A key calculation done by Inmarsat was determining the “Doppler shift” in the ping, or the slight change in the frequency of the signal caused by the movement of the aircraft relative to the satellite in space.
“From that process – a compression or an expansion of the wavelengths – you can determine whether the aircraft is getting closer or farther away,” McLaughlin explains. “It’s been a groundbreaking but traditional mathematics-based process that was then peer-reviewed by others in the space industry, and indeed contributed to by Boeing.”Unfortunately for the concept of "peer review", it seems that not all of the peer reviewers in the comments section of Aviation Week imediatelly agreed with the validity of Inmarsats calculations. The following is from Mr.
MylesFmNowhere
5:30 PM on 3/25/2014
It seems to me that the key time period where the doppler readings would determine North vs South route is in the portion of the journey shortly after the left-turn and crossing Malaysia. At that point, if the plane headed north, it already being in the North several degrees, the GEO satellite would see a faster 'approaching' plane if it were heading South than if it were heading North. But what gets me is the numbers. Even if you assumed a net approaching speed of 300meters/second reltative to the sattelite (probably a stretch), the amount of frequency shift would be only 1ppm. Are we to believe that this satellite resolves EACH and EVERY frequency source it's monitoring to 1ppm? Are they saving raw signals or does the satellite itself count cycles per second & save it? Curious.