Although some of the arguments blaming the flight crew for this accident sound convincing on the surface, whenever I review either Vette's or Mahon's writings it seems that they are entirely justified for their conclusions.  Most of the counter arguments do not seem to address the evidence in its totality.  Some of the most disturbing is Captain Keesing's testimony as the ANZ Flight Operations Director up until April 1979, stating that in fact it WAS ANZ company policy to fly below 2000 feet, that that was always the policy AND practice of ANZ, and that Ian Gemmel's correspondence with the Civil Aviation Dept regarding 6000 and 16000 foot MSA's was without his knowledge, furthermore that CAD's letter confirming these altitudes was only received by ANZ one week AFTER the fatal accident, and when the previous arrangement had been to fly well below these altitudes.  

Mahon was to refer to this timing somewhat charitably as "unfortunate".  Keesing was the only then current or former senior person from ANZ who admitted seeing and authorizing one of the MANY well publicized articles indicating clearly that the flights were ALWAYS carried out well below 6000 feet, with the possible exception of the first two, when passengers in fact complained about the poor viewing at high level.  Search Youtube and you will find footage filmed by the National Film Unit on the flight previous to the fatal one with the pilot clearly announcing the altitude as 3000 feet, so much for the argument that Collins and his crew were breaking company policy by descending below 6000.  Company "policy" is not formed by what is contained in obscure correspondence that almost no one seems to have been aware of, it is formed by what the company practices actually are at the time.

Other aspects of company policy that remain disturbing to me are treating these flights as "rewards" to pilots, and sending them in charge without previous Antartic flying experience, and possibly in the case of Flight 901 providing an "experienced" Antarctic traveller such as Peter Mulgrew who effectively added to the crew's visual mindset by apparently confirming their mistaken VFR position fixes.  

I note other contributers here have blamed the crew for "navigating" at low level using the INS, and for not checking the INS waypoints.  In fact at the last waypoint they could have checked, the INS was operating perfectly, and of course it continued to perform within spec right up until impact.  The crew's faith in the dual redundant INS was not misplaced, only in the printout they received from the company that morning which should have notified them of any changes made to the coordinates, but didn't.  The crew certainly didn't navigate at low level on the INS, they navigated to the let-down point on the INS, then may have continued to check their position on the INS at low level while flying VMC.  And if there is anything wrong with that, just imagine what we would be saying about them if they had flown into the mountain while flying VMC, while their correctly programmed INS was warning them they were off track.

As for the potential let down by ground based radar, I think most correspondents have no idea as to the limitations of the McMurdo radar, effectively a portable precision approach radar of some age.  A lot of people seem to have an unshakeable faith in radar without knowing anything of the various types and their limitations.  The McMurdo radar was not designed for surveillance and had a pencil beam, not a wide vertical beam.  When ANZ tried to argue that their intention was for the aircraft to fly direct over Mount Erebus and then descend orbitally under radar control, the senior US controller claimed this had never been done and would have been virtually impossible with the radar equipment they had, which would have required the radar controller to continually adjust beam angle while trying to track the aircraft, and he would never have authorized such a procedure.  The whole story was thus a fabrication by ANZ witnesses.

While the DC-10 INS RNAV had limitations in terms of how it displayed the information, it was by far the best system available at the time for navigating a pre-planned route in the area of the flights with or without radar surveillance, as long as the coordinates loaded in WERE the same as those the pilots were expecting.

When you add the likelihood of the visual clear-air white-out illusion and Mulgrew's verbal confirmation of the pilots' VFR observations to the mindset that they clearly expected the INS to have lined them up to be flying down McMurdo Sound just like every other previous flight, I think the trap they fell into is entirely understandable.  That the trap was set at all is a fact that surely ANZ must take full responsibility for.

Although some of the arguments blaming the flight crew for this accident sound convincing on the surface, whenever I review either Vette's or Mahon's writings it seems that they are entirely justified for their conclusions.  Most of the counter arguments do not seem to address the evidence in its totality.  Some of the most disturbing is Captain Keesing's testimony as the ANZ Flight Operations Director up until April 1979, stating that in fact it WAS ANZ company policy to fly below 2000 feet, that that was always the policy AND practice of ANZ, and that Ian Gemmel's correspondence with the Civil Aviation Dept regarding 6000 and 16000 foot MSA's was without his knowledge, furthermore that CAD's letter confirming these altitudes was only received by ANZ one week AFTER the fatal accident, and when the previous arrangement had been to fly well below these altitudes. 

Lermy - even if it was company policy to fly to 2000 ft, not even Collin's most ardent supporters would suggest that it was "company policy" to do so under any conditions. Collins flew through a hole in the cloud, nearly to ground level, without reference to the horizon ahead of him using the only the INS to confirm his position. To anyone who actually understands the role of the INS in flying (GPS in 2013), this is utterly reckless.

The crew's faith in the dual redundant INS was not misplaced
Not misplaced at 30,000 feet; but most definitely misplaced at 1500 ft. If flying by the INS was that safe, pilots today would be able to fly from Auckland right into Queenstown through the mountains with one press of an INS switch. They can't now, and they certainly couldn't in 1979. Think about it. If the INS - in 1979 - was that crucial a navigational tool, to the extent that a single misprogrammed digit could cause an aircraft to crash - do you think the crew would've been permitted to input the co-ordinates manually?

A lot of people seem to have an unshakeable faith in radar without knowing anything of the various types and their limitations.
The fact is that Collins was offered a radar let down by McMurdo, and in fact stated "crikey, that's what we want to hear". But then on seeing a hole in the cloud cover below him, he elected to descend through that - almost on a whim it would seem. The fact that he verbalised himself "very hard to tell the difference between cloud and ice" and then descened below cloud to almost ground level without visual reference to the horizon ahead, is nothing short of macabre.

The crew certainly didn't navigate at low level on the INS, they navigated to the let-down point on the INS, then may have continued to check their position on the INS at low level while flying VMC
Mahon wrote something to the effect of "Collins was so good a pilot, that he double checked his position during the descent using the INS", which in fact is nonsense. A flight should be conducted using EITHER VMC or IMC techniques, not a hybrid of both, which is what Collins was doing. He was almost certainly still using the INS during the descent because he realised he was not truly in visual conditions.

Other aspects of company policy that remain disturbing to me are treating these flights as "rewards" to pilots, and sending them in charge without previous Antartic flying experience, and possibly in the case of Flight 901 providing an "experienced" Antarctic traveller such as Peter Mulgrew who effectively added to the crew's visual mindset by apparently confirming their mistaken VFR position fixes
Now here I do agree with you. You're taking pilots who are used to flying at high altitude between Auckland and Honolulu, and asking them to fly to Antarctica and provide a sightseeing experience. And also waiving the usual requirement of one crew member having flown the route before, because these flights are considered routine. What person in their right mind would consider such a situation "routine"???

[QUOTE]-TC- wrote

oranmore wrote< Even Vette accepts that the DC-10's INS was not to be used to descend below the Minimum Safe Altitude.

It wasn't. The circling descent was visual all the way - the NAV track may have been used as a starting point, but the first visual fix was likely taken off Cape Bernacchi/Taylor Valley as they made the initial right turn

The INS was rearmed several times during the descent. Had they been truly VMC, there would have been no need. At no time during the descent did they get a view in the direction they were ultimately heading - ie. south. Look up the requirements for VFR flight. Those conditions did not exist.

Gasman wrote Collins flew through a hole in the cloud, nearly to ground level, without reference to the horizon ahead of him using the only the INS to confirm his position.

Wrong. Collins and crew flew a circling turn to the west, then a long racetrack pattern to the north-east [i.e. back to where they had just come from], where the cloud was patchy and visibility was much better, making a final left turn to rejoin the NAV track and take visual fixes below the cloud. Again, once a decision is made to conduct VFR flight, the INS should become redundant. 

The fact is that Collins was offered a radar let down by McMurdo, and in fact stated "crikey, that's what we want to hear". But then on seeing a hole in the cloud cover below him, he elected to descend through that - almost on a whim it would seem.

Again, no. The undoctored transcript ascribes only the words "that's what we want" to Collins, the rest is unintelligible. Far from descending through a "gap" in the cloud, he circled and flew a pattern taking the aircraft back to where it had come from (and where the cloud cover was patchy) in order to get a better set of visual fixes. Furthermore, he did not turn down a directed radar let-down - MacCentre acknowledges a request for a "VMC ... radar let-down", meaning that the descent is visual but the aircraft is expecting the radar station to keep an eye on them. If a directed radar let-down was offered or requested, then the words "radar vectors" would appear in the transcript, which they don't.

Incorrect again. From the transcript MCO:"Within range of 40 miles of McMurdo we have radar that will, if you desire, let you down to 1500ft on radar vectors". Cassin replied "Roger that's acceptable" and Collins "that's what we want to hear". Collins had also earlier remarked "very difficult to tell the difference between cloud and ice"

A flight should be conducted using EITHER VMC or IMC techniques, not a hybrid of both, which is what Collins was doing. He was almost certainly still using the INS during the descent because he realised he was not truly in visual conditions.

That's rubbish. I'm one of those who is very careful to avoid making a "Chippindale vs. Mahon" argument out of things - I much prefer to work from the facts independently, but Chippindale was grasping at straws here. If you follow his argument (i.e. that engaging NAV track/FMS guidance implies that the crew are unaware of their location) to its logical conclusion, then approximately 99% of all airline operations are conducted with the crew not knowing where they are.

I suspect you're not a pilot?? That's absolutely correct - most of the time, during cruise (the phase that INS (GPS in 2015) is designed for) pilots do not know with any accuracy where they are in relation to the ground!! They know where they are in relation to high altitude airways and waypoints, but the ground - no. 

A cautious and meticulous pilot will use all the tools at their disposal to operate the flight in the safest manner possible, and such a pilot, upon being briefed to go to an unfamiliar location, might be inclined to take the document with the nav co-ordinates home and find a map upon which they can visually plot their course, even if it takes them two weeks to get hold of the maps - that way, they'll have a visual reference as to where they are in relation to the NAV track at all times. Jim Collins did just that.

The point which most layman, including yourself seem to miss is that that "course" is relevant above the minimum safe altitude only. Which means, because you are supposedly above obstacles, the exact course track doesn't matter much. Once you enter the descent phase, an entire new set of rules come in to play. On TE901, those rules went out the window.

Which specific documented requirements should I be looking at? Because they tend to vary - as I understand it the VMC requirement for Antarctic flights at the time was 20 miles visibility. MacCentre told them they had 40. To state "At no time during the descent did they get a view in the direction they were ultimately heading" is inaccurate - if you look at the descent profile in the Chippindale report you can see that when they rolled out of the final left turn at the north-east extremity of the racetrack pattern they were at around 7,500ft, so in the course of the descent from FL075 to around FL018 - over a period of around five minutes - they were looking *straight ahead* in the direction they were going. This gave them plenty of time to confirm the visual fixes they needed, and let's not forget that Mulgrew confirmed every one of them. As far as "Had they been truly VMC, there would have been no need" goes...
Right, but "should become redundant" is not the same thing as "should not be used under any circumstances", much less that "if INS is being used, one is entitled to conclude that the crew don't know where they are or are in IMC".

Obviously, if you're flying VFR in an area with which you are familiar, then there's no point in engaging the INS in NAV mode whatsoever. However, would you not agree - as I said before - that it is at least plausible that a cautious, conscientious and methodical pilot expecting to operate VFR in an area with which they are not previously familiar may be inclined to use every tool at their disposal to provide extra layers of safety and confirmation? Of course, the only way to make the NAV track work in that sense and be able to refer to it quickly in conjunction with the more usual VFR map work would be to plot the co-ordinates on a map, so you could refer to that plot and be extra sure of where you are - and as it happens, Capt. Collins did precisely that - in fact he went to quite a lot of trouble to procure the maps required in the weeks leading up to the flight. Us engineers would call that a "belt-and braces" approach, and radar supervision would be the icing on the cake - in effect three discrete layers of safety.

This being the case, while using INS in NAV mode under VFR would be considered somewhat unusual in the majority of flight operations, in this case the INS is being used as a third-level safety backup, adding an extra layer of redundancy in the truest sense of the word.

OK, so Collins makes that comment when they are at FL180 or thereabouts, prior to making any decisions or requesting any arrangements with MacCentre for descent. The fact that Collins and his crew backtracked to the NE for some considerable distance where - as the passengers' photos prove - there was little or no cloud cover and could satisfy VFR requirements before making their final descent, arguably renders the remark moot.

As far as drawing conclusions from the CVR/ATC transcripts go, there are some very important caveats that must be borne in mind:

Firstly, the transcript that appears in the Chippindale report has to be considered at best somewhat unreliable - not out of any antipathy to the man himself or the conclusions he drew, but simply because in the act of going to RAE Farnborough and re-transcribing the tapes alone he broke several distinct rules and protocols that CVR experts consider must be observed in the process of transcription. The result of this is that the only transcript known to be valid by the standards of accident investigation is the original "Washington Transcript".

Secondly, the Chippindale transcript is an amalgam of the CVR and the ATC recordings whereas the Washington transcript is of the CVR alone. An interesting side-effect of this is that if you read the Chippindale transcript out of context it quite quickly becomes apparent that transmissions from MacCentre practically disappear for approximately 5 minutes leading up to impact; from that one could reasonably wonder why the crew didn't take this sudden radio silence as a warning. The necessary context required to explain this is that those 5 minutes (or thereabouts) were inexplicably erased from the ATC tape prior to Chippindale's arrival.

Getting back to your point, as I understand it, the first important call from MacCentre in the Washington transcript is at 12:03/00:32 and contains the phrase "VMC descent is approved", then later at 16:00/00:33 (approaching the top of descent) "one zero thousand requesting a radar letdown below this cloud" to which Cassin replies "that is affirmative nine zero one". As best I can understand, these exchanges give the crew the impression that MacCentre are OK with the descent in VMC, and that they are visible on radar. The continued flashing of the transponder coding indicator will reinforce the latter impression. The missing 5 minutes of ATC tape make determinations of any degree of accuracy with regard to what the radar/VHF operator said - and crucially what the HF operator did or did not pass on to the crew - practically impossible.


You would be correct in your surmise - I am in fact a software engineer. But to provide the necessary context I'll point out that I've been an aviation nerd since the age of 5, was interested in the Erebus accident from the age of 11 and was an Air Cadet for several years. Somewhere in a box in my Mum's house are the prospectus materials and application forms for RAF Officer training at Cranwell which ended up gathering dust as a result of a combination of long hair, rock music and pacifism diverting my life path in my mid-teens. :)

[If you'll permit a minor diversion, at this point I reckon it's worth mentioning that while Capt. Collins was an ex-RNZAF pilot, he was not of the more usual "officer" type (as Gemmell and Chippindale were) - he was selected for flight training from the ranks where he worked as a mechanic/engineer. His oft-remarked tendency towards prolific note-taking, meticulous planning and never taking the slightest detail for granted may well at least in part have stemmed from that - because it's a given that the brass keep a much more watchful eye on those in that position.]

That being said, I didn't say anything about the crew knowing where they were "in relation to the ground". As you say, at any altitude INS shows you where you are in relation to the waypoints. Collins' solution to that was to take the briefing co-ordinates home, get hold of the necessary topographical maps and plot the route on them (remember - he was a qualified navigator). Presumably he reasoned that on the day he would be VFR under radar supervision if conditions allowed - but just in case things looked dicey, he could refer to his AINS readout and cross-reference his position between waypoints against the plot on the map. All else being well, this is an elegant solution to provide an extra redundant level of safety very much in character for an engineering pilot with nuts-and-bolts experience.


I thought we'd get to that. In a general sense, the MSA applies primarily to IFR operations - the briefing stated that under VFR conditions, the aircraft was permitted to descend to any altitude offered by MacCentre under radar supervision. In the procedural sense, the approvals from MacCentre obviated the MSA requirement, and the evidence from the passenger photos proving that the aircraft was in clear air during the last stage of descent contradict Chippindale's assertion that the aircraft was in cloud for some time prior to impact. This was the state of knowledge at the time of Mahon's Royal Commission.

MacFarlane's work has uncovered quite a bit of contextual information in the intervening time, and Capt. Keesing is central to the suppositions one might draw. The Antarctic flights were Capt. Gemmell's brainchild, and he had been laying the groundwork for them since 1969. Prior to the first flights in 1977, Capt. Keesing (then Chief Pilot) submitted to NZCA his protocol for the Antarctic flights, which were more-or-less derived from the military procedures plus 2,000ft terrain clearance (unless invited to do otherwise) and a requirement of 20 miles visibility for VFR operations - the inherent limitation being that lack of such conditions required a return to base. Gemmell subsequently - and without Keesing's knowledge - submitted to NZCA his own protocol for the Antarctic flights, which allowed for IMC operations by overflying Erebus with an MSA of 16,000ft, then descending to an MSA of 6,000ft over McMurdo Station using a combination of the NDB (later TACAN) ground aid and radar. It would appear that NZCA had developed a tendency to simply rubber-stamp anything ANZ sent them provided that it fell within existing regulations, and as a result NZCA effectively had two separate and contradictory approved protocols for the Antarctic flights. According to some sources, the commander at MacCentre received only the first protocol (Keesing's) for approval, which he did. It is alleged that he stated that had he received Gemmell's, he would have dismissed it out of hand on the basis that overflying an active volcano is inherently dangerous at any altitude, and furthermore that the radar setup at McMurdo was unsuitable for the let-down procedure being proposed.

Gemmell takes the first flight down in 1977 and subsequently hands over the running of the Antarctic operation, but not before rescinding the requirement that Captains on their first trip down must be supervised - allegedly on the basis that the briefing he developed was sufficient. By 1979 Gemmell has ousted Keesing as Chief Pilot, and both that year and the previous year's Antarctic flights operated without his input. In the intervening time, not only has Chief Navigator Hewitt's "finger trouble" affected the computerised flight plan (and consequently the briefing materials), but the briefings have become an odd mixture of Keesing's and Gemmell's protocols - furthermore, the briefings were being given by pilots who had never flown the route themselves.

Apologies for long-windedness, but the upshot of this was that the AV presentation (which was shown to Chippindale) specified an MSA of 16,000ft whilst operating in the vicinity of Erebus, but according to the statements of every pilot who attended the same briefing as Collins, the track they were shown and the co-ordinates they were given indicated a track well to the west of Erebus where it would be logical to conclude that the MSA did not apply.

So, to recap - Chippindale's claim that Collins and crew busted MSA was predicated on the procedural information he got from ANZ, most if not all of which was supplied by or via Capt. Gemmell. For argument's sake, let's say that Gemmell might not even have been aware of how the procedures and briefings had changed since he handed things over, so Chippindale ends up drawing all his conclusions from the assumption that Gemmell's protocol (routing, MSA and all) was the only one in existence and operating at the time - an assumption which is erroneous.

After all that bumph, what we're left with is Jim Collins and crew turning back on track at FL075 and descending after heading some distance NE where there is little to no cloud (i.e. fully VMC from FL100). Radio reception is good at this point and both RT and the transponder light indicate they are under radar supervision. They start to take multiple visual fixes in earnest with Peter Mulgrew confirming what they see. If all else fails, Collins' "homework" is there to provide an extra layer of safety. They're aware of an MSA of 16,000ft in the vicinity of Erebus, but as far as they know they are nowhere near Erebus, nor do they intend to be unless they can see it. They've got 40 miles of visibility under the overcast surrounding Erebus, so - again, as far as they know - the base of the mountain should be clearly visible were it ahead of them.

Unfortunately, what they don't know is that Hewitt's attempt to change the co-ordinates surreptitiously defeats Collins' extra layer of safety. Thanks to the erasure of the ATC tape, whatever happened with the radar and RT is lost to posterity - defeating the next layer. At no point are they ever aware of this, but they are entirely reliant on the evidence of their eyes to keep them out of trouble - and for whatever reason, there was a deliberate decision made not to brief them or any of the other crews following Gemmell and Grundy's proving flights on the dangers of sector whiteout and how it can make their eyes deceive them. The VFR rules at the time weren't written to take optical illusions into account.

“OK - first off, an airline pilot has every right and entitlement to expect that the flight plan they receive on the day will be exactly the same as that they were briefed on;”

There is no such “right and entitlement”. In any event, the track that they were briefed on was to McMurdo Station.

 “That aside, what exactly would you expect them to check the waypoint co-ordinates against?”

The chart, being NZMS135, provided on the morning of the fllght (and which was also in the captain’s atlas.)

“Every pilot who attended the same briefing as Collins stated under oath that whatever the AV presented, the subsequent verbal presentation and supporting materials indicated a route down McMurdo Sound to the west of Erebus.”

Some more thinking required: Is “west of Erebus” inconsistent with being told that the waypoint is at McMurdo Station? As a matter of fact, a track from Cape Hallett to McMurdo Station would go over the summit of Erebus - but what if you didn’t know that? To know, you need a map showing Cape Hallett, McMurdo Station, and Erebus. NZMS135 showed all those features but at the briefing they were only provided with a photocopy of the inset from that chart, which did not show Cape Hallett. So it is a distinct possibility that the attendees at the briefing did not know that a track to McMurdo Station would go over the summit of Erebus. As regards the evidence, no witness said that they were told that the final waypoint was out in the middle of the sound, 27 miles to the west. All they said was that they were not told that the track went over Erebus.

IT is a plain and undisputed fact that those at the briefing were told that the final waypoint was at McMurdo Station. Mahon’s way around this unpalatable piece of evidence was to point to other allegedly contradictory material and say that it would have outweighed the audio. That was one of the more ridiculous of Mahon’s findings. If one thing says A and another thing says B, do you take your pick or do you ask which one it is? You ask, obviously. Was any such question asked at the briefing? No, because there was no issue as to where the final waypoint was located. That only became an issue the night before, when the captain plotted the position of the final waypoint“

"“So think about that: The captain was told, at the briefing, that the waypoint was at McMurdo Station.” I find that rather hard to believe.”

Try harder. Mahon Report, page 60, paragraph 164(b): “The pictorial representations showed the observers that the flight was down McMurdo Sound and these displays would, not unnaturally, take precedence over the spoken words indicating a direct track from Cape Hallett to McMurdo Station and indicating the NDB co-ordinates as the destination waypoint.”