Underfire:
It was the analysis of the INMARSAT data that lead them to the current search area, much the same as AF447.
Not entirely correct.
While the Imarsat data gave the general area to look in, the final ping-ring still gives us an arc a thousand nautical miles or so long. All the search zones were along that arc - but which segment of that arc is correct?
There needs to be a method of fine-tuning the information.
Ian W
Silvertate:
It was the realisation that the aircraft was flying lower.
I think the reasoning was the reverse. Assume last ping was at out-of-fuel. That gives a fuel burn rate which can then be reverse looked up to find the cruise levels/speeds that would result in that burn rate now fit those speeds/levels to the several INMARSAT rings the one that most closely fits tells you where on the final ring the aircraft was out-of-fuel.
Amended due to post 9719 by Ian W below.
You are right that the fuel burn remains the same, for all the various speed/altitude scenarios, because the fuel load is known and the time is known. But we still have various potential tracks depending on the speed/altitude of the aircraft.
You can have a high-TAS-speed flight at high altitude - and end up south and west of Australia. (Yellow track.)
You can have a low-TAS-speed flight at low altitude - and end up north and west of Australia. (Purple track.)
So we come back to the same question I posed above - what made the search teams decide that the low level scenario was the more likely? Although the burn-rate is a known factor, there are still numerous speed/altitude combinations that will achieve that same fuel burn. Thus there still must be some other information that prompted a look towards lower and thus altitudes/slower TAS speeds (the purple line). Perhaps it became likely that the aircraft never climbed again, after its assumed descent to to low level while skirting Malaysia.
Note how each hourly ping lies on the same ping-ring, but end up hundreds of miles apart - even though the fuel burn is the same for each hour flown.
These tracks below are for illustration only, and are not to any exact speed-scale. *
