dissenting:
The majority holds (1) that the Federal Energy Regulatory Commission (“FERC”) did not err in granting the Virginia Electric and Power Company (‘VEPCO”) an amendment to its license for the Lake Gaston power project authorizing the withdrawal of sixty million gallons per day of project water without requiring VEPCO to obtain a water quality certification from the State in which the power project’s “discharge” originates, and (2) that FERC’s decision to issue this license amendment was not arbitrary and capricious. I disagree with both of these propositions.
A. North Carolina’s Clean Water Act § 401(a)(1) Certification Rights
The first paragraph of the Federal Water Pollution Control Act, popularly known as the Clean Water Act (“the Act”), declares that its goal is to “restore and maintain the chemical, physical, and biological integrity of the Nation’s waters.” 33 U.S.C. § 1251(a). The second subsection sets forth Congress’ overarching policy in the Act, of “recogniz[ing], preserving], and protecting] the primary responsibilities and rights of States to prevent, reduce, and eliminate pollution____” Id. at § 1251(b). The Act charges each State with the duty of instituting comprehensive standards establishing quality goals for all intrastate waters. See 33 U.S.C. §§ 1311(b)(1)(C), 1313.
Section 401(a)(1) of the Act, 33 U.S.C. § 1341(a)(1), implements the policy of empowering States to protect their water quality programs by authorizing them to veto federal licenses or permits that threaten to undermine the quality of their waters. Specifically, § 401(a)(1) requires “[a]ny applicant *1195for a Federal license or permit to conduct any activity ... which may result in any discharge into the navigable waters” to provide the licensing or permitting agency with “a certification from the State in which the discharge originates or will originate” that the discharge does not threaten the water quality standards that the State has implemented pursuant to other provisions of the Act. Id. I disagree with the majority’s conclusion that North Carolina had no such certification right with regard to this license amendment, because I believe it ignores the language, structure, and purpose of the Act, and wrests away from North Carolina the very power to make water quality judgments about its own waterways that Congress expressly meant for that State alone to have.
The § 401 certification right is an essential component of the Act’s state-oriented regulatory scheme. See Keating v. FERC, 927 F.2d 616, 622 (D.C.Cir.1991) (“The States remain, under the Clean Water Act, the prime bulwark in the effort to abate water pollution ... Congress intended that the States would retain the power to block, for environmental reasons, local water projects that might otherwise win federal approval.”) (citation and internal quotation marks omitted). Obviously, how courts construe the Act’s language specifying those situations in which States may wield this powerful weapon can critically affect their ability to protect the integrity of their waters in the way Congress intended. Too stingy an interpretation of § 401(a)(1) will severely burden each State’s ability to implement the water quality program mandated by the Act, and will frustrate Congress’ clear intent to institute a system under which States carry both the responsibility of setting their own water quality standards and a corresponding power to protect those standards from federally-licensed activities that threaten to undermine them. It is precisely because the § 401(a)(1) certification right is so vital to each State’s ability to maintain its water quality that thirty-nine States signed an amicus brief strongly opposing FERC’s assertion that North Carolina has no certification rights over this license amendment. See Brief of Amici Curiae States of Alabama, Alaska, Arizona, Arkansas, California, Connecticut, Delaware, Florida, Hawaii, Idaho, Illinois, Indiana, Iowa, Kansas, Kentucky, Louisiana, Maine, Maryland, Massachusetts, Michigan, Minnesota, Mississippi, Missouri, Montana, Nevada, New Hampshire, New Mexico, New York, North Dakota, Ohio, Oklahoma, Pennsylvania, South Dakota, Texas, Utah, Vermont, Washington, West Virginia, and Wyoming.
The majority assumes arguendo that the flow of water through the power project’s turbines in North Carolina is a § 401(a)(1) “discharge” implicating that State’s certification rights, and this assumption is in fact required under the precedents set by the Supreme Court,1 this court,2 and FERC.3 Furthermore, it appears that North Carolina will wield a § 401(a)(1) certification right in 2001, when the project’s license will expire and VEPCO will request a new license. See Brief for Appellant Roanoke River Basin Association at 18-22. Yet the majority reaches the paradoxical conclusion that FERC may approve this license amendment, which all parties concede will alter the North Carolina' “discharge,” without requiring any certification from North Carolina. It is candidly difficult for me to get past this conundrum: Logically, if a State must consent before a new discharge is introduced into its waters, then a change in that discharge must require *1196a new consent.4 A question might, of course, be raised as to whether some license amendments are so clearly outside of the scope of § 401(a)(1) as to create no certification rights; license amendments that propose to alter a project in a way that will have no effect on an existing discharge, or that will have at most a de minimis effect on the discharge, may perhaps be granted without requiring a new certification from the State in which the discharge originates.5 But the majority imposes a far more sweeping and arbitrary limitation on State certification rights, by declaring that no alteration of a discharge triggers a certification right unless it represents an “addition ... of a substance or substances” to the discharge. Majority opinion at 1187.
I reject the majority’s “substance-adding” threshold condition on the exercise of State certification rights. It badly distorts the definitions section of the Act, and improperly usurps crucial judgments on water quality that Congress reserved for the States into which the discharges flow. The initial flaw is the majority’s characterization of the Act’s definition of “discharge of a pollutant” as “the nearest evidence we have of definitional intent by Congress” as to when States have certification rights under § 401(a)(1). Id. The Act’s legislative history says otherwise. The language creating certification rights in States in which a “discharge” may “originate” as a result of federally-licensed activities first appeared as part of the Water Quality Improvement Act of 1970, Pub.L. No. 91-224, 84 Stat. 91 (1970). The phrase “discharge of a pollutant,” on which the majority leans so heavily, was nowhere to be found in that early version of the Act. It was only added with the passage, two years later, of the Federal Water Pollution Control Act Amendments of 1972, Pub.L. No. 92-500, 86 Stat. 816 (1972), which introduced it in conjunction with a set of provisions requiring permits for the “discharge of pollutants.” 33 U.S.C. § 1342. Because the bare-bones term “discharge” was part of the Act before the 1972 amendments, and because it wasn’t until two years later that Congress organized its new permit program around the separately coined and defined phrase “discharge of a pollutant,” the notion that the latter phrase provides compelling evidence of the meaning of the former term is totally unpersuasive.
An even stranger facet of the majority’s statutory analysis is its insistence that “near[er]” evidence of Congress’ “definitional intent” as to what constitutes a discharge can be found by looking at the Act’s definition of the phrase “discharge of a pollutant” than by relying on the Act’s own definition of the very term under discussion — “discharge.” The Act expressly states that the term “ ‘discharge’ when used without qualification includes a discharge of a pollutant, and a discharge of pollutants.” 33 U.S.C. § 1362(16) (emphasis added). The only inference I can discern from this clear expression of nonexclusivity in the definition of “discharge” is that it was meant to encompass a wider array of interferences with the navigable waters than would qualify as “discharges of *1197pollutants.” See National Wildlife Federation v. Gorsuch, 693 F.2d 156, 171-72 (D.C.Cir.1982) (distinguishing between statutory definitions that specify what a term “means” and those that offer a non-exelusive listing of what a term “includes”).6 How the majority manages to limit the scope of the unqualified term “discharge” to one subset that is “inelude[d]” in the broader term is beyond my ken.
Further evidence that Congress intended for States to have the power to protect their waters from interferences created by changes to a “discharge” that do not comprise the “addition” of any “substance or substances” is observable in the Act’s treatment of the term “pollution.” Section 304 of the Act “expressly recognizes that water ‘pollution’ may result from ‘changes in the movement, flow, or circulation of any navigable waters____’” PUD No. 1, 511 U.S. at 719-20, 114 S.Ct. at 1913 (quoting 33 U.S.C. § 1314(f)). The Environmental Protection Agency’s regulations implementing the Act likewise express a concern with interferences that have “flowage effects.” Id. I had thought it beyond dispute that even alterations of existing discharges that do not “add” any “substance or substances” to the water may yet affect the water’s “movement, flow, or circulation,” that this effect may constitute “pollution,” and that “[i]t is the policy of the Congress to recognize, preserve, and protect the primary responsibilities and rights of States to prevent, reduce, and eliminate pollution.” 33 U.S.C. § 1251(b) (emphasis added).7 Yet the majority is comfortable in wrenching away from a State this very power to protect its water quality from the “pollution” threatened by the alteration of an existing discharge.
The majority’s niggardly construction of the term “discharge” threatens a congressional policy permeating the general structure of the Act, a policy specifically declared in the Act’s second subsection: States were *1198to be put in charge of making and enforcing the crucial judgments surrounding water quality within their borders. See 33 U.S.C. § 1251(b). When a federally-licensed activity would “result in” a “discharge” into the State’s waters, the decision whether that interference with the State’s waters would be tolerated was to belong solely to the State in which that discharge originated, not to a federal agency or to a federal court. Given the relative competency of State water quality agencies and federal courts, this policy seems quite sound. The waterways that we judges see only as lines on a diagram are in reality extraordinarily complex and sensitive systems in which water, oxygen, dissolved minerals, sediment, and other materials intermingle and interact. Interferences with the integrity of a waterway need not take the archetypal form of sludge spewing from an outfall; other changes can be just as deadly to fish and other wildlife as are toxic chemicals. For example, alterations that cause the temperature of the water in a discharge to rise, or that rearrange the spatial distribution of oxygen, sediment, or minerals in the downstream water, could be catastrophic from the perspective of water quality, see Gorsuch, 693 F.2d at 163-64.8 Yet the effect of this court’s decision will be to bar States in which altered discharges originate from acting to protect their waters from such pernicious effects, whenever the alteration does not “add” any “substance or substances” to the water. I do not believe Congress intended to give this court’s essentially ad hoc and inevitably inexpert judgments precedence over the expert judgments of the States’ water quality authorities, in the making of such crucial determinations.
The majority itself recognizes, in expressing “serious reservations” about FERC’s ability to limit by regulation States’ certification rights to license amendments that might have a “material adverse impact” on the water quality in a discharge, that judgments intrinsically related to water quality in the context of the alteration of existing discharges are the sole province of the States in which the discharges originate. See majority opinion at 1186 (quoting 18 C.F.R. § 4.38(f)(7)(iii) (1995)). By empowering itself to make a threshold determination regarding whether an alteration in a discharge threatens a “material adverse impact” on the State’s water quality, FERC is taking from the State a power that Congress intended for the State alone to have. Section 401(a)(1) enables States to raise an absolute bar to federal licenses that threaten their waters; Congress completely stripped federal agencies of the power to issue covered licenses without the State’s certification. Indeed, FERC is assigning itself a function which makes no sense in light of the structure of the Act, which puts each State in charge of authoritatively defining the term “water quality” within its borders — why, in this one context, should a federal agency be able to stand in a State’s shoes and make the State’s “water quality” determinations for it? Thus, I share the majority’s concerns about FERC’s regulation, but I can’t help but wonder why the majority doesn’t feel the same discomfort about its own arrogation of a similar power to preempt a State’s water quality determinations. Congress did not intend for this court to usurp North Carolina’s power to manage its water quality program any more than it intended to give FERC such a role, and surely we are as ignorant as FERC on the subject of what “water quality” is optimal for the citizens of North Carolina.
The Supreme Court and our own court have previously exercised appropriate reluctance to read the Clean Water Act in the arbitrary and artificial fashion proffered today. When a city and a utility district urged the Supreme Court in PUD No. 1 to adopt a formalistic approach very much like the majority’s, the Court categorically refused. The PUD No. 1 petitioners asserted that the Clean Water Act is concerned only with water “quality,” and does not regulate water “quantity.” See PUD No. 1, 511 U.S. at 717-19, 114 S.Ct. at 1912. The Supreme Court *1199rejected.this “artificial distinction,” id, noting that water quantity is closely related to water quality, and that both the definitions section and § 304 of the Act set forth a “broad conception of pollution” which “expressly evinces Congress’ concern with the physical and biological integrity of water.” Id. at 719, 114 S.Ct. at 1913. And in National Wildlife Federation this court likewise recognized that the Clean Water Act is not to be constrained by artificial limitations such as the majority’s “substance-adding” standard. In that case, we found that the “backing up” of water by a dam constituted a § 401(a)(1) “discharge,” see National Wildlife Federation, 912 F.2d at 1483-84; in so doing, we apparently recognized that an interference with a river that results in the alteration of the “movement, flow, or circulation” of water (and thus causes “pollution”) is cause for the recognition of a certification right in the State where this discharge originates.9 Since the “backing up” of water does not satisfy the majority’s “substance-adding” standard, it is clear that today’s holding is in tension with our own precedent.10
Instead of crafting an arbitrary and exceedingly formalistic new standard governing the scope of State certification rights under § 401(a)(1), we should be traversing the main channel charted by the Act, the Supreme Court, and our own precedent. These authorities require us to find that the license amendment at issue in this case is one that triggers a certification right in North Carolina, the State in which the project’s discharge originates. Thus I would suspend consideration of the other challenges to these orders, and require FERC to await a § 401(a)(1) certification from North Carolina before considering whether to grant the amendment.
B. FERC’s Finding of a Need to Withdraw Sixty Million Gallons of Water per Day
Even if North Carolina’s failure to certify the license amendment were not contrary to the Act, however, I would vacate the challenged orders and remand this matter to FERC on the ground that FERC’s estimate of the “need” for the fifty-four million gallons per day to be supplied by the pipeline project was arbitrary and capricious.
FERC had already reached the conclusion that Virginia Beach needed to withdraw sixty million gallons of the Roanoke River per day by October of 1994, when it declared that “[n]o new analysis” of this estimate was necessary. Scoping Document 2, FERC No.2009-003; Deferred Appendix “D.A.” at 2547, 2551. It is hardly surprising, therefore, that FERC’s July, 1995 Final Environmental Impact Statement arrived at this precise estimate of Virginia Beach’s “need.” Nor is it surprising that FERC’s attempt to disguise the fact that it was “backing into” this number fails rather badly.
I agree with my colleagues that the first few steps in FERC’s analysis were tolerably reasonable. FERC organized its analysis into separate projections of the demand for water in the year 2030 and of regional supply in that year. The basic structure of the inquiry was sound, as was the projection of the region’s 2030 demand for water. But when FERC turned to the supply side of the equation, its logic became curiouser and curiouser. On the apparent assumption that the water supply from existing sources would remain constant through 2030, FERC’s pro*1200jection of 2030 supply was based on the water available from existing sources. Specifically, FERC’s supply estimate was based on the amount the residents of the region could withdraw from the existing water sources during the worst drought of record since 1930; this drought occurred in 1930, and the U.S. Geological Survey estimates that such droughts may recur at intervals of between thirty and eighty or more years. The estimated “Water Supply Deficit” for 2030, based on this drought-adjusted supply, came out to 31.8 million gallons per day.
Common sense would seem to indicate that the obvious way to prepare for short-term periods of special water need is to create water storage systems that can be filled during nondrought periods, exhausted to meet these short-term needs when they arise, and refilled when the droughts recede. But FERC rejected all of the storage options pressed by participants in the proceedings below, arguing that storage techniques were irrelevant to its “need” calculation because storage systems do not constitute “sources” of water, and because they serve only short-term need. See majority opinion at 1192. Of course, the fact that storage systems aren’t new “sources” of water in the sense of meeting long-term, continuous demand hardly justifies FERC’s refusal to consider the availability of these systems, because they just as clearly are precisely suited to serve short-term periods of special demand that occur at intervals more than long enough to permit the storage systems to be replenished, and thus could be used to obviate any increment in the quantity of water to be constantly supplied to the region to protect it from droughts of extraordinary severity.
A similar illogic permeates FERC’s attempt to explain its refusal to take account of the water available from the region’s wells. FERC specifically refused to factor in the water available from two Norfolk wells on the ground that these wells were meant to be used only during droughts that threaten military readiness, and FERC likewise thought it unnecessary to consider the water available from unspecified “emergency wells” in the region on the ground that their use was restricted to municipally-declared emergencies.11 The Corps, FERC noted, “considers these wells as protection against extreme droughts.” D.A. at 3166. I can only speculate as to what the terms “municipal emergency” and “extreme drought” are meant to cover, if not droughts more severe than the worst since 1930. FERC’s reasoning falls far short of providing any rational or even vaguely comprehensible explanation for its refusal to integrate into its need calculation the water available from “emergency” wells and water storage systems — the most intuitively obvious strategies for dealing with infrequent, short-term, emergency water needs.
Having rejected the commonplace strategies that would enable the region to weather droughts, but would not enable FERC to claim that a pipeline continuously withdrawing sixty million gallons of water per day was needed, FERC’s next step was to ratchet its “need” estimate up to that amount. For this final adjustment, FERC set out five reasons for its conclusion that a second “drought adjustment” was necessary, explaining that merely covering the 31.8 million gallon per day “deficit” was insufficient because:
• it provides no protection against droughts of greater severity than those that occurred in the past;
• there is no guarantee that water sharing would occur among the cities;
• severe water use restrictions would continue to place local users at a public health and safety risk;
• future water demands may exceed projections; and
• loss of a water supply in this large metropolitan area would be catastrophic.
D.A. at 3167, see also majority opinion at 1193. For these “reasons,” FERC concluded *1201that it was necessary to jack up the “need” estimate to include another thirteen to sixty-eight million gallons per day of continuous supply. But these “reasons” simply amount to FERC’s generalized worry that its entire analysis leading up the 31.8 million gallon per day “deficit” might be flawed in such a way as to make its need estimate too low. If water were an unlimited, free resource, it might seem appropriate to double the estimate of water “need” in ease actual need were to exceed all expectations, but if water were an unlimited, free resource, nothing much would turn on whether FERC’s “need” estimate satisfied any baseline standard of rationality. FERC itself recognizes that the amount of water to be removed from the river must correspond to a rational calculation of actual “need,” and thus makes a show of reaching this estimate by using the available data and reasonable assumptions; it cannot at the end of the game simply knock the pieces off the chessboard in the name of unforeseen contingencies and unprecedented emergencies, and still claim that it did not act arbitrarily or capriciously because it played by the rules almost to the end.
Because, finally, I believe that the license amendment FERC approved by its challenged order may not be granted consistently with § 401(a)(1) of the Clean Water Act unless and until North Carolina has certified that the activity to be conducted pursuant to the amendment does not threaten that State’s water quality standards, I would vacate FERC’s order and rehearing order, and remand the matter to FERC with the instruction that it withhold the amendment pending North Carolina’s certification. Even if I did not believe that FERC’s failure to require VEPCO to obtain North Carolina’s certification was illegal, I would still vacate these orders on the ground that they were arbitrary and capricious, and would remand the matter to FERC for an estimate of Virginia Beach’s “need” for water that gives due consideration to storage options and “emergency” sources. On both of these grounds, I respectfully dissent from the majority opinion.
.See PUD No. 1 of Jefferson County v. Washington Dep’t of Ecology, 511 U.S. 700, 711, 114 S.Ct. 1900, 1908, 128 L.Ed.2d 716 (1994) (“There is no dispute that petitioners were required to obtain a certification from the State pursuant to § 401. Petitioners concede that, at a minimum, the project will result in two possible discharges — the release of dredged and fill material during the construction of the project, and the discharge of water at the end of the tailrace after the water has been used to generate electricity.“).
. See National Wildlife Federation v. FERC, 912 F.2d 1471, 1484 (D.C.Cir.1990) (“Common sense supports FERC’s conclusion that the discharge in this case would occur at the dam, where the flow of water would be blocked and consequently the water would be backed up.... ”).
. See City of Fort Smith, 42 F.E.R.C. ¶ 61,362 at 62,047 (1988) ("In the case of the Lee Creek project, the discharge will occur over the dam and from the powerhouse tailrace ____”), on reh’g, 44 F.E.R.C. ¶ 61,160 at 61,513 (1988), aff'd, 912 F.2d 1471 (D.C.Cir.1990).
. The majority questions the logic of this assertion by arguing that it would be irrational as applied to statutes that refer to “conduct resulting in injury” and "transactions resulting in financial loss.” See majority opinion at 1186 n.2. But these analogies are clearly inapt. The more appropriate analogy would be to an argument that a statute referring explicitly to conduct that threatens to cause one type of injury or loss might be construed also to cover closely related types of injury or loss, rather than to an argument that a statute referring to conduct that causes some injury or loss should be construed also to cover conduct that causes the reduction of these ills. After all, the record before us involves a situation in which a reduction in a discharge may increase the very sort of “injury” to a State's water qualily that the Act seeks to empower States to prevent. See infra.
. Additionally, there is in all cases the threshold requirement that a discharge covered by the license in question must "originate” within the State claiming a § 401(a)(1) certification right. The rights of “affected” States that cannot satisfy this threshold condition are defined in § 401(a)(2). See 33 U.S.C. § 1341(a)(2). Since there is no question that a discharge covered by the license that FERC amended originates in North Carolina, my interpretation would not by any means make § 401(a)(2) superfluous, cf. majority opinion at 1186 (summarizing FERC's arguments); if North Carolina were not the locus of a discharge covered by the license to be amended, its rights would be defined by § 401(a)(2), as are the rights of downstream "affected” States in regard to this license amendment.
. I cannot fathom the majority’s reasoning that, since the phrase "discharge of a poEutant” describes the addition of substances to the water, Congress meant for the term "discharge” also to refer only to the addition of substances to the water. See majority opinion at 1187 n.4. I find the majority's reasoning untenable — it seems far more reasonable that, by retaining the term "discharge” unmodified by the words "of a poEutant,” Congress must have intended that the term would have a broader meaning when used without these restrictive modifying words than it does when used with these restrictive modifying words. And this broader meaning should encompass the subtraction of part of an existing discharge because, as I explain infra, the subtraction of water from a discharge may be just as harmful to a waterway as the addition of substances to the water. Under the majority’s reasoning, even the initial construction of this power project — which created a new discharge in North Carolina — would not have triggered any certification rights in North Carolina if §. 401(a)(1) had then been in effect, provided that the project did not add any substance to the water that passed through the dam turbines. Such a result, of course, would be at odds with the majority's assumption arguendo that the flow of water through a power project’s turbines is a discharge. See majority opinion at 1187.
. Employing the currently popular “dictionary jurisprudence," the majority reasons that the “activity” here — the diversion of water to Virginia Beach — wiE not “result in” a discharge, as required by § 401(a)(1), see majority opinion at 1188, because the activity authorized by the license amendment gives rise only to an "altered” discharge, rather than a brand new one. I faff to see how this dictionary exercise provides any compelling reason to ignore the implications of the Act's State-enforcement policy and central concern with preventing and eliminating "pollution.” To the contrary, it would block a State's certification rights even in'the case of a license amendment authorizing the "addition” of a "substance or substance?" to an existing discharge, since the "addition” of substances merely "alters” a discharge, and does not cause the discharge to "arise as a consequence” of the addition. Id. I find the majority's use of the dictionary to support its assertion that there is no hint of ambiguity in § 401(a)(1), see id. at 1183 n. 1, unacceptable; to my mind, there is certainly enough ambiguity to justify consideration of the purpose and structure of the Act in determining the scope of States' certification rights. After all, the Act defines the term "discharge" in a purposely open-ended fashion, see 33 U.S.C. § 1362(16), and § 401(a)(1) doesn’t even require concrete certainty that a discharge will “result” from the activity to be licensed — it requires State certification for any activity that "may result in” a discharge. 33 U.S.C. § 1341(a)(1) (emphasis added). Imposing dictionary-derived limitations on the Act is particularly dangerous in this context and invites egregious abuses — for example, a licensee could obtain a State’s certification of a minor, non-threatening discharge and then transform the discharge into a poEution-causing nightmare, evading the State’s power to prevent the change by characterizing it as merely as merely an "alteration” of an existing discharge.
. In Gorsuch, this court observed that the term "pollutant” might reasonably include such "Dam-Induced Water Quality Changes” as low dissolved oxygen, dissolved minerals, heat, sediment, and oxygen "supersaturation,” id. at 161, 166, though we deferred to the EPA’s interpretation of the term “pollutant,” in a different section of the Act, as not including these effects. See id. at 183.
. In that case, we declined to recognize a certification right in the upstream State, however, because we found that the discharge "originate[d]" at the dam, which was located downstream in another state. See id. at 1484.
. The majority disclaims tension between this case and National Wildlife Federation on the grounds that the parties in that case did not dispute the question of whether the construction and operation of the dam would create a discharge, and that the discharge at issue related to a new dam, rather than to the altered operations of an existing dam. See majority opinion at 1188. But the National Wildlife opinion refers to the expulsion of water through the dam's turbines as a § 401 "discharge,” and describes the "discharge” as follows: "|T]he discharge in this case would occur at the dam, where the flow of water would be blocked and consequently the water would be backed up ....” National Wildlife Federation, 912 F.2d at 1484 (emphasis added). The fact that the “backing up” of water resulted from the construction of a new dam, rather than from the alteration of an existing dam, played no part in the National Wildlife Federation court’s analysis.
. The majority asserts that FERC’s refusal to consider the water in these wells in its estimate of the regional supply was appropriate because these wells may not be available for use by Virginia Beach after next year. See majority opinion at 1193. But this portion of FERC’s analysis was based on the supply of water available to the region, rather than to Virginia Beach alone, and therefore it is irrelevant that this emergency supply may become unavailable to one of the cities in the region.