Pan American World Airways flight 845
Boeing introduced the Boeing 747 to airline operations beginning with Pan American World Airways on January 21, 1970 (Zhang, 2016). The airplane weighed significantly more than the popular Boeing 707 (maximum takeoff weight 333,680 pounds) or Douglas DC-8 (maximum takeoff weight 325,000 pounds), with a maximum takeoff weight of 713,000 pounds (Plane & Pilot, 2009). Only 19 months after the Boeing 747’s introduction to passenger service, the crew and passengers onboard N747PA underwent a harrowing ordeal traced back to dispatcher errors during preflight performance and maximum weight planning (National Transportation Safety Board [NTSB], 1972, p. 1).
N747PA was a historic aircraft for Pan Am. It was the same aircraft Pan Am had used for its first commercial service in 1971 (Glionna, 2010) On July 30, 1971, dispatchers John Pepin, Francis Keithey, and Edward Anderson (NTSB, 1972, p. 36), stationed in the Western Area Headquarters Dispatch Office at San Francisco International Airport, prepared dispatch release documents (NTSB, 1972, p. 6). The dispatcher assigned to Pan Am flight 845, scheduled to fly from San Francisco to Tokyo, planned the takeoff on runway 28L which was 10,600 feet long. He selected runway 01R as an alternate runway for planning purposes. He failed to check airport conditions; had he checked, he would have discovered that runway 28L closed earlier in the day (NTSB, 1972, p. 7).
The crew of Pan Am 845 radioed the dispatch office to request a departure runway change to runway 01R (NTSB, 1972, p. 7). The dispatcher needed to recompute the maximum available takeoff weight based on a different runway with a different length and set of obstacles off the end of the runway. Further reducing the available takeoff distance on runway 01R was a one thousand foot stretch at the beginning of the runway designated as a blast overrun area (NTSB, 1972, pp. 7-8). Unfortunately, the dispatcher did not realize that the 9,500-foot runway distance shown in his Pan American Route Manual was incorrect. The distance available was only 8,500 feet because of the first 1,000 feet was rendered unusable by the blast overrun area. The dispatcher checked the performance weight calculations against the clearway computations kept in the Pan American Route Manual at the dispatch office and determined that there were no takeoff limitations at the planned takeoff weight of 708,002 pounds (NTSB, 1972, pp. 3-4). 14 CFR §1.1 defines a clearway as an area beyond the departure end of a runway that is free of obstacles protruding into an upward sloping area of 1.25 percent.
Prior to this exchange with the dispatcher, the crew set takeoff speed “bugs” for the V-1 takeoff decision speed, V-R rotation speed, and V-2 takeoff safety speed. These speeds were set with the assumption of a flap setting of 10 degrees which would have applied to a departure from runway 28L (NTSB, 1972, p. 3). During the conversation with the dispatcher about the departure runway change, the dispatcher informed the crew that the takeoff would require a 20 degree flap setting and a 3A-Wet power setting (NTSB, 1972, p. 4). The crew reset the flaps to the appropriate 20 degree setting, but they did not recheck the speed computations for a flap setting of 20 degrees. The speed “bugs” remained set at the speeds calculated for the initial setting of 10 degrees of flaps (NTSB, 1972, pp. 4-5).
The crew completed all other preflight preparations and began the takeoff at 3:28 PM. The takeoff seemed normal until the first officer noticed the end of the runway “coming up at a very rapid speed” (NTSB, 1972, p. 5). He called for the captain to rotate the aircraft to the takeoff attitude not because the appropriate V-R speed was reached, but because the flight was rapidly reaching the end of the runway. Passing through 165 knots, the crew felt a bump (NTSB, 1972, p. 5). Horrified passengers in the rear of the cabin saw three pieces of metal from the runway’s approach lighting system penetrate up through the aircraft’s floor, severely injuring two passengers. One of the pieces impaled four seats, but the seats were unoccupied. The third piece flew through rear seats and the rear lavatories (NTSB, 1972, p. 14).
With the fuselage seriously damaged, the flight engineer found hydraulic systems one, three, and four unusable due to fluid leaks. The captain maneuvered the aircraft to check remaining flight control systems and flew over the ocean to begin dumping fuel in preparation for an emergency landing (NTSB, 1972, p. 5). Dumping the necessary 180,000 pounds of fuel took about 45 minutes, after which time the crew headed back to San Francisco to attempt a landing. Due to the damaged hydraulic systems, the captain did not have full elevator control authority, and the aircraft touched down hard and bounced back into the air. After landing again, it veered off the right side of the runway and came to a rest in the intersection of four runways (NTSB, 1972, p. 6).
During the subsequent evacuation, chaos ensued. The flight deck crew did not make a public-address announcement directing the cabin crew to evacuate the aircraft. The evacuation commenced after the crew came down the stairs from the upper deck and shouted to the cabin crew to begin the evacuation (NTSB, 1972, p. 14). Several exit slides in the forward cabin failed, and passengers rushed to usable exits in the rear of the aircraft. The movement of the passengers to the rear combined with partial landing gear failures caused the airplane to seesaw back and forth, eventually settling on its tail with the nose gear completely off the ground (NTSB, 1972, p. 24). As the tail settled, the first exit slides in the forward cabin pointed straight to the ground in a nearly vertical position. Eight passengers were seriously injured as they plunged down on these now-ineffective escape slides (NTSB, 1972, p. 15).
In its final analysis of the accident, the National Transportation Safety Board (1972) calculated that given the existing takeoff conditions, the maximum takeoff gross weight for the available 8,400 feet of runway with clearway should have been limited to 697,400 pounds (p. 18). The aircraft’s loaded weight that day of 708,000 pounds required a runway length for takeoff of 8,675 feet plus clearway (NTSB, 1972, p. 18). The NTSB cited these causal factors directly relating to dispatching:
The dispatcher prepared a flight release for the longest runway on the airport without ascertaining the status of the runway (NTSB, 1972, p. 17).
The dispatcher was “lulled into a sense of complacency” (NTSB, 1972, p. 17).
There were no formal procedures for briefing the dispatcher on abnormalities of the operation in effect or expected to happen (NTSB, 1972, p. 17).
The Pan American Route Manual contained errors (NTSB, 1972, p. 21).
The dispatcher failed to propose use of runway 28R to the flight crew even though wind favored it. He simply reverted to his original alternate runway plan of 01R (NTSB, 1972, p. 20).
The probable cause determined by the NTSB (1972) was: "…the pilot’s use of incorrect takeoff reference speeds. This resulted from a series of irregularities involving (1) the collection and dissemination of airport information; (2) aircraft dispatching; and (3) crew management and discipline; which collectively rendered ineffective the air carrier’s operational control system. (p. 1)."
The accident report notably listed the three Pan Am dispatchers’ names and dispatcher certificate information. No previous aircraft accident reports included dispatcher names along with the flight crewmember names. Ultimately, dispatchers and pilots share a tremendous amount of responsibility for operational control and safe operations. This accident clearly illustrates the potentially fatal consequences that can occur when operational control breaks down even in the preflight planning process.
Image credit: Edward Marmet / CC-BY-SA-3.0
Watch Pan Am 845's emergency landing
This vintage airline training video, while of poor quality, dramatically illustrates emergency landing and chaos during the subsequent evacuation.