Mount Erebus disaster

Captain Jim Collins, 45, and co-pilot Greg Cassin, 37, had never flown to Antarctica before (while flight engineer Gordon Brooks had flown to Antarctica only once previously), but they were experienced pilots and were considered qualified for the flight. On 9 November 1979, 19 days before departure, the two pilots attended a briefing in which they were given a copy of the previous flight's flight plan.[3]

The flight plan that had been approved in 1977 by the Civil Aviation Division of the New Zealand Department of Transport was along a track directly from Cape Hallett to the McMurdo non-directional beacon (NDB), which, coincidentally, entailed flying almost directly over the 12,448-foot (3,794 m) peak of Mount Erebus, but because of a typing error in the coordinates when the route was computerised, the printout from Air New Zealand's ground computer system presented at the 9 November briefing corresponded to a southerly flight path down the middle of the wide McMurdo Sound, about 27 miles (43 km) to the west of Mount Erebus.[8] The majority of the previous 13 flights had also entered this flight plan's coordinates into their aircraft navigational systems and flown the McMurdo Sound route, unaware that the route flown did not correspond with the approved route.[9]

Captain Leslie Simpson, the pilot of a flight on 14 November and also present at the 9 November briefing,[10] compared the coordinates of the McMurdo TACAN navigation beacon (about 5 km (3.1 mi) east of McMurdo NDB), and the McMurdo waypoint that his flight crew had entered into the inertial navigation system (INS), and was surprised to find a large distance between the two. After his flight, Captain Simpson advised Air New Zealand's navigation section of the difference in positions. For reasons that were disputed, this triggered Air New Zealand's navigation section to resolve to update the McMurdo waypoint coordinates stored in the ground computer to correspond with the coordinates of the McMurdo TACAN beacon, despite this also not corresponding with the approved route.[8]

The navigation section changed the McMurdo waypoint co-ordinate stored in the ground computer system around 1:40 am on the morning of the flight. Crucially, the flight crew of Flight 901 was not notified of the change. The flight plan printout given to the crew on the morning of the flight, which was subsequently entered by them into the aircraft's INS, differed from the flight plan presented at the 9 November briefing and from Captain Collins' map mark-ups, which he had prepared the night before the fatal flight. The key difference was that the flight plan presented at the briefing corresponded to a track down McMurdo Sound, giving Mount Erebus a wide berth to the east, whereas the flight plan printed on the morning of the flight corresponded to a track that coincided with Mount Erebus, which would result in a collision with Mount Erebus if this leg were flown at an altitude less than 13,000 feet (4,000 m).[9]

The computer program was altered such that the standard telex forwarded to US air traffic controllers (ATCs) at the United States Antarctic science facility at McMurdo Station displayed the word "McMurdo", rather than the coordinates of latitude and longitude, for the final waypoint. During the subsequent inquiry, Justice Mahon concluded that this was a deliberate attempt to conceal from the United States authorities that the flight plan had been changed, and probably because it was known that US Air Traffic Control would lodge an objection to the new flight path.[11]

The flight had earlier paused during the approach to McMurdo Sound to carry out a descent, via a figure-eight manoeuvre, through a gap in the low cloud base (later estimated to be at around 2,000 to 3,000 feet (610 to 910 m)) while over water to establish visual contact with surface landmarks and give the passengers a better view.[12] The flight crew either was unaware of or ignored the approved route's minimum safe altitude (MSA) of 16,000 feet (4,900 m) for the approach to Mount Erebus, and 6,000 feet (1,800 m) in the sector south of Mount Erebus (and then only when the cloud base was at 7,000 feet (2,100 m) or better). Photographs and news stories from previous flights showed that many of these had also been flown at levels substantially below the route's MSA. In addition, preflight briefings for previous flights had approved descents to any altitude authorised by the US ATC at McMurdo Station. As the US ATC expected Flight 901 to follow the same route as previous flights down McMurdo Sound, and in accordance with the route waypoints previously advised by Air New Zealand to them, the ATC advised Flight 901 that it had a radar that could let them down to 1,500 feet (460 m). However, the radar equipment did not pick up the aircraft, and the crew also experienced difficulty establishing VHF communications. The distance-measuring equipment did not lock onto the McMurdo Tactical Air Navigation System (TACAN) for any useful period.[13]^:7

Cockpit voice recorder transcripts from the last minutes of the flight before impact with Mount Erebus indicated that the flight crew believed they were flying over McMurdo Sound, well to the west of Mount Erebus and with the Ross Ice Shelf visible on the horizon, when in reality they were flying directly toward the mountain. Despite most of the crew being engaged in identifying visual landmarks at the time, they never perceived the mountain directly in front of them. About 6 minutes after completing a descent in visual meteorological conditions, Flight 901 collided with the mountain at an altitude around 1,500 feet (460 m), on the lower slopes of the 12,448-ft-tall (3,794 m) mountain. Passenger photographs taken seconds before the collision removed all doubt of a "flying in cloud" theory, showing perfectly clear visibility well beneath the cloud base, with landmarks 13 miles (21 km) to the left and 10 miles (16 km) to the right of the aircraft visible.[14]

The crew put the coordinates into the plane's computer before they departed at 7:21 am from Auckland International Airport. Unknown to them, the coordinates had been modified earlier that morning to correct the error introduced previously and undetected until then. The crew evidently did not check the destination waypoint against a topographical map (as did Captain Simpson on the flight of 14 November) or they would have noticed the change. Charts for the Antarctic were not available to the pilot for planning purposes, being withheld until the flight was about to depart. The charts eventually provided, which were carried on the aircraft, were neither comprehensive enough nor large enough in scale to support detailed plotting.[13]^:29

These new coordinates changed the flight plan to track 27 miles (43 km) east of their understanding. The coordinates programmed the plane to overfly Mount Erebus, a 12,448-foot-high (3,794 m) volcano, instead of down McMurdo Sound.[3]

About four hours after a smooth take-off, the flight was 42 miles (68 km) away from McMurdo Station. The radio communications centre there allowed the pilots to descend to 10,000 ft (3,000 m) and to continue "visually". Air-safety regulations at the time did not allow flights to descend to lower than 6,000 ft (1,800 m), even in good weather, although Air New Zealand's own travel magazine showed photographs of previous flights clearly operating below 6,000 ft (1,800 m). Collins believed the plane was over open water.[3]

Flight path of Flight 901

Collins told McMurdo Station that he would be dropping to 2,000 feet (610 m), at which point he switched control of the aircraft to the automated computer system. Outside, a layer of clouds blended with the white of the snow-covered volcano, forming a sector whiteout – no contrast between the two was seen to warn the pilots. The effect deceived everyone on the flight deck, making them believe that the white mountainside was the Ross Ice Shelf, a huge expanse of floating ice derived from the great ice sheets of Antarctica, which was in fact now behind the mountain. As it was little understood, even by experienced polar pilots, Air New Zealand had provided no training for the flight crew on the sector whiteout phenomenon. Consequently, the crew thought they were flying along McMurdo Sound, when they were actually flying over Lewis Bay in front of Mt. Erebus.[3]

At 12:49 pm, the ground proximity warning system (GPWS) began sounding a series of "whoop, whoop, pull up" alarms, warning that the plane was dangerously close to terrain. The CVR recorded the following:[nb 2]

• GPWS: "Whoop whoop. Pull up. Whoop whoop..."
• F/E: "Five hundred feet."
• GPWS: "...Pull up."
• F/E: "Four hundred feet."
• GPWS: "Whoop, whoop. Pull up. Whoop whoop. Pull up."
• CA: "Go-around power please."
• GPWS: "Whoop, whoop. Pull-"
• CAM: [Sound of impact]

The go-around power was immediately applied, but it was too late.[15][16] No time remained to divert the aircraft, and 6 seconds later, the plane crashed into the side of Mount Erebus and exploded, instantly killing everyone on board. The accident occurred at 12:50 pm at a position of 77°25′30″S 167°27′30″E and an elevation of 1,467 feet (447 m) above mean sea level.[13]

McMurdo Station attempted to contact the flight after the crash, and informed Air New Zealand headquarters in Auckland that communication with the aircraft had been lost. United States search-and-rescue personnel were placed on standby.[3]

Air New Zealand had not lost any passengers to an accident or incident until this event took place.[17] The nationalities of the passengers and crew included:[3][18]

At 2:00 pm, the United States Navy released a situation report stating:

Air New Zealand Flight 901 has failed to acknowledge radio transmissions. ... One LC-130 fixed-wing aircraft and two UH-1N rotary-wing aircraft are preparing to launch for SAR effort.[19]^:1

Data gathered at 3:43 pm were added to the situation report, stating that the visibility was 40 miles (64 km). Also, six aircraft had been launched to find the flight.[19]^:2

Flight 901 was due to arrive back at Christchurch at 6:05 pm for a stopover including refuelling and a crew change before completing the journey back to Auckland. Around 50 passengers were also supposed to disembark at Christchurch. Airport staff initially told the waiting families that the flight being slightly late was not unusual, but as time went on, it became clear that something was wrong.[20]

At 9:00 pm, about half an hour after the plane would have run out of fuel, Air New Zealand informed the press that it believed the aircraft to be lost. Rescue teams searched along the assumed flight path, but found nothing. At 12:55 am, the crew of a United States Navy aircraft discovered unidentified debris along the side of Mount Erebus.[19]^:4 No survivors could be seen. Around 9:00 am, 20 hours after the crash, helicopters with search parties managed to land on the side of the mountain. They confirmed that the wreckage was that of Flight 901 and that all 237 passengers and 20 crew members had been killed. The DC-10's altitude at the time of the collision was 1,465 feet (447 m).

The vertical stabiliser section of the plane, with the koru logo clearly visible, was found in the snow.[21] Bodies and fragments of the aircraft were flown back to Auckland for identification.[22] The remains of 44 of the victims were not individually identified. A funeral was held for them on 22 February 1980.

The recovery effort of Flight 901 was called "Operation Overdue".

Efforts for recovery were extensive, owing in part to the pressure from Japan, as 24 passengers had been Japanese. The operation lasted until 9 December 1979, with up to 60 recovery workers on site at a time. A team of New Zealand Police officers and a mountain-face rescue team were dispatched on a No. 40 Squadron C-130 Hercules aircraft.

The job of individual identification took many weeks, and was largely done by teams of pathologists, dentists, and police. The mortuary team was led by Inspector Jim Morgan, who collated and edited a report on the recovery operation. Recordkeeping had to be meticulous because of the number and fragmented state of the human remains that had to be identified to the satisfaction of the coroner. The exercise resulted in 83% of the deceased passengers and crew eventually being identified, sometimes from evidence such as a finger capable of yielding a print, or keys in a pocket.[23]

The fact that we all spent about a week camped in polar tents amid the wreckage and dead bodies, maintaining a 24-hour work schedule says it all. We split the men into two shifts (12 hours on and 12 off), and recovered with great effort all the human remains at the site. Many bodies were trapped under tons of fuselage and wings and much physical effort was required to dig them out and extract them.

Initially, there was very little water at the site and we had only one bowl between all of us to wash our hands in before eating. The water was black. In the first days on site, we did not wash plates and utensils after eating, but handed them on to the next shift because we were unable to wash them. I could not eat my first meal on site because it was a meat stew. Our polar clothing became covered in black human grease (a result of burns on the bodies).

We felt relieved when the first resupply of woollen gloves arrived because ours had become saturated in human grease, however, we needed the finger movement that wool gloves afforded, i.e., writing down the details of what we saw and assigning body and grid numbers to all body parts and labelling them. All bodies and body parts were photographed in situ by U.S. Navy photographers who worked with us. Also, U.S. Navy personnel helped us to lift and pack bodies into body bags, which was very exhausting work.

Later, the skua gulls were eating the bodies in front of us, causing us much mental anguish, as well as destroying the chances of identifying the corpses. We tried to shoo them away, but to no avail; we then threw flares, also to no avail. Because of this, we had to pick up all the bodies/parts that had been bagged and create 11 large piles of human remains around the crash site in order to bury them under snow to keep the birds off. To do this we had to scoop up the top layer of snow over the crash site and bury them, only later to uncover them when the weather cleared and the helos were able to get back on the site. It was immensely exhausting work.

After we had almost completed the mission, we were trapped by bad weather and isolated. At that point, NZPO2 and I allowed the liquor that had survived the crash to be given out and we had a party (macabre, but we had to let off steam).

We ran out of cigarettes, a catastrophe that caused all persons, civilians and police on site, to hand in their personal supplies so we could dish them out equally and spin out the supply we had. As the weather cleared, the helos were able to get back and we then were able to hook the piles of bodies in cargo nets under the helicopters and they were taken to McMurdo. This was doubly exhausting because we also had to wind down the personnel numbers with each helo load and that left the remaining people with more work to do. It was exhausting uncovering the bodies and loading them and dangerous, too, as debris from the crash site was whipped up by the helo rotors. Risks were taken by all those involved in this work. The civilians from McDonnell Douglas, MOT, and U.S. Navy personnel were first to leave and then the Police and DSIR followed. I am proud of my service and those of my colleagues on Mount Erebus.[24]

— Jim Morgan

In 2006, the New Zealand Special Service Medal (Erebus) was instituted to recognise the service of New Zealanders, and citizens of the United States of America and other countries, who were involved in the body recovery, identification, and crash investigation phases of Operation Overdue. On 5 June 2009, the New Zealand government recognised some of the Americans who assisted in Operation Overdue during a ceremony in Washington, DC. A total of 40 Americans, mostly Navy personnel, are eligible to receive the medal.[25]

The accident report compiled by New Zealand's chief inspector of air accidents, Ron Chippindale, was released on 12 June 1980. It cited pilot error as the principal cause of the accident and attributed blame to the decision of Collins to descend below the customary minimum altitude level, and to continue at that altitude when the crew was unsure of the plane's position. The customary minimum altitude prohibited descent below 6,000 feet (1,800 m) even under good weather conditions, but a combination of factors led the captain to believe the plane was over the sea (the middle of McMurdo Sound and few small low islands), and previous flight 901 pilots had regularly flown low over the area to give passengers a better view, as evidenced by photographs in Air New Zealand's own travel magazine and by first-hand accounts of personnel based on the ground at NZ's Scott Base.

In response to public demand, the New Zealand government announced a further one-man Royal Commission of Inquiry into the accident, to be performed by Justice Peter Mahon. This Royal Commission was 'handicapped' in that the deadline was extremely short; originally set for 31 October 1980, it was later extended four times.[28]

Mahon's report, released on 27 April 1981, cleared the crew of blame for the disaster. Mahon said the single, dominant, and effective cause of the crash was Air New Zealand's alteration of the flight plan waypoint coordinates in the ground navigation computer without advising the crew. The new flight plan took the aircraft directly over the mountain, rather than along its flank. Due to whiteout conditions, "a malevolent trick of the polar light", the crew were unable to visually identify the mountain in front of them. Furthermore, they may have experienced a rare meteorological phenomenon called sector whiteout, which creates the visual illusion of a flat horizon far in the distance. A very broad gap between cloud layers appeared to allow a view of the distant Ross Ice Shelf and beyond. Mahon noted that the flight crew, with many thousands of hours of flight time between them, had considerable experience with the extreme accuracy of the aircraft's inertial navigation system. Mahon also found that the preflight briefings for previous flights had approved descents to any altitude authorised by the US ATC at McMurdo Station, and that the radio communications centre at McMurdo Station had indeed authorised Collins to descend to 1,500 feet (460 m), below the minimum safe level of 6,000 feet (1,800 m).

In his report, Mahon found that airline executives and senior pilots had engaged in a conspiracy to whitewash the inquiry, accusing them of "an orchestrated litany of lies" by covering up evidence and lying to investigators.[29]^:¶377[30] Mahon found that, in the original report, Chippindale had a poor grasp of the flying involved in jet-airline operation, as he (and the New Zealand CAA in general) was typically involved in investigating simple light aircraft crashes. Chippindale's investigation techniques were revealed as lacking in rigour, which allowed errors and avoidable gaps in knowledge to appear in reports. Consequently, Chippindale entirely missed the importance of the flight-plan change and the rare meteorological conditions of Antarctica. Had the pilots been informed of the flight plan change, the crash would have been avoided.

On 20 May 1981, Air New Zealand applied to the High Court of New Zealand for judicial review of Mahon's order that it pay more than half the costs of the Mahon inquiry, and for judicial review of some of the findings of fact Mahon had made in his report. The application was referred to the Court of Appeal, which unanimously set aside the costs order. The Court of Appeal, by majority, though, declined to go any further, and in particular, declined to set aside Mahon's finding that members of the management of Air New Zealand had conspired to commit perjury before the inquiry to cover up the errors of the ground staff.[28]

Mahon then appealed to the Privy Council in London against the Court of Appeal's decision. His findings as to the cause of the accident, namely reprogramming of the aircraft's flight plan by the ground crew, who then failed to inform the flight crew, had not been challenged before the Court of Appeal, so were not challenged before the Privy Council. His conclusion that the crash was the result of the aircrew being misdirected as to their flight path, not due to pilot error, therefore remained.

Regarding the issue of Air New Zealand stating a minimum altitude of 6,000 feet for pilots in the vicinity of McMurdo Base, the Privy Council stated, "Their Lordships accept unreservedly that ... the evidence given by several of the executive pilots at the inquiry was false. But, even though false ... it cannot have formed part of a predetermined plan of deception. Those witnesses whom the Judge disbelieved on this issue were, as their Lordships must accept, being untruthful ... they were also being singularly naive. [Q]uite apart from the mass of evidence of flights at low altitudes and the publicity given to them ... it is not conceivable that individual witnesses falsely disclaimed knowledge of low flying on previous Antarctic flights in a concerted attempt to deceive anybody".[28]

But the Law Lords of the Privy Council under the chair of Lord Diplock effectively agreed with some of the views of the minority in the Court of Appeal in concluding that Mahon had acted in breach of natural justice in making his finding of a conspiracy by Air New Zealand management and it was not supported by the evidence. In its judgment, delivered on 20 October 1983, the Privy Council therefore dismissed Mahon's appeal.[28][31] Aviation researcher John King wrote in his book New Zealand Tragedies, Aviation:

They demolished his case (Mahon's case for a cover-up) item by item, including Exhibit 164, which they said could not "be understood by any experienced pilot to be intended for the purposes of navigation" and went even further, saying there was no clear proof on which to base a finding that a plan of deception, led by the company's chief executive, had ever existed.

"Exhibit 164" was a photocopied diagram of McMurdo Sound showing a southbound flight path passing west of Ross Island and a northbound path passing the island on the east. The diagram did not extend sufficiently far south to show where, how, or even if they joined, and left the two paths disconnected. Evidence had been given to the effect that the diagram had been included in the flight crew's briefing documentation.