Erebus 25 years on

[Graybeard]

1979 was a very bad year for the DC-10 and its maker, McDonnell-Douglas. The AA-191 crash and subsequent grounding was followed by a Western Airlines crash in Mexico City in August, with about 50 fatals. Navigation played a part in that one, as they landed on a closed runway and hit a dump truck lost in the fog. There was no evidence they had switched from INS to ILS navigation, for example.

Certainly, there would have been lots more attention given the Erebus crash in the US media, save for Iran taking US hostages in the embassy in Teheran within days before. The DC-10 would have been cast in a bad light once more.

If ANZ were into a cover-up early on, it seems they would have tried first to blame McDonnell-Douglas and the DC-10's sophisticated AINS-70, Area Inertial Navigation System. Was there no evidence of this?

GB

[prospector]

"If ANZ were into a cover-up early on,"

Not that was made general public knowledge.

But that that is assuming that Mahon's theory of conspiracy held any water. Not everybody believed that there was a conspiracy, least of all the Appeal Court and the Privy Council.

[ampan]

Greybeard #377:

Of all the various conspiracies that AirNZ are supposed to have entered into, I’m not aware of any suggestions that they attempted to pin the blame on McDonnell Douglas, although I’m sure they would have liked to. McDonnell Douglas had two representatives on the team that went down to the ice, so they were on the ball very quickly.


Brian Abraham #364:

We’ll never know exactly what happened, but we have the last 30 minutes of what was said, and what was done. We also have photos of what was seen (out the side). So it’s possible to have a reasonably well-informed guess, isn’t it? After all, that’s what Mr Justice Mahon purported to do – and he came back with a finding that implied that the flight was doomed from the outset. Against this there are at least three potential errors made by the crew, each of which, if established, can be described as a cause of the accident: (1) Not clarifying the actual nav track before departure; (2) Descending from 18000 feet to 10,000 feet before getting the radar fix; (3) Failing to identify Beaufort Island. There is no disputing that these three events occurred. The issue is whether any of them amounted to an error.

I’m not suggesting that I’ve got any of the answers. In a previous post, I suggested that there would have been nothing wrong with staying at 18000 feet until the final waypoint, getting a radar fix, and then descending through the cloud to McMurdo Station. But in making that suggestion, I overlooked the fact that the cloudbase at McMurdo Station was at 2000 feet. So there would be plenty wrong with going to the final waypoint 27 miles west of McMurdo Station, and then doing a high-speed descent, with the flaps in and gear up, from 18000 feet to 1500 feet – even if you knew nothing about sector whiteout. That could never have been the plan.

So what was the plan? There must be some wiser heads out there who can provide the answer.

[Desert Dingo]

Well said ampan.
I think a lot can be gained by the "what would I have done if I was in that situation?" type of reasoning.

I have no difficulty concluding that I would probably have done exactly the same and also finished spread all over the mountain. As you have pointed out before, it is hard to find one instance where the crew did something that did not seem reasonable at the time.

It is unfortunate that Chippindale's report was made public before Mahon's enquiry. Up until this disaster, Chippindale had not investigated a major airline disaster. His experience was of the "reckless light aircraft pilot flying at low level in poor visibility and hits hill" type investigation.
It seems he formed the opinion that this was just a larger version of his previous investigations as evidenced by his unwarranted changes to the CVR transcript to suit his pre-conceived idea that the crew were flying recklessly at low level, in cloud, and unsure of their position and ignoring protestation of alarm from others on the flight deck. All of that has been shown to be nowhere near the truth. However, once published, he was pretty well "locked in" to defending that position.

Just think: there may have been a much less disputed outcome, if there had only been one (public) enquiry into the disaster.

[prospector]

Desert Dingo,
""reckless light aircraft pilot flying at low level in poor visibility and hits hill" type investigation.

I would suggest you read a bit more widely than Mahons utterings before you print crap like that.

Ron Chippendale was an ex RNZAF transport pilot, instructor,

"He had formally been involved in military flight safety and accident investigation for several years before being seconded to similar duties in Civil Aviation, and his standing in the world community is such that as someone scrupulously fair and independent, he has been asked to investigate some accidents of a particularly delicate political nature. (Recently, for instance he was appointed to the five-strong United Nations team to investigate the loss of Korean Airlines Boeing 747 airliner, Flight KAL007, shot down by a Soviet fighter near the island of Sakhalin in 1983).

Anyway, what is the difference between a reckless small aircraft pilot flying low level in reduced visibility and hitting a hill, than reckless flying low level, in reduced visibility, in a large aircraft and hitting a mountain???. Apart from the fact the large aircraft had all the latest nav gear, top of the line equipment, multi crew, weather radar, etc.

I believe I have an answer, but would be interested to hear if you have.


"I have no difficulty concluding that I would probably have done exactly the same and also finished spread all over the mountain. As you have pointed out before, it is hard to find one instance where the crew did something that did not seem reasonable at the time."

What is reasonable about descending VMC to do a sightseeing run, when conditions at the base were below the authorised descent minima, and the crew had already been advised to give McMurdo a miss due weather and proceed to Taylor Valley area.

[ampan]

AINS

18000 to 10000 without the radar fix is the main issue, isn't it? (Note that the Starlifter waited for the radar fix.)

In making that descent, the only available positional information was from the AINS. I don't think there is any argument about that.

So was that an error? It depends on the the state-of-the-play re the AINS in November 1979.

The three major accidents that I'm aware of involving the AINS are Erebus, Korean Air (mid 80s), and that dreadful c8ck-up in Columbia in the mid 90s.

What about the 1970s? Were there any AINS c8ck-ups during that decade?

[prospector]

Ampan,

It was by all accounts very accurate, have a look at the impact position as against the AINS position, just that nobody bothered to plot the position on a chart before they commenced the descent.

[Graybeard]

"So was that an error? It depends on the the state-of-the-play re the AINS in November 1979."

"The three major accidents that I'm aware of involving the AINS are Erebus, Korean Air (mid 80s), and that dreadful c8ck-up in Columbia in the mid 90s."

The CFIT crash in Columbia was an AAL 757. It had FMS, a Boeing/Honeywell copy of the AINS-70, with some advancements.

The AINS-70 that was factory equipped in about a third of DC-10s was the only RNAV system I'm aware of that used the letters AINS. It was part of the configuration of DC10-30 bought by the KSSU group of airlines, KLM, SAS, Swissair and UTA. Related airlines such as Philipine and Garuda flew it also. I don't know if or how ANZ was related to KSSU, but all five of their DC-10 had the AINS-70.

KAL also had five DC10-30, only one of which was KSSU configuration, it having been repossessed from a small Asian airline. KAL had a real training problem with that single plane equipped with AINS-70. About a month after 007 was shot down, that plane departed from Seoul, destination Honolulu. Somehow, pilot error or pilot misinterpretation induced the AINS-70 to show a desired track of about 45 degrees, instead of about 90 degrees toward HNL.. . 45 degrees from Seoul would have put them on course over Sakhalin Island...

KAL management then inquired about training on the AINS-70, but instead fired a bunch of pilots. A few years later, that plane crashed in fog on an NDB approach a couple of miles short of the runway in Libya. I never heard anything of a crash investigation.

The KSSU/AINS-70 DC-10 fleet had an impeccable record with regard to navigation. In fact, it was in some regard too good, as the ANZ pilots had come to trust it implicitly.

I read in "Whiteout" that the only real chart the pilots had was the brochure given to the pax; in other words, not even as good as the National Geographic map that came out in the magazine about that time. Is that true?

GB

[ampan]

Greybeard:

Possibly, but in this particular case, the captain had his own collection of maps and atlases, and is said to have used them to plot the nav track the night before the flight. The problem is that no such map/atlas was found at the crash site, or at the captain's house.

The reputation of the AINS as at 1979 is important to the pilot error issue. If there had never been any incident of a data-entry mistake in the ten years up until then, it makes it more difficult to argue that that the crew made an error in descending without first verifying their position. (On the other hand, I'm reasonably sure that they would not have made a landing approach without doing so.)

[prospector]

Ampan.
Let's try again, this I have printed in this thread at least once and probably more times.

Fom John King's publication.

"The actual track and distance chart supplied to flight crews was a subject of considerable debate during the enquiry. Pilots giving evidence on behalf of NZALPA insisted it was exhibit 164, but Joe Lawson in his evidence said that three copies of Annex J of the Chippendale report, which clearly showed the Southern waypoint as being Williams Field had been placed in the aircraft documents. From the evidence given none of the pilots recalled definitely using either Exhibit 164 or Annex J. Also found in the wreckage was Exhibit 78. a copy of the NV90 computer printout structured as a skeleton flight plan with waypoints names, tracks and distances. BUT THAT WAS NEVER ELABORATED UPON DURING THE ENQUIRY.

Weighing the evidence presented before him- and the credibility of the witnesses themselves-Peter Mahon concluded that exhibit 164 was the track and distance chart supplied to Antarctic crews, and Annex J was something the airline would rather had been found in the cockpit of the wrecked DC10. THAT CONCLUSION WAS ONE OF MANY LATER DISMISSED AT BOTH THE COURT OF APPEAL AND PRIVY COUNCIL HEARINGS.

Surely that covers whether a positive position could be established by just reading the AINS position readout and transferring it onto a chart before any descent was contemplated.

To have position confirmed by radar before descent would also have been very desirable, even if, as Desert Dingo keeps on about, not required, in Mahon's view.

But it still gets back to the fact. who wants to go sightseeing at 250 kts plus, with a reported cloud base of 2,000ft???. and so reported by trained observers on the ground at FLT901's intended destination???. Who also recommended going elsewhere??

[Graybeard]

It's a basic FAA rule that the pilot must have a chart appropriate to the flight.

There were no Jepps or other civilian charts ever printed for Antartica, AFAIK. Whose responsibility was it for the ANZ crews to have appropriate charts or maps? That should have been a priority for all concerned.

Was there or was there not a chart with sufficient legitimacy and accuracy to provide a valid cross check of the MCMDO waypoint? If there is no definite documentation of it, you would have to determine there was not. That should have been part of the route approval.

From what I remember at the time, there was a waypoint named McMurdo out on the Ross ice shelf, whose position, of course, could be confused with the position of McMurdo Station. I believe that was the waypoint used by prior flights.

Awareness that pilots could confuse VOR locations with airports of the same name was highlighted in the late 1970's when a plane hit a mountain in Arkansas on whose VOR was the same name as the airport in the valley a few miles away. The FAA changed the VOR name to Ozark Mountain VOR, or some such.

They also renamed all VORs not located on airports, but having the same name. ONT, Ontario, California, VOR, for example was renamed PDZ, Paradise VOR, as it was 5 miles from the field. That also meant the audio ident was no longer the same as the airport ident. I know another case where a commuter plane hit the ridge 5 miles south of Elko, Nevada, as the pilots used the VOR DME distance instead of the ILS DME distance.

In short, if the MCMDO reporting point on the Ross ice shelf had been named ROSSI, or PENGN, or whatever, instead of MDMDO, the confusion would never have existed.

Further, what good reason was there to use McMurdo Station as the waypoint instead of MCMDO? If the flight were ever to be in a single engine driftdown condition, it would be far safer to use the Ross ice shelf waypoint rather than the route over the volcano. And, there was also always the possibility of undetected eruption or ash from the volcano.

You have to admit it was stupid to have McMurdo Station or Scott Base as a waypoint without MCMDO on the Ross ice shelf instead, or in addition.

I'll get back to the AINS-70 in a later post.

GB

[ampan]

Back to the Facts

Prospector #386:

As Prospector suggests, the plan must have been to get under the cloud early, and fly fast (because no flaps were allowed) and low to McMurdo Station. That does not sound like a very good plan, given potential altimeter problems and given that it was “very hard to tell the difference between the cloud and the ice”. Also, how much viewing time of McMurdo Station would you get at 1500 feet going 250 knots?

The problem is that it’s difficult to label the plan as a clear-cut error. The same applies to the descent from 18000 feet to 10000 feet. It would have been a clear-cut error if it was through cloud, because if the AINS was wrong, they could have planted the aircraft on the top of Ross Island. But they descended through the hole, VMC, believing they could see the terrain. It would have been desirable to get the radar fix before going down. It would have been desirable to at least do one check the current AINS coordinates. But it’s difficult to say that it was an error to not do either.

And it’s difficult to label the failure to identify Beaufort Island as an error. They were already down in the sucker hole when that happened.

The only clear-cut error that I can see depends on what was said at the briefing. If Capt. Wilson told the crew that the nav track went over Erebus, then it was an error to not clarify the situation when, the night before the flight, the captain discovered that the print-out showed a nav track down McMurdo Sound. To do so would have been simple: Check the coodinates as they were manually entered before departure. The problem with that alleged error is the controversy over what was actually said at the briefing.

If you put the briefing issue to one side, there doesn’t seem to be a single isolated incident that can be pointed to and labelled as a clear-cut error. On the other hand, the flight was definitely not doomed from the start, as Mahon suggests. All that can really be said is that the crew “could have done better”.

[ampan]

Greybeard #387:


Maps

Apart from the captain’s own atlas, the crew were given a folder of maps at despatch, one of which was Chart NZMS135. Although this chart was large scale, it had a smaller scale inset of McMurdo Sound. (The map and inset are at pages 90, 91 & 37 of McFarlane’s book). This map was sufficient to estimate the position using the current AINS coordinates and was sufficient to enable identification of Beaufort Island.



Change of Waypoint

There were actually four different waypoints used. (They’re shown at p176 of McFarland’s book). Three of the waypoints were at McMurdo Station, and produced a nav track going over Erebus. The other produced a nav track down McMurdo Sound.

For the first two flights, the waypoint was at Willams Field, the airstrip: 77.53S & 166.48E. On these flights, 16000 feet was the minimum altitude. AirNZ then obtained approval to descend to 6000 feet in a sector behind Ross Island. A cloud-break procedure was adopted, using the NDB – and an officer in the Nav Section made a slight alteration to the waypoint, changing the coordinates to that of the NDB.

Later, another officer in the Nav Section was doing a computer changeover of all company routes. This officer mistakenly had the waypoint as 77.53S & 166.48E; the original Williams Field waypoint. When this officer typed in the coordinates, he made a one-digit typing error, and the waypoint became 77.53S & 164.48E. This is the point when the nav track changed from over Erebus to down McMurdo Sound.

About six further flights occurred with the McMurdo Sound nav track but the error was not noticed, as they had blue sky conditions. In November 1979 the NDB started to fail. About two weeks before the fatal flight, a particularly attentive captain, while overhead the TACAN at McMurdo Station, noted that he was well away from the nav track, which he found to be surprising. So he spoke to the exec pilot responsible for navigation, who spoke to someone, who spoke to someone else, etc. By the time the message came out the other end, it was “Change the waypoint from the NBD to the TACAN”. So the coordinates were changed to those of the TACAN, which moved the nav track back over Erebus. But given that the Nav Section thought the nav track was always over Erebus, the crew were not told.

Mahon thought most of the above was rubbish. He believed that the change of the nav track to McMurdo Sound was deliberate. Mahon thought that the Nav Section secretly changed the nav track, and that the naming of the waypoint was connected to this.

The primary reason for Mahon’s belief that the Nav Section changed the track to McMurdo Sound deliberately was his refusal to accept that AirNZ would deliberately have a nav track going over an “active volcano”. Greybeard has similar doubts. Putting the active volcano to one side, McMurdo Station was the logical position for the final waypoint. That’s where the airfield, NDB and TACAN were located. A nav track from Cape Hallett to McMurdo Station necessarily crossed Erebus, the “active volcano”. An active volcano is not in a constant state of eruption. An eruption is a 1 in 20 year event, or thereabouts. (And if the volcano happened to be erupting, what about heading select? ) There was a permanent plume, but this was mainly water vapour, the stuff clouds are made of. I think a DC10 could have handled it.

[Brian Abraham]

All that can really be said is that the crew “could have done better”

Afraid I can't let that slide by. The correct statement is the whole airline from top to bottom could have done better.

Graybeard - you make some good observations. Yes there was an RNC chart for the area and what has me stumped is why the airline did not follow that route. Military aircraft were forebidden to overfly Erebus (know not why but the issues you raise would be more than enough reason in my book). Another question I have is why on the flight plans prior to the fateful flight did they not have the NDB at McMurdo as a waypoint on the flight plan, if as prospector maintains ALL FLIGHTS were to make their descent in the airline stipulated sector.

Edited - got my facts wrong

[Graybeard]

"Putting the active volcano to one side, McMurdo Station was the logical position for the final waypoint. That’s where the airfield, NDB and TACAN were located."

Didn't the military flights go around Erebus, rather than over it? I don't believe I'm hallucinating that there was a reporting point named McMurdo on the Ross ice shelf. I don't have any data to prove that, however, as it was told to me.

Since this was a sightseeing flight, in which they might descend below the elevation of Erebus when obscured, there was no reason not to make the approach over the Ross ice shelf as part of every flight plan.

Thanks for the history of the floating waypoints. I still would tend to believe somebody punched in the "erroneous" waypoint over the Ross ice shelf intentionally. You check and double check when you make a change like that. Given the rest of the operation, however, a lucky typo could have happened.

There's more.

GB

[ampan]

It might assist if I set out the co-ordinates for each of the four waypoints;

(1) Williams Field - 77.53S 166.48E

(2) NDB - 77.51S 166.41E

(3) Somewhere in
McMurdo Sound - 77.53S 164.48E

(4) TACAN - 77.53S 166.58E



If the Nav officers were telling the truth, it was still a monumental balls-up.

[ampan]

Greybeard #391:

You're correct: The military route went around Ross Island, by first overflying the "Byrd Reporting Point", which was in the centre of McMurdo Sound, and then turning left.

Most of the AirNZ flights arrived to find perfect weather, so they did exactly as you suggest: Fly close to Victorialand on the run down, then turn left, over the general vicinity of the Byrd Reporting Point, and on to McMurdo Station. The nav track wasn't an issue, in what the pilots thought were perfect VMC conditions. (although it now seems that Antarctica can always bite you, even in "gin clear" conditions)

[Graybeard]

Navigation by Weather Radar

The DC-10 was equipped with dual Bendix RDR-1F weather radar, kind of an old clunk, but about as good as was available at the time. There were two PPI, plan position indicators, outboard of each pilot's position. There was a weather mode, a "contour" mode, and a Map mode, any of which could do a decent job of ground mapping.

I have seen nothing written that any of the pilots for the Antarctic trips had ever been briefed on using the radar for fixing position. Had it been for a briefing for any other route, this would not be particularly surprising to me, as I have talked to a lot of pilots, and most of the airlines glossed over weather radar training, assuming the pilots had got it elsewhere. For this trip, it should have been stressed.

I believe there was not any indication from the CVR tape that either pilot had ever turned on the wx radar, let alone looked at the display. Just a little bit of knob dicking anytime in the last half hour would have shown their position vs. Erebus.

This was both a systemic fault and a pilot fault. Neither the airline nor the pilots escape serious blame for this.

GB

[prospector]

Brian Abraham.

"if as prospector maintains ALL FLIGHTS were to make their descent in the airline stipulated sector."

This from Maurice McGreal publication. "In relation to the criticism that the flight was planned over Mt Erebus, an active volcano, Bolt/Kennedy responded by saying that the duty of CAD was not to specify the route, that was for the airline, and CAD's job was to set the conditions that would ensure the safety of the flight.This clearly related to to the fact that the minimum safe altitude named in the Civil Aviation requirements for the flight was 16,000ft until the aircraft had passed over the McMurdo base station at which time it could descend to 6,000ft provided that the visibility was satisfactory, that is in visual conditions."

Because Air New Zealand took it upon themselves to treat these requirements as they did does not make them any less the mandatory requirements from the regulating body.

[prospector]

Greybeard,

The question of radar was dealt with by Mahon, he and his assisting counsel made a special trip to the States, to the manufacturer, to gather facts.

He came up with the conclusion that there would be no return from Erebus, because it was covered by dry ice, the weather radar only reflected back from moisture,ie rain. Take that as you will, but that was his finding.

" This aspect has been hotly debated. Many pilots flying the DC10's to Antarctica used the weather radar in mapping mode, which clearly confirmed the outline of Ross Island and its high ground that they could see through the cockpit windows in the clear Antarctic conditions."

You will note that "many pilots" used the weather radar in mapping mode.

But the commissioner in his wisdom, after a taxpayer funded trip to the US, claimed that they were wrong, the radar would not show up Erebus.

Once again, the commissioner knew better than professionals who do this for a living.