Nightfall comes early in that part of the world during the winter and it was already dark with light sleet falling as 09C began servicing on the ramp. The aircraft was sprayed down with heated ethylene glycol to remove ice that had accumulated as it sat in the freezing precipitation. With the passengers loaded and the 10 crew members in place, 09C taxied away from the terminal to runway 06R shortly before 5pm. The wind was light and visibility was just under 5 miles with and overcast layer at 1600ft. Downs would be handling the controls for the takeoff, but Reid taxied the aircraft into position on the icy runway and informed the crew that he would handle the brakes throughout the takeoff. Even though the aircraft was near maximum gross weight, with some 10,000ft of available runway, they were well within limits of making their Vr speed of 153kts.

When cleared for takeoff, Reid held the brakes while advancing the power to 80%. He then released the brakes and Downs maintained directional control as Reid brought the engines up to full takeoff power. The aircraft appeared to accelerate normally, but V1 came a few moments late. Acceleration the became increasingly sluggish, but the aircraft achieved Vr with over 1,500ft of runway remaining. Downs brought the nose of the aircraft up into the takeoff attitude, but for some reason, the aircraft failed to climb away from the runway. Six seconds later, still in the nose-up attitude, the aircraft ran out of runway. Reid closed the throttles and attempted to stop the aircraft, but it ploughed through navigation structures and hit a drainage ditch which caused the aircraft to break up. Fuel from both wings ignited and the aircraft slid to a stop 3,000ft from the end of the runway.

 Reid, uninjured, exited through his cockpit window and immediately began evacuating passengers. Finding the flight deck door blocked, he went back outside and entered Downs' window and assisted Downs, Fink, and Leonard out to safety. Despite the flight crew's efforts and the quick response of fire crews, 46 people died in the post-crash fire.

......Several of the survivors reported that about halfway through the takeoff roll, there were several loud explosions which they believed to be the main tires blowing out. These sounds were not audible from the flightdeck and there was no accompanying vibration. Examination of the runway area revealed evidence to support the passengers claims however. The tire tracks appeared normal from the taxiway to the runway holding point. There was a clear "footprint" where the warm tires had melted through the ice while the aircraft waited in position for takeoff on the runway. From this point on, however, skid marks were found along the entire length of the runway and evidence of tire deterioration increased along the takeoff path.

The first piece of tire rubber was found only about 500ft down the runway. By about 4,000ft, all of the port tires had blown out and all the starboard tires by 8,000ft. Examination of the wheels still intact after impact showed that they had all been ground down on only one side indicating that they were not rotating during the takeoff roll. The brakes all appeared to be in normal working order and the parking brake was found in the "off" position. No evidence of any malfunction of the system was found.

......The NTSB requested that NASA (National Aeronautics and Space Administration) conduct tests on the rolling and sliding friction caused by aircraft tyres. The concern was that the tire could produce a sliding coefficient of friction so low on ice that they would fail to spin up after brake release. NASA concluded in all cases that the tires would release and spin up normally. However, further testing revealed an interesting discovery. The initial tests were conducted on frosted and glazed ice which resulted in friction coefficients which were still considerably higher than the normal wheels-rolling friction coefficient. 09C, however, had been sitting in position which allowed the ice directly in contact with the tires to melt.

This liquid water on top of the ice reduced the friction coefficient to such a degree that it was almost identical to the normal wheels-rolling friction coefficient. Under these circumstances, the aircraft could initially accelerate without the wheels rolling with little hesitation. Analysis of the FDR showed that the aircraft accelerated almost normally up to about 100kts. After this, the deterioration of the tire surface began to increase the friction coefficient and progressively decrease performance.

The aircraft only reached a maximum speed of 1 knot below Vr. Investigators believed, however, that the wheels should have begun to spin normally unless brake pressure was held prior to the takeoff role and not released. No evidence was ever found to support this and both crew members were sure that the brakes had been released prior to the aircraft's takeoff role. Although the FDR showed evidence of decreased performance after passing through 100kts, the progressive nature of the problem hindered the crew's ability to make a prompt and correct analysis and correction of the difficulties. The crew's actions were not cited as contributing to the accident, but the NTSB suggested that FAA should revise takeoff procedures to allow crews to evaluate takeoff performance at all points in the takeoff roll.

Editorial Note
...... As I stated above, investigators believed that brake pressure had to be applied AND HELD prior to the takeoff roll in order for the tires to have slid throughout the takeoff roll. It was clear that Reid had applied the toe brakes prior to bringing up power, but he then released them. Downs supported this by saying that he felt no pressure on the pedals as he controlled the aircraft on takeoff. Barring a failure of the hydraulic system, which was damaged by fire, there was no evidence in the mechanical brake system to indicate that they were applied. This was a highly experienced crew, each with nearly 14,000 hours, all having over 2,000 hours on the DC-8 itself. They showed good cockpit organization and adherence to checklists and procedures.

I believe that further study of the frictional data may indicate what actually happened. Breakaway friction coefficient is determined by thrust to weight ratios. With the aircraft at its takeoff weight, only 65% power would be required to get the aircraft to skid with the brakes ON. We know that Reid brought the power to 80% prior to brake release. He immediately released the brakes and applied full takeoff power.

It seems likely that the aircraft actually started to skid when power was brought up above 65%. With the smooth application of power by Reid, this skidding was probably not detectable until he had already released the brakes and the crew would then believe themselves to be accelerating normally. At this point, with the wheels already skidding on the water-covered ice, releasing the brakes would not be sufficient to get the wheels to spin up.