A flight crew begins work by meeting in the flight planning and despatch area at an airport some 45 minutes to an hour prior to departure. They will review company notices, air traffic control notices, and the weather forecasts and reports. They will calculate the minimum amount of fuel required for their flight. Often there are special issues to be considered; the carriage of cargo requiring special handling and notification, or the presence of VIPs. On reaching the aircraft, they will check the aircraft’s maintenance records, and review any allowed technical defects that may require alternative procedures to be applied by the crew.
The horizontal path
Despite advances in technology, this procedure has changed little in decades.
The flight crew of TE901 had an additional chore to perform, to check the pre-programmed flight route provided by the Navigation section which was stored within the aircraft on board library. . This involved checking a series of positions along the route,, known as waypoints, between the flight plan and the on-board computer. One pilot would load the route then both pilots would check that the on-board computer entries matched the hard copy co-ordinates on the flight plan, which were in the latitude and longitude format. They expected that, provided they had performed this data checking task with no error that the aircraft’s computer brain would, if left to itself, fly down the 40 mile wide McMurdo Sound.
They expected this because 19 days earlier, in a briefing given to them by flight operations personnel, they were shown via an audio-visual presentation scenes depicting flight down McMurdo Sound. They were also shown a copy of the current computer generated flight plan. Later, using the co-ordinates from that flight plan, Captain Jim Collins would plot the route on a chart at home so he could show his daughters where he would be going. He described McMurdo Sound, and its western side, saying that the aircraft would keep close to that side. He did not mention Mt Erebus.
On the night before the flight, flight planners made what they thought was a small correction to an earlier mistake made some months previously, when the flight plans were computerised. In their mind, they were only shifting the final Antarctic waypoint about two miles, which was about the expected error usually found in flights of a similar duration. Now, according to the airline witnesses, the navigation staff always knew the track ran more or less directly over Mt Erebus, and the shifting of the waypoint some two miles would still run the track more or less over the volcano. But in fact, what they had done was to shift the route from McMurdo Sound, to over Mt Erebus, a change of nearly 30 nautical miles. So how did this happen?
Originally, the route was over Mt Erebus, and the first few flights were planned and operated that way. Most of these initial flights were able to operate away from the planned nav track in brilliant weather conditions ensuring good surface and horizon definition. Then, some 18 months after Antarctic flights began, 14 months prior to the accident, the airline computerised its method of storing and producing flight plans. At this time, an error was made; the data entry operator keyed a 4 instead of a 6 into the machine, and this error had never been noticed as such. The effect of this erroneous keystroke was to shift the route nearly 30 miles to the west, so that it ran down McMurdo Sound. One of the reasons that no-one saw this as an error was that it seemed a logical change to make; shifting the route away from an active volcano and down over the flat sea ice of McMurdo Sound actually produced a better, safer result. This produced an end waypoint near the Dailey Islands, and was sometimes referred to as the “Dailey Islands” waypoint. But previous aircraft seldom had to fly all the way to that point. It became almost normal practice to disengage this mode of navigation, known as “navigation (or nav.) mode” and fly in “heading mode” which, rather than fly along the “nav. track,” allowed the crew to steer simply by selecting directions to fly at random, that is, pointing the aircraft in the direction they wanted to go. This facilitated sightseeing.
For the next 14 months flights flew this altered route, and flew sightseeing missions along it, and it became such that crews would be briefed on this route that terminated at the Dailey Islands. Hence Capt. Collins’ comment to his daughters the night before, that that would be where the aircraft was to operate.
If the airline’s testimony was correct, the airline, its flight planners, and management staff all still believed the track lay over Mt Erebus. Yet the flight crews, briefing officers, and publicity staff all knew that the track ran down McMurdo Sound. Thus, the night before the accident, when those same flight planners thought they were making a small correction, they were in fact making a much larger alteration to the flight path. The crews’ minds were programmed for a path down McMurdo; but the aircraft was now programmed on a path towards Mt Erebus. This was a fundamental disconnect between man and machine; a loss of what aviators call “situational awareness.” Through little fault of their own, this crew lost situational awareness before the aircraft had left the ground, because the pre-programmed route loaded into the aircraft’s navigation system was not the same as that presented to them at their route briefing days before.
The crew believed the DC10’s guidance systems were taking the aircraft on the “nav track” along the expected flight path on the left (diagram above) leading into the safety of McMurdo Sound. The figure of eight orbit was the procedure which the crew employed in order to descend through one of the large breaks in the clouds to reach clear air beneath the cloud layer. No cloud penetration was required. Capt. Collins continued to lock the aircraft back onto “nav track” during his descent. That flight path would have taken them up the safety of McMurdo Sound, close to the US military flight path, to Air New Zealand’s de facto waypoint near West Dailey Island which the airline had used for 14 months. The pilots also believed they were located on the expected left hand track (diagram above) because the observations they were making of geographical features corresponded with those features they would be expecting to observe. Those features were on the points of the compass they would be expecting. Furthermore the experienced commentator on board, Peter Mulgrew, made his own observations of features, in which he also identified the aircraft’s location as being on this expected track in McMurdo Sound. His comments on those landmarks follow. The dotted lines show where Mulgrew was most probably looking as he made his comments and the times show the time he made each comment. 43:27 means 0043:27 GMT.
Because the airline altered the last leg of the flight plan by 26-28 miles to the east at 2.10 am on the morning of the flight and failed to tell the crew, the DC10 was in reality flying the actual flight path on the right (diagram above). Instead of flying into McMurdo Sound on the plotted flight path as the crew believed, they were in fact flying into Lewis Bay . The actual observation of landmarks which Mulgrew was making is shown by the dotted lines on the right hand side of the diagram. For example, Cape Bird appeared to be Cape Bernacchi because estimation of distance is difficult in polar conditions and the lower height of Bird compensated for its shorter distance. Cape Tennyson looked the same as the crew would expect Cape Royds to look. A fog “ramp” concealed the cliffs in Lewis Bay which now was directly ahead on the DC10’s track. They were flying beneath cloud and the insidious optical illusion of sector whiteout made the terrain in front appear to be the flat open expanse of McMurdo Sound. At each side the black rock of Cape Bird and Cape Tennyson which the crew mistakenly identified as Bernacchi and Royds respectively gave the illusion that their observations of landmarks confirmed the DC10 was exactly where they expected it to be while established on the nav. track.