ROGERS, Circuit Judge:

James C. Robinson petitions for review of the order of the National Transportation Safety Board affirming the emergency revocation by the Administrator of the Federal Aviation Administration of his airline transport pilot certificate. The revocation was based on Robinson's operation of a twin-engine helicopter with only one operable engine, in violation of Federal Aviation Regulations. Robinson maintains first, that the Administrator's emergency action was arbitrary and capricious, an abuse of discretion, and not in accordance with law, and second, that the Board's order upholding the decision of the Administrative Law Judge is similarly arbitrary and capricious and unsupported by substantial evidence. The first issue is moot. On the second issue, we remand the case to the Board to explain its finding that Robinson operated his helicopter with only one operable engine.

I.

On October 18, 1993, the Administrator of the Federal Aviation Administration (FAA) issued an Emergency Order revoking Robinson's airline transport pilot certificate because of the unsafe operation of his helicopter. The Emergency Order stated that on May 16, 1993, Robinson executed a takeoff under the power of a single engine and departed with only one operable engine, contrary to the pilot action called for in the helicopter's Aircraft Flight Manual. The Order recited that Robinson had violated four Federal Aviation Regulations — 14 C.F.R. §§ 91.7(a) (b), 91.9(a), and 91.13(a) (1993) — and that by reason of his lack of care, judgment, and responsibility required of one holding an Airline Transport Pilot Certificate, an emergency existed requiring that the Order take effect immediately. Robinson appealed to the National Transportation Safety Board, and a hearing was held before an Administrative Law Judge (ALJ).

The undisputed evidence showed that at approximately 6:26 p.m. on May 26, 1993, Robinson landed the helicopter at the Wall Street Heliport in New York City to meet his passenger, Robert E. Brennan. Before takeoff, a chip light went on in the helicopter indicating a problem with one of the engines. After a takeoff was aborted, the passenger took off in another aircraft, and Robinson thereafter took off in his helicopter, landing at Allaire Airport around 8:30 p.m. Repair records showed that one of the helicopter's engines had experienced a chip light and loss of power.

The Administrator presented four witnesses to show that Robinson had taken off from the Wall Street Heliport with only one operable engine. John P. Licciardi, the senior airport operations agent at the Wall Street Heliport, testified that on May 26, Robinson parked the helicopter, shut it down, and entered the terminal building, where Licciardi informed Robinson that his passenger was on his way and that Robinson should start his aircraft. After observing Robinson start his helicopter, Licciardi informed Robinson that he could move the helicopter to the transient area in front of the terminal. According to Licciardi, Robinson said that because of trouble with engine number two he could not move the helicopter to the transient. Robinson then asked Licciardi to escort the passenger to the parking place. Once the passenger was on board the helicopter, Robinson attempted to take off. Licciardi testified that the helicopter "skidded around the heliport, rather than hover[ing] up off the ground," and it looked like Robinson was unable "to get lift." Licciardi characterized the maneuver in his heliport log entry at 7:07 p.m. as an "attempted . . . single engine takeoff using the barge, to no avail," but he conceded that the statement in the logbook about the single engine takeoff was simply his own conclusion. Licciardi also admitted that his logbook entry that Robinson was going to attempt a second single engine takeoff was supposition on his part.

Three FAA investigators testified based on their review of records, personal interviews, and knowledge of the helicopter's chip light system. Anthony B. Winton, an FAA aviation safety inspector, opined on the basis of Licciardi's testimony about his own observations, that Robinson's attempted Wall Street takeoff was "indicative of an under-power takeoff, i.e., the running takeoff profile, and that there was a problem with the engine." He further opined, however, that "the engine may have been in idle power or completely shut down." Winton had no information whether the chip light was on or off at the time of the takeoff, and he agreed that it was possible that Robinson could have extinguished the chip light before taking off, and that somewhere between Wall Street and the Allaire Airport Robinson could have experienced a second chip light, taken appropriate action, and landed as soon as possible at Allaire. But, based on the other investigators' testimony, Winton concluded that because particles and not fuzz were in the oil system "the chip light most likely stayed on at Wall Street." In Winton's opinion, Robinson's conduct in taking off with a single engine was surprising and foolhardy.

Robinson presented four witnesses, including himself. One witness testified consistently with the FAA witnesses about the safety of flying the helicopter once the chip light was powered off, and two witnesses — Robert E. Brennan (the passenger) and John H. Ford (president and director of maintenance at Raco Helicopters) — corroborated Robinson's testimony that both engines were operable when he took off. Brennan, in addition to describing the attempted running takeoff and what Robinson had told him, testified that Robinson had both engines running, that the levers for both engines were up, and that there was no chip light problem while he was in the helicopter at the Wall Street Heliport. Ford testified that when Robinson landed the helicopter at Allaire Airport on May 26th, Robinson informed Ford that he had experienced a problem with one of the engines, that his chip light had come on, and that he had shut down the engine. Ford opened the cowling, and from the outside the engine looked normal. However, Ford explained, he was unable to identify which part of the engine was incurring the metal or chip problem because "the engine was a bit too warm to be playing with it. So, I elected not to do that at that time. * * * It was too warm to be handling." He further explained that a non-functioning engine would not generate heat and that the fire walls between the helicopter's two engines prevented any heat transfer from one engine to the other. Upon disassembling the engine the following day, Ford discovered that the oil separator's shaft had sheared; he testified that as a result of the structure and manner in which the unit was manufactured, metal particles were present in large quantities in the lubricating oil.

Robinson testified that he had full power available at all times on May 26th until he was about a mile away from Allaire Airport. He admitted that he experienced a chip light illumination for his number two engine while at the Wall Street Heliport, but he testified that the light went out after he pushed the chip pulse, making the aircraft safe to fly. Thereafter, according to Robinson, he attempted a running takeoff with his passenger on board because the helicopter was too heavy with fuel to fly if one of his engines failed. By attempting a running takeoff, Robinson thought that he would be able to stay safely outside of the "height velocity envelope." However, the unsuccessful running takeoff indicated that he would not be able to do so and, because he knew it would take some time to burn off enough fuel to be outside the height velocity envelope, he requested another helicopter to take his passenger. Once Robinson had burned off enough fuel, he took off, using a running takeoff with reduced power to simulate a single-engine takeoff in order to verify that he would have sufficient power in the event of another engine problem. Robinson claimed that he simulated the single-engine operation only during the early portion of the takeoff and switched to normal two-engine power for the takeoff itself. He testified that the chip light did not come on until about a mile away from Allaire Airport, when the engine started to spool down. Robinson secured the engine and landed on the taxi at the Allaire Airport with only the number one engine running. According to Robinson, the number two engine was inoperative for about two minutes during the flight from the Wall Street Heliport.

The ALJ upheld the Administrator's emergency order of revocation. While commenting that the Administrator's case was "far from being airtight," the ALJ concluded in an oral decision and order that the Administrator was not arbitrary and capricious in instituting the Emergency Order of Revocation and had sustained his burden of proof. The ALJ found that Robinson took off under the power of a single engine and that departing with only one operable engine contravened the pilot action that is called for in the aircraft flight manual. The ALJ also found that "safety and air commerce and air transportation and the public interest . . . require the affirmation of the Administrator's Emergency Order of Revocation. . . ." Crediting Licciardi's version of events, as supported by the FAA Inspectors' investigation and their opinion that Robinson's helicopter was not airworthy or safe to fly, the ALJ concluded that Robinson's response was inconsistent with the care, judgment, and responsibility of an airline transport pilot certificate holder.

The Board denied Robinson's appeal and by order affirmed the ALJ's oral decision. The Board relied on Licciardi's testimony and logbook notations as well as FAA Investigator Winton's opinion that the circumstances surrounding Robinson's takeoff from the Wall Street Heliport indicated that he had lost power in one of his engines. The Board also relied on repair records that indicated that one of the helicopter's engines suffered a power failure. The Board rejected Robinson's challenge to the ALJ's credibility determinations, in light of the "abundant circumstantial evidence to support the law judge's conclusion that, despite his denials, [Robinson] experienced engine failure at Wall Street but took off nonetheless." Robinson now petitions the court for review of the Board's order.

II.

Robinson contends that the FAA's decision to invoke emergency proceedings was arbitrary and capricious and not in accordance with law because the FAA did not have a reasonable basis for declaring an emergency. We do not reach this issue because we conclude, sua sponte, that it is moot inasmuch as Robinson can point to no consequence of the emergency nature of the revocation that the court could now undo. See St. Paul Fire Marine Ins. Co. v. Barry, 438 U.S. 531, 537, 98 S.Ct. 2923, 2927, 57 L.Ed.2d 932 (1978); see generally Tucson Medical Ctr. v. Sullivan, 947 F.2d 971, 977 (D.C. Cir. 1991); 13A WRIGHT, supra note 5, § 3533.3 at 273-74; cf. Schering Corp. v. Shalala, 995 F.2d 1103, 1105 (D.C. Cir. 1993).

The Board's order affirmed the Administrator's determination that safety in air commerce or air transportation and the public interest require revocation of Robinson's certificate. Consequently, the ultimate relief that Robinson seeks — the return of his pilot's certificate without recertification — can only be accomplished at this point if the court reverses the decision of the Board. Reversing the Administrator's use of emergency power would not provide that relief in view of the Board's subsequent determination that air safety and the public interest require the revocation. Indeed, the emergency underlying immediate issuance of the Administrator's order has ceased to exist because Robinson is now deprived of his pilot certificate by reason of the Board's order. See Air East, Inc. v. National Transp. Safety Bd., 512 F.2d 1227, 1231 (3d Cir.) ("under [46 U.S.C. app.] § 1429(a), emergency revocation has an effect limited in time to a period of sixty days"), cert. denied, 423 U.S. 863, 96 S.Ct. 122, 46 L.Ed.2d 92 (1975). Thus, the only "live" issue before the court is whether the Board's revocation was lawful. Cf. Doe v. Sullivan, 938 F.2d 1370, 1375 (D.C. Cir. 1991).

Although Robinson seeks an order striking the Administrator's designation of the revocation as an emergency, he has suggested no consequences to him that would follow from obtaining this relief or from the failure to obtain such relief. See Penthouse Int'l, Ltd. v. Meese, 939 F.2d 1011, 1019-20 (D.C. Cir. 1991), cert. denied, ___ U.S. ___, 112 S.Ct. 1513, 117 L.Ed.2d 650 (1992). Moreover, because Robinson would have the right to seek judicial review of any future emergency revocation order as soon as it was issued, see 49 U.S.C. app. § 1486(a) (1988), the emergency revocation does not fall within the capable-of-repetition-yet-evading-review exception to mootness. Cf. Christian Knights of the Ku Klux Klan v. District of Columbia, 972 F.2d 365, 369-71 (D.C. Cir. 1992). Nor has Robinson presented evidence of a policy or a continuing practice to which he objects. See City of Houston, Texas v. H.U.D., 24 F.3d 1421 (D.C. Cir. opinion circulated 1994) at 1429-30, 1431-32 (citing Payne Enters. v. United States, 837 F.2d 486, 490-92 (D.C. Cir. 1988); Better Gov't Ass'n v. Department of State, 780 F.2d 86, 90-92 (D.C. Cir. 1986); Super Tire Eng'g Co. v. McCorkle, 416 U.S. 115, 121-25, 94 S.Ct. 1694, 1697-1700, 40 L.Ed.2d 1 (1974)). Indeed, dispositive for our purposes is Robinson's counsel's candid admission at oral argument that he was seeking an advisory opinion from the court in light of the FAA's recent increased use of its emergency power. Cf. Alton S. Ry. Co. v. International Ass'n of Machinists Aerospace Workers, 463 F.2d 872, 877 (D.C. Cir. 1972).

Accordingly, we hold that Robinson's challenge to the Administrator's Emergency Revocation Order is moot.

III.

The question remains whether the Board's order upholding the revocation is arbitrary and capricious and unsupported by substantial evidence. Robinson contends that the ALJ's findings were arbitrary and capricious because they were ambiguous, cursory, and inappropriate in light of the overwhelming weight of the evidence supporting Robinson's version of what happened on May 26, 1993. He also contends that the Board's reliance on "abundant circumstantial evidence" is erroneous because none of that evidence sufficiently supported the FAA's contention that Robinson left the Wall Street Heliport under the power of only one engine. Specifically, the issue is whether the findings in the Board's order upholding the oral decision of the ALJ are adequate to explain why it rejected or disbelieved Ford's testimony that the second engine was too warm to handle when Robinson landed the helicopter at Allaire Airport, and whether the Board actually found that the second engine was inoperable, as distinct from not operating, when Robinson took off from the Wall Street Heliport.

Under the Administrative Procedure Act, 5 U.S.C. § 706(2)(A), the court is required to set aside agency findings that are unsupported by "substantial evidence." Throckmorton v. National Transp. Safety Bd., 963 F.2d 441, 444 (D.C. Cir. 1992). "`The substantial evidence test is a narrow standard of review' requiring only `such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.'" Id. (quoting Chritton v. National Transp. Safety Bd., 888 F.2d 854, 856 (D.C. Cir. 1989) (quotations omitted)). An agency conclusion "`may be supported by substantial evidence even though a plausible alternative interpretation of the evidence would support a contrary view.'" Id. (quoting Chritton v. National Transp. Safety Bd., supra, 888 F.2d at 856 (quotations omitted)); see also Consolo v. Federal Maritime Comm'n, 383 U.S. 607, 620, 86 S.Ct. 1018, 1026, 16 L.Ed.2d 131 (1966) (substantial evidence is "enough to justify, if the trial were to a jury, a refusal to direct a verdict when the conclusion sought to be drawn from it is one of fact for the jury") (citation omitted). The court's function is to determine only "whether the agency . . . could fairly and reasonably find the facts that it did." Throckmorton v. National Transp. Safety Bd., supra, 963 F.2d at 444 (quoting Chritton v. National Transp. Safety Bd., supra, 888 F.2d at 856 (in turn quoting Western Air Lines, Inc. v. Civil Aeronautics Bd., 495 F.2d 145, 152 (D.C. Cir. 1974))).

Robinson contends that the FAA's claim that he took off from the heliport with only one operable engine rested solely on the testimony of Licciardi, since other witnesses had no first-hand knowledge and were merely interpreting Licciardi's observations. He points out that Licciardi's observations corroborated Robinson's explanation for his actions, and that Licciardi admitted that his conclusions — on which the other witnesses, the ALJ, and the Board relied — were based on supposition, not fact. Specifically, Licciardi admitted he had no knowledge of whether Robinson actually took off with a single engine, if both engines were running when the helicopter was at the Wall Street Heliport, and what type of engine trouble Robinson was experiencing.

Robinson maintains that the ALJ, having noted that Licciardi's testimony left a lot to be desired, did not credit Licciardi's testimony. Further, Robinson maintains that credibility was not a relevant issue with regard to the other three FAA witnesses because they had no direct evidence about whether Robinson flew with only one engine operating. By contrast, Robinson points out that his testimony was corroborated by Carroll's (the salesman) testimony as well as the FAA witnesses' testimony that it was safe to fly the helicopter if the pulse system extinguished the chip light, and by Brennan's (the passenger) testimony about what Robinson told him on May 26th. In addition, Robinson maintains that Ford's (Raco Helicopter's president and director of maintenance) testimony "completely corroborated Robinson's version of what transpired on May 26, 1993," because Ford testified that when the helicopter landed at the Allaire Airport, the number two engine was too warm to handle. Ford explained that a nonoperating engine generates no heat and that fire walls prevent heat transfer from one engine to the other. In other words, Ford provided direct evidence, Robinson argues, that the number two engine was in operation during the flight from the Wall Street Heliport.

The latter point is critical. While the Board's order could properly rest on circumstantial evidence, see Sorenson v. National Transp. Safety Bd., 684 F.2d 683, 685 (10th Cir. 1982), the absence of eyewitness testimony from the FAA, combined with the absence of documentary evidence to pinpoint the time and number of chip lights, is significant because of the corroboration of key parts of Robinson's account. Brennan testified that he saw both engine levers up, indicating that the two engines were operating before Robinson left Wall Street. Ford testified that the second engine was too hot to handle when Robinson landed at Allaire Airport, indicating that it had been operating during the flight from Wall Street. No evidence in the record contradicts either witness' testimony.

The Board's order does not mention Ford's testimony, much less explain how the Board evaluated it. According appropriate deference to credibility determinations by the ALJ as between Robinson and Licciardi, the Board still could not ignore Ford's testimony. See Burlington Truck Lines, Inc. v. United States, 371 U.S. 156, 168, 83 S.Ct. 239, 245, 9 L.Ed.2d 207 (1962). Ford's testimony presented direct corroboration of a critical fact contrary to the basis for the revocation of Robinson's certificate, namely that he took off with only one operable engine. The evidence came, moreover, from a third party with direct knowledge of the heat of the number two engine at a critical time. So far as we can determine, unless it had decided to discredit Ford, there was no reason for the Board not to conclude that the second engine was "operable" at some point during the flight from the Wall Street Heliport. See Blackman v. Busey, 938 F.2d 659, 662 (6th Cir. 1991) (citing Chirino v. National Transp. Safety Bd., 849 F.2d 1525, 1532 (D.C. Cir. 1988)).

The Board concluded that the ALJ's credibility determination was bolstered by "abundant circumstantial evidence to support the law judge's conclusion that, despite his denials, [Robinson] experienced engine failure at Wall Street but took off nonetheless." The Board explained that the FAA's case did not rest solely on Licciardi's conclusions because "the undisputed sequence of events in this case (the aborted takeoff with the passenger after an admitted chip light, the discharge of the passenger, the running takeoff, and the subsequent repair of an engine due to `power failure') is strong circumstantial proof that [Robinson] experienced engine failure at Wall Street." However, the ALJ found both that Robinson left the Wall Street Heliport "under the power of a single engine," and that he departed from the Heliport "with only one operable engine." These are different findings, but their difference is not acknowledged in either the ALJ's findings or the Board's order. Neither finding is supported by more than Licciardi's observations and suppositions and the FAA's investigators' conclusions about them. The Board's reliance on the repair records was misplaced unless it also determined that Ford's testimony was in some manner deficient.

Furthermore, no witness for the FAA testified that one of the engines was "inoperable" at the critical time. FAA Investigator Winton testified only that the engine was either "inoperative" (i.e., not operating) or "at idle" at the time of the takeoff. Licciardi, the operations agent at the Wall Street Heliport, never testified whether the engine was operable at the time of the takeoff; he only observed what he concluded from a distance was Robinson executing a single engine takeoff. This appears consistent with Robinson's testimony that he initially began the takeoff with reduced power that would simulate a single engine, but had appropriate power thereafter, with both engines operating during the flight to Allaire Airport.

Thus, the circumstantial evidence on which the Board relied does not provide the missing evidence that is required to support the finding that the number two engine was inoperable when Robinson left Wall Street. The Board's approval of the ALJ's reliance on Licciardi's speculation and the FAA investigators' surprise that Robinson would take off after the chip light came on is flawed. While the ALJ acknowledged the extensive evidence about the operation of the chip light system, the ALJ gave no weight to the fact that if the chip light goes out the helicopter is safe to fly. Instead, the ALJ focused on Robinson's response to the chip light indication at the heliport that FAA inspectors characterized as surprising and foolhardy. The inspectors' conclusions about Robinson's conduct, however, rested on the assumption that he took off with only one operable engine.

Accordingly, we grant Robinson's petition for review and we remand the case to the Board to explain the basis for its finding that Robinson operated his helicopter with only one engine operable.

Remanded.