January 13, 1982 was a bitter winter's day at Washington's National Airport with temperatures reaching only minus five degrees Celsius by the afternoon. Cloud ceilings were 400ft with visibility limited to just over half a mile in snow showers. Among the many aircraft scheduled to depart National that day was Air Florida's 737 service to Ft. Lauderdale, call-sign Palm 90. Piloting that day's 2:15pm flight was Captain Larry Wheaton and First Officer Roger Pettit, both fairly young pilots enjoying their new jobs with Air Florida, a fairly new airline which expanded rapidly as the result of deregulation.

......Just before 1:40pm, the airport was closed so crews could plough off National's one instrument runway, scheduled to re-open at 2:30pm. Air Florida elected to begin boarding regardless and by 2:30pm, all of the 74 passengers were on board the aircraft. Meanwhile, Wheaton had asked for Palm 90 to be de-iced in preparation for departure. The tower told the aircraft that there would be further delay and that Palm 90 was number eleven in priority when the airport did re-open. Wheaton then instructed the de-icing to be discontinued, the port side have just been started. Half an hour later, Wheaton again called for de-icing as the airport was about to re-open.

By 3:10pm, de-icing was complete and ground personnel reported to Wheaton that there was only a "light dusting" or snow on the wings. Snow was still falling at national at the time. At 3:23pm, Palm 90 was cleared to push from the gate. The tug tried to push the 737, but snow that had accumulated on the ground caused it's tires to spin. Wheaton then suggested that, contrary to policy, they would use the aircraft's reverse thrust to assist in the push. The reversers were engaged for about a minute and a half, but were only successful in throwing up slush and snow. Another tug was brought in with chains and the aircraft was successfully pushed back.

......Palm 90 then taxied into position behind a New York Air DC-9 which was the last of sixteen aircraft in line for takeoff. Fifteen minutes later the New York Air aircraft was cleared for takeoff and Palm 90 was instructed to taxi into position and hold, being prepared for immediate takeoff. At 3:59pm, Palm 90 was cleared for takeoff as it was still positioning itself on the runway, visibility now down to a quarter mile.

......The departure from runway 36 requires aircraft to make a left 40 degree turn shortly after becoming airborne so as to follow the Potomac River and avoid flying over the Washington monument and the White House. The tower lost sight of Palm 90 during it's roll due to the reduced visibility, but radar showed it airborne and the tower controller instructed Palm 90 to contact the departure controller. Less than a minute after taking off, Palm 90 descended at low airspeed into the Rochambeau bridge and ploughed through into the Potomac river, only it's tail not submerged. Rescue crews attempted to reach the survivors, but icy conditions prevented the passengers from being reached except by helicopter. In the end, only six people survived.

......Crews were able to recover both the FDR and CVR from the bottom of the river and both devices proved critical in solving the puzzle of Palm 90. Knowing that it had been nearly 50 minutes between the aircraft's de-icing and takeoff, investigators were curious as to what had gone on in the aircraft in that span. Their first clue came when, during the after engine start checklist, the captain replied "off" to the First Officer's call for anti-ice. Though it seems hard to believe that the captain would reply "off" to anti-ice, extensive audio enhancement has given validity to the tape.

While waiting in line for takeoff, Wheaton positioned the aircraft behind the New York Air DC-9, attempting to use the aircraft's exhaust to melt the ice off Palm 90's fuselage and wings. Though Wheaton thought this a sound practice, in reality the exhaust will just melt the ice and blow it back over the wing, allowing it to re-freeze further back in areas which the aircraft's anti-ice system can not clear. While this information gave evidence of airframe icing, further analysis showed other problems with Palm 90. Shortly before takeoff, the crew have a brief discussion concerning anomalies in the engine instrument readings.

Pettit suggested that the hot (less dense) exhaust from the DC-9 ahead was causing a lower than normal reading on one of the EPR gauges. The indications seemed to return to near normal as Palm 90 got closer to takeoff. Just before takeoff, Pettit began the brief, calling out takeoff power as EPR 2.04, V1 as 138kts, Vr as 140kts, and V2 as 144kts. As Palm 90 was cleared for takeoff, Pettit advanced the throttle and immediately remarked at the abnormal indications from the EPR gauges again. Pettit remarked several time that it was "real cold", indicating that the engines indicated the takeoff EPR of 2.04 quickly before the throttles were fully advanced. Spectrum analysis of the engine sounds from the CVR indicated that the engines were actually running at an approximate EPR of 1.70 throughout the takeoff.

In studying the engines for signs of the anomaly, investigators found that the engine de-ice system was turned off. In re-creating the conditions, investigators confirmed that ice on the compressor inlet pressure probe would cause a higher than actual thrust reading on the EPR gauges. First Officer Pettit seemed to be aware of the anomaly during takeoff, but did not appear to have any idea what was causing it. Pettit remarked several times that "that doesn't seem right", meaning that the low throttle setting was producing a high EPR reading while the aircraft was not accelerating properly.

Still, 45 seconds into the takeoff roll, Palm 90 reached it's rotation speed and pitched up abruptly, causing Captain Wheaton to exclaim "Easy!" and then, as the stall warning came on, "Forward! Forward!", indicating to Pettit to lower the nose to prevent the stall. Investigators found that ice build up in the wing leading edge and slats could cause an abrupt nose up pitch on takeoff. Pettit apparently believed that the engines were producing max thrust because at no time during the 30 second flight were the throttles advanced to provide more power to prevent stall.

The recording ended with the crew's final acknowledgement of the severity of their situation. "Larry-we're going down Larry!". "I know it!".