over time; and (6) “improve water quality and enhance the vegetation aspect of aquatic/riparian areas to provide for long-term sustainability of resident and anadro-mous fisheries by reducing stream temperatures and lowering sedimentation.”

In January 2004, the Forest Service issued another ROD and selected modified Alternative C, which includes commercial timber harvest on 6261 acres, pre-commer-cial thinning on 9957 acres, 6.1 miles of new and temporary road construction, 16.3 miles of road reconstruction, and fuel-reduction treatments on 5379 acres. Modified Alternative C would allow logging of 12.8 million board feet of timber, primarily through ground-based tractor logging. The ROD concludes that this alternative presents “the best balance of activities suited for meeting the identified needs of the Deep Watershed at this time” and “balances water quality issues while improving uplands and riparian vegetation conditions and reducing susceptibility to moderate and high severity fires.”

C. Procedural History

Following issuance of the 2004 ROD, LOWD filed an administrative appeal, which the Forest Service denied. LOWD then filed a complaint in the District of Oregon pursuant to the APA, alleging that the Forest Service’s approval of the Project violated NEPA, NFMA, and the applicable Ochoco LRMP, and seeking declaratory and injunctive relief. In adopting the magistrate judge’s findings and recommendations, the district court denied LOWD’s motion for summary judgment, granted the Forest Service’s cross-motion for summary judgment, and dismissed the action with prejudice. LOWD timely appealed. We have jurisdiction over LOWD’s appeal pursuant to 28 U.S.C. § 1291 .

STANDARD OF REVIEW

We review the district court’s summary judgment ruling de novo. Nw. Envtl. Advocates v. Nat'l Marine Fisheries Serv., 460 F.3d 1125 , 1132 (9th Cir.2006). Thus, “ ‘[vjiewing the evidence in the light most favorable to the nonmoving party,’ we must determine ‘whether there are any genuine issues of material fact and whether the district court correctly applied the law.’ ” Pension Trust Fund for Operating Eng’rs v. Fed. Ins. Co., 307 F.3d 944, 948-49 (9th Cir.2002) (quoting State Farm Mut. Auto. Ins. Co. v. Davis, 937 F.2d 1415, 1417 (9th Cir.1991)).

In making this determination, “we must remember that the APA provides the authority for our review of decisions under NEPA and NFMA.” Lands Council v. McNair (Lands Council II), 537 F.3d 981, 987 (9th Cir.2008) (en banc). Under the APA, the court may set aside only Forest Service actions that are “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” 5 U.S.C. § 706 (2)(A).

“Review under the arbitrary and capricious standard ‘is narrow, and [we do] not substitute [our] judgment for that of the agency.’ ” Lands Council II, 537 F.3d at 987 (quoting Earth Island Inst. v. U.S. Forest Serv., 442 F.3d 1147, 1156 (9th Cir.2006)) (alterations in original). “Rather, we will reverse a decision as arbitrary and capricious only if the agency relied on factors Congress did not intend it to consider, ‘entirely failed to consider an important aspect of the problem,’ or offered an explanation ‘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 Earth Island Inst., 442 F.3d at 1156 ). In other words, there must be “ ‘a clear error of judgment.’ ” Marsh v. Or. Natural Res. Council, 490 U.S. 360, 378 , 109 S.Ct. 1851 , 104 L.Ed.2d 377 (1989) (quoting Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 416 , 91 S.Ct. 814 , 28 L.Ed.2d 136 (1971)).

DISCUSSION

LOWD argues that, in developing the Project, the Forest Service violated NEPA because the FSEIS fails to consider adequately the cumulative effects of the Project, as required by 40 C.F.R. § 1508.7 . LOWD also argues that the Forest Service violated NFMA and the applicable Ochoco LRMP by: (1) failing to assure that Forest Service employees properly mark large-diameter trees for avoidance of harvest; and (2) failing to maintain connective habitat corridors in the Project planning area.

A. Cumulative Effects Analysis

LOWD first claims that the Forest Service failed to consider adequately the cumulative environmental effects of other timber sales and grazing in the Project planning area. The Forest Service responds that its responsibility is to consider cumulative effects in the aggregate rather than on an individual basis, and that the FSEIS contains a sufficient discussion of cumulative effects to satisfy the agency’s duty under NEPA. We conclude that the Forest Service may aggregate its cumulative effects analysis, and that it properly did so with respect to grazing and future timber sales. However, we find the aggregated cumulative effects analysis of past timber sales deficient because the Forest Service failed to include the relevant inputs in the FSEIS itself.

NEPA provides that a federal agency that proposes a “major Federal action! ] significantly affecting the quality of the human environment” must prepare a detailed EIS on the proposed action, including an analysis of “alternatives to the proposed action” and a discussion of the significant environmental impacts. 42 U.S.C. § 4332 (2)(C). To comply with this requirement, the Forest Service must consider, among other things, the “cumulative impacts” of the proposed action, which NEPA’s implementing regulations define as the “impact on the environment which results from the incremental impact of the action when added to other past, present, and reasonably foreseeable future actions regardless of what agency (Federal or non-Federal) or person undertakes such other actions.” 40 C.F.R. § 1508.7 .

1. Timber Sales

a. Past Timber Sales

LOWD contends that the cumulative effects analysis in the FSEIS regarding past timber sales is insufficient because it “only mentions one ... past timber sale, the Summit timber sale,” and otherwise generally “states that timber harvest has occurred in the past.” The Forest Service counters that, under its reading of Department of Transportation v. Public Citizen, 541 U.S. 752 , 124 S.Ct. 2204 , 159 L.Ed.2d 60 (2004), “agencies can conduct an adequate cumulative effects analysis by focusing on the current aggregate effects of past actions.”

i. Aggregation

In rejecting LOWD’s past-timber-sales arguments, the district court agreed with the Forest Service’s interpretation of Public Citizen :

The district court then concluded, without elaboration or citation to the record, that “the FSEIS discusses cumulative effects from past vegetation management activities in the watershed to a sufficient degree.”

This reading of Public Citizen is inaccurate. In that case, the Court addressed the narrow issue of whether NEPA “require[d] the Federal Motor Carrier Safety Administration (FMCSA) to evaluate the environmental effects of cross-border operations of Mexican-domiciled motor carriers” when FMCSA merely promulgated administrative rules implementing a presidential order “allowing] such border-crossing activities to occur.” Public Citizen, 541 U.S. at 756, 766 , 124 S.Ct. 2204 . In concluding that NEPA lacked such a requirement, the Court did state that “[t]he ‘cumulative impact’ regulation required FMCSA to consider the ‘incremental impact’ of the safety rules themselves, in the context of the [presidential order] and other relevant circumstances,” not “to treat the [presidential order] itself, or consequences from the [order].” Id. at 669-70, 124 S.Ct. 2204 (quoting 40 C.F.R. § 1508.7 ). Nevertheless, the Court emphasized twice that its analysis and holding were limited to the “critical feature” of the case — ie., that FMCSA lacked authority to countermand the presidential order allowing Mexican carriers into the United States, id. at 766, 770 , 124 S.Ct. 2204 — and subsequent Ninth Circuit cases have limited their application of Public Citizen on that basis. See, e.g., Or. Natural Res. Council Fund v. Brong, 492 F.3d 1120 , 1134 n. 20 (9th Cir.2007) (holding that “Public Citizen’s limitation on NEPA does not apply” where an agency has statutory authority to prevent the relevant effects). Accordingly, because the Forest Service has statutory authority to regulate the environmental consequences of the Project, Public Citizen does not support the agency’s position.

We conclude, however, that a different source does permit the Forest Service to consider cumulative effects in the aggregate. During the summary judgment proceedings, the Forest Service introduced a June 24, 2005 memorandum issued by the Chairman of the Council on Environmental Quality (CEQ), entitled “Guidance on Consideration of Past Actions in Cumulative Effects Analysis.” This CEQ memorandum advises that “[a]gencies are not required to list or analyze the effects of individual past actions unless such information is necessary to describe the cumulative effects of all past actions combined.” Instead, the memorandum explains, “agencies can conduct an adequate cumulative effects analysis by focusing on the current aggregate effects of past actions without delving into the historical details of individual past actions.”

The magistrate judge struck the CEQ memorandum from the record, characterizing it as a “-hoc rationalization” with no relevance to judicial review of the Forest Service’s compliance with NEPA. Based on Supreme Court precedent, however, this determination constituted an abuse of the magistrate judge’s discretion. See Golden Gate Hotel Ass’n v. City & County of S.F., 18 F.3d 1482, 1485 (9th Cir.1994) (“We review a district court’s decision to grant a motion to strike unscheduled supplementary material for abuse of discretion.”). In Auer v. Robbins, 519 U.S. 452, 461 , 117 S.Ct. 905 , 137 L.Ed.2d 79 (1997), the Court granted deference to an agency’s interpretation of its own regulation even though the agency offered the interpretation for the first time as a litigation position. The Court noted that the agency’s interpretation was “controlling unless plainly erroneous or inconsistent with the regulation,” and that “[t]here [was] simply no reason to suspect that the interpretation d[id] not reflect the agency’s fair and considered judgment on the matter in question.” Id. at 461-62 , 117 S.Ct. 905 (internal quotation marks omitted).

Similarly here, CEQ’s interpretation that 40 C.F.R. § 1508.7 permits consideration of all past impacts in the aggregate is not plainly erroneous or inconsistent with the language of the regulation, and CEQ is the agency charged with interpreting NEPA and that adopted the regulation. See Jones v. Gordon, 792 F.2d 821, 827 (9th Cir.1986). In addition, LOWD points to no evidence to suggest that CEQ’s interpretation does not reflect the agency’s fair and considered judgment on the cumulative effects issue, and CEQ’s interpretation is just as plausible as LOWD’s position that analysis of “past effects” requires evaluation of discrete past events. The CEQ memorandum is therefore entitled to Auer deference and, as a result, we hold that the Forest Service may aggregate its cumulative effects analysis pursuant to 40 C.F.R. § 1508.7 .

Our circuit’s precedent further supports this conclusion. Although LOWD argues that Lands Council v. Powell (Lands Council I), 395 F.3d 1019, 1028 (9th Cir.2005), requires a complete cataloguing of all prior timber sales in all cases, this interpretation is incorrect. Lands Council I instead merely reaffirms the general rule that “NEPA requires adequate cataloguing of relevant past projects in the area.” Id. at 1027 (citing Muckleshoot Indian Tribe v. U.S. Forest Serv., 177 F.3d 800 , 809-10 (9th Cir.1999)) (emphasis added). An aggregated cumulative effects analysis that includes the relevant past-timber-sale inputs comports with this standard, and also furthers NEPA’s purpose of “concentrating] on the issues that are truly significant to the action in question.” See 40 C.F.R. § 1500.1 (b).

Moreover, sitting en banc, this court recently clarified that Lands Council I does not require the Forest Service to conduct any particular test or to use any particular method, so long as “the evidence ... provided to support[the Forest Service’s] conclusions, along with other materials in the record,” ensure that the agency “made no clear error of judgment that would render its action arbitrary and capricious.” Lands Council II, 537 F.3d at 993 (internal quotation marks omitted). In Lands Council II, the en banc court reasoned that this approach not only “respects our law that requires us to defer to an agency’s determination in an area involving a high level of technical expertise,” but also acknowledges that “[w]e are not free to impose on the agency [our] own notion of which procedures are best or most likely to further some vague, undefined public good.” Id. (internal quotation marks omitted) (alterations in original).

Accordingly, to the extent that 40 C.F.R. § 1508.7 does not explicitly provide otherwise, the Forest Service is free to consider cumulative effects in the aggregate or to use any other procedure it deems appropriate. It is not for this court to tell the Forest Service what specific evidence to include, nor how specifically to present it.

ii. Tiering

Notwithstanding our conclusion that the Forest Service may aggregate its cumulative effects analyses, the Forest Service’s almost complete failure to include the relevant past-timber-sale inputs in the FSEIS itself fails to survive arbitrary and capricious review. See Lands Council II, 537 F.3d at 987 (noting that “we will reverse a decision as arbitrary and capricious ... if the agency ... ‘entirely failed to consider an important aspect of the problem’”) (quoting Earth Island Inst., 442 F.3d at 1156 ).

As LOWD observes, the FSEIS itself mentions only one past timber sale, the Summit Timber Sale, and otherwise notes generally that other timber sales occurred in the past. The Forest Service counters that no deficiency exists because the omitted information about the past timber sales appears in the Watershed Analysis:

The FSEIS states, in turn, that “[t]he purpose and need for action is based on the analysis and conclusions regarding the conditions described in the Deep Creek Watershed Analysis,” and that the Watershed Analysis “is tiered to the[FSEIS] ... and its applicable [ROD]. The management direction, along with standards and guidelines, for activities proposed are based, in part, on these documents.”

The problem with the Forest Service’s approach, however, is that the FSEIS cannot “tier” to the Watershed Analysis to analyze sufficiently the cumulative effects of the Project. The NEPA implementing regulations define “tiering” as:

40 C.F.R. § 1508.28 . Meanwhile, this court “ha[s] previously interpreted the regulations to allow tiering only to another environmental impact statement.” Muckleshoot Indian Tribe, 177 F.3d at 810; see also Or. Natural Res. Council v. U.S. BLM, 470 F.3d 818 , 823 (9th Cir.2006) (holding similarly proposed tiering impermissible because “the Watershed Analysis is not a NEPA document”); Kern v. U.S. BLM, 284 F.3d 1062 , 1073 (9th Cir.2002) (holding that “tiering to a document that has not itself been subject to NEPA review is not permitted”).

Accordingly, although we in no way doubt what the Forest Service says about the Watershed Analysis, the FSEIS’s cumulative effects analysis is insufficient because it cannot tier to a non-NEPA document that the FSEIS fails to include. We therefore reverse the district court’s grant of summary judgment in favor of the Forest Service, and remand so that the agency can reissue its NEPA documentation to include the omitted, but clearly relevant, information regarding past timber sales contained in the Watershed Analysis. See Kern, 284 F.3d at 1073 (“While NEPA empowers neither the plaintiffs nor this court to second-guess the [agency’s] management decisions, it does require the [agency] to articulate, publicly and in detail, the reasons for and likely effects of those management decisions, and to allow public comment on that articulation.”).

b. Future Timber Sales

LOWD also argues that “there is simply no mention of future timber sales in the FSEIS.” Regarding such projects, however, the FSEIS states: “No other timber vegetation management activities are planned within this watershed within the foreseeable future.” As the district court noted, “this distinguishes the ... Project from other timber harvest cases where courts found that multiple planned timber sales in one area required consideration of cumulative effects as to timber management activities.” See, e.g., Blue Mountain Biodiversity Project v. Blackwood, 161 F.3d 1208 , 1214-16 (9th Cir.1998). Accordingly, the district court properly concluded that the FSEIS need say nothing more regarding future timber sales to satisfy the cumulative effects standard.

2. Grazing

LOWD further asserts that the FSEIS fails to consider adequately the cumulative environmental effects of grazing in the Project planning area. This argument lacks merit for the following reasons. First, the FSEIS adequately evaluates the cumulative effects of past grazing. For example, the FSEIS discusses how unregulated sheep grazing dating back to the 1880s, before the conversion of the allotments to cattle, contributed to “the loss of [soil], stream bank degradation, and channel erosion”; explains the past development of an allotment “to draw cattle from riparian areas and improve live-stock distribution” and its resulting “effects[on] sensitive aquatic habitats”; and notes “[t]he cumulative effects of fire suppression, roads, conifer invasion, and browsing by cattle and native ungulates have had detrimental effects on hydrology ... and riparian vegetation.”

Second, the FSEIS adequately evaluates the cumulative effects of present grazing. In a section entitled “Cumulative Effects on Range,” for instance, the FSEIS states that “[livestock plays no role in the overstory stand composition, structures, or density aspects of the Purpose and Need” of the Project, “[n]or does it have much influence on the risk of moderate and high intensity fires.” Although the Forest Service acknowledges in that section that limited areas exist where grazing could affect the vegetative conditions of riparian communities, the agency explains that limited thinning and no-treatment buffers along streams will mitigate any such effects. The FSEIS also discusses grazing effects in the cumulative-effects-analysis sections for noxious weeds, soils, threatened, endangered, and sensitive species, riparian habitats, and habitat-indicator species like deer and elk.

Finally, regarding future grazing, the FSEIS explains that “Allotment Management Plan updates ... were initiated in 2002 and alternative grazing regimes will be developed and considered under a separate environmental analysis and decision that is scheduled for completion in 2004.” A cumulative effects analysis of those future grazing regimes was therefore impracticable because the Forest Service had not yet designated the specific grazing allotments at the time it issued the FSEIS. Cf. N. Alaska Envtl. Ctr. v. Kempthorne, 457 F.3d 969, 976-77 (9th Cir.2006) (“We conclude that the government[, in its EIS for oil and gas leases,] was not required at this stage to do a parcel by parcel examination of potential environmental effects. Such effects are currently unidentifiable, because the parcels likely to be affected are not yet known.”). Accordingly, the district court properly held that the FSEIS adequately considers the cumulative effects of grazing.

B. Marking of Large-Diameter Trees

Next, LOWD contends that the Forest Service’s tree markings will result in the logging of trees equal to or greater than twenty-one inches in diameter, in violation of NFMA and the applicable Ocho-co LRMP. LOWD also asserts that the Forest Service violated NFMA because agency personnel allegedly did not mark the trees, and because the timber sale contract language is insufficient to prevent harvesting of the large trees.

At oral argument in the district court, the court urged the Forest Service to provide further assurances of the correct marking of trees greater than twenty-one inches in diameter at breast height. The Forest Service responded with the declarations of two Forest Service employees. The Declaration of Lori Blackburn, “a certified silviculturist for the Forest Service for over 12 years,” states that she “wrote the marking guides for the [Project]; participated in training the District marking crew; and monitored their work during the implementation of [her] marking guides on the sale.” The “marking crew consisted of seasonal Forest Service employees supervised by a ... permanent Forest Service employee,” and Blackburn “formally reviewed the work of the marking crew three times,” ensuring that the crew properly measured and marked the trees to avoid harvesting those over twenty-one inches in diameter. Her declaration concludes, “my oversight and monitoring indicate that the marking crew for the [Project] has marked the sale in accordance with the prescriptions and design criteria that have been adopted for the ... [P]roject.”

Meanwhile, the Declaration of Dennis R. Dietrich, “a Timber Sale Contracting Officer for the Ochoco ... National Forest[ ],” provides the following language from the timber sale contract regarding “Individual Tree Designation”:

Dietrich concludes that “[t]his combination of the marking on the trees and the description in the contract, along with professional contract administration, makes [him] confident that the Forest Service has taken every reasonable measure to protect against the harvest of trees that are not authorized by contract to be cut.” He also notes that, “[t]o further protect against the possibility that a tree over 21 inches in diameter will be cut during execution of the [Project], the contract that will be awarded for the sale will provide for liquidated damages to be assessed against the purchaser if any such unauthorized harvest were to occur.”

Like the district court, we hold that these declarations sufficiently indicate that Forest Service employees marked the trees, as required by 16 U.S.C. § 472a(g), and that the marking complied with NFMA’s and the applicable Ochoco LRMP’s requirements.

C. Connective Habitat Corridors

Finally, LOWD alleges that the Project fails to maintain connective habitat corridors in the planning area, as required by the NFMA and the applicable Ochoco LRMP. LOWD points to the fact that many of the connective corridors in the Project planning area already fail to meet the Ochoco LRMP, and that further timber harvest in those corridors will only exacerbate the problem. The Forest Service responds that LOWD incorrectly focuses the inquiry on each corridor as a whole rather than on the particular stands within a connective corridor.

The Ochoco LRMP, as amended by the Eastside Screens interim standards, directs the Forest Service, in pertinent part, to:

Standard 6(d)(3)(a)(l)-(2). The Ochoco LRMP explains that “[hjarvesting within connectivity corridors is permitted if all the criteria in [subparagraph] (2) can be met, and if some amount of understory (if any occurs) is left in patches or scattered to assist in supporting stand density and cover.” Standard 6(d)(3)(a)(4). The Ocho-co LRMP also states that, “[t]o reduce fragmentation of LOS stands, or at least not increase it from current levels, stands that do not currently meet LOS that are located within, or surrounded by, blocks of LOS stands should not be considered for even-aged regeneration, or group selection at this time.” Standard 6(d)(3)(a)(4)(b).

Meanwhile, the FSEIS states that, “[o]f the 3,966 acres identified in connectivity corridors, 23% currently are in the upper one-third of site potential,” meaning that 77% of the planning does not comply with the connectivity requirements. The Forest Service — interpreting the language of Standard 6(d)(3)(a)(2) to mean that “harvesting within a particular stand is permissible whe[n] the stand has a canopy closure within the top one-third of site potential and if the stand is at least 400 feet at its narrowest point” — asserts that all of the individual stands proposed for harvest in the Project, regardless of which corridor they are in, meet the necessary criteria. LOWD, by contrast, interprets Standard 6(d)(3)(a)(2) to mean that all stands within the corridor as a whole must be in the top one-third of site potential for harvesting in that corridor. Accordingly, LOWD argues, “[bjecause none of the corridors meet the requirement, none of the timber stands — which comprise the corridors — -meet the connectivity corridor requirements.”

We find the Forest Service’s interpretation persuasive for the following reasons. First, as noted above, an agency’s interpretation of its own regulations is entitled to substantial deference, and our review is limited to ensuring that the agency’s interpretation is not “plainly erroneous or inconsistent with the regulation.” Auer, 519 U.S. at 461-62 , 117 S.Ct. 905 (internal quotation marks omitted). Because Standard 6(d)(3)(a)(2) refers to “Stand” or “Stands” instead of every stand or all stands, and because Standard 6(d)(3)(a)(4) refers to “[hjarvesting within connectivity corridors” — as opposed to “harvesting a connectivity corridor” or simply “harvesting connectivity corridors” — the Forest Service’s interpretation does not appear plainly erroneous or inconsistent. Rather, LOWD’s interpretation appears plainly erroneous, as it would prohibit harvesting in any watershed governed by the connectivity requirements, an impractical result almost certainly unintended by the Forest Service.

Second, the FSEIS relies on data collection and findings made by the Deep Creek EIS Interdisciplinary Team, which identified and mapped the connection acreage, assessed the harvesting proposed within each corridor, and certified that the silvi-cultural prescriptions will ensure that the stands remain within the top one-third of their site potential.

Finally, the Forest Service points out— and LOWD does not dispute — that no change in the connectivity-compliance percentages will result from implementation of the Project, thereby fulfilling the connectivity requirements’ directive that the agency “maintain” connective corridors. See Standard 6(d)(3)(a). As a result, the district court properly concluded that the Project does not violate the connective-corridor requirements of NFMA and the Ochoco LRMP.

CONCLUSION

The Forest Service’s approval of the Project violates NEPA because the FSEIS may not tier to the non-NEPA Watershed Analysis to consider adequately the aggregate cumulative effects of past timber sales. We reverse the district court’s grant of summary judgment in favor of the Forest Service, and we remand this case so the agency can reissue its NEPA documentation to include the omitted information regarding past timber sales contained in the Watershed Analysis. Each party shall bear its own costs on appeal.

REVERSED and REMANDED.

. Although LOWD does not focus on "ongoing” timber sale activities in the Project planning area, the record reflects that the FSEIS's cumulative effects analysis is also sufficient in that regard. . See 16 U.S.C. § 1604 (i) (requiring consistency between site-specific projects and the applicable LRMP). . A set of interim wildlife, ecosystem, and riparian standards known as the "Eastside Screens” and adopted in 1993 amended the Ochoco LRMP, in relevant part, to proscribe "the logging of green trees larger than 21 inches at breast height.” .LOWD relies on 16 U.S.C. § 472a(g), which states: Designation, marking when necessary, and supervision of harvesting of trees ... shall by conducted by persons employed by the Secretary of Agriculture. Such persons shall have no personal interest in the purchase or harvest of such products and shall not be directly or indirectly in the employment of the purchaser thereof. . See 16 U.S.C. § 1604 (i) (requiring consistency between site-specific projects and the applicable LRMP).