Accident Overview

Photo of snow plow operations
Photo of snow plow operations
Photo copyright Kevin Wachter - used with permission

Photo of AirFlorida 737 in a snow storm
Photo of AirFlorda 737 in a snow storm

Air Florida Flight 90, a Boeing Model 737-200 series airplane, was a scheduled flight from Washington, DC to Fort Lauderdale, Florida, with an intermediate stop in Tampa.

There were 74 passengers and a crew of four on board. Approximate time of departure for Flight 90 was originally scheduled for 1415 local time. Snow had been falling at the airport all morning and into the afternoon. At 1338 the airport closed all operations for snow removal and was reopened at 1453.

This resulted in Flight 90 being delayed approximately one hour and 45 minutes. During this delay snow continued to fall, accumulating on all aircraft being held, including Flight 90.

During the airport closure, deicing operations were being conducted on all aircraft being prepared for departure, which resulted in a need for flightcrews and deicing crews to schedule deicing services in the general sequence of the aircraft being cleared for departure. There were several aircraft cleared for pushback ahead of Flight 90. The flightcrew of Flight 90 had communicated their desire for the deicing to begin just prior to the estimated reopening of the airport, being at approximately 1430. Since the opening had been delayed from the original estimated time, deicing that had begun on Flight 90 prior to 1430 was requested to be stopped by the flightcrew. Once the airport opening was announced, the flightcrew requested the deicing be resumed, which began again at approximately 1450. The snowfall rate was very heavy during the deicing process. At the conclusion of the deicing process, at approximately 1510, the ground crew noticed about two to three inches of snow on the ramp around Flight 90 at Gate 12 where the airplane was parked.

Series of photos of aircraft being de-iced 
Photos of aircraft being de-iced
Photos copyright Richard Barsby, Olaf Juergensmeier, Robin Schmitt-Opitz (Rotate), and Vasco Garcia - Used with permission

Photos of thrust reversers deployed
Photos of thrust reversers deployed
Bottom photo copyright Kazutaka Yagi
- used with permission

At 1515, the airplane main entry door was closed and the jetway retracted. At this time an airline station manager, who was on the jetway, was asked by the pilot how much snow was on the aircraft. The manager responded that there was a dusting on the left wing from the engine to the tip and the other areas were clean. At 1525 the tug attempted to push Flight 90 away from the gate. A combination of the slope of the ramp and the snow and slush accumulated around the tires of the tug (not equipped with tire chains) resulted in the tug being unable to move the airplane. At that point the flightcrew started the engines and deployed the reversers in an attempt to assist the tug in moving the airplane.

Prior to this attempt, the operator of the tug advised the flightcrew that use of reversers was against company policy. The operator of the tug was from a contract service provider and not an employee of Air Florida. It is estimated that the engines were in reverse thrust for 30 to 90 seconds, during which a significant amount of snow and ice was observed blowing around the aircraft from the engines.

At approximately 1533, the initial tug was disconnected from Flight 90 and another tug equipped with chains was requested. At this time an assistant station manager for Air Florida, who was inside the terminal between Gates 11 and 12, stated that he saw a small amount of snow on top of the fuselage and radome up to the bottom of the windshield and a light dusting on the left wing. At 1535, Flight 90 was pushed away from the gate and both engines started. The airplane moved into a taxi line up of several other airplanes at approximately 1539.

During the post engine start check list, the cockpit voice recorder (CVR) recorded the captain respond "off" to a checklist item "anti-ice." The CVR also recorded several comments between the captain and first officer concerning the subject of ice on the airplane. During this conversation, at 1546, the first officer made a comment, "Well, all we need is the inside of the wings anyway, the wingtips are going to speed up on eighty anyway, they'll shuck all that other stuff." There was then an exchange concerning the amount of ice they were able to observe on their respective wings. The captain commented, "I got a little on mine." The first officer replied, "A little?  This one's got about a quarter to half an inch on it all the way."

Diagram of Engine Pressure Ratio Indicating System
Diagram of Engine Pressure Ratio Indicating System
( View Large Image)
There was also conversation concerning the engine-indication instruments. At 1549, the first officer asked, "See this difference in that left engine and right one?" It was believed by the investigators that the first officer was observing a difference in the EPR being displayed between the left and right engines while the engines were operating at idle thrust.  EPR is an indication of engine thrust on the Pratt and Whitney JT8D engines installed on the Boeing Model 737-200, and the difference in the displayed values was believed to be due to one of the engine's inlet probes having been plugged with ice.

The first officer then further commented, "I don't know why that's different-less, 'less it's hot air going into that right one.  That must be it--from his exhaust; it was doing that at the chocks awhile ago... Ah."

At 1552, the captain said, "Don't do that-Apple, I need to get the other wing done." It was believed that both the reference to the "hot air going into the engine" and the need to "get the other wing done" was making reference to the hot engine exhaust from the New York Air McDonnell Douglas Model DC-9 airplane that Flight 90 was behind on the taxiway. "Apple" was an air traffic controller term used to identify the New York Air airplane. There appeared to be an expectation by the flightcrew of Flight 90 that the ice that had accumulated on the Air Florida Model 737-200's wings would benefit by holding behind the Model DC-9 that was in front of them, and the hot air being ingested by the engines of Flight 90 was influencing the EPR reading.

At 1553, the first officer remarked, "Boy...this is a losing battle here on trying to deice those things. It (gives) you a false feeling of security; that's all it does." The captain and first officer continued a general discussion of deicing until 1554. At 1558, the New York Air Model DC-9 was cleared for takeoff, and the flightcrew of Flight 90 completed the pre-takeoff checklist. The CVR recorded verification that the takeoff EPR target thrust setting was 2.04 and the V1 (critical engine failure recognition speed), VR (rotation speed), and V2 (minimum safe speed in the second segment of a climb following an engine failure) speeds to be 138 knots (kns), 140 kns and 144 kns, respectively. It was later confirmed that these were the appropriate takeoff thrust and speed schedules for the Model 737-200, based upon the prevailing airport conditions and the takeoff weight of Flight 90.

Photos of Air Florida Flight 90 recovery scene
Photos of Air Florida Flight 90 recovery scene

At 1559, Flight 90 was cleared to "taxi into position and hold" and to "be ready for an immediate takeoff." At 1559:28, Flight 90 was cleared for takeoff and told not to delay due to landing traffic which was 2 ½ miles out on approach. At 1559:45 Flight 90 turned on to runway 36 and the pilot told the first officer "your throttles." At 1559:56 the CVR recorded the sound of the engines spooling up and the captain saying "real cold, real cold." It was believed by the investigators that the captain may have been referring to the temperature of the engines' exhaust gas temperature (EGT) indications for both engines, since it was later established that the engines were producing significantly less than the target thrust due to the EPR gauges reading erroneously high due to plugged inlet probes. This resulted in the EGT readings being significantly lower than what would normally be expected for takeoff thrust settings. Since the EPR gauges were both set to 2.04 and were apparently indicating 2.04, the crew believed that takeoff thrust was being produced.

(View Air Florida PT2 Probe Animation)

During the early part of the takeoff roll at 1559:58, the first officer, who was performing the takeoff, commented "God, look at that thing! That don't seem right, does it?" At 1600:05, the first officer again remarked, "...that's not right...," to which the captain responded, "Yes it is, there is eighty." The first officer reiterated, "Naw, I don't think that's right." It was believed by the investigators that the first officer was referring to one or more of the following conditions: (1) the general slow acceleration of the airplane, (2) the lower than normal engine noise, and (3) the position of the engine thrust levers, which would have been significantly less forward than for a normal takeoff using normal takeoff thrust settings. At 1600:20, the first officer commented, "...maybe it is," but then, two seconds later, after the captain called, "Hundred and twenty," the first officer said, "I don't know."

Photo of the 14th Street Bridge
Photo of the 14th Street Bridge

The CVR recorded the voice of the captain calling out V1 followed by V2 six seconds later. The next sound recorded, two seconds after the call out of V2, was the sound of the stick shaker. At 1600:45, the captain said "Forward, forward," and at 1600:48, "We only want five hundred." At 1600:50, the captain continued, "Come on, forward, forward, just barely climb." At 1601, the first officer said "Larry, we're going down, Larry," to which the captain responded, "I know it."

(View Air Florida takeoff animation)

Diagram of flight path
Diagram of flight path
(View Larger Image)
At approximately 1601, Air Florida Flight 90 struck the northbound span of the heavily congested 14th Street Bridge, which connects the District of Columbia with Arlington County, Virginia, and plunged into the ice-covered Potomac River.

The initial impact was near the west end of the bridge, 3/4 mile from the departure end of Runway 36. Heavy snow was continuing to fall and visibility at the airport was between 1/4 mile and 5/8 mile.

Of the 74 passengers and five crewmembers on board, 70 passengers and four crewmembers were killed, including both the captain and first officer. There were also four occupants in vehicles on the bridge killed and several injured.

Post Accident Testing

Following the Air Florida accident, Boeing conducted flight testing of a Model 737-200, which had been modified to simulate snow and ice contamination on the wing leading edges and upper surfaces (see photographs below) to better understand the effects of snow and ice. Extensive aerodynamic testing was accomplished which confirmed the importance of adhering to the "Clean Airplane" concept for dispatch (lift curve). It was determined that even moderate amounts of contamination, especially on the leading edges, would have a significant impact on climb performance. Although this testing was conducted on a Model 737-200, the same conclusions were believed to generally apply to all airplanes.

Photos of Boeing Model 737-200 snow and ice contamination simulation
Photos of Boeing Model 737-200 snow and ice contamination simulation
(view large photos)

Back to top