Opinion for the Court filed by Chief Judge HARRY T. EDWARDS.
The City of Centraba, Washington (“Centraba”) brings this petition for review to challenge an order of the Federal Energy Regulatory Commission (“FERC” or “Commission”) requiring Centraba to conduct a study of the effects of the Yelm Hydroelectric Project on the anadromous fish in the Nisqually River. Centraba filed an application in 1989 for a bcense to operate the existing Yelm Hydroelectric Project. The National Marine Fisheries Service (“NMFS”) recommended that FERC require Centraba to construct a tailrace barrier to prevent harm to the river’s anadromous fish population. The Acting Director of Hydropower Licensing determined that the cost of constructing a tailrace barrier could not be justified, because the benefits of the barrier had not been demonstrated. He held, however, that a study should be undertaken to determine whether, and to what extent, the Project was harmful to the fish. FERC, with two Commissioners dissenting, denied Centraba’s petition for rehearing. Centraba now petitions for review.
Under sections 4(e) and 10(a) of the Federal Power Act (“FPA” or “Act”), 16 U.S.C. §§ 797(e), 803(a), as amended by the Electric Consumers Protection Act (“ECPA”), Pub.L. No. 99-495, 100 Stat. 1243 (1994), FERC.must balance power and non-power values when deciding whether to issue hydropower licenses. In this case, Centraba contends that the Commission’s order requiring a study should be vacated, because FERC failed to accurately weigh the high cost of the study against the negligible benefits to be derived from the study. We agree. FERC does not dispute that the study will cost Centraba up to $300,000 to determine whether a tailrace barrier costing $1,000,-000 should be constructed. FERC also concedes that the study could prove inconclusive. In addition, FERC has no meaningful hard evidence to prove that the hydroelectric project is harmful to fish. In contrast, the Nisqually Indian Tribe *387(“Tribe”), which operates a fishery on the river, has submitted concrete data to show that no real harm to the fish results from the Project. In short, the record in this case does not support either the construction of a tailrace barrier or a study to determine its feasibility. Accordingly, the petition for review is granted.
I. Background
Centraba operates the Yelm Hydroelectric Project along the Nisqually River. In the late 1970s, the Nisqually Indian Tribe, which operates a fishery along the Nisquaby River, filed a complaint claiming that the Yelm Project was harming the fishery. Centraba and the Tribe subsequently commenced negotiations in an effort to settle the Tribe’s claims. In 1985, in a separate proceeding, FERC determined that Centraba, which had been operating the Yelm Project since 1930 without a bcense, was required to file for a bcense. Centraba submitted a bcense appbcation in 1989.
The Tribe and Centraba finaby reached a settlement in 1991. See Stipulation and Settlement Agreement between the City of Centraba and the Nisquaby Indian Tribe, Executed Feb. 28, 1991, reprinted in Joint Appendix (“J.A.”) 88. Centraba agreed to provide the Tribe with money, land, and other concessions in exchange for the Tribe’s support in its bcense appbcation with FERC. Centraba also agreed to achieve a minimum flow in the river.
In the original settlement between Centraba and the Tribe, Centraba agreed to construct a tailrace barrier. Not long after the settlement was signed, however, the Tribe submitted a letter to Centraba stating that it did not believe that a tab-race barrier was either necessary or desirable. See Letter from Dorian S. Sanchez, Chairman, Nisquaby Indian Tribe to William C. Cummings, Director, Centraba City Light (July 25, 1991), reprinted in J.A. 103. The letter noted that, “since minimum flows were estabbshed in 1977,” the Tribe had “not documented any delay problems at the powerhouse.” Id. at 2, reprinted in J.A. 104. Moreover, the letter stated that a tailrace barrier would “have an immediate and ongoing negative impact on fisheries habitat.” Id. Finally, the letter acknowledged that, in beu of constructing a tailrace barrier, Centraba had committed to contribute an amount equal to one-half the cost of a tailrace barrier to support fisheries enhancement projects within the Nisqually River basin. Id. The Tribe viewed these other fisheries enhancements as “much more beneficial” than a tailrace barrier. Id. On this last point, the letter noted that
[t]he Nisqually Tribe’s 1990 settlement agreement with Centraba calls for Centraba to construct a tabrace barrier. This portion of the agreement was based on the assumption that a tailrace barrier would be required of Centraba by FERC; it was not based on a lengthy analysis of the need for the barrier and should not be used by FERC or other agencies as a basis for requiring such a barrier. If the tabrace barrier is not imposed by FERC, Centraba and the Nisquaby Tribe wbl meet to modify the settlement agreement to provide additional funding for fisheries enhancement.
Id. at 2-3, reprinted in J.A. 104-05. Subsequently, in a response to FERC’s request for additional information, Centraba maintained that a tailrace barrier would not be in the public interest, both because of its “high cost” and the “absence of any reliable data to support the need for the barrier.” See Centraba’s Response to Enclosure B of FERC Staff Letter Dated April 12, 1991, reprinted in J.A. 111.
Meanwhbe, as Centraba and the Tribe moved to avoid construction of a tailrace barrier, officials from NMFS pressed a different view. NMFS recommended to FERC that, as a condition of any bcense, Centraba should be required to bubd a tailrace barrier to protect the fish. See Letter from Dean L. Shumway, Director, Division of Project Review, FERC to Merritt E. Tuttle, Division Chief, NMFS (Mar. 23, 1992), reprinted in J.A. 169 (noting NMFS’ recommendations). On March 16, *3881992, however, FERC issued an Environmental Assessment for the license and determined that a tailrace barrier was not justified. See Environmental Assessment for Hydropower License, Yelm Project, Mar. 16, 1992, 22-23, reprinted in J.A. 115, 141-42. The Assessment noted that fish could certainly be attracted to “the high velocity of the tailrace discharge flows” which could delay migration or could injure or kill the fish, and that a tailrace barrier would prevent those harms. Id. at 22, reprinted in J.A. 141. But, the Assessment stated, no “site-specific studies have been completed that quantify the numbers, if any, of salmon delayed in their upstream migration ... or the numbers that are injured or killed.” Id. The Assessment also noted the Tribe’s findings that minimum flow had reduced the problem and that large numbers of salmon had successfully migrated past the Project without a tailrace barrier. The Assessment concluded:
Since no information exists regarding the magnitude of delay, false attraction, or turbine blade mortality of salmon that the project could be causing, if any, we cannot conclude that tailrace barriers would significantly reduce delay of upstream migrants or turbine runner mortality.
Therefore, we cannot conclude that significant enhancement to the river’s salmon populations or fishery would be achieved with construction of the tail-race barrier. Consequently, a tailrace barrier cannot be justified.
Id. at 23, reprinted in J.A. 142.
Because FERC had rejected NMFS’ recommendation of a tailrace barrier, an effort was made pursuant to § 10(j) of the FPA, 16 U.S.C. § 803(j), to reconcile the disagreement between the agencies. FERC allowed NMFS time to substantiate the need for a tailrace barrier. See Letter to Merritt E. Tuttle (Mar. 23, 1992), reprinted in J.A. 169. In response to FERC’s inquiry, NMFS claimed that the Environmental Assessment was “seriously flawed.” Letter from Merritt E. Tuttle, Division Chief, NMFS to Dean L. Shumway, Director, Division of Project Review, FERC (Apr. 30, 1992), reprinted in J.A. 173. NMFS argued that there was not enough information about harm to the fish, because Centraba had faded to do a study on the impact of the Yelm Project. It noted that NMFS’ general pobcy is to require tailrace barriers, because fish can be injured in the tailrace. It described how that happened and noted that “[djelay and injury has been documented at a number of sites.” Id. at 174.
NMFS also asserted that sufficient evidence demonstrated potential harm to the fish. It claimed that, even without a study, injury and false attraction were potential problems, because the fish in the river were definitely strong enough to swim up to the turbine. NMFS contended that “FERC’s failure to order Centraba to conduct [a] tailrace study ... has created the informational void present at this time.” Id. at 6, reprinted in J.A. 178. NMFS thus recommended that FERC either order the barrier to be built or order Centraba to conduct a study evaluating the impacts of the Project on the fisheries.
Two meetings between FERC and NMFS officials (along with other interested parties) were held during the summer of 1992, and NMFS officials submitted further arguments in support of their recommendation. The meetings did not resolve any issues, however. Everyone agreed that salmon can be found in the tailrace area, but the parties disagreed about whether the fish were harmed. NMFS continued to insist that either a barrier should be built, or a study done. It submitted a so-called “Summary of Key Evidence” to make the case for its recommendation. It noted that fish are commonly observed in the tailrace; that someone witnessed a fish trying to swim into the tubes and one fish jumped and bkely injured itself; and that false attraction, delay and injury “has commonly occurred at other powerhouse tailraces in the Pacific Northwest.” Summary of Key Evidence, submitted by Merritt E. Tuttle, Division *389Chief, NMFS, Aug. 18, 1992, reprinted in J.A. 197.
In April 1996, FERC issued a Final Environmental Impact Statement (assessing the cumulative impacts of the Yelm Project with another hydroelectric project on the river) that concluded that the operation of the two projects would result in minor impacts on the environment, which would be outweighed by the projects’ benefits. The report also stated that the fish stocks at issue were listed “healthy.” A “healthy stock” listing means that a “stock óf fish [is] experiencing production levels consistent with its available habitat and within the natural variations in survival for the stock.” Final Environmental Impact Statement, Nisqually Hydroelectric Project, July, 1996, reprinted in J.A. 206, 226-27 (listing chinook salmon, coho salmon, chum salmon, pink salmon, and winter steelhead trout stock as “healthy”).
Finally, on March 7, 1997, the Acting Director of the Office of Hydropower Licensing issued an order rejecting NMFS’ recommendation that Centralia build a tailrace barrier, but accepting the recommendation that Centralia be required to conduct a study on the potential need for a tailrace barrier. The order concluded that “the information is not sufficient to quantify” the harm to the fish “caused by the Yelm Project’s tailrace, or the consequences of these potential project-induced effects on the stream’s fishery.” City of Centralia Light Dep’t, 78 F.E.R.C. ¶ 62,-171, at 64,636 (1997). Thus, the incremental benefits of the barrier were not demonstrated sufficiently to justify the cost of the barrier. However, the Director agreed with NMFS that “a study is necessary to determine how many fish are injured or killed by contact with the project turbine blades ... [and] also ... [that] a study is needed to determine how much false attraction and delay” is occurring. Id.
Centralia sought rehearing, claiming that a study was not justified. While the petition for rehearing was pending, the Tribe continued to support Centralia’s contention that a study was neither necessary nor justified. It reiterated its earlier recommendation that a tailrace barrier not be constructed, and stated that, in its view, “further studies are not needed and would be a waste of resources.” Letter from David A. Troutt, Natural Resources Director, Nisqually Indian Tribe to Director, Office of Hydropower Licensing, FERC 4 (Apr. 11, 1997), reprinted in J.A. 317, 320. The Tribe also elaborated on the value to be derived from the alternative fishery enhancement measures that Centralia would fund if Centralia were not required to build the barrier. Specifically, the Tribe noted that
[t]hese alternative measures address the single greatest threat to the salmon resources along the Nisqually River and tributaries, the threat of increasing development pressure due to urbanization along the basin’s shorelines. They will focus on the acquisition for permanent protection and rehabilitation of critical habitat areas along the river and major tributaries. This would further the adaptive management goals discussed above and required under current approaches to salmon management, and would be substantially more beneficial than imposing a costly tailrace barrier that addresses a minimal or nonexistent problem.
Id. The Tribe also submitted additional support in 1998 and 1999, stating that, in its annual harvest management activities, there were “no indications through [its] surveys that the fish are being delayed at any point during their migration.” Letter from David A. Troutt, Natural Resources Director, Nisqually Indian Tribe to Bill Tobin, Natural Resources Attorney, Nisqually Indian Tribe (May 12, 1998), reprinted in J.A. 327; Letter from David A. Troutt, Natural Resources Director, Nisqually Indian Tribe to Bill Tobin, Natural Resources Attorney, Nisqually Indian Tribe (Feb. 11, 1999), reprinted in J.A. 333.
*390Before FERC acted on Centralia’s petition for rehearing, Centralia and the Tribe concluded an amendment to their settlement agreement. They agreed:
• to use their best efforts to obtain concurrence of FERC [and other agencies] that the tailrace barrier should not be a condition of Centralia’s FERC license.
• if ... Centralia is not required to construct the tailrace barrier, then Centralia and the Tribe shall equally share Centralia’s savings from not having to construct the barrier.
• money paid by Centralia to the Tribe ... shall be segregated by the Tribe and used exclusively for the protection, rehabilitation or enhancement of the anadromous fisheries resources of the Nisqually River.
Stipulation and Settlement Agreement between the City of Centralia and the Nisqually Indian Tribe, Amendment No. 1, reprinted in J.A. 328.
FERC then denied rehearing. It noted that there is evidence of fish in the Project’s tailrace, and that, .“[b]ased on the record before us, during parts of the salmonid migration season, the tailrace flows nearly equals flow in the bypassed reach,” and that fish could be attracted to it, “especially during August and September.” City of Centralia Light Dep’t, 87 F.E.R.C. ¶ 61,383, at 62,421 (1999). The Commission thus concluded that fish might be harmed. FERC noted that this sparse evidence, while not enough to support an order requiring the construction of a tail-race barrier, was enough to require Centralia to conduct a study.
Two Commissioners dissented. Commissioner Bailey stated that she was not convinced that the evidence in the record was sufficient to require a study, and she did not believe that a study was justified under the balancing required by Part I of the Act. She pointed out that “the record submissions ... both before and after the date of the [environmental assessment] do nothing more than suggest, based on isolated observations and flow-based understandings, the presence of fish (including a single jumping adult salmon) in, or the attraction of fish to, the tailrace area of the project.” Id. at 62,423. She cited with approval the Tribe’s studies and stated that they demonstrated that there was no need for Centralia to conduct a study. She stated that the Commission had failed to engage in the balancing of costs and benefits required by the Act and that the costs of a study outweighed any perceived benefits. She cited Amendment No. 1 of the Centralia-Tribe settlement and stated that any concerns she might have had about not ordering the barrier were alleviated by the parties’ agreement to spend the savings on other fishery enhancements. See id. at 62,424. Commissioner Hérbert agreed with Commissioner Bailey, adding that “funds are more efficiently utilized if spent on species enhancement measures, as stipulated between the city and the tribe, rather than the study prescribed in this matter.” Id. “Let us remember,” he added, “it is the survival and enhancement of the species which is paramount.” Id. This appeal followed.
II. Analysis
This case presents an odd conjunction of statutory provisions, which were fully explored in United States Dep’t of Interior v. FERC, 952 F.2d 538 (D.C.Cir.1992):
Under the FPA, FERC may license hydroelectric projects on federal lands and on waterways that are subject to congressional regulation under the Commerce Clause. 16 U.S.C. § 797(e) (1988). Under sections 4(e) and 10(a) of the FPA, 16 U.S.C. §§ 797(e), 803(a), as amended by the Electric Consumers Protection Act (“ECPA”), Pub.L. No. 99-495, 100 Stat. 1243 (1986), FERC must consider environmental issues when deciding whether to issue hydro-power licenses.
In deciding whether to issue any license under [the FPA] for any project, the Commission, in addition to the power and development purposes *391for which licenses are issued, shall give equal consideration to the purposes of energy conservation, the protection, mitigation of damage to, and enhancement of, fish and wildlife (including related spawning grounds and habitat), the protection of recreational opportunities, and the preservation of other aspects of environmental quality-
16 U.S.C. § 797(e) (1988).
All licenses issued under this subchapter shall be on the following conditions: That the project adopted ... will be best adapted to a comprehensive plan ... for the adequate protection, mitigation, and enhancement of fish and wildlife (including related spawning grounds and habitat)....
16 U.S.C. § 803(a) (1988).
Additionally, under section 10(j) of the FPA, 16 U.S.C. § 803(3), FERC must impose conditions on licenses “based on recommendations received pursuant to the Fish and Wildlife Coordination Act (16 U.S.C. § 661 et seq.) from the National Marine Fisheries Service, the United States Fish and Wildlife Service, and State fish and wildlife agencies.” § 803(j)(l). FERC retains ultimate authority, however, to decide whether any recommended conditions are “inconsistent with the purposes of’ the FPA or other laws. § 803(j)(2). When it acts contrary to a recommendation received from a wildlife agency, FERC must make an appropriate finding on the record to justify its decision. Id.
sH ^4
The ECPA amendments to the FPA, which added the “equal consideration” language to section 4(e) and created the section 10(j) process, were aimed primarily at increasing FERC’s sensitivity to environmental concerns....
[However], the ECPA amendments do not give environmental factors preemptive force.... FERC still is charged with determining the “public interest,” ie., balancing power and non-power values. Even where the fish and wildlife
agencies make formal section 10(j) recommendations, those agencies have no veto power. See National Wildlife Fed’n v. FERC, 912 F.2d 1471, 1480 (D.C.Cir.1990) (“While the Commission must address each recommendation, the discretion ultimately vests in the Commission as to how to incorporate each recommendation. If we read the statute any other way, the Commission would be held hostage to every agency recommendation, and the Commission’s role of reconciling all competing interests would be compromised.”).
Id. at 543-45. See also City of Oconto Falls v. FERC, 204 F.3d 1154, 1160 (D.C.Cir.2000) (“[T]he Commission still is charged with determining the ‘public interest,’ ie., balancing power and non-power values. Even where the fish and wildlife agencies make formal § 10(j) recommendations, those agencies have no veto power.”) (citations and internal quotation marks omitted).
The question raised in this petition for review is whether FERC’s order, “requiring Centralia to conduct a study to determine the extent of [Yelm] project impacts on the upstream migration of steel-head trout and salmon,” City of Centralia, 87 F.E.R.C. at 62,421-22, is supported by substantial evidence and is not arbitrary and capricious. See United States Dep’t of Interior, 952 F.2d at 545 (stating the substantial evidence and arbitrary and capricious standards of review); City of Oconto Falls, 204 F.3d at 1159 (same); Texaco, Inc. v. FERC, 148 F.3d 1091, 1095 (D.C.Cir.1998) (same). Because FERC’s order is devoid of “reasoned decision making,” Texaco, 148 F.3d at 1095, and lacks substantial evidence to justify requiring Centralia to conduct a study, we are constrained to grant the petition for review.
FERC is required by sections 4(e) and 10(a) of the Act to give “equal consideration” to energy and environmental considerations. In other words, it must “balanc[e] power and non-power values.” United States Dep’t of Interior, 952 F.2d *392at 545. The statute does “not give environmental factors preemptive force.” Id. FERC’s order in this case makes no attempt whatsoever to “balance” power and non-power values in justifying the need for a study. The reason seems obvious: there are no material environmental considerations that weigh in favor of a study.
FERC is clear, and the record surely supports its conclusion, that there are no facts sufficient to require Centralia to construct a tailrace barrier. See City of Centralia, 87 F.E.R.C. at 62,421. The agency’s Environmental Assessment report made it plain that there was nothing in the record to support the construction of a tailrace barrier. Environmental Assessment at 23, reprinted in J.A. 142. In other words, FERC’s own studies could uncover no non-power concerns. And during the § 10(j) exchanges between FERC and NMFS officials, FERC never varied from the position taken in the Environmental Assessment that the costs of a barrier could not be justified.
Having acknowledged that a barrier is not justified, FERC stumbles badly in concluding that the costs of a study could be justified. FERC is certainly empowered to require an applicant to conduct a study when there is some evidence of a problem and a study is necessary to determine the extent of the harm. But not even FERC is suggesting that an applicant has a duty to determine if a problem exists. Yet, that is the result of the disputed order in this case.
FERC’s order articulates no evidence of harm. Rather, FERC relies on the fact that fish are observed in the Project’s tailrace and the assertion that “during parts of the salmonid migration season, the tailrace flows nearly equals flow in the bypassed reach.” Thus, FERC argues, “salmonids could be attracted to the tail-race discharge or the bypassed reach” and “either be delayed in their migrations or injured or killed if they enter the draft tubes and are struck by the turbines.” City of Centralia, 87 F.E.R.C. at 62,421. FERC’s conclusion is based on sheer speculation. Therefore, it cannot be said that there is substantial evidence justifying a study. See Bangor Hydro-Elec. Co. v. FERC, 78 F.3d 659, 663 (D.C.Cir.1996) (rejecting characterization that an agency conclusion was a “finding” where it was merely a prediction based on opinions). That fish could be attracted to the flow is not evidence of a problem that warrants a study. Indeed, FERC’s only evidence that fish are harmed is that someone saw a single jumping fish that hit a concrete barrier and may have been hurt. City of Centralia, 87 F.E.R.C. at 62,421 n. 7. This is not evidence enough to support the disputed conclusion in this case.
Not only is there no substantial evidence to support FERC’s order, there is substantial evidence to support the opposite position endorsed by Centralia and the Tribe. As already noted, the record contains abundant submissions from the Tribe showing that, based on the Tribe’s surveys, no harm results from the Project. See Letter to Director of the Office of Hydropower Licensing (Apr. 11, 1997) (stating that the Tribe believed that both the barrier and a study were unnecessary and describing the benefits of the other fishery enhancement projects), reprinted in J.A. 317. See also Letter to William C. Cummings (July 25, 1991), reprinted in J.A. 103 (“[S]ince minimum flows were established in 1977 we have not documented any delay problems at the powerhouse,” and “we have studied chinook and pink migration and. determined that these species were found far upstream of the powerhouse at an early date.”); Letter to Bill Tobin (May 12,1998), reprinted in J.A. 327 (noting that after conducting field surveys on foot, boat, raft or helicopter, the Tribe had “not noticed any delays caused by the power plant associated with the Centralia hydroelectric project”); Letter to Bill Tobin (Feb. 11, 1999), reprinted in J.A. 333 (same). Neither NMFS in its submissions nor FERC in its order point to record evidence to refute the Tribe’s submissions.
*393Finally, Centraba contends, and we agree, that FERC’s order is arbitrary and capricious for want of reasoned decision-making. On the record here, “the costs of [FERC’s] prescription far outweigh any benefits to fish or the general environment and is therefore unreasonable.” Bangor Hydro-Elec., 78 F.3d at 663. FERC acknowledges that “a study could be rather expensive [i.e., as much as $300,000] and could prove inconclusive.” City of Centralia, 87 F.E.R.C. at 62,421 n. 3. And Centraba and the Tribe convincingly argue that the more than $500,000 that Centraba will contribute to pay for alternative fishery enhancement measures is a far superi- or way to protect the fish and river environment than a $300,000 expenditure for what will likely be an inconclusive study on the feasibibty of a tailrace barrier.
FERC has no reasonable answer to Centralia’s argument, because it has failed to weigh the relative costs and benefits of the proposed study. Standing alone, the study is arguably too expensive, for it is difficult to justify a $300,000 expenditure for an inconclusive study to determine whether to spend another $1,000,000 to construct a tailrace barrier to address a problem that has not been identified. When weighed against the alternative remedies proposed by Centraba and the Tribe, however, the order requiring a study borders on absurd.
Given the record in this case, we must conclude that FERC has not provided reasonable support, ie., “substantial evidence,” either for the construction of a tailrace barrier or a study to determine its feasibibty; and we also must conclude that FERC has failed to show, in the required statutory balancing of power and non-power values, that a study is “reasonably related to its goal” of enhancing the fishery of the Nisqually River. Bangor Hydro-Elec., 78 F.3d at 664.
III. Conclusion
For the aforementioned reasons, the petition for review is granted. FERC’s order requiring Centraba to conduct a study is hereby reversed and vacated.