Thank you Michael for the fascinating link to the radar track. I'd forgotten to look for that but what an interesting picture it shows.
I guess it demonstrates that if turns are flown smoothly they're not easy to discern from the comfort of the cabin, as there were clearly plenty of turns on that approach and subsequent landing! As BOAC points out there is some corruption of the tracking data, but when zoomed in significantly the genuine track is easy to make out.
Bearing in mind the previous discussion about runway 04 being the normal choice (even with a small tailwind) it's curious that with the wind approximately 090 the initial approach was to 22 (as my friend had originally told me), but I'm confused by the actual track shown. It appears that the aircraft was never even near to being set up for an approach to 22 as, unless I'm missing something obvious (perfectly possible ...
), they seemed to be tracking almost due west towards the threshold of 22 before the g/a. and tight left turn. Even after that point they don't seem to follow the prescribed missed approach procedure for 22, heading roughly 165 then 180 before joining the pattern for a 04 approach, although as Tagron states, presumably this could have been to avoid conflicting traffic. As a result I guess they never needed to get to the 3000' hold at NERAS before making the right turns necessary to set up for an approach to 04. This certainly seems to bear out the reason for the crew announcement to the passengers of the attempt to land in the opposite direction.
Thanks Tagron for your suggestion about the weather being below minima for the 22 approach. I'm kicking myself that I no longer have the METAR data for 1900Z on Saturday. Does anyone know if there is a site where archived raw METAR data from 3 days ago is still available?
So, if we accept there were good (albeit unknown to us) reasons for the initial approach to be on 22 and that the weather might have played a part in the missed approach (hence the curious PA announcements) then the only question remaining for me is why the initial track was almost 090 towards the runway 22 threshold instead of aligned with it? Could it simply be that there would still have been time for a left turn at a late stage to align with 22 had the approach continued as originally planned?
I've found this to be a very interesting exercise in trying to piece together all the facts to explain what was clearly a somewhat unusual scenario. Thanks all for helping me do so.