A few days ago Capt Kremin pointed out that if the autopilot was in a mode that maintained a constant magnetic heading, there would be a significant turn to the East at the latitudes of Australia due to the large change in magnetic deviation. I looked at this in the context of the modelling that AAIB, NTSB and others are doing, to understand if the differences could be seen in the INMARSAT ping data that it is believed the authorities have. The answer seems to be not, as the two model tracks in the diagram (great circle in grey and constant magnetic heading in yellow) have very similar ping signatures with average differences of less than 15km (r.m.s. error) in the distances for each ping from the sub-satellite point. This may be less than the noise in the INMARSAT data. The two final southerly points of the tracks are 640km apart (one hour South of the last ping arc). Since a single search area was used by AMSA at the start of their activity, some method must have been used to decide between the track options.
Note: The only publically released data point on the map is the red final arc. The other ping arcs (only two shown) are just modelled results showing they match both tracks. The track direction is just an example that roughly matches the data in the AMSA maps.
