Save Our Sound OBX, Inc. v. NC Dept of Transportation

PUBLISHED UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT No. 18-1649 SAVE OUR SOUND OBX, INC.; MARK HAINES; JER MEHTA; GLENN STEVENS; DAVID HADLEY; THOMAS ASCHMONEIT; RICHARD AYELLA, Plaintiffs – Appellants, v. NORTH CAROLINA DEPARTMENT OF TRANSPORTATION; FEDERAL HIGHWAY ADMINISTRATION; JOHN F. SULLIVAN, III, in his official capacity as Division Administrator for the Federal Highway Administration; JAMES H. TROGDON, III, in his official capacity as Secretary of the North Carolina Department of Transportation, Defendants – Appellees, and DEFENDERS OF WILDLIFE; NATIONAL WILDLIFE REFUGE ASSOCIATION, Intervenors/Defendants – Appellees. Appeal from the United States District Court for the Eastern District of North Carolina, at Elizabeth City. Louise W. Flanagan, District Judge. (2:17-cv-00004-FL) Argued: December 11, 2018 Decided: January 23, 2019 Before NIEMEYER, DUNCAN, and QUATTLEBAUM, Circuit Judges. Affirmed by published opinion. Judge Duncan wrote the opinion, in which Judge Niemeyer and Judge Quattlebaum joined. ARGUED: David Ari Schnitzer, GIBSON, DUNN & CRUTCHER LLP, Washington, D.C., for Appellants. Thekla Hansen-Young, UNITED STATES DEPARTMENT OF JUSTICE, Washington, D.C.; Colin Justice, NORTH CAROLINA DEPARTMENT OF JUSTICE, Raleigh, North Carolina, for Appellees. ON BRIEF: Michael K. Murphy, Kyle N. Guest, GIBSON, DUNN & CRUTCHER LLP, Washington, D.C., for Appellants. Jeffrey H. Wood, Acting Assistant Attorney General, Eric Grant, Deputy Assistant Attorney General, Andrew C. Mergen, Robert J. Lundman, Carter F. Thurman, Appellate Section, Environment and Natural Resources Division, UNITED STATES DEPARTMENT OF JUSTICE, Washington, D.C.; Josh Stein, Attorney General, OFFICE OF THE ATTORNEY GENERAL OF NORTH CAROLINA, Raleigh, North Carolina; Scott T. Slusser, Special Deputy Attorney General, Mollie Cozart, Assistant Attorney General, NORTH CAROLINA DEPARTMENT OF JUSTICE, Raleigh, North Carolina, for Appellees North Carolina Department of Transportation, Federal Highway Administration, John F. Sullivan, III, and James H. Trogdon, III. Kimberley Hunter, Nicholas S. Torrey, SOUTHERN ENVIRONMENTAL LAW CENTER, Chapel Hill, North Carolina, for Appellees Defenders of Wildlife and National Wildlife Refuge Association. 2 DUNCAN, Circuit Judge: Appellants Save Our Sound OBX, Inc. and its members, residents and vacationers from North Carolina’s Outer Banks, (collectively “SOS”) challenge the decision of the North Carolina Department of Transportation (the “NCDOT”), the Federal Highway Administration (the “FHWA”), and their administrators (collectively “the Agencies”) to replace a segment of North Carolina Highway 12 (“NC-12”) with a bridge across the Pamlico Sound (the “Jug-Handle Bridge”). The district court granted the Agencies’ motion for summary judgment, finding that they did not violate the National Environmental Policy Act (the “NEPA”), 42 U.S.C. § 4321 et seq., or the Department of Transportation Act (the “DTA”), 49 U.S.C. § 301 et seq., when they approved the bridge. SOS challenges that ruling on appeal. For the reasons that follow, we affirm. I. This case involves a segment of NC-12, which is the main roadway passing through the Outer Banks of North Carolina. State and federal agencies have been working for several years to update and improve NC-12 because of its susceptibility to weather damage and erosion. Like many highway construction projects, the NC-12 project required cooperation among state and federal agencies. For instance, in North Carolina, NCDOT has authority over highway construction, while FWHA supplies federal funds for highway projects. The agencies tasked with improving NC-12, including NCDOT and FHWA, formed a Merger Team to coordinate decisionmaking and regulatory compliance for the NC-12 3 project. 1 The Merger Team was responsible for ensuring that the NC-12 project complied with the requirements of NEPA and the DTA, among other regulations. SOS challenges certain agency decisions in the NC-12 project under NEPA and the DTA. Accordingly, we first provide a brief overview of the requirements of those statutes before turning to the specific facts and procedural history of this appeal. A. The first statute at issue in this appeal is NEPA. Pursuant to NEPA, 42 U.S.C. § 4321 et seq., and its implementing regulations, government agencies considering certain projects must evaluate whether the project would have a significant impact on the environment by preparing an Environmental Assessment (an “EA”). Id. § 4332(C); 40 C.F.R. § 1508.9; see id. § 1508.18 (defining the types of federal actions to which NEPA applies). If the project would have a significant impact, the agency must prepare an Environmental Impact Statement (an “EIS”). 42 U.S.C. § 4332(C). The agency is responsible for ensuring that the EIS complies with various regulatory requirements. See 40 C.F.R. § 1502.1 et seq. For instance, the EIS must “provide full and fair discussion” of any significant environmental impacts of a proposed action. 40 C.F.R. § 1502.1. 1 The Merger Team was comprised of representatives from NCDOT, FHWA, the U.S. Fish and Wildlife Service, the U.S. Army Corps of Engineers, the U.S. Environmental Protection Agency, the National Marine Fisheries Service, the National Park Service, the North Carolina Department of Cultural Resources, the North Carolina Wildlife Resources Commission, and three divisions of the North Carolina Department of Environmental and Natural Resources (the “NCDENR”) (now named the North Carolina Department of Environmental Quality)--the Division of Water Resources, the Division of Coastal Management, and the Division of Marine Fisheries. 4 Additionally, the agency must “[r]igorously explore and objectively evaluate all reasonable alternatives” that could “avoid or minimize adverse impacts.” Id. §§ 1502.1, 1502.14. The agency is permitted, however, to identify a preference among alternatives based on non-environmental considerations, such as economic factors or the agency’s statutory mission. Id. § 1505.2(b). If the agency has a preferred alternative, NEPA requires the agency to identify that preference in the EIS. Id. § 1502.14. NEPA also requires the agency to prepare a supplemental EIS (an “SEIS”) if significant new information or environmental changes come to light after the agency prepares an EIS. Id. § 1502.9(c). After the agency makes its final decision about which alternative to pursue, it publishes a record of decision (an “ROD”) explaining its choice. Id. § 1505.2. We now turn to the second statute at issue: the DTA. The DTA contains substantive requirements for government transportation projects. 49 U.S.C. § 301 et seq. Relevant here, the so-called “§ 4(f)” requirements 2 concern transportation projects that require the use of publicly owned land of a wildlife refuge or a significant historic site. Id. § 303(c). Historic sites include structures “included in, or eligible for inclusion in,” the National Register of Historic Places. 23 C.F.R. § 774.17. The Secretary of Transportation may only approve such projects if there is no “feasible and prudent” alternative to using that land and the “project includes all possible planning to minimize harm . . . resulting from the use.” Id. § 303(c). If there is no feasible and prudent 2 The DTA explains that “[t]he requirements of [49 U.S.C. § 303] are commonly referred to as section 4(f) requirements” in reference to a previous version of the statute. 49 U.S.C. § 303(f)(1). 5 alternative, the Secretary may only approve the alternative that “[c]auses the least overall harm in light of the [DTA’s] preservation purpose.” 23 C.F.R. § 774.3(c). Relevant considerations in selecting the least harmful alternative include whether harm to the land can be mitigated, whether harm to the land affects the attributes qualifying that land for protection, and whether the alternative meets the needs of the project. Id. B. Having established the relevant framework, we consider the facts of this case. SOS’s claims in this litigation concern the Agencies’ environmental analysis under NEPA and the DTA with respect to a segment of NC-12 that passes from the southern edge of Bodie Island to the village of Rodanthe. For this segment, the Merger Team was responsible for preparing EAs and EISs pursuant to NEPA and for determining which proposed plan for the project was the least environmentally-damaging practicable alternative (the “LEDPA”) pursuant to section 404 of the Clean Water Act, 33 U.S.C. § 1344, among other requirements. 3 See 40 C.F.R. § 230.10 (setting out the LEDPA requirement). In 2008, the Merger Team issued an EIS and § 4(f) evaluation (the “2008 EIS”) for improving NC-12. For the segment at issue here, the 2008 EIS included discussion of several alternatives, including a proposed bridge in the Pamlico Sound near Rodanthe 3 Section 404 of the Clean Water Act regulates activities involving “the discharge of dredged or fill material” into navigable waters. Id. 6 (the “Bridge South alternative”) and proposals involving beach nourishment. An EA in 2010 (the “2010 EA”) further developed these alternatives. The Merger Team released an updated EA in 2013 (the “2013 EA”) to account for environmental changes after the 2010 EA, including the effects of Hurricane Irene in 2011. The 2013 EA identified four alternatives for the segment at issue: (1) the so-called Jug-Handle Bridge, a bridge extending out into the Pamlico Sound (also referred to in the environmental analyses as the “Bridge on New Location”); 4 (2) an easement bridge on the existing NC-12 location; (3) beach nourishment; and (4) an easement bridge combined with beach nourishment. The Merger Team did not study the beach nourishment alternatives in depth in the 2013 EA because, at a 2011 meeting, it had already determined not to pursue them after experts reported on a “high erosion rate and a lack of sand supply.” J.A. 843. In the 2013 EA, the Merger Team identified the easement bridge as its preferred alternative under NEPA. However, certain members of the Merger Team objected to finding that the easement bridge was the LEDPA under the Clean Water Act, citing concerns about its location within the surf zone, additional permits associated with erosion setback requirements, and its impact on a nearby wildlife refuge. See J.A. 1745–47. In the meantime, environmental groups Defenders of Wildlife and the National Wildlife Refuge Association (intervenors here, collectively the “Environmental Groups”) brought suit in federal court to challenge the Agencies’ NEPA and § 4(f) determinations 4 The Jug-Handle Bridge was a modified version of the Bridge South alternative that was introduced in the 2008 EIS and further developed in the 2010 EA. 7 with respect to a different segment of NC-12 in the Outer Banks--the Bonner Bridge, which connects Bodie Island and Hatteras Island to the north of Rodanthe. See Defs. of Wildlife v. N.C. Dep’t of Transp., 762 F.3d 374 (4th Cir. 2014). The Environmental Groups and the Agencies eventually reached an agreement in 2015 (the “Settlement”). 5 The Settlement required NCDOT to identify the Jug-Handle Bridge as its preferred alternative for the segment of NC-12 at issue in this case. It also required NCDOT to seek Merger Team concurrence that the Jug-Handle Bridge was the LEDPA. In doing so, it stated that nothing in the Settlement “requires or should be interpreted to predetermine the choice” of the Jug-Handle Bridge as the final selected alternative. J.A. 1090. In exchange, the Environmental Groups dismissed the Bonner Bridge suit and agreed not to challenge the Agencies in court if the Jug-Handle Bridge was determined to be the LEDPA and was ultimately selected in the ROD for this project. After a 2015 meeting, the Agencies identified the Jug-Handle Bridge as their preferred alternative. Following public comment and detailed studies, the Merger Team determined that the Jug-Handle Bridge was the LEDPA. In 2016, the Merger Team released a revised EA (the “2016 EA”) to evaluate the environmental impacts of the Jug- Handle Bridge and its associated construction activities. Later that year, the Merger Team issued an ROD (the “2016 ROD”) formally approving the Jug-Handle Bridge. In addition to authorizing construction of the Jug-Handle Bridge, the 2016 ROD also addressed concerns relating to a shipwreck in the proposed bridge’s path known as 5 NCDENR was also a party to the Settlement and is represented on the Merger Team, though it is not a defendant here. 8 the Pappy’s Lane Wreck. The shipwreck is eligible for listing on the National Register of Historic Places. Because of the shipwreck’s historical significance, the 2016 ROD ordered a data recovery project on the shipwreck, which later uncovered evidence that the shipwreck was a World War II assault vessel. The Merger Team has not yet determined how it will respond to this new information. C. We now turn to the procedural history of this litigation, which began when SOS sued the Agencies on February 2, 2017. A month later, the Environmental Groups intervened in the suit in support of the Agencies. In its complaint, SOS alleged, as relevant here, that the Agencies’ approval of the Jug-Handle Bridge violated NEPA because that decision was predetermined by the Settlement. To show evidence of predetermination, SOS moved to supplement the administrative record before the district court to include documents related to the negotiation of the Settlement on the grounds that those documents were before the Agencies when they made their decision to select the Jug-Handle Bridge. The district court denied this motion. Later, SOS moved to amend its complaint to add claims related to construction impacts in Rodanthe and to the Pappy’s Lane Wreck. Specifically, SOS contended that the Agencies’ environmental analyses did not adequately consider the effects of construction or the historic significance of the Pappy’s Lane Wreck. The district court granted the motion as to construction impacts but denied it as to the shipwreck. 9 SOS, the Agencies, and the Environmental Groups each filed cross-motions for summary judgment on the issue of whether the Agencies’ environmental analyses violated NEPA or the DTA. The district court granted the Agencies’ and Environmental Groups’ motions for summary judgment and denied SOS’s motion. SOS appeals the district court’s grant of the Agencies’ motion for summary judgment and its denial of SOS’s motion to amend its complaint. SOS contends that the Agencies were not entitled to summary judgment because the Agencies’ environmental analyses violated NEPA. It also contends that the district court erred in denying SOS’s motion to amend its complaint with claims related to the Pappy’s Lane Wreck. We discuss SOS’s NEPA claims and its motion to amend its complaint, in that order. II. SOS contends that the Agencies violated NEPA when they selected the Jug- Handle Bridge for the NC-12 project. “We review a grant of summary judgment de novo,” Defs. of Wildlife, 762 F.3d at 392, and, accordingly, review the Agencies’ actions directly, pursuant to the Administrative Procedure Act (the “APA”). The APA directs us to overturn agency actions if they are “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” 5 U.S.C. § 706(2)(A). Our precedent also guides the scope of our review of the sufficiency of an agency’s NEPA analysis. Nat’l Audubon Soc’y v. Dep’t of Navy, 422 F.3d 174, 185, 199 (4th Cir. 2005). In conducting this review, “we must make a searching and careful inquiry into the facts” and examine whether the decision was based on consideration of the relevant factors. Id. at 185 10 (citation and internal quotation marks omitted). However, we do not “second-guess agency decisions, so long as the agency has given a hard look at the environmental impacts of its proposed action.” Id. at 199. Under this standard of review, we affirm because the Agencies did not violate NEPA in approving the Jug-Handle Bridge. We consider SOS’s arguments that the Agencies’ environmental analyses violated NEPA because (1) the Agencies failed to prepare an SEIS with regards to the Jug-Handle Bridge and beach nourishment alternatives before issuing the 2016 ROD, (2) the Agencies failed to adequately consider the impacts of construction, and (3) the Settlement impermissibly predetermined the Agencies’ choice of the Jug-Handle Bridge. A. First, SOS contends that the Agencies’ NEPA analysis was deficient because the Agencies should have prepared an SEIS before approving the Jug-Handle Bridge in the 2016 ROD. NEPA requires an agency to prepare an SEIS if “[t]he agency makes substantial changes” to a proposed action that implicate environmental concerns or if “[t]here are significant new circumstances or information” that would affect the environmental impacts of the proposed action. 40 C.F.R. § 1502.9(c). To merit an SEIS, the changes “must present a seriously different picture of the environmental impact of the proposed project.” Hughes River Watershed Conservancy v. Glickman, 81 F.3d 437, 443 (4th Cir. 1996) (citation and alterations omitted). 11 We review an agency’s decision not to prepare an SEIS in two steps, consistent with our standard of review under NEPA. Id. At step one, we “determine whether the agency took a hard look at the proffered new information.” Id. If the agency concludes after a preliminary inquiry that the “environmental effect of the change is clearly insignificant,” its decision not to prepare an SEIS satisfies the hard look requirement. Hodges v. Abraham, 300 F.3d 432, 446 (4th Cir. 2002). For instance, in Hodges, we held that an agency took a hard look at the environmental consequences of its change in plans when it explicitly evaluated those consequences and determined that they were not significantly different from those discussed in a prior EIS. Id. at 447. We also noted that the plaintiff in that case “failed to identify any particular risk arising from” the change that had not already been addressed in the agency’s environmental analyses. Id. Next, at step two of the SEIS inquiry, we review whether the agency’s decision not to prepare an SEIS after taking a hard look was arbitrary or capricious. Hughes River, 81 F.3d at 443. SOS contends that an SEIS was necessary (1) to evaluate the environmental effects of the Jug-Handle Bridge alternative because it was different from options that had previously been evaluated and (2) to reconsider alternatives that involved beach nourishment pursuant to new information about sand availability and beach erosion rates. We address each contention in turn and affirm the district court’s determination that no SEIS was required. Neither of these changes are sufficiently different from the circumstances initially evaluated in the EIS to merit an SEIS. The Agencies’ decision not to prepare one after taking a hard look at the changes was therefore not arbitrary or capricious. 12 i. Contrary to SOS’s contention, the Agencies were not required to prepare an SEIS with respect to the Jug-Handle Bridge alternative. SOS argues that the final alignment of the Jug-Handle Bridge alternative (that is, its path across the Pamlico Sound and its connection points with the shore) was significantly different from previously evaluated alternatives. Specifically, SOS contends that the bridge’s alignment changed significantly because the alternative evaluated in the 2008 EIS was “amorphously defined” and because the alternative evaluated in the 2010 EA (to which the 2013 EA, 2016 EA, and 2016 ROD refer) was “not comparable” to the Jug-Handle Bridge. Appellant’s Br. at 35, 36. It argues that these differences render prior environmental analyses of those alternatives insufficient to evaluate the environmental effects of the Jug-Handle Bridge as approved in the 2016 ROD. We disagree and conclude that the Agencies complied with NEPA when they decided that an SEIS was not warranted by these changes. At step one, the Agencies took a hard look at changes in the bridge’s alignment in the 2016 EA. The 2016 EA describes the similarities and differences between the Jug- Handle Bridge as proposed in 2016 and the versions evaluated in the 2013 EA, the 2010 EA, and the 2008 EIS. For instance, it explains that the Agencies decided to shift the alignment of the bridge to avoid areas of “dense submerged aquatic vegetation” that fell in the path of previous versions of the bridge. J.A. 1198. The 2016 EA also explains that the Jug-Handle Bridge’s alignment reduces effects on the community as compared with 13 previous versions because it requires a narrower right-of-way. Because the Agencies went into detail in their comparison between the Jug-Handle Bridge and previous versions of the bridge, their coverage satisfies the hard look requirement. See Hodges, 300 F.3d at 446–47. And, at step two, because the Agencies took the requisite hard look and neither their environmental analyses nor SOS identify any particular differences that would merit an SEIS, their decision not to prepare an SEIS was not arbitrary or capricious. See Hughes River, 81 F.3d at 443. An SEIS is only required when changes to a project present a “seriously different picture of the environmental impact.” Id. (alterations and citation omitted). Based on the discussion presented in the 2016 EA, which compared the different versions of the bridge, it was not arbitrary or capricious for the Agencies to determine that the Jug-Handle Bridge’s final alignment did not present a “seriously different picture,” id., than the 2008 EIS, the 2010 EA, or the 2013 EA. The Agencies explained how the Jug-Handle Bridge was different from previous versions of the bridge, and their explanations do not implicate any significant environmental concerns. Accordingly, their decision not to prepare an SEIS was not arbitrary or capricious. ii. Similarly, the Agencies were not required to prepare an SEIS to reevaluate the feasibility of alternatives that involved beach nourishment. SOS contends that the Agencies failed to adequately reconsider beach nourishment after new erosion projections were released and after a 2014 emergency beach nourishment project in the area 14 successfully repaired damage from Hurricane Sandy. According to SOS, this new information showed that coastal conditions had changed such that erosion would no longer threaten beach nourishment and that adequate sand was in fact available to complete the project. However, the Agencies took the requisite hard look at these new circumstances in the 2016 EA. See Hughes River, 81 F.3d at 443. The 2016 EA notes that while erosion projections in the Rodanthe area did decrease, erosion rates in Rodanthe remained “amongst the highest rates along the North Carolina coast.” J.A. 1218. It also discusses the 2014 emergency beach nourishment project in detail and considers updated information about coastal conditions in the area. Specifically, it notes that the 2014 emergency beach nourishment project was “essentially one round of nourishment in one part of [NC-12].” J.A. 1227. Finally, it concludes that the Agencies thoroughly considered the environmental effects of beach nourishment when they originally evaluated those alternatives in the 2008 EIS. This detailed discussion satisfies the hard look requirement. See Hodges, 300 F.3d at 446–47. And after taking that hard look, the Agencies’ decision not to prepare an SEIS was not arbitrary or capricious. See Hughes River, 81 F.3d at 443. SOS cites improved erosion rates and an increased supply of sand as new factors that the Agencies should have considered when deciding whether to prepare an SEIS. However, erosion and sand supply were not the Agencies’ only reasons for initially rejecting beach nourishment in the 2008 EIS. The Agencies also cited independent concerns such as inadequate protection against future breaches, risks of overwash, and incompatibility with the 15 mission of a neighboring wildlife refuge. It was not arbitrary or capricious for the Agencies to decline to reconsider beach nourishment alternatives in an SEIS when the new information proffered by SOS did not implicate all of the Agencies’ independently adequate reasons for initially rejecting beach nourishment. Accordingly, the Agencies were not required to prepare an SEIS to consider the alignment of the Jug-Handle Bridge or to consider beach nourishment alternatives. Their environmental analyses therefore do not violate NEPA in this respect. B. SOS also contends that the Agencies’ environmental analyses violate NEPA because they do not adequately consider the environmental effects of construction in the Rodanthe area. Specifically, SOS argues that the effects of construction traffic and haul roads are not adequately addressed in the 2016 EA and that any discussion of these issues in the 2008 EIS is irrelevant because that document focuses on a larger area. NEPA requires agencies to consider all “significant environmental impacts,” 40 C.F.R. § 1502.1, which can include impacts caused by the use of construction roads and increased traffic. See, e.g., Arlington Coal. on Transp. v. Volpe, 458 F.2d 1323, 1332 (4th Cir. 1972) (explaining that NEPA applies to highway construction projects). Accordingly, if construction roads and traffic are expected to affect the environment near an agency project, the agency must provide “full and fair discussion” of those effects and incorporate them into its comparison of alternatives. 40 C.F.R. § 1502.1. 16 Here, the Agencies adequately considered the effects of construction traffic as a result of the Jug-Handle Bridge in the 2016 ROD. Specifically, the Agencies explained that a temporary construction easement would be necessary to support traffic during construction but noted that “[t]he land affected will be restored after construction is completed.” J.A. 1559. While the 2016 ROD does not compare the construction traffic effects of the Jug-Handle Bridge with those of other proposed alternatives, the Agencies did perform that comparison in the 2013 EA and the 2016 EA and still found the bridge to be the LEDPA. The Agencies also specifically considered the environmental effects of haul roads in the 2008 EIS. 6 We therefore cannot say that the Agencies did not provide “full and fair discussion” of the effects of construction traffic or haul roads when they selected the Jug- Handle Bridge among alternatives. See 40 C.F.R. § 1502.1. Accordingly, the Agencies did not violate NEPA here, and we affirm the district court. 6 To be sure, that discussion did not explicitly consider haul roads in the Rodanthe area. However, it is unclear from the record whether the Agencies actually plan to use haul roads for the Jug-Handle Bridge project. See, e.g., J.A. 214 (containing a declaration from an NCDOT engineer stating that NCDOT and its contractor decided not to pursue a previously-discussed haul road in connection with a barge because shallow water depths would make it impracticable). Regardless, NEPA does not compel the Agencies to specifically consider the environmental impacts of haul roads so long as they adequately examine the environmental consequences of the Jug-Handle Bridge project, including construction impacts, as a whole. See 40 C.F.R. § 1502.2(b) (“Impacts shall be discussed [in the EIS] in proportion to their significance.”). 17 C. Finally, SOS contends that the Agencies violated NEPA because their choice of the Jug-Handle Bridge did not follow from their NEPA analysis but, rather, was a predetermined result of the Settlement. NEPA “prohibits agencies from preparing an EIS simply to ‘justify [] decisions already made.’” Nat’l Audubon Soc’y, 422 F.3d at 199 (quoting 40 C.F.R. § 1502.2(g)) (alteration in original). That is because, under NEPA, agencies that have yet to issue a final decision may not “commit resources prejudicing selection of alternatives,” 40 C.F.R. § 1502.2(f), nor may they take any action that would “[l]imit the choice of reasonable alternatives,” 40 C.F.R. § 1506.1(a). An EIS based upon a predetermined choice that a certain alternative will be selected would violate these principles because it would not allow the agency to fully consider all alternatives. We conclude that the Agencies’ choice of the Jug-Handle Bridge was not impermissibly predetermined. We consider, in turn, the Agencies’ objective environmental analyses, the language of the Settlement, and the role of documents generated during negotiation of the Settlement. i. First, SOS contends that flaws in the Agencies’ environmental analyses reveal that their choice of the Jug-Handle Bridge was predetermined by the Settlement. In alleging that the environmental analyses were inadequate, it primarily relies upon arguments that we have already discussed--namely, that the Agencies failed to fully consider the new alignment of the bridge, the feasibility of beach nourishment alternatives, and the effects 18 of construction traffic and haul roads. SOS contends that these flaws exist because the Agencies prematurely decided to choose the Jug-Handle Bridge after the Settlement. Under our precedent, “the evidence we look to in determining whether [predetermination] has taken place consists of the environmental analysis itself.” 7 Nat’l Audubon Soc’y, 422 F.3d at 199. In National Audubon Society, in reviewing allegations that an agency had settled on a certain outcome before preparing its EIS, we declined to look to the agency’s internal documents and emails. Id. Instead, we limited the scope of our inquiry to the agency’s objective environmental analysis, reasoning that “[w]here an agency has merely engaged in post hoc rationalization, there will be evidence of this in its failure to comprehensively investigate the environmental impact of its actions and acknowledge their consequences.” Id. Accordingly, our analysis focuses on whether an agency’s objective environmental analyses demonstrate evidence of predetermination. Here, when we examine Agencies’ environmental analyses, we cannot say that they violated NEPA by selecting the Jug-Handle Bridge following the Settlement. To be sure, the Agencies changed their preferred alternative from the easement bridge to the Jug-Handle Bridge following the Settlement. But that change alone does not mean that the Agencies’ choice was predetermined, particularly where members of the Merger Team had expressed concerns about the easement bridge as far back as 2013. And when we look to the Agencies’ environmental analyses here, those analyses satisfied NEPA’s 7 We note that some circuits look beyond the objective environmental analysis for evidence of predetermination. See, e.g., Forest Guardians v. U.S. Fish & Wildlife Serv., 611 F.3d 692, 716–17 (10th Cir. 2010) (declining to adopt the Fourth Circuit’s rule). 19 requirements. See discussion supra Sections II.A., II.B. We do not see, and SOS cannot identify, any evidence that the Agencies “fail[ed] to comprehensively investigate” and evaluate the environmental impacts of the Jug-Handle Bridge or the other alternatives. Nat’l Audubon Soc’y, 422 F.3d at 199. The EAs and EIS prepared by the Agencies do not, on their faces, show evidence of predetermination. SOS cites to cases from the Ninth and Tenth Circuits to contend that courts have found predetermination based on contractual commitments made by an agency before the NEPA process was complete. See Davis v. Mineta, 302 F.3d 1104, 1112 (10th Cir. 2002) (holding that a contract requiring a consultant to recommend that a project had no significant environmental impact before actually preparing an EA violated NEPA); Metcalf v. Daley, 214 F.3d 1135, 1145 (9th Cir. 2000) (same, regarding a contract requiring an agency to make a proposal for a specific alternative before preparing an EA). However, these cases are distinguishable because they involved contracts that agencies entered into before conducting any environmental analysis at all; in contrast, the Settlement here, which the Agencies and the Environmental Groups entered into in 2015, was preceded by a number of environmental analyses, including the 2008 EIS, the 2010 EA, and the 2013 EA. And when we examine each of these environmental analyses, we find no evidence of predetermination warranting reversal. ii. Although circuit precedent limits our predetermination inquiry to the “environmental analysis itself,” id., SOS nonetheless urges us to consider the Settlement 20 as evidence that the Agencies’ choice was predetermined. But even if we look to the Settlement, that document does not support SOS’s claim. The Settlement only required NCDOT to identify the Jug-Handle Bridge as its preferred alternative and to seek Merger Team concurrence that the Jug-Handle Bridge was the LEDPA. These conditions do not require the Agencies to select the Jug-Handle Bridge as the final approved alternative for this project. For instance, it remained possible that the Agencies’ environmental analyses would demonstrate that the Jug-Handle Bridge was not the LEDPA. Additionally, the Agencies’ preferences alone could not bind the entire Merger Team, which was ultimately responsible for approving the final alternative, because the parties to the Settlement comprise only three of the ten state and federal agencies represented on the Merger Team. 8 This does not constitute predetermination. iii. SOS also urges us to consider external documents generated during the negotiation of the Settlement as evidence of predetermination and to reverse the district court’s denial of SOS’s motion to include these documents in the record. We decline to do so because the district court did not abuse its discretion in denying this motion. See Fort Sumter Tours, Inc. v. Babbitt, 66 F.3d 1324, 1335–36 (4th Cir. 1995) (explaining that we review 8 To be sure, the parties to the Settlement represented three of the four Lead Agencies tasked with decisionmaking if the Merger Team could not reach consensus. But, as discussed, the Settlement only required those agencies to seek concurrence that the Jug-Handle Bridge was the LEDPA; it did not require them to go beyond the objective environmental analysis to choose the Jug-Handle Bridge if it was not, in fact, the LEDPA. 21 a district court’s decision whether to expand the administrative record for abuse of discretion). Federal courts base their review of agency decisions on “the ‘whole record’ compiled by the agency.” Citizens to Pres. Overton Park, Inc. v. Volpe, 401 U.S. 402, 419 (1971) (quoting 5 U.S.C. § 706), abrogated on other grounds by Califano v. Sanders, 430 U.S. 99 (1977). We may supplement the record as presented by the agency if the “bare record” does not reveal the agency’s reasoning or if it appears that the agency acted in bad faith. Id. at 420. Otherwise, “inquiry into the mental processes of administrative decisionmakers is usually to be avoided.” Id. Here, the record as presented by the Agencies reveals their reasoning, as we have discussed, and the district court found no evidence that the Agencies acted in bad faith. The district court also determined that there was no evidence that the Agencies actually reviewed the documents put forth by SOS as part of their decision to approve the Jug- Handle Bridge. We hold that the district court did not abuse its discretion by denying SOS’s motion to supplement the record, and accordingly, our review is limited to the record as compiled by the Agencies. SOS contends that, even if we cannot consider these documents as part of the administrative record, we should nevertheless consider them as extra-record evidence. It is true that “in the NEPA context, courts generally have been willing to look outside the record when assessing the adequacy of an EIS or a determination that no EIS is necessary.” Ohio Valley Envtl. Coal. v. Aracoma Coal Co., 556 F.3d 177, 201 (4th Cir. 2009). However, we have explained that extra-record evidence in that context is 22 primarily useful “to inform the court about environmental factors that the agency may not have considered.” Id. (emphasis added). Here, the proffered documents would do nothing to aid the court’s understanding of the environmental concerns at issue, as they concern the legal terms of the Settlement between the Agencies and the Environmental Groups. Therefore, the district court properly excluded them, and we decline to consider them now. There is superficial appeal in suggesting that the Agencies’ choice was predetermined because the Settlement yielded a preferred alternative. However, that conclusion is not legally or factually supported by the record. Legally, our analysis is cabined to the environmental analysis itself, with which we find no issues meriting reversal. And factually, we see no evidence that the Merger Team was forced to approve the Jug-Handle Bridge. In sum, the Agencies did not violate NEPA when they approved the Jug-Handle Bridge because they were not required to prepare an SEIS, they adequately evaluated the effects of construction traffic, and they did not rely on a predetermined choice among alternatives. We therefore affirm the district court’s grant of summary judgment in favor of the Agencies. III. Finally, SOS contends that the district court erred when it denied SOS’s motion to amend its complaint with claims related to the Pappy’s Lane Wreck. Specifically, SOS alleges that the Agencies violated § 4(f) by approving a transportation project that 23 threatens to harm the land of a site of historic significance because the Jug-Handle Bridge will pass above the wreck--the site of a World War II vessel. We review a district court’s denial of a motion to amend the complaint for abuse of discretion. See Laber v. Harvey, 438 F.3d 404, 428 (4th Cir. 2006) (en banc). However, if the district court denied such a motion on the grounds that the amendment would have been futile, we review its legal conclusions de novo. United States ex. rel. Ahumada v. NISH, 756 F.3d 268, 274 (4th Cir. 2014). A party “may amend its pleading once as a matter of course” before the opposing party files a responsive pleading. Fed. R. Civ. P. 15(a)(1). After that time, a party may only amend its pleading with the consent of the opposing party or with leave of the court, and “[t]he court should freely give leave when justice so requires.” Fed. R. Civ. P. 15(a)(2). Courts may deny leave to amend a pleading if the amendment would have been futile. Laber, 438 F.3d at 426. A proposed amendment is futile when it is “clearly insufficient or frivolous on its face.” Johnson v. Oroweat Foods Co., 785 F.2d 503, 510 (4th Cir. 1986). A proposed amendment is also futile if the claim it presents would not survive a motion to dismiss. Perkins v. United States, 55 F.3d 910, 917 (4th Cir. 1995). Here, the district court found that SOS’s proposed amendment would be futile for two alternative reasons. First, it determined that, to the extent that SOS sought to assert that the 2016 EA was no longer valid in light of new information about the shipwreck, that claim would be unripe where the Agencies had not yet made a decision concerning the significance of the new information. Second, it determined that, to the extent that SOS sought to assert that the 2016 EA was inadequate when issued, that claim would fail 24 because it would ask the Agencies to consider information that was not discovered until after the 2016 EA was issued. For the reasons that follow, we affirm the district court’s denial of SOS’s motion to amend its complaint. First, to the extent that SOS’s proposed amendments contend that the Agencies failed to consider new information about the shipwreck, those amendments would be futile because they would raise a claim that is not yet ripe for review. 9 Courts may only review “final agency action,” 5 U.S.C. § 704, and challenges to agency decisions that are yet to be made are not ripe for review. See Pashby v. Delia, 709 F.3d 307, 317 (4th Cir. 2013) (holding that a claim is only ripe when it involves a formalized administrative decision). Here, the Agencies have not yet made a final decision about the significance of new information revealing that the shipwreck was a World War II assault vessel. Therefore, any claim that the Agencies did not properly account for this information would be unripe, and any amendment seeking to add such a claim would be futile. Second, SOS contends on appeal that it is challenging the adequacy of the 2016 ROD’s treatment of the Pappy’s Lane Wreck rather than the Agencies’ failure to consider new information about the shipwreck’s significance. Specifically, SOS contends that, because the Agencies knew that the shipwreck was eligible for listing on the National 9 SOS appears to have abandoned this argument on appeal. See Appellants’ Br. at 53 (contending that the district court “fundamentally misread [SOS’s] motion” when it denied it on ripeness grounds because “[t]he [new] allegations were about the inadequacies of the [2016] ROD and the process leading up to that document”) (emphasis added). Nevertheless, we discuss it because it implicates concerns related to the argument that SOS does raise on appeal regarding the adequacy of the 2016 ROD when issued. 25 Register of Historic Places, they were not authorized to build a bridge in its path unless they determined that there was “no prudent and feasible alternative to using that land” and that the “project include[d] all possible planning to minimize harm . . . resulting from the use.” See 49 U.S.C. § 303(c). However, as the district court explained, SOS cannot contend that it was challenging the adequacy of the 2016 ROD’s treatment of the shipwreck when its claims “rest[] upon recent discovery that the Pappy[’s] Lane Wreck contains a World War II vessel.” Save Our Sound OBX, Inc. v. N.C. Dep’t of Transp., No. 2:17-CV-4-FL, 2017 WL 7048561, at *5 (E.D.N.C. Dec. 5, 2017). Instead, the Agencies satisfied regulatory requirements when, based on what they knew about the Pappy’s Lane Wreck, they ordered data recovery and decided that the shipwreck did not merit preservation in place. Arguments that the Agencies should have gone further are grounded in new information and, therefore, not reviewable until the Agencies issue a final decision, as discussed above. We therefore affirm the district court’s denial of SOS’s motion to amend its complaint. IV. We affirm the district court because the Agencies did not violate NEPA or the DTA when they issued the 2016 ROD selecting the Jug-Handle Bridge and because the district court did not abuse its discretion when it denied SOS’s motions to supplement the administrative record and to amend the complaint. AFFIRMED 26