I have been struggling with the various arguments for choosing the southerly arc in preference to the northern one. It seems to me that the doppler/BFO shift is due to the instantaneous rate of change of the path length between the aircraft and the satellite combined with the instantaneous rate of path length between the ground station and the satellite, This same path length rate of change is what has driven the distance between the arcs at the hourly time intervals, ie I can't see how the two bits of information are sufficiently independent of each other to tell whether the southern or northern arc is a better fit.
Also, I have not seen any mention of windspeed or direction. If the aircraft has flown for 6 hours from it's last known positive position would not the speed (and probably the direction) have been influenced by the path passing through different weather patterns thus messing up the "assumed" speed and track constancy?
Lastly, can someone point me to some figures for the actual speed of the satellite relative to it's nominal geostationary position? Is the speed significant relative to the airspeed of a 777 or even relative to the likely winds seen at the assumed altitude of the aircraft in that area of the world?