Paul Holmes and Erebus

[prospector]

------------------- Amen ----------------

[chris lz]

Now before I forget, can someone fill me in on the alleged possible discontent of F/E Lucas? I can't find a thing on this.

[Dream Land]

If you were to fly into a hill whilst conducting an RNAV approach because Jeppesen accidentally shifted all the waypoint coordinates by a few miles, then that would be your fault would it?

It would be if you chose to disregard the MDA of 16000 FT.

[ampan]

#1AHRS: The difference between what you did and what Capt Collins did is that you actually had visual meteorological conditions, whereas he didn't. The AINS didn't improve his vision, nor did it assist in distinguishing between the cloud and the ice. The transcript gives a good indication of what he was thinking: 'Problem with vision below the cloud / No NDB and no contact with TACAN yet / Might have to go somewhere else / Radar assist! Yes please.' Yet 10 minutes later he dives down under the cloud without any radar assist.

[DozyWannabe]

Originally Posted by ampan

#1AHRS: The difference between what you did and what Capt Collins did is that you actually had visual meteorological conditions, whereas he didn't.

Under the cloud he did - in "by the book" terms anyway, the Mac Central controller told him so - 40 miles plus visibility (the maximum in aviation terms meaning anything from literally 40 miles to the horizon). The fact that the book was not sufficiently informed to take sector whiteout into account was not something Collins was aware of, and not something he should carry the can for.

[#1AHRS]

Ampan, My point being that Collins was fooled by sector whiteout, a phenomenon that he had no knowledge or training on. Both pilots obviously believed that they had adequate visual reference by what they saw out of the window and by what ATC told them. They believed, based on what they were briefed upon, that the RNAV track was to drop them over the entrance of McMurdo Sound, clear of mountains.

Current practice by RNZAF, RAAF and the yanks at the time was for commanders to make three entries under supervision. Air NZ had achieved a dispensation from that on the basis that their training (or briefings) were of such a high standard that this was not required.

Sector whiteout was no mystery to regular operators in Antarctica at that time and yet it was not a topic that was included in any of the ANZ briefings. Even their CP (Shredder) showed only rudimentary knowledge of the whiteout illusion during the inquiry. This barely reflects a high standard of training.

There is no argument that the crew didn't play a part in this accident. They did as they most often do. However that is not what Chippendales report portrays. It conveniently heaps nearly all of the blame onto the dead crew. That certainly curbed the litigation against New Zealanders and saved Muldoons career, but it did nothing toward identifying the latent and active failures that were clearly buried well into Air New Zealand's operations management system. In particular the Navigation section.

Mahons report did.

[FGD135]

It would be if you chose to disregard the MDA of 16000 FT.

Dream Land, you are missing the point in more ways than one.

You're way off base to say that he "disregarded" the MDA. If you read back over the last 4 or so pages of this thread, you will find that:

1. It was not an S.O.P to observe this "MDA". The real S.O.P was to descend where/when they actually did;

2. The crew were briefed to descend where/when they did.

And, the whole point of the RNAV approach analogy is to demonstrate that the actions of other persons on the ground - before you take off - can bring about your demise through no fault of your own.

If you were conducting an RNAV approach as per my example then you would have descended below the MDA in the course of conducting it. So this response is really quite dopey:

It would be if you chose to disregard the MDA of 16000 FT.

#1AHRS, very good post that summarises the circumstances very well. DozyWannabe gave a very similar summary much earlier.

[Dream Land]

My point is that I am trying to make is that the tour itself was not something conducted by use of any company supplied fixes if I am understanding the information correctly.

The waypoints brought you to a point where you would then begin a "visual tour" your statement implies to me that they actually conducted the tour by the use of some waypoints, one waypoint which ANZ changed considerably the night before departure.

After they transition to a VFR tour, it is then the responsibility of the flight crew to understand that they are solely responsible for identifying their location before conducting any descents, regardless of some visibility report given to them, and what airline crew would assume they are under radar control based on the the fact that a light on the xponder is flashing, these aren't ppl's, these are seasoned airline pilots.

[ampan]

I think the transponder light is a bit of a red herring. It only came on for a brief period, and he was already going down at that point. There's no indication that the crew believed that they were under radar control.

The same applies to the advice that visibility was 40 miles. The captain wasn't led astray by that, because his cloud/ice comment came after, not before.

As to sector whiteout, #1AHRS has confirmed that it was well-understood by polar aviators. In other words, the phenomena was not discovered by Vette. In fact, it was referred to in the Chippendale report. Given that, why couldn't Capt Collins have had a general understanding of the problem, either from his own private reading or from his time in the RNZAF, or both? The point is that he was obviously aware of the potential issue re VMC, despite not having been briefed on it.

[DozyWannabe]

Originally Posted by Dream Land

My point is that I am trying to make is that the tour itself was not something conducted by use of any company supplied fixes if I am understanding the information correctly.

That was the case in the days before the computerised flight plans, certainly - but from 1978 onwards, when the computerised flight plans were followed, things became less certain. With the passage of time the Antarctic flights stopped being a preserve of the management captains who had drawn up the original process, and passed to the line. Prospector, who appears to venerate the abilities of the management pilots of the day as being somehow superior, sees this as the point the rot began to creep in (and blames the union) - whereas I see the fact that the transition was not properly managed as a failure of management in terms of putting a process in place to transfer the knowledge. With the computerised flight plans in place it seems to have become the de facto practice to follow the computer track for as far as necessary before visual flight can be begun, and the fixes on the computer track did indeed come from the company - they were entered by Chief Navigator Hewitt himself.

Captain Simpson noticed that the waypoint labelled "MCMURDO" was a significant distance from the station - prior to this the assumption had been that the waypoint had been selected to follow the military track and did not refer to the station itself. He reported this upon his return and the message was passed up the chain to Hewitt, whose testimony states that he reset the waypoint to the TACAN, believing he was making a correction of 2 miles of latitude rather than the 27 mile correction he was actually making. Whilst he never went into specifics we can only assume that he made the change whilst looking at his original notes rather than what was in the database itself, which would have revealed not the 2 mile error he was expecting due to mistranscribing, but a 27 mile error which appeared as a consequence of his mis-typing the co-ordinates a little over a year previously. In any case he did not notify ops of the change and the fact that the co-ordinates had been changed was never communicated to the pilots who he would (or at least should) have been aware would be flying the route that morning.

The ethics of Hewitt's actions are debatable depending on how sympathetic one is to his account, but in strict terms his actions were a clear violation of company procedure as well as common sense and they were performed before the flight crew even awoke that morning.

What follows is a classic case of organisational entropy, which in this case had many tendrils manipulating events.

Firstly, the de facto procedure of following the computer track until visual flight was possible meant that in a situation where visual flight below an overcast was the only available option (I don't buy the assertion that every preceding flight occurred in "gin-clear" conditions throughout - even if the flights were scheduled for the best weather possible, we're talking about Antarctica here and the weather is anything but predictable), the pilots would be relying to some extent on the AINS and radar to confirm their position before descending - and would not be aware that the topography of Lewis Bay could from certain angles be mistaken for that of McMurdo Sound. Because these flights had been going on for some time they had become perceived amongst the flight crews as somewhat routine, and this led to a further perception that the computer would keep you away from high ground if the programmed track was followed.

Next we have the clarity of the briefing becoming diluted over time - rules became guidelines and the potential dangers that may have been well understood by the early crews were not properly communicated to those that followed. The audio/visual and written materials that dated from the earlier flights were verbally contradicted during the briefing itself, with the verbal instructions taking precedent - and this was explicitly confirmed by the briefing pilot's testimony.

Finally, we have the breakdown in communication between management, nav section and ops. Nav section clearly did not query the fact that the returning flights from Antarctica were following a track that was considerably different from that which they testified was intended. The error was only brought to light by chance when a line captain (Simpson) informed his superior on his return of the apparent discrepancy. The message certainly reached Chief Navigator Hewitt, but it would appear that neither ops nor management heard of it, nor was there a co-ordinated effort to solve the problem. Hewitt therefore made his change without notification or supervision in the early hours of the morning and no thought was given to the potential consequences of such a change.

These circumstances combined in such a manner that only a large dose of luck might have been able to help the crew of TE901 realise their predicament, and sadly luck was certainly not with them that day.

@ampan - How often would one have watched a transponder light in those days? It only needed to be on during the few seconds during which a decision was made to have proved fatal. Sector whiteout was understood by polar aviators, but not by line crews - or indeed even the management pilots that prospector speaks so highly of, because Gemmell certainly did not appear to understand the phenomenon judging by his testimony. Let me check the transcripts on the "40 miles" and get back to you.

[henry crun]

DozyWannabe: In your last paragraph in the post above you refer again to the transponder light being on for a few seconds.

You appear to believe that is justification for Collins to believe he was identified and and radar control, is that correct ?

[FGD135]

DozyWannabe,

Another highly articulate and thoughtful summary of the circumstances. We are grateful that you are putting so much effort into this thread.

... only a large dose of luck might have been able to help the crew of TE901 realise their predicament, ...

If ever anybody needed a one-sentence summary, that would have to be it.

You appear to believe that is justification for Collins to believe he was identified and and radar control, is that correct ?

henry crun, you appear to be reading way more into Dozy's words than he has written. Dozy was speculating on the significance of the XPNDR reply light. If the light happened to be flashing during those few seconds that Collins was assessing whether he should continue, then it is very likely that the light contributed to his belief that everything was safe.

Just speculation.

You seem to prefer the strictly black or white concept of being positively-identified or NOT, but if you could see inside any aviators mind while he is on the job you would find that his reasoning, rather than resting on a set of black-or-white foundations, rests instead on a large set of foundations that are comprised of shades of grey!

Aviators are human beings after all, and there often isn't time or justification to turn shades-of-grey into black-or-white.

This is the reasonable view of Collins and his copilot's performance that day. Some here insist their performance should have been robot-like but that is just not reasonable.

There's that word again: reasonable.

[henry crun]

FGD135: With the experience of being a radar controller for many years, and knowing the dangers that mis-identification can bring, it would be unthinkable of me to beleive in identification in any terms other than black or white.

Collins was a senior captain, widely experienced in being under radar control in many countries, and you will have some trouble convincing me he would have been influenced by the very brief flickering of the transponder light in the manner which Dozy Wannabe suggests.

Imho there was no room for any shade of grey in his (Collins) situation.

[DozyWannabe]

Originally Posted by henry crun

...and you will have some trouble convincing me he would have been influenced by the very brief flickering of the transponder light in the manner which Dozy Wannabe suggests...

How do you know it was "brief", or indeed how brief it may have been? It could have been coding for several minutes while the decision-making process was happening.

Imho there was no room for any shade of grey in his (Collins) situation.

Let's talk about mental models for a second. As far as Collins is aware, the radar controller will not approve a let-down until they are positively identified - also as far as he is aware, the radar controller will have an exact copy of the flight plan that he was given along with the co-ordinates, so even if identification is sporadic, the controller will know where he is at a given point in time. He's also used to radar control as it exists in the civilised world, where things are indeed "black-and-white". This is Collins' mental model, and as such he was entitled to take the controller's assent to the let-down as confirmation that everything is safe for him to do so.

The controller's mental model is somewhat different, because unbeknownst to Collins, the flight plan sent to Mac Central by ANZ was incomplete and in fact lacked the all-important changed co-ordinates for the McMurdo waypoint. So it's possible that the controller approved the let-down based on a calculated position in turn based on the flight plan, which may have been the plan from the previous flight which had the (original, "false") co-ordinates used as a reference (because the co-ordinates were missing from the new one). In "black-and-white" terms the controller should not have approved the let-down for this reason, but this is not your usual radar-controlled airspace - it is in fact a sparsely-populated bit of airspace on the edge of the known world and things tend to be a little less rigid here. The controller is also aware that TE901 is not the average hoary old military propliner with decades-old nav equipment that he is used to handling, it is a state of the art jetliner festooned with gadgetry that is probably better at calculating their current position than he is.

The difference in these mental models could well carry the genesis of the physical reasons for the let-down being approved when perhaps it should not have been. As I did before, I must ask myself - if the controller had done things by the book and the aircraft crashed, then why destroy a segment of the radar trace that would have erased all doubt?

[ampan]

Capt Collins was advised that the radar's range was 40 miles.

He started his descent below MSA at 43 miles out (ie, that's when he pulled out the knob and began the first orbit using Heading Select).

[DozyWannabe]

They're going a little over four nautical miles per minute when they make the request at 12:02, which means that they should have shown up when clearance is granted at 12:03, and would be well within range when they begin their descent at 12:11.

(All timings from the left-hand column of Washington Transcript)

[ampan]

It still doesn't fit: (1) There's no reference to the transponder before the descent commenced; (2) Just before descending, the F/O told Mac Station that they would be descending VMC, and the captain was careful to maintain VMC all the way down.

PS - Where were able to get the Washington transcript from?

[DozyWannabe]

@ampan - Do you think they'd have started the descent without it? They appear to lose it for a time, but it returns at timestamp 15:21. Unfortunately there are no publicly-available FDR traces to help confirm one way or the other (not that such information would likely be available on FDR traces of that vintage).

Both the Washington Transcript and the version from the Chippindale Report are downloadable as PDFs from :

Transcript

Note that while they make their position clear (the website being run by NZALPA), they at least provide a significant amount of the raw material from both the Chippindale Report and the Mahon Royal Commission for people to make up their own minds, which does them credit as far as I'm concerned.

The descent may have been VMC, but they tried to remain in contact with Mac Central, advising of their position and altitude as they went - remember that the crew were perceiving Lewis Bay as McMurdo Sound due to a combination of the track they were expecting to be on and the visual illusions that provided false confirmation. Perhaps they were expecting some kind of warning if they were off-course and even in the ATC transcripts (including those that were not heard by the crew), at no point did the radar controller try to warn them that they were approaching Lewis Bay and Erebus. If the transponder was coding at 15:21 - 26 miles east of where he was expecting it (which means that they probably did show up) - why? It's not like he was distracted by heavy traffic...

Because of the visual illusions and the fact that they did not receive any kind of warning before it was too late (in this case the GPWS) it is entirely understandable that the crew believed that things were proceeding according to the plan they had been briefed on because all the information they were receiving confirmed their mental model and nothing contradicted it other than a single-digit change that was not required to be checked against the flight plan they had based their work on the previous night.

[ampan]

The transponder light could only have come on once, being the occasion referred to in the transcript. Firstly, the radar's range was 40 miles or thereabouts. Secondly, the track, directly towards Erebus, was right in the middle of the radar's blind sector (see The Erebus Papers, p434). The only time the transponder light came on was when he went off to the right of track, on the first descending orbit.

If the descent was caused by the illumination of the transponder light, someone would have said something. What caused the descent, according to the transcript, was sighting the ground. Thereupon, he abandoned the radar assist and went VMC. If he assumed that his position had been confirmed by radar, he would have gone straight down, through the cloud base. No need to muck around with orbits.

[henry crun]

DozyWannabe: I will try and answer your questions and points in the order you have made them..

I don't know how long the transponder light was on for, and neither do you, so we are both speculating.

You say that "As far as Collins is aware, the radar controller will not approve a let-down until they are positively identified".
You are completely wrong in this assumption, the radar controller cleared Collins for a VMC descent. In doing so he handed complete responsibilty for cloud avoidance and terrain clearance to the crew, without any suggestion of radar identification or flight following.
I have issued many hundreds, maybe thousands, of VMC descent clearances to aircraft on IFR flight plans and it was never understood by any commercial pilot to mean anything other what I have stated above.
In that knowledge, Collins was not entitled to take the controller's assent to the let-down as confirmation that everything is safe for him to do so, nor do I believe he would have assumed it to be so.

You mention again your belief that a brief part of the radar trace was deliberately destroyed.
To do so you are making another assumption, neither you nor I know for certain how that part of the tape came to be erased.
Do you know what playback equipment was used ? do you know if the tape was frozen post accident knowledge, and if so, what safeguards were there to preserve it's integrity ? do you know who was authorised to make any playback, pre or post accident ? do you know if accidential erasure was possible on the type of playback equipment used ?

In reply to ampan's point you appear to making the mistake of believing that because the aircraft was just within the radar maximum range it would have shown up.
Surveillance radar does not necessarily work like that, there can variations, holes, and gaps in coverage for a variety of reasons. Particularly as the radar at McMurdo was, as I understand it, by no means new.
When I was working I could have shown you several instances where an aircraft well within the range limit of coverage, would be on one track at a specific range painting loud and clear, but if it had been on another track only 8-10 miles away at exactly the same range it would not show at all, not a glimmer of a paint
This was, I will add, using a radar of similar vintage but longer range capabilty to that I understand was in use at McMurdo.


Secondly, in your reply to ampan you say "the transponder was coding at 15:21 - 26 miles east of where he was expecting it (which means that they probably did show up) - why? It's not like he was distracted by heavy traffic..."

It may or may not have shown on the radar, regardless of whether it did or did not show, by then Collins was accepted a VMC descent and, as I have already explained, had accepted complete responsibilty for his flight safety.
Knowing this, the controller might well have looked away from the scope to talk to a colleague, or perhaps perform some other admin task.