287 F.3d 1151 (2002)
FPL ENERGY MAINE HYDRO LLC, Petitioner,
FEDERAL ENERGY REGULATORY COMMISSION, Respondent.
American Rivers, et al., Intervenors.
United States Court of Appeals, District of Columbia Circuit.
Argued March 14, 2002.
Decided May 3, 2002.
1152*1152 1153*1153 Catherine R. Connors argued the cause for petitioner. With her on the briefs was Matthew D. Manahan.
1154*1154 Cynthia L. Amara was on the brief for amici curiae Clifton Power Corporation and New England Legal Foundation in support of petitioner.
Laura J. Vallance, Attorney, Federal Energy Regulatory Commission, argued the cause for respondent. With her on the brief were Cynthia A. Marlette, General Counsel, and Dennis Lane, Solicitor.
Daniel H. Squire and IJay Palansky were on the brief for intervenors American Rivers, et al.
Before: SENTELLE, ROGERS and GARLAND, Circuit Judges.
Opinion for the Court filed by Circuit Judge SENTELLE.
SENTELLE, Circuit Judge:
Petitioner FPL Energy Maine Hydro LLC (FPL), a hydroelectric facility, petitions this Court for review of two orders by the Federal Energy Regulatory Commission (FERC), the first determining that FPL is subject to licensing because the Messalonskee Stream on which FPL is located is “navigable” under 16 U.S.C. § 796(8), the second denying FPL’s request for rehearing. Because we find that FERC’s interpretation of the statute governing navigability is a reasonable one and that its navigability finding was supported by substantial evidence, we deny the petitions.
The Messalonskee Stream (Stream) is a tributary of the Kennebec River (Kennebec) located in central Maine. It runs approximately ten miles from the Messalonskee Lake to the Kennebec, with four dams located along its stretch. The southernmost dam is the Union Gas Project, located approximately one mile up the Stream from the confluence of the Stream and the Kennebec. In between the dam and the Kennebec, beginning from the dam and progressing downstream, are two sets of rapids or “rips,” a bridge, a third set of rips, and two islands that together span approximately 200 feet downstream with a shallow shoal on the east side of the islands and a narrow and rocky channel on the west side. Below the islands the Stream widens and deepens as it encounters the backwater of the Kennebec. The Kennebec, itself a navigable water, empties into the Atlantic Ocean.
Pursuant to section 23(b)(1) of the Federal Power Act (FPA), 16 U.S.C. § 817(1), a non-federal hydroelectric project must be licensed if it is located on a navigable water of the United States, as defined by 16 U.S.C. § 796(8), or if other criteria not relevant to this case are met. Because the four dams constitute one development unit, if one project requires a license, then they all must be licensed. See Kennebec Water District, 80 FERC ¶ 61,208, 61,828, 1997 WL 438885 (1997). Thus if the Stream between the Union Gas Project and the Kennebec is deemed navigable pursuant to 16 U.S.C. § 796(8), all four projects require a license.
Section 3(8) of the FPA defines navigable waters as
those parts of streams … which either in their natural or improved condition notwithstanding interruptions between the navigable parts of such streams or waters by falls, shallows, or rapids compelling land carriage, are used or suitable for use for the transportation of persons or property in interstate or foreign commerce….
16 U.S.C. § 796(8). This means that, to be navigable for purposes of the FPA, a waterway must form a highway for commerce with other states or with foreign countries, by itself or by connecting with other waters. See The Montello, 87 U.S. (20 Wall.) 1155*1155 430, 439, 22 L.Ed. 391 (1874). Courts have determined a waterway to be navigable if “(1) it presently is being used or is suitable for use, or (2) it has been used or was suitable for use in the past, or (3) it could be made suitable for use in the future by reasonable improvements.” Rochester Gas & Electric Corp. v. FPC, 344 F.2d 594, 596 (2d Cir.1965) (emphasis in original), see also Marine Stevedoring Corp. v. Oosting, 398 F.2d 900, 908 n. 15 (4th Cir.1968), rev’d on other grounds, Nacirema Operating Co. v. Johnson, 396 U.S. 212, 90 S.Ct. 347, 24 L.Ed.2d 371 (1969); Sierra Pacific Power Co. v. FERC, 681 F.2d 1134, 1137-38 (9th Cir.1982). Navigability can be established based on any of these three requirements; each alone is sufficient. Rochester Gas, 344 F.2d at 596.
II. Proceedings Below
The Union Gas Project is currently licensed by FPL (as successor in interest to Central Maine Power Company — the original licensee of the project). The original license for the project was issued in 1968 and expired in 1993. Since that time, the project has been operating on annual licenses. As part of a jurisdictional examination of several projects for which licensing might not have been required, the Office of Hydropower Licensing conducted a navigation report on the Stream in 1996. This report indicated that the Stream was not navigable because there was “no evidence of usage of the stream as a water highway, a continuous link for interstate commerce, either commercial or recreational, from above the project sites, past the projects, to the Kennebec River.” Following a review of comments to the report, the Acting Director of the Office of Hydropower Licensing issued an order finding that the Union Gas Project was located on a navigable waterway and therefore required a license. See Kennebec Water District, 79 FERC ¶ 62,041, 1997 WL 362917 (Apr. 21, 1997). On rehearing, FERC concluded the evidence submitted was inadequate to support a finding of navigability. See Kennebec Water District, 80 FERC ¶ 61,208, 1997 WL 438885 (Aug. 6, 1997). Following petitions for rehearing on that order, FERC set the issue of navigability for a hearing before an Administrative Law Judge (ALJ) to determine, among other things, the physical characteristics of the Stream, the difficulty associated with navigating the Stream, and the nature and frequency of actual use of the river for recreational boating. See Kennebec Water District, 81 FERC ¶ 61,073, 61,306, 1997 WL 698330 (Oct. 21, 1997). The ALJ, who did not address the physical characteristics of the Stream as they relate to navigability, found that the Stream was not navigable despite three “successful” and two “unsuccessful” canoe trips made for the purpose of litigation. See Kennebec Water District, 82 FERC ¶ 63,004, 1998 WL 9486 (Jan. 14, 1998). The ALJ also held that there was no evidence of “regular and substantial recreational use” to serve as a proxy for the simpler types of commercial navigation as allowed under United States v. Appalachian Electric Power Co., 311 U.S. 377, 61 S.Ct. 291, 85 L.Ed. 243 (1940). FERC trial staff and several intervenors in the proceeding below filed exceptions to the ALJ’s initial decision, which Central Maine Power opposed. Upon review, FERC concluded that the ALJ applied an incorrect legal standard by requiring evidence of “regular and substantial recreational use” for a finding of navigability. It therefore reversed the non-navigability finding and 1156*1156 required FPL to obtain a license. See Kennebec Water District, 84 FERC ¶ 61,027, 1998 WL 514602 (July 16, 1998). FERC based its navigability finding on the three “successful” canoe trips taken for purpose of litigation, as well as the physical characteristics of the Stream. Id. at 61,126. FERC denied FPL’s rehearing request and these petitions for review followed.
This case requires us to answer two questions: first, whether FERC’s interpretation of “navigability” under the FPA was reasonable; second, if so, whether FERC’s navigability finding was supported by substantial evidence. We answer both questions affirmatively.
A. “Navigability” Interpretation
FPL argues that FERC departed from the statutory “suitable for use … in … commerce” test set forth at 16 U.S.C. § 796(8) (see also Rochester Gas, 344 F.2d at 596) and instead applied a mere “possibility of passage” test to determine whether the Stream is navigable. FPL contends that FERC conducted its navigability determination by looking only to three non-commercial, non-recreational test canoe trips that indicated it was possible to navigate downstream in unusual conditions, rather than looking to historical and present commercial or recreational use by average canoeists. FPL further contends that, in determining navigability, FERC deviated from its past precedent that requires more than just a showing of specialized, recreational boating where historical evidence of commercial use was lacking. FPL also faults FERC for relying on the flow created when the hydroelectric facility is generating to create navigability, as well as for failing to identify the commercial use to which the Stream could realistically be put.
We reject each of FPL’s arguments. Where an administrative agency is tasked with interpreting an ambiguous statute that it administers, a court will defer to that agency’s interpretation so long as it is reasonable. See Chevron U.S.A., Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 843, 104 S.Ct. 2778, 2782, 81 L.Ed.2d 694 (1984). As the statute does not define when a waterway is “suitable for use … in … commerce,” we assume that Congress intended FERC to address the ambiguity in the statute and develop an appropriate test. See United States v. Mead Corp., 533 U.S. 218, 229, 121 S.Ct. 2164, 2172, 150 L.Ed.2d 292 (2001). We find that FERC’s interpretation of navigability under the FPA, which was based on test canoe trips and the Stream’s physical characteristics in the absence of any commercial or recreational use, was reasonable and entitled to deference.
At the outset, we may quickly reject FPL’s argument that it was improper for FERC to rely on the flow created when the Union Gas Project is generating to create navigability. FPL confuses a finding of whether a waterway has always been navigable because an improvement could be made if reasonable in cost, with a finding of whether a waterway, together with its improvements, is presently “suitable for use.” See Rochester Gas, 344 F.2d at 596. The question before this Court is whether the Stream, with the presence of the Union Gas Project and the flow created when there is generation, is presently navigable, see Washington Water Power Co. v. FERC, 775 F.2d 305, 332 (D.C.Cir.1985), not whether the Stream was navigable prior to the Project’s construction. See id. at 331-32.
1157*1157 Turning next to FPL’s primary argument, FPL contends that FERC’s “possibility of passage” test: 1) ignored the fact that the Stream has never been used, either historically or presently, for commercial or private purposes, and 2) deviated from past decisions wherein FERC denied a finding of navigability when the only evidence of actual use was specialized, recreational boating.
All parties agree that the Stream has never been used for commercial traffic. But just because a body of water has not been used for commercial use does not mean that it is not susceptible to commercial use. See United States v. Utah, 283 U.S. 64, 82, 51 S.Ct. 438, 443, 75 L.Ed. 844 (1931). In Appalachian Power, 311 U.S. at 416, 61 S.Ct. at 303, the Supreme Court held that a lack of commercial traffic is not “a bar to a conclusion of navigability where personal or private use of boats demonstrates the availability of the stream for the simpler types of commercial navigation.” Thus without evidence of commercial use, FERC may look to other types of use to establish navigability.
In the past, FERC has often relied on evidence of recreational use as a proxy for commercial suitability. Here, however, there is no evidence of recreational use of the Stream. In fact, the only evidence indicating actual use of the Stream comes from the three trips made for the purpose of litigation. Echoing the ALJ, FPL argues that this evidence is insufficient to serve as a proxy for commercial suitability. Instead, FPL argues that FERC precedent requires a navigability finding to be based on “regular and substantial recreational use” in the absence of actual commercial use. We disagree. The statute and the case law make clear that evidence of actual use is not necessary for a navigability determination. “[T]he test [is] whether the river … is used, or capable of being used as a highway for commerce, over which trade and travel is or may be conducted in the customary modes of trade and travel on water.” Economy Light & Power Co. v. United States, 256 U.S. 113, 121-22, 41 S.Ct. 409, 412, 65 L.Ed. 847 (1921) (emphasis added). To determine whether a waterway is capable of being used as a highway for commerce, evidence other than recreational use may be considered. That is, recreational use may be the most common proxy evidence FERC relies on for findings of commercial suitability, but FERC has never taken the position, nor will this Court take the position now, that recreational use is the only proxy evidence on which FERC may rely. “Recreational boating is … of interest because it is boating, not because it is recreational. Any similar personal or private use not involving recreation, such as use of a river as a means of personal transportation, would be equally relevant to a determination of suitability for commercial navigation.” Kennebec Water District, 88 FERC ¶ 61,118, 61,304, 1999 WL 549672 (July 28, 1999). FERC may therefore, in the absence of commercial use, rely on evidence other than recreational use if that evidence is relevant to a finding of navigability. The Supreme Court has identified at least two types of evidence, declaring that the “capacity [of a waterway to meet the needs of commerce] may be shown by physical characteristics and experimentation as well as by the uses to which the streams have been put.” United States v. Utah, 283 U.S. at 83, 51 S.Ct. at 444. In the absence of any “uses to which the [Stream] ha[s] been put,” FERC acted consistently with Supreme Court precedent in relying on the three test canoe trips.
FPL nonetheless argues that the test trips were of a type of specialized, recreational boating that FERC has previously 1158*1158 disregarded when making navigability determinations. The cases on which FPL relies, however, are distinguishable from the facts before us. For example, in Pennsylvania Elec. Co., 56 FERC ¶ 61,435, 62,549-50, 1991 WL 297437 (1991), FERC determined a river was non-navigable because a substantial reach of the river could only be navigated by a kayak (or comparably specialized sporting craft designed for river running) maneuvered by an expert paddler. Similarly, in PacifiCorp Elec. Operations, 73 FERC ¶ 61,365, 62,140, 1995 WL 756395 (1995), rehearing denied, 74 FERC ¶ 61,262, 1996 WL 89781 (1996), FERC determined that a waterway was non-navigable because “all of the evidence of use or suitability for use for recreation concerns use by skilled kayakers or whitewater rafters.” FERC held that “[t]his is not the sort of recreational boating that [it] has recognized as demonstrating the suitability of a river for the simpler types of commercial navigation.” Id. at 62,140-41. The river in PacifiCorp included Class IV rapids that could not be navigated easily without a specialized boat. In the present case, the Stream at most contains Class II rapids that were successfully crossed by a canoe. FERC has repeatedly found waterways to be navigable that may be traversed by a canoe — a simpler type of commercial transportation. See, e.g., Appalachian Power, 311 U.S. at 415-16, 61 S.Ct. at 303. We therefore find that FERC did not depart from precedent when it relied on successful test trips taken in canoes.
FPL further argues that FERC’s navigability test was flawed because FERC failed to identify the possible commercial use to which the Stream may be put. We see no reason why FERC must identify the precise commercial use to which a previously unused waterway may be put in order for the Commission’s finding of navigability to be upheld. The test is whether the waterway is presently “suitable for use for the transportation of persons or property in interstate or foreign commerce,” not whether the waterway is presently suitable for a specific type of commercial activity named by FERC and approved of by an opposing party. 16 U.S.C. § 796(8); see also PacifiCorp, 73 FERC at 62,140, 1995 WL 756395 (“[I]n order to demonstrate that the [waterway] at the project site is a navigable water, [FERC] need only find that it was or is used or suitable for use to transport persons or property between the project site and [another navigable water].”). Consequently, if the evidence in the record supports a finding that the Stream is suitable for transporting persons or property to the Kennebec, then this Court will uphold such a finding without adding a requirement that FERC identify a specific type of commerce associated with the transportation of persons or property. The language of the statute does not require such a finding; neither do we.
Even more important to our analysis, FERC did not rely on the test canoe trips alone when finding that the Stream was navigable. FERC also looked to the Stream’s physical characteristics when making its navigability determination. See United States v. Utah, 283 U.S. at 83, 51 S.Ct. at 443-44. Thus absent evidence of commercial or recreational use, FERC properly relied on both “physical characteristics and experimentation” to determine whether the Stream was suitable for use in commerce. Id.
For the reasons stated, we find that FERC’s interpretation of the FPA’s navigability test was reasonable insofar as it necessarily relied on test canoe trips and the Stream’s physical characteristics in the absence of past or present commercial and 1159*1159 recreational use of the waterway. It is therefore entitled to deference.
B. Substantial Evidence
As our discussion of FERC’s deference-worthy interpretation of the navigability test may suggest, we also conclude that FERC’s finding of navigability is supported by substantial evidence. 16 U.S.C. § 825l(b) (“The finding of the Commission as to the facts, if supported by substantial evidence, shall be conclusive.”).
The “experimental” test canoe trips provide sufficient evidence that the Stream is navigable. Three witnesses, all with differing interests in the litigation, successfully navigated downstream without incident, and two attempted and succeeded in navigating upstream, albeit with some difficulty. Although FPL made much of the difficulty associated with this upstream travel both in its brief and at oral argument, FPL failed to provide any explanation as to why an upstream trip — either made with ease or with difficulty — is necessary for a navigability finding when the evidence of successful downstream trips is clear. Nowhere in the statute or accompanying case law does it state that the transport of persons or property in interstate or foreign commerce must include two-way transport. We do not view this case as an opportunity to suggest otherwise.
In addition to relying on the three test trips, FERC made a separate determination that the physical characteristics of the Stream rendered it suitable for commercial navigation. The Supreme Court has held that a water’s “capacity [for commercial navigation] may be shown by physical characteristics and experimentation as well as by the uses to which the streams have been put.” United States v. Utah, 283 U.S. at 83, 51 S.Ct. at 444. And in Montana Power Co. v. Federal Power Commission, 185 F.2d 491, 495 (D.C.Cir.1950), this Court declared that “[i]f the stream’s flow, depth, gradient, width and capacity make it `suitable for use’ in interstate commerce, it is subject to the licensing authority of [FERC].” In other words, a waterway may never have been used for transportation, commerce, or recreation, but nonetheless may be suitable for interstate commerce (and subject to licensing by FERC) based on its physical characteristics. See Loving v. Alexander, 745 F.2d 861, 864 (4th Cir.1984) (“The extent and manner of use of a navigable river is not important as long as it is usable as an actual avenue of commerce.”).
The record includes sufficient evidence regarding the Stream’s physical characteristics on which FERC relied in making its navigability determination. FERC noted that the Stream has a very slight gradient, has a depth of approximately three and a half to four feet when the dam is generating (although slightly shallower around the two islands), is wide enough to support passage up and down the Stream and around the two islands, and has few obstacles (e.g., boulders and fallen trees) that a canoeist may steer around without difficulty. Kennebec Water District, 84 FERC ¶ 61,027, 61,125-26, 1998 WL 514602 (July 16, 1998). There is nothing in the record to suggest that the Stream’s physical characteristics preclude navigability when the dam is generating other than a water depth around the islands that is lower than the rest of the Stream. But the same record includes evidence that the three canoeists successfully navigated down the Stream and past the islands. For reasons already stated, the difficulty experienced by one of the canoeists traveling upstream past the islands does not negate the otherwise sufficient evidence. Moreover, even though the dam does not operate all the time, thereby rendering the water level too low 1160*1160 to navigate year round, navigability need not be available “at all seasons of the year, or at all stages of the water.” Economy Light & Power, 256 U.S. at 122, 41 S.Ct. at 412.
FPL attempts to bolster its argument with evidence that the test trips upon which FERC relied were made during periods of unusual water conditions. FPL argues that this should negate FERC’s navigability finding. That is, FPL argues that since “susceptibility of use as a highway for commerce should not be confined to exceptional conditions or short periods of temporary high water,” the canoe test trips do not show the stream to be navigable. Loving, 745 F.2d at 865; see also United States v. Utah, 283 U.S. at 87, 51 S.Ct. at 445. True, the evidence indicates that two of the test trips took place during periods of high water, but that same evidence indicates that the remaining test trip took place during a period of low water. FPL’s witness navigated up and down the Stream when the Kennebec’s flow was 4030 cubic-feet per second (cfs) (down from its average flow of 5000 cfs), an event which would have reduced the backwater effect from the Kennebec and lowered the already low depth of water around the islands. This is sufficient evidence that the Stream can be navigated both up and downstream during non-optimal water levels.
We acknowledge that the evidence of navigability is not overwhelming. But to uphold FERC’s navigability determination, we need only find that the evidence on which the finding is based is substantial. 16 U.S.C. § 825l(b); see Consolidated Hydro, Inc. v. FERC, 968 F.2d 1258, 1261 (D.C.Cir.1992). The “substantial evidence” standard requires more than a scintilla, but can be satisfied by something less than a preponderance of the evidence. See Whitmore v. AFIA Worldwide Ins., 837 F.2d 513, 515 n. 4 (D.C.Cir.1988). We therefore find that the test trips, together with the Stream’s physical characteristics, constitute substantial evidence to support FERC’s finding of navigability.
FERC’s reliance on test canoe trips and the physical characteristics of the Stream, in the absence of historical or present commercial or recreational use, is a reasonable interpretation of the navigability test set forth in the Federal Power Act and is therefore entitled to deference by this Court. Moreover, the record includes substantial evidence to support FERC’s navigability finding. Three witnesses were able to successfully navigate down the Stream, and the physical characteristics of the Stream support a finding of navigability. For these reasons, the petitions for review are denied.
 Union Gas was previously required to be licensed pursuant to FERC’s incorrect interpretation of a different section of the FPA.