Full opinion text
OPINION M. SMITH, Circuit Judge: This case arises out of three applications by the Tongue River Railroad Company, Inc. (TRRC) to build a 130-mile railroad line in Southeastern Montana to haul coal. The Surface Transportation Board (Board), or its predecessor, the Interstate Commerce Commission (ICC), approved each of the three applications (individually, TRRC I, II, and III). Northern Plains Resource Council, Inc. (NPRC), Mark Fix, the City of Forsyth, Native Action, Inc. (Native Action), and United Transportation Union-General Committee of Adjustment (UTU-GCA) (collectively, Petitioners) challenge TRRC II and III. Petitioners challenge the approval of TRRC II and III on a number of environmental and public convenience and necessity grounds. We hold that the Board failed to take the requisite “hard look” at certain material environmental impacts inherent in TRRC II and III in the manner required by the National Environmental Policy Act (NEPA) prior to approving those applications. We further hold that the Board did not err in its public convenience and necessity analyses, except with respect to its reliance on the viability of TRRC II during the approval of TRRC III. Accordingly, we reverse and remand in part, and affirm in part. BACKGROUND I. Statutory Framework A. NEPA Requirements NEPA requires that federal agencies prepare “a detailed statement by the responsible official on ... the environmental impact” of any federal actions “significantly affecting the quality of the human environment.” 42 U.S.C. § 4332(C); Ctr. for Biological Diversity v. Nat’l Highway Traffic Safety Admin., 538 F.3d 1172, 1185 (9th Cir.2008). NEPA’s purpose is twofold: (1) to ensure that agencies carefully consider information about significant environmental impacts and (2) to guarantee relevant information is available to the public. Robertson v. Methow Valley Citizens Council, 490 U.S. 332, 349, 109 S.Ct. 1835, 104 L.Ed.2d 351 (1989); Ctr. for Biological Diversity, 538 F.3d at 1185. Regulations governing how NEPA is implemented have been promulgated by the Council of Environmental Quality, at 40 C.F.R. §§ 1505.1-1508.28. Regulations governing how NEPA applies to the constructions of railroads have also been promulgated by the Board, at 49 C.F.R. §§ 1105.1-1105.12. An environmental impact statement (EIS) is normally prepared for new railroad construction projects by the Board’s Section of Environmental Analysis (SEA). See 49 C.F.R. §§ 1105.2, 1105.6(a). An EIS must analyze the direct, indirect, and cumulative impacts from a proposed action. 40 C.F.R. § 1508.25(c). The SEA first invites public comment on the scope of the issues to be analyzed in the EIS. 49 C.F.R. § 1105.10(a)(2). After finalizing the scope of the issues to be considered, and completing consultations and site visits, the SEA issues a draft EIS (DEIS). Id. § 1105.10(a)(3). Public comments may be submitted to the SEA after the DEIS has been issued. Id. § 1105.10(a)(4). The SEA’s final EIS (FEIS) discusses comments received, and any changes made to the DEIS in response to those comments. Id. The FEIS, and any comments and responses to the DEIS, are part of the record considered by the Board in deciding whether to grant an application. Id. § 1105.10(f). B. Licensing of New Railroad Lines Under 49 U.S.C. § 10901, the Board has exclusive licensing authority for the construction and operation of new railroad lines. A proceeding to grant authority begins when an application is filed with the Board. 49 U.S.C. § 10901(b). Under Section 10901(c), as amended by the Interstate Commerce Commission Termination Act (ICCTA) of 1995, the Board “shall issue a certificate authorizing activities for which such authority is requested in an application filed under subsection (b) unless the Board finds that such activities are inconsistent with the public convenience and necessity.” 49 U.S.C. § 10901(c). II. Factual Background The TRRC seeks to construct and operate a railroad line in the Tongue River Valley in southeastern Montana. See end-note 1, fig. 1-1. In 1983, the TRRC filed an application with the ICC to construct an 89-mile railroad line between Miles City and Ashland, Montana. See Tongue River R.R. Co. — Rail Construction and Operation — in Custer, Powder River and Rosebud Counties, MT, Finance Docket No. 30186, 1986 ICC LEXIS 314, at *1 (ICC 1986). The purpose of the line was to serve new coal mines in the Ashland area. The plan was to connect the new line to the main line railroad at Miles City, presently owned by Burlington Northern Santa Fe Railroad (BNSF). Id. The ICC approved the TRRC I application on May 1, 1986. Id. The issuance of the TRRC I application is not before us in this litigation. We note, however, that to date the TRRC I line has not been constructed. In 1989, the TRRC filed an application with the ICC to construct and operate an approximately 41-mile-long railroad line from Ashland to Decker, Montana. This line, TRRC II, was intended to connect with TRRC I to create a combined railroad line of approximately 130 miles. The purpose of TRRC II was to bring coal from Wyoming’s Powder River Basin to the BNSF main line in Miles City, and then on to other destinations in the Midwest. The ICC was concerned about potential environmental impacts that would be caused by the construction of TRRC II’s “preferred route,” as submitted in its application. As a result, the ICC also evaluated an alternate route called the “Four Mile Creek Alternative,” which was approximately ten miles longer than the TRRC’s preferred route requested in its TRRC II application. In April 1996, the Board issued its FEIS for the TRRC II project. In October 1996, the Board issued its decision approving TRRC II, utilizing the Four Mile Creek Alternative with numerous mitigation conditions. The Board also required that the TRRC build the entire line from Ashland to Decker and the 89-mile line from Miles City to Ash-land, approved in TRRC I, within three years of the date of its TRRC II decision. Petitioners sought reconsideration of TRRC II, which the Board denied. Petitioners NPRC, UTU-GCA, and Native Action all filed appeals to our court. Subsequently, the TRRC (and BNSF through intervention) also petitioned for reconsideration of TRRC II, claiming that new studies showed the Four Mile Creek Alternative was not viable from an engineering and operational perspective, and, instead, proposed a new alternative, the 17.4-mile “Western Alignment.” The Board denied the TRRC petition for reconsideration in December 1997 on the grounds that its objections could have been raised earlier, but left open the possibility that the TRRC could file a new application for the Western Alignment. In 1998, Petitioners’ appeals in this matter were ordered held in abeyance pending the Board’s decision concerning TRRC III, discussed infra. In 1999, the Board revoked the condition that the entire line be built within three years of its TRRC II decision, and did not impose any future time limit on construction of the line. In April 1998, the TRRC filed a new application to build the Western Alignment rather than the Four Mile Creek Alternative. That proceeding is called TRRC III in this appeal. Subsequently, the NEPA process was delayed for almost three years due to TRRC’s financial problems. In October 2004, the Board’s contractor issued a Draft Supplemental Environmental Impact Statement (DSEIS) for TRRC III. The Final Supplemental Environmental Impact Statement (FSEIS) was completed in 2006 for TRRC III, and included updates to the environmental reviews in TRRC I and TRRC II. On October 5, 2007, the Board approved the construction and operation of the Western Alignment along with numerous mitigation measures. Petitioners timely appeal. STANDARD OF REVIEW AND JURISDICTION I. Review of NEPA Claims Section 706 of the Administrative Procedure Act (APA) governs judicial review of agency decisions made pursuant to NEPA. 5 U.S.C. § 706; City of Sausalito v. O’Neill, 386 F.3d 1186, 1205-06 (9th Cir. 2004). An agency’s action must be upheld unless it is “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 v. McNair, 537 F.3d 981, 987 (9th Cir.2008) (en banc) (quoting Earth Island Inst. v. U.S. Forest Serv., 442 F.3d 1147, 1156 (9th Cir.2006), abrogated on other grounds by Winter v. Natural Res. Def. Council, Inc., 555 U.S. 7, 129 S.Ct. 365, 172 L.Ed.2d 249 (2008)) (internal quotations marks and brackets omitted). [W]e 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. at 987 (citations and internal quotation marks omitted). We look to the evidence the agency has provided to support its conclusions, along with materials in the record, to make this determination. Id. at 993. A court generally must be “at its most deferential” when reviewing scientific judgments and technical analyses within the agency’s expertise. See Balt Gas & Elec. Co. v. Natural Res. Def. Council, Inc., 462 U.S. 87, 103, 103 S.Ct. 2246, 76 L.Ed.2d 437 (1983). The court is not to “act as a panel of scientists that instructs the [agency] ..., chooses among scientific studies ..., and orders the agency to explain every possible scientific uncertainty.” Lands Council, 537 F.3d at 988. And “[w]hen specialists express conflicting views, an agency must have discretion to rely on the reasonable opinions of its own qualified experts even if, as an original matter, a court might find contrary views more persuasive.” Id. at 1000 (quoting Marsh v. Or. Natural Res. Council, 490 U.S. 360, 378, 109 S.Ct. 1851, 104 L.Ed.2d 377 (1989)). While we afford deference to the judgment and expertise of the agency, the agency must, at a minimum, support its conclusions with studies that the agency deems reliable. Lands Council, 537 F.3d at 994. The agency must “explain the conclusions it has drawn from its chosen methodology, and the reasons it considered the underlying evidence to be reliable.” Id. The agency will have acted arbitrarily and capriciously when “the record plainly demonstrates that [the agency] made a clear error in judgment in concluding that a project meets the requirements” of NEPA. See id. NEPA imposes a procedural requirement on federal agencies to “take[ ] a ‘hard look’ at the potential environmental consequences of the proposed action.” Or. Natural Res. Council v. Bureau of Land Mgmt., 470 F.3d 818, 820 (9th Cir.2006) (quoting Klamath-Siskiyou Wildlands Ctr. v. Bureau of Land Mgmt., 387 F.3d 989, 993 (9th Cir.2004)). “Judicial review of agency decision-making under NEPA is limited to the question of whether the agency took a ‘hard look’ at the proposed action as required by a strict reading of NEPA’s procedural requirements.” Bering Strait Citizens for Responsible Dev. v. U.S. Army Corps of Eng’rs, 524 F.3d 938, 947 (9th Cir.2008) (citing Churchill Cnty. v. Norton, 276 F.3d 1060, 1072 (9th Cir. 2001)). Through these procedural requirements, NEPA seeks to make certain that agencies “will have available, and will carefully consider, detailed information concerning significant environmental impacts, and that the relevant information will be made available to the larger [public] audience.” N. Idaho Cmty. Action Network v. U.S. Dep’t of Transp., 545 F.3d 1147, 1153 (9th Cir.2008) (citations and internal quotation marks omitted). II. Review of the Railroad Application “Our review of the construction given by an agency to a statute it administers is limited and deferential.” Wash. State Dep’t of Game v. I.C.C., 829 F.2d 877, 879 (9th Cir.1987). We review the agency’s statutory construction under the analytical framework set out in Chevron U.S.A, Inc. v. Natural Resources Defense Council, 467 U.S. 837, 842-45, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984). DHX, Inc. v. Surface Transp. Bd., 501 F.3d 1080, 1086 (9th Cir. 2007). First, we inquire whether Congress has addressed directly the issue before the court. Chevron, 467 U.S. at 842, 104 S.Ct. 2778. If Congress has clearly spoken on the issue, our inquiry ends and the agency “must give effect to the unambiguously expressed intent of Congress.” Id. at 842-43, 104 S.Ct. 2778. We are compelled to overrule an agency’s interpretation if it is “contrary to clear congressional intent.” Id. at 843 n.9, 104 S.Ct. 2778. However, if the statute does not address the specific issue before us, or does so ambiguously, “the question for the court is whether the agency’s answer is based on a permissible construction of the statute.” Id. at 843, 104 S.Ct. 2778. The agency’s construction must therefore only be “reasonable” and need not be the same as the construction the court itself would have embraced had it reviewed the statute de novo. Id. at 843-44, 104 S.Ct. 2778. “[T]he court does not simply impose its own construction on the statute....” Id. at 843,104 S.Ct. 2778. We review decisions by the Board on railroad application approvals under the APA, 5 U.S.C. § 706(2). DHX 501 F.3d at 1086. An agency’s decision must be upheld unless it is “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” 5 U.S.C. § 706(2)(A). The Board’s factual findings must be upheld if they are supported by substantial evidence. In re Transcon Lines, 89 F.3d 559, 564 (9th Cir.1996). III. Jurisdiction. We have jurisdiction under 28 U.S.C. §§ 2321(a), 2342(5), and 2344. DISCUSSION I. Environmental Claims A. Cumulative Impacts “NEPA requires that where several actions have a cumulative ... environmental effect, this consequence must be considered in an EIS.” Te-Moak Tribe of W. Shoshone of Nev. v. U.S. Dep’t of Interior, 608 F.3d 592, 602 (9th Cir.2010) (citation and internal quotation marks omitted). The pertinent regulation defines cumulative impact as follows: Cumulative impact is 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. Cumulative impacts can result from individually minor but collectively significant actions taking place over a period of time. 40 C.F.R. § 1508.7. A cumulative impact analysis “must be more than perfunctory; it must provide ‘a useful analysis of the cumulative impacts of past, present, and future projects.’ ” Kern v. U.S. Bureau of Land Mgmt., 284 F.3d 1062, 1075 (9th Cir.2002) (quoting Muckleshoot Indian Tribe v. U.S. Forest Serv., 177 F.3d 800, 810 (9th Cir.1999)). To be useful to decision makers and the public, the cumulative impact analysis must include “some quantified or detailed information; ... general statements about possible effects and some risk do not constitute a hard look absent a justification regarding why more definitive information could not be provided.” Ocean Advocates v. U.S. Army Corps of Eng’rs, 402 F.3d 846, 868 (9th Cir.2005) (quoting Neighbors of Cuddy Mountain v. U.S. Forest Serv., 137 F.3d 1372, 1379-80 (9th Cir.1998)). Federal agencies may “aggregate [ ] cumulative effects analysis” for NEPA purposes. League of Wilderness Defenders-Blue Mountains Biodiversity Project v. U.S. Forest Serv., 549 F.3d 1211, 1217 (9th Cir.2008) (“[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.” (quoting Council on Environmental Quality Memorandum, “Guidance on Consideration of Past Actions in Cumulative Effects Analysis” (June 24, 2005))). Petitioners contend that the Board’s cumulative impact analysis in TRRC III ignores the combined impacts of future coal bed methane (CBM) well development and coal mining projects that will also come into being in Southeastern Montana. Petitioners further contend that the Board failed to account for the combined effects of the referenced projects and the likely effects on air quality, wildlife, and water quality of the proposed construction and operation of the TRRC railroad. We agree with Petitioners’ contentions concerning the cumulative foreseeable effects of CBM wells and the Otter Creek Coal Mine. 1. Coal Bed Methane In 2003, the Bureau of Land Management (BLM) and the State of Montana finished a programmatic EIS (Methane EIS) that evaluated the future impacts of CBM development in the Powder River Basin. Bureau of Land Management, Final Statewide Oil and Gas Envtl. Impact Statement (Jan. 2003) (Methane EIS), available at http://www.blm.gov/mt/st/en/ fo/miles_city_field_office/og_eis.html. The Methane EIS determined that a reasonably foreseeable development scenario in Montana over the next 20 years would involve drilling between 10,000 and 26,000 new CBM wells and 250 and 975 conventional oil and gas wells. Id. at 4-5. The Tongue River Valley is an area of concentration for such wells. Id. at Map 4-1. The Board incorporated the Methane EIS in its preparation of the FSEIS in TRRC III. The Board also evaluated the cumulative impact of nine approved sites for CBM development near the TRRC sites. The BLM and the State of Montana had previously issued Environmental Assessments (EAs) for each of the projects and found that “no cumulative impacts are expected to occur” because adverse impacts were likely to occur only if there was simultaneous construction of the CBM wells and the TRRC railroad, and that project construction would not overlap. Notably, the Board found that “[t]he simultaneous construction of CBM gas wells could result in cumulative effects on land use.” Ultimately, however, the Board concluded that because “impacts of current CBM proposals would not occur within the same temporal parameters as construction of the Tongue River Railroad, no cumulative impacts would result.” As a result, the Board determined that there would be no cumulative impacts from the nine approved CBM projects. a. Five-Year Time Frame Importantly, in order to reach its no cumulative impacts conclusion, the Board limited its analysis to projects that would be constructed within a five-year period. Petitioners challenge this limitation. The Board defends its decision to use a five-year period to analyze cumulative effects because that period accounts for the time period necessary to construct the railroad (estimated to be three years) and to operate the railroad for two years thereafter. Further, the Board argues that it often uses a five-year period when examining environmental effects. Moreover, the Board maintains that any analysis focused on a time period after five years is wholly speculative because of likely changes in economic growth, industry growth, and shifts in energy transportation policies. Viewed in light of the facts presented by this case, we disagree with the Board. Each project is different, and the agency is required to rationally explain its decision in the context of project-specific effects. See Council on Envtl. Quality, Considering Cumulative Effects Under the National Environmental Policy Act, Office of NEPA Policy and Compliance, 16 (Jan. 1997), available at http://energy.gov/sites/ prod/files/nepapub/nepa_documents/Red Dont/G-CE Q-ConsidCumulEffects.pdf (“The time frame of the project-specific analysis should also be evaluated to determine its applicability to the cumulative effects analysis.”) (emphasis added); see also Selkirk Conservation Alliance v. Forsgren, 336 F.3d 944, 962 (9th Cir.2003) (recognizing that the scope of the EIS is a “delicate choice” and should be entrusted to the agency but the agency must have “considered the relevant factors and articulated a rational connection between the facts found and the choice made”) (citation omitted). The Board partially justifies its use of a five-year limitation period because it has used a similar time framework in the past. However, in this case, the BLM and State of Montana’s study projects a significant growth of CBM wells over the next 20 years. Methane EIS at 4-5. Thus, they anticipated that building CBM wells will continue well after the termination of a five-year time framework. The Board does not explain why its default five-year time frame should necessarily apply under these circumstances. The Board also reasoned that five years was appropriate because it would allow for approximately three years of construction time, followed by two years of operation time. The Board’s assumption that the railroad would be constructed and operational within five years of approval in 2007 is clearly flawed in light of the history of this railroad project, which has been under way for almost three decades, starting with TRRC I’s approval in 1986. To date, construction has not begun on any section of the railroad. Accordingly, there is little reason to presume that the railroad would be constructed in five years, which — importantly for the Board’s analysis — means that we can place little confidence in the Board’s assumption that the project will not overlap with reasonably foreseeable CBM projects. Moreover, time restrictions for the construction of TRRC I and TRRC II were withdrawn, and TRRC III contains no construction time requirements. If we credit the 20-year BLM analysis and actual railroad construction is pushed out beyond five years, there may be overlapping CBM well construction. However, the Board’s environmental analysis — which admits that cumulative impacts may result from simultaneous development — does not address this issue or account for this possibility. In its decision, the Board worked around the Methane EIS by concluding that it would be too speculative to determine the effects from wells that are not already approved. In other words, because the Board does not know specific construction time lines or locations, it will await the development of site-specific EISs before including the known projects into its analysis. We acknowledge that the Board is not required to engage in speculative analysis. As we have explained, “[i]t is not appropriate to defer consideration of cumulative impacts to a future date when meaningful consideration can be given now, nor do we require the government to do the impractical, if not enough information is available to permit meaningful consideration.” Envtl. Prot. Info. Ctr. v. U.S. Forest Serv., 451 F.3d 1005, 1014 (9th Cir.2006) (internal quotation marks and citation omitted). However, projects need not be finalized before they are reasonably foreseeable. “NEPA requires that an EIS engage in reasonable forecasting. Because speculation is ... implicit in NEPA, [ ] we must reject any attempt by agencies to shirk their responsibilities under NEPA by labeling any and all discussion of future environmental effects as crystal ball inquiry.” Selkirk, 336 F.3d at 962 (internal quotation marks and citation omitted). As the Environmental Protection Agency (EPA) also has noted, “reasonably foreseeable future actions need to be considered even if they are not specific proposals.” EPA, Consideration of Cumulative Impact Analysis in EPA Review of NEPA Documents, Office of Federal Activities, 12-13 (May 1999), available at http://www. epa.gov/compliance/resources/policies/ nepa/cumulative.pdf. The TRRC is not operating in a vacuum. The route for the railroad is known, including the terrain and counties in which it would operate. Likewise, the Methane EIS has described in some detail the likely scope of CBM development in the future. Although the Methane EIS is a programmatic document, it is specific to the number of wells, field compressors, roads, and pipelines for each of the three counties that the TRRC railroad would cross. Methane EIS, Minerals Appendix at 17-22. For example, in Big Horn County (where Decker, Montana, the end of the railroad in TRRC III, is located), an estimated 2,500 to 7,000 CBM wells are reasonably foreseeable, and most of the wells would be located in the southeastern part of the county, where Decker is located. Id. at 17. The EIS also projected that 100 to 250 field compressors and 1,450 to 4,200 miles of gathering lines would be built in this area. Id. at 17. Similarly, in Rosebud County (where Ashland, Montana, the other end of the railroad in TRRC III is located), an estimated 1,000 to 2,800 CBM wells will be drilled, along with 40 to 100 field compressors and 600 to 1,650 miles of gathering lines. Id. at 19. The greatest concentration of CBM development is anticipated to occur in the Rosebud, Big Horn, and neighboring Powder River Counties, which is proximate to where the TRRC railroad will run. Id. at Map 4-1. In short, the BLM and the State of Montana have described a time frame and a reasonably foreseeable development plan for CBM development in areas that overlap with the TRRC railroad plans. Under the circumstances, the Board has not sufficiently explained why it cannot or should not incorporate this available data concerning likely future development into its environmental impact analysis. We do not ask the Board to peer into a crystal ball. See Selkirk, 336 F.3d at 962. The Methane EIS report contained actual numbers, broken down by counties, about development over the next 20 years. We conclude, therefore, that in this case, an adequate cumulative impact analysis necessarily requires that such information be included. Accordingly, we hold that the Board arbitrarily and capriciously relied on the five-year time frame, which resulted in a faulty analysis of the possible cumulative impacts from reasonably foreseeable CBM projects that could overlap construction of the railroad line. b. Nine Approved CBM Sites Petitioners also argue that the Board’s EIS did not properly consider the effects of the nine already-approved CBM wells. These projects were either under consideration or had already been approved by the time the TRRC III environmental study and Plans of Development (PODs) for the projects were submitted to the Board for consideration. The Board concluded that any potentially significant cumulative impacts would be limited to the construction period, and that there would be no overlap between the well construction and the railroad construction activities. Petitioners contend that the Board failed to consider the significant long-term cumulative operational impact of the railroad and these wells. We conclude that this finding by the Board was not arbitrary and capricious. First, Petitioners do not present any evidence that the construction of the approved projects will overlap with TRRC construction. The Board concluded that the construction for the approved CBMs typically occur within two to sixth months after their approval and in some cases, construction had already begun at the time of the analysis. Thus, there would be no overlap with railroad construction, which would not occur until much later. We cannot overturn this finding without evidence to the contrary. Second, the Board did address both construction and operational impacts in many areas, including land use, biological resources, soil and geology, hydrology and water quality, cultural and paleontological resources, transportation and safety, air quality, noise and vibration, socioeconomics and environmental justice, recreation, aesthetics, and energy. Based on these considerations, the Board reasonably concluded that the TRRC projects would not result in cumulative effects with respect to the nine approved CBM sites. Petitioners argue that this conclusion is far-fetched and belied by the record, but again fail to present evidence that contradicts the findings of the Board. c. Air Quality and Wildlife Beyond their attack on temporal limitation, see supra, Petitioners argue that the Board failed to consider the cumulative effect of the CBM wells and the railroad on air quality and wildlife. We disagree. The Board analyzed air quality assuming an operational railroad and concluded that such effects would be minor, particularly when mitigation measures are applied. The Board considered two types of emissions that would affect air quality— fugitive dust emissions and combustion emissions. Fugitive dust emissions include wind-blown dust which could result from exposed soils from devegetation along the railroad right of way. The SEA estimated that 10% of the right of way would have exposed soils and estimated the emissions based on EPA emission factors. Fugitive dust emissions also include coal dust from traveling coal cars. Combustion emissions from the operation of the railroad locomotives were also considered using emission standards released by the EPA. The Board is afforded deference in choosing its scientific method for modeling data, and it chose to limit the scope to the Air Quality Control Region in Eastern Montana. See Lands Council, 537 F.3d at 988 (explaining that the court is not to impose its own scientific judgment on the agency); id. at 1000 (‘When specialists express conflicting views, an agency must have discretion to rely on the reasonable opinions of its own qualified experts even if, as an original matter, a court might find contrary views more persuasive.”) (citation and quotation marks omitted). Thus, Petitioners do not demonstrate how the Board here failed to take a “hard look,” even if they did not include the air quality model preferred by Petitioners or air quality data from Wyoming. Additionally, the Board analyzed effects of construction and operation of the railroad and the CBM wells on wildlife and reasonably concluded that most of the effects were likely to occur during construction, rather than during operation of the railroad. For example, the Board found that most of the effects, including removal of vegetation and habitat, increased sedimentation, and loss of flood plain, are localized and temporary in nature. The EIS addresses these impacts and creates mitigation measures. Additionally, Petitioners do not cite any specific deficiencies in this analysis, but instead point to generalized wildlife cumulative impacts addressed in the Methane EIS for the entire state. Therefore, there is no basis for finding this particular Board decision irrational, and it is entitled to deference. 2. Other Coal Mines Petitioners further argue that the Board failed to address any cumulative impacts of the railroad with existing coal mines — in particular, the Ashland and Otter Creek Mines. a. Ashland Coal Mines Petitioners contend that the Board also erred by relying on the analysis of the Ashland Mines from TRRC I when doing its analysis for TRRC II and III. Petitioners are correct that the TRRC I data was incorporated. Wholesale adoption of TRRC I’s data into TRRC II and III could conceivably raise some questions because TRRC I pre-dated the other projects by many years, and the data could have changed in the meantime. However, the Board evaluated whether there would be any increases in coal production in the Ashland area mines for TRRC II and TRRC III. The Board concluded that no material changes are predicted, beyond those already analyzed in TRRC I. Accordingly, because coal production has remained stable, as noted in the TRRC I study, and was reevaluated for the two subsequent projects, the Board did not unreasonably continue to rely on the Ash-land mines coal production conclusions in TRRC I in doing its analysis for TRRC II and TRRC III. b. Otter Creek Coal Mines However, we do not reach the same conclusion as to the Otter Creek mines. Petitioners argue that the Board failed to address any impacts from the Otter Creek mines, notwithstanding the fact that the financial justification for the entire line included hauling Otter Creek coal. The Board first responds that Petitioners have waived this argument because Petitioner Mark Fix agreed in comments sent to the Board that the Otter Creek development was unforeseeable. Second, the Board claims that mining development in Otter Creek was not reasonably foreseeable when the Supplemental EIS in TRRC III was completed. We disagree with both the Board’s contentions. Parties must alert an agency to their position and contentions. Vermont Yankee Nuclear Power Corp. v. Natural Res. Def. Council, Inc., 435 U.S. 519, 553, 98 S.Ct. 1197, 55 L.Ed.2d 460 (1978). A party waives arguments that are not raised during the administrative process. See Dep’t of Transp. v. Public Citizen, 541 U.S. 752, 764-65, 124 S.Ct. 2204, 159 L.Ed.2d 60 (2004). Here, however, Petitioner NPRC specifically commented that development of the Otter Creek coal mines must be addressed. Specifically, on December 6, 2004, NPRC stated that the DSEIS was deficient because: [i]n its preliminary determination that the TRR is in the public need, the [Board] relies, at least in part, on the development of coal tracts along Otter Creek and the development of coal-fired power plants near Ashland, Montana. The [Board] contends that the TRR is needed to provide a means of hauling coal from yet-to-be leased much less permitted coal tracts. Yet in the SDEIS, the [Board] concludes that the proposed coal mines and power plants are too speculative to be considered reasonably foreseeable future actions. Consequently, the [Board] failed to evaluate the potential cumulative impacts of this development on the human environment. This [Board] decision is disingenuous, implausible, and a textbook example of arbitrary and capricious decision-making. Earlier, on November 16, 2004, Petitioner Mark Fix told the Board that he did not see Otter Creek as foreseeable. However, in this case, the subsequent comment by Petitioner NPRC brought the issue to the Board’s attention. The Board had notice of the concern and was afforded the opportunity to bring its expertise to bear. Thus, the argument is not waived. Moreover, the Board’s own explanation that the Otter Creek mines were not foreseeable is arbitrary and capricious. The Board knew that in 2002 the federal government transferred the land to the State of Montana for coal development, and the DSEIS included a map with the sites of future coal mines. Even more significantly, the Board relied on the coal mine development in Otter Creek to justify the financial soundness of the proposal, since it included the tonnage forecasts in its 2007 final decision. We therefore hold that, because the Otter Creek mine was reasonably foreseeable at the time of TRRC III, the FSEIS violated NEPA by failing to address the mine’s impacts. 3. Water Quality Petitioners’ last cumulative effect argument is that the FSEIS does not analyze the combined effects of the railroad, CBM, and coal mines as to water quality. The Board responds that, because there is really no impact from CBM development (because construction will not occur simultaneously) and there are virtually no impacts from coal mine development (as mitigated), there are no effects to add to the virtually non-existent impact on water quality from the railroad. We agree with the Board that where a proposed project has “virtually no effect” on water quality, the agency is not required to examine cumulative impacts from other projects because it would not provide an informed analysis. See Nw. Envtl. Advocates v. Nat’l Marine Fisheries Serv., 460 F.3d 1125, 1140 (9th Cir. 2006). However, we disagree with the Board’s analysis in this instance because it is largely based on the reasoning that the railroad construction will not occur simultaneously with the construction of the CBM wells. This argument does not address the concern expressed above, see supra Section LA.l.a, that there may well be simultaneous construction. Construction of the railroad is estimated to increase sentiment loading in the Tongue River. Thus, were simultaneous construction to occur, the Board must address this water quality issue. The Board also fails to address the cumulative effects of railroad construction combined with the water quality degradation from operational CBM wells, which are estimated to be substantial. The Board’s position is more persuasive as it relates to operation of the railroad, since the railroad is not expected to discharge water during operation. Mitigation measures such as revegetation are estimated to reduce the sediment delivery during operations to “near zero.” Accordingly, we hold that the Board’s FSEIS in TRRC III is deficient as to its water quality analysis related to construction, but sufficient as to operations. Furthermore, because the Board failed to account for the Otter Creek coal mines, the Board should perform cumulative water quality analysis, taking this reasonably foreseeable project into account on remand. B. Adequacy of the Baseline Data Petitioners also contend that the TRRC II and III EIS documents do not provide adequate baseline data to assess the impacts of the railroad. Petitioners take issue with the Board’s analysis concerning the pallid sturgeon, sage grouse, fish and aquatic resources, other wildlife, and sensitive plants. Because the TRRC III FSEIS does not provide baseline data for many of the species, and instead plans to conduct surveys and studies as part of its post-approval mitigation measures, we hold that the Board did not take a sufficiently “hard look” to fulfill its NEPAimposed obligations at the impacts as to these species prior to issuing its decision. NEPA requires that the agency provide the data on which it bases its environmental analysis. See Lands Council, 537 F.3d at 994 (holding that an agency must support its conclusions with studies that the agency deems reliable). Such analyses must occur before the proposed action is approved, not afterward. See LaFlamme v. F.E.R.C., 852 F.2d 389, 400 (9th Cir.1988) (“[T]he very purpose of NEPA’s requirement that an EIS be prepared for all actions that may significantly affect the environment is to obviate the need for speculation by insuring that available data is gathered and analyzed prior to the implementation of the proposed action.”) (internal citation and quotation marks omitted). “[0]nce a project begins, the ‘pre-project environment’ becomes a thing of the past” and evaluation of the project’s effect becomes “simply impossible.” Id. Petitioners cite the pallid sturgeon as an example of a wildlife resource with insufficient baseline data. The pallid sturgeon was listed in 1990 as being endangered throughout its range by the U.S. Fish and Wildlife Service (FWS). During the EIS process for TRRC III, the FWS submitted comments to the Board indicating that the Board failed to determine whether vibrations and noise from the railroad would be harmful to the fish at a nearby hatchery. The TRRC had hired an engineering firm, Womack & Associates, Inc. (Womack) in 1998 to study the effects of the railroad at the Miles City Fish Hatchery, with supplemental reports being prepared in 2004 and 2006. The FWS, however, disagreed with the analysis by Womack and expressed interest in doing its own studies as needed to assure that there would be no impacts from the construction and operation of the TRRC railroad. As the TRRC III decision itself recognized: As set out in the SEIS, the parties also have had a number of discussions on the potential vibration issues. [FWS] has asked the applicant to conduct additional baseline studies to more fully understand the potential long-term effects that vibration may have on the fish. TRRC is concerned that long-term studies of the sort [FWS] has sought could significantly delay its construction schedule. However, in April 2006, the railroad agreed to implement a work plan (included as Appendix G to the Final SEIS) for additional vibration monitoring at the hatchery. The work plan, implemented as a mitigation measure, was developed by Womack and the noise and vibration program would include measurements and analysis to “Measure baseline conditions at the [Miles City Fish Hatchery],” to “[p]redict and assess future sound pressure levels from construction and operation of the TRR near the [Miles City Fish Hatchery] and compare to baseline conditions,” and to “Measure actual noise and vibration during the construction and operation of the TRR to compare actual levels to predicted levels.” Submitted, as to sage grouse, the only data the Board collected was the number of acres of potential sage grouse habitat within the 200-foot railroad right of way. The BLM submitted comments to the Board indicating that there was insufficient information as to the sage grouse: With the increasing importance of sage grouse, more discussion on sage grouse is needed, including discussion on wintering areas and impacts of West Nile virus on sage grouse. Need to check statewide data base for more information on sage grouse strutting grounds which have been found in the past few years.... Sage grouse inventories need to be conducted at least two miles from any proposed disturbance. In response, the Board agreed to conduct sage grouse surveys at least two miles from the proposed disturbance as a mitigation measure and that “pre-construction surveys would be conducted to determine the extent of sage grouse habitats and activity in the project area.” Further, the Board acknowledged that “[l]ocations of leks [communal grounds used by the grouse for courtship, breeding, and wintering] from Birney to the terminus of the proposed Western Alignment at the Spring Creek spur are not well known, so potential impacts on grouse in this portion of the line are difficult to determine.” However, in response to the BLM’s comments, the Board did not collect this data before approval of the railroad application and instead instituted mitigation measure 26 to conduct surveys on sage grouse at a later date. Finally, as to fish and aquatic resources, other wildlife, and sensitive plants, Petitioners also contend that baseline data is only being gathered as part of mitigation measures. For fish, mitigation measure 34 purports to “conduct a three-part study plan to identify aquatic resources” by conducting a fish and fish habitat survey to “estimate population and to monitor potential mortality or emigration due to construction impacts.” For other wildlife, other mitigation measures call for data reconnaissance to “locate habitat areas and nesting sites” for several species of animals for the entire rail line. For plants, the SEA was unable to conduct ground-level surveys to identify which state and federally protected plant species existed along the railroad’s right of way. Thus, the Board prepared a mitigation measure to conduct a field search to “identify plant species of concern (Federal and state) and to implement appropriate mitigation measures during construction activities if such species are found.” We recognize the Board’s extensive mitigation efforts. However, such mitigation measures, while necessary, are not alone sufficient to meet the Board’s NEPA obligations to determine the projected extent of the environmental harm to enumerated resources before a project is approved. Mitigation measures may help alleviate impact after construction, but do not help to evaluate and understand the impact before construction. In a way, reliance on mitigation measures presupposes approval. It assumes that — regardless of what effects construction may have on resources- — -there are mitigation measures that might counteract the effect without first understanding the extent of the problem. This is inconsistent with what NEPA requires. NEPA aims (1) to ensure that agencies carefully consider information about significant environmental impacts and (2) to guarantee relevant information is available to the public. Robertson, 490 U.S. at 349, 109 S.Ct. 1835; N. Idaho Cmty. Action Network, 545 F.3d at 1153. The use of mitigation measures as a proxy for baseline data does not further either purpose. First, without this data, an agency cannot carefully consider information about significant environment impacts. Thus, the agency “fail[s] to consider an important aspect of the problem,” resulting in an arbitrary and capricious decision. See Lands Council, 537 F.3d at 987 (citation omitted). Second, even if the mitigation measures may guarantee that the data will be collected some time in the future, the data is not available during the EIS process and is not available to the public for comment. Significantly, in such a situation, the EIS process cannot serve its larger informational role, and the public is deprived of their opportunity to play a role in the decision-making process. See Robertson, 490 U.S. at 349, 109 S.Ct. 1835. While the Board must prepare a detailed statement on “any adverse environmental effects which cannot be avoided should the proposal be implemented,” 42 U.S.C. § 4332(C)(ii), this requirement does not contravene the Board’s other obligation to ensure that data exists before approval so that the Board can understand the adverse environment effects ab initio. The Board responds that it was unable to conduct some of the surveys due to the rough terrain in the area, and that it was unable to access some portions of the land in order to conduct the surveys because the access points were on private property. We are unpersuaded that these excuses can relieve the Board of its requirement under NEPA to gather information before it can make an informed decision. The Board presented no evidence indicating that it even attempted to contact any private land owners about gaining access to their land to conduct the surveys. As to rugged terrain, the Board does not explain how waiting until post-approval mitigation to conduct the survey alleviates this issue. Had the Board attempted to obtain the baseline data for these plants and wildlife through some scientific study or methodology that its experts deemed reliable, we would evaluate the situation from a different perspective. After all, we do not “act as a panel of scientists that instructs the [agency] ..., chooses among scientific studies ..., and orders the agency to explain every possible scientific uncertainty.” Lands Council, 537 F.3d at 988. The problem here, however, is that the Board did not collect this data in the first place, and was therefore unable to consider it during the EIS process. We hold that the Board violated NEPA by not taking a sufficiently “hard look” when it deferred gathering baseline data discussed in this Section I.B. C. Stale Data Petitioners also contend that the Board relied on stale data in making its TRRC III environmental impacts analysis. Board admits that it was unable to conduct on-the-ground surveys as part of the EIS process. The Board cites the rough terrain, rural location, and limited access due to private property as the reasons that it was unable to conduct on-the-ground surveys. The Board instead relied on aerial surveys and photography, along with data from TRRC I and TRRC II. We agree with Petitioners that the Board’s reliance on this data does not constitute a “hard look” under NEPA. Many of the aerial surveys relied upon were conducted many years ago — in 1985 for TRRC I, in 1992 for TRRC II, and in 1997 for TRRC III. At the time of the 2007 decision in TRRC III, even the most recent aerial survey was ten years old. Reliance on data that is too stale to carry the weight assigned to it may be arbitrary and capricious. See Lands Council v. Powell, 395 F.3d 1019, 1031 (9th Cir.2005) (finding that six-year-old data, without updated habitat surveys, was too stale). For the reasons infra, we find the Board’s reliance on stale aerial surveys was arbitrary and capricious under the circumstances in this case. The Board contends that it is entitled to rely on this outdated data because “the physical environment of the area at issue here is substantially the same.” However, the Board does not cite any scientific studies or testimony in the record that supports this conclusion. See Lands Council, 537 F.3d at 994. Moreover, even assuming this was true, we fail to understand how this necessarily and logically leads to the conclusion that the information regarding habitat and populations of numerous species remains the same as well. The Board also did not adequately update the aerial surveys with more recent data. The Board’s 2004 helicopter surveys were limited to black-tailed prairie dog colonies and bald eagle wintering and nesting sites. The Board also contends that it conducted site visits, including one in 2003. However, these site visits were conducted from the local public roads. The Board has acknowledged that much of the TRRC railroad line would traverse private property that is not near public roads. The Board fails to explain how its surveys from the public roads allowed it to adequately update the resource data when it admits that much of the prospective railroad land is not even accessible from public roads. Given the dearth of other data in this case, the Board fails to show that its reliance on several-years-old aerial photographs can support its conclusions. The Board’s decision, citing Mid States Coalition for Progress v. Surface Transportation Board, 345 F.3d 520 (8th Cir.2003), contends that environmental surveys often use aerial photography and aerial surveys. In Mid States Coalition, the Board used aerial photographs to identify noise-sensitive receptors — which was defined as any residence. 345 F.3d at 538. The objection raised was that a single building that could be identified via the aerial photograph could in fact contain multiple residences. Id. The Eighth Circuit held that the agency is not required to maximize precision at all costs and the method used was a reasonable way to approximate the number of affected receptors. Id. The use of the aerial photographs in Mid States Coalition stands in stark contrast with its use by the Board in TRRC III. Here, the Board purported to use aerial photographs and surveys to identify habitats and populations of fish, plants, and other wildlife. The Board fails to cite to any explanation in the record of what reliable methodology allowed it to determine the population of fish in rivers or identify sensitive plant species from these aerial surveys. We do not conclude that an agency cannot rely on aerial surveys in certain situations for studying habitats and populations. However, in this case, we find it inherently illogical to credit the evidence the Board has submitted to support its conclusion that the aerial surveys conducted were sufficient to establish the habitat and population for the numerous plants and wildlife potentially at risk. In summary, the Board relied on stale data during the environment impact analysis process of TRRC III and failed to properly update the data with additional studies and surveys. We hold that such faulty rebanee does not constitute the “hard look” required under NEPA. D. Geographic Scope Petitioners also argue that the Board improperly limited the geographic scope of the Board’s direct impacts analysis and focused only on the railroad’s right of way (ROW). “The task of selecting the geographic boundaries of an EIS requires a complicated analysis of several factors, such as the scope of the project considered, the features of the land, and the types of species in the area.” Selkirk, 836 F.3d at 958. The determination of the geographic boundary where environmental impacts may occur is often “a task assigned to the special competency of the appropriate agencies.” Kleppe v. Sierra Club, 427 U.S. 390, 414, 96 S.Ct. 2718, 49 L.Ed.2d 576 (1976). The Board did not arbitrarily limit the geographic scope to the railroad’s ROW. The impact analysis was limited to the area surrounding the ROW for the study of land use, noise and vibration, and cultural resources because these disturbances are limited and not expected to travel far. For each resource, the Board determined the scope of its impact analysis — for example, 300 meters on each side of the center of the ROW for soil surveys and 1,500 feet on each side of the center of the ROW for cultural resources. For other resources such as fish and wildlife, however, the Board did not limit the scope to the area surrounding the ROW. Some studies the Board reviewed covered a 6,000-square-mile area for mammals and birds and another covered a 7,500-square-mile area for birds. Further, for water and air quality analysis, the Board’s review was not limited to the ROW and covered much more of the neighboring area. We hold that the Board did not err in limiting the scope of the studies. E. Single EIS Petitioners also claim that the Board erred by not creating a single comprehensive EIS to cover the various TRRC applications. Petitioners contend that a single EIS is required because the TRRC applications were closely related, and under the regulations, should be discussed under a single impact statement. See 40 C.F.R. § 1508.25(a)(1) (“Connected actions ... should be discussed in the same impact statement.”). Actions are “connected” if they “[e]annot or will not proceed unless other actions are taken previously or simultaneously.” 40 C.F.R. § 1508.25(a)(l)(ii). We have explained that “[p]roposals or parts of proposals which are related to each other closely enough to be, in effect, a single course of action shall be evaluated in a single impact statement.” Klamath-Siskiyou Wildlands Ctr., 387 F.3d at 998 (citing 40 C.F.R. § 1502.4(a)). “The purpose of this requirement is to prevent an agency from dividing a project into multiple actions, each of which individually has an insignificant environmental impact, but which collectively have a substantial impact.” Great Basin Mine Watch v. Hankins, 456 F.3d 955, 969 (9th Cir.2006) (citation and quotation marks omitted). In determining whether there is a connection between projects, this circuit employs an “independent utility” test. Wetlands Action Network v. U.S. Army Corps of Eng’rs, 222 F.3d 1105, 1118 (9th Cir.2000), abrogated on other grounds by Wilderness Soc. v. U.S. Forest Serv., 630 F.3d 1173 (9th Cir.2011). The test asks whether “each of two projects would have taken place with or without the other.” Id. (citation omitted). If the answer is yes, then the projects have “independent utility” and do not require the same EIS. Id. “To prevail, plaintiffs must show that [the agency] was arbitrary and capricious in failing to prepare one comprehensive environmental statement.” Native Ecosystems Council v. Dombeck, 304 F.3d 886, 894 (9th Cir.2002) (citing Kleppe, 427 U.S. at 412, 96 S.Ct. 2718). Petitioners argue that all the projects lack independent utility. In particular, because TRRC II and TRRC III connect to TRRC I on one end, neither railroad would lead anywhere without TRRC I. Conversely, the Board argues that the projects should not be considered as one single project, because TRRC II and TRRC III were additions and modifications to the TRRC I project. The Board contends that it had no way of knowing, at the time of TRRC I, that TRRC II or TRRC III would be later proposed. Thus, the Board contends that it did what was required of it, and supplemented the EISs for TRRC II and TRRC III to consider the impact of the entire railroad line. Both parties make thought-provoking arguments. However, the timing of the TRRC applications precluded them from being filed as a single EIS. The railroad approved in TRRC I had independent utility in and of itself. The TRRC II and TRRC III applications do not have much utility outside of TRRC I. Nevertheless, when the Board approved TRRC II and TRRC III, it did incorporate the findings of the previous EISs and looked at the total environmental impact of the entire 130-mile railroad line even though it did not prepare one single comprehensive EIS. While we hold that the Board’s EIS preparation is arbitrary and capricious in other ways, see supra Section I.A-C, we do not find that its incorporation of the previous EISs is per se in error for reasons already discussed. In other words, we believe that the Board’s method of incorporating the EISs — which considered the entire railroad line — may have been appropriate had it cured the other deficiencies cited supra in Section I.A-C. F. Tiering Petitioners’ final argument is that the Board violated NEPA when it tiered its EIS to five other site-specific EISs. Tiering is the process of incorporating by reference coverage of general matters in broader environmental impact statements, such as national program or policy statements, into subsequent narrower environmental analysis, such as site-specific statements. 40 C.F.R. § 1508.28. Tiering under NEPA is permitted in two circumstances: (1) “a program, plan, or policy impact statement” may be tiered to “a program, plan, or policy statement or analysis of lesser scope or to a site-specific statement or analysis,” and (2) “an environmental impact statement on a specific action at an early stage” may be tiered “to a supplement ... or a subsequent statement or analysis at a later stage.” Id. § 1508.28(a), (b). Only documents that have undergone NEPA analysis may be tiered. Kern, 284 F.3d at 1073. The Board contends that it did not impermissibly tier any other site-specific EISs but relied on the EISs only for general background information. We agree. In the TRRC III FSEIS, the Board stated: Additionally, the Tongue River region has been studied extensively, not only by SEA in conducting the environmental analysis for Tongue River I and Tongue River II, but also by BLM and MT DNRC in the preparation of EISs for Powder River I, Monteo Mine, CX Ranch, and the Tongue River Reservoir Dam reconstruction, as well as the analysis of coal-bed-methane-production wells. Where appropriate, SEA has relied on these other environmental analyses. The Powder River I, Monteo Mine, CX Ranch, Tongue River Reservoir Dam, and coal-bed-methane production wells EISs are site-specific EISs that do not fall into either situation where tiering is permitted. However, Petitioners fail to explain what aspect of the TRRC environmental analysis directly relies on the incorporation of these other EISs. Thus, we reject Petitioners’ contention that the Board engaged in illegal tiering. To summarize our holdings in Section I, we reverse and remand on (1) the Board’s cumulative impact analysis in TRRC III as to the reasonably foreseeable coal bed methane projects, the Otter Creek Coal Mine, and water quality analysis; (2) the adequacy of the baseline data in TRRC III as to the pallid sturgeon, sage grouse, fish and aquatic life, other wildlife, and sensitive plants; and (3) the Board’s reliance on stale data, consistent with the analysis outlined above in Section I. We affirm the Board as to Petitioners’ other environmental claims. II. Railroad Claims A. The 49 U.S.C. § 10901 Standard Petitioners contend that the Board did not apply the correct statutory standard when it granted permission to construct the rail lines in TRRC II and TRRC III. We disagree. 49 U.S.C. § 10901 governs when the Board may authorize the construction and operation of railroad lines. As the statute existed prior to 1980, the Board could authorize construction of a rail line only if it found that “the present or future public convenience and necessity require or will be enhanced by the construction or acquisition (or both) and operation of the railroad line.” 49 U.S.C. § 10901 (1976) (emphasis added). The Staggers Act of 1980 changed the language to authorize construction of a rail line if the Board finds that “the present or future public convenience and necessity require or permit the construction or acquisition (or both) and operation of the railroad fine.” 49 U.S.C. § 1090