Wizofoz:
You cannot express speed without specifying speed
relative to what.
Absolutely right Wizofoz. I was trying to avoid being too technical. Now I was too fuzzy I guess.
The reference frame I meant is Earth Centered Earth Fixed, so speed relative to the (non rotating) earth.
Rephrased:
Groundspeed is the absolute value of
(a) the horizontal part of the ECEF velocity
(b) of the differential of the aircraft's position projected on the earth's surface (modeled by the WGS84 ellipsoid)
Feel free to shoot holes in the new definitions.
Or see it like this:
Assume the earth is a perfect sphere with a circumference of 40.000 km. When flying halfway around this earth at 10 km high you travel 20.000+PI*10 = 20031.415... km through the air, or 20.000 km measured on the surface. If you manage to do this in 20 hours, what is the average ground speed?
Definition (a): 20.031 /20 = 1001.57 km/h
Definition (b): 20.000 /20 = 1000 km/h
Whenever I've used or calculated ground speed in issues related to aircraft performance or particularly calibration of pitot-static system pressure error corrections, the definition used has always been horizontal component of absolute speed (although I don't recall any document where that is explicitly stated).
Genghis, I use that same approach in my aircraft performance related simulations. In the context of pitot-static corrections it perfectly makes sense to use velocity, since that is what drives the physics of the system.
In addition to that, the accuracy of the pitot-static system is such that the 0.16% difference between the two definitions (at 10km height) is probably within the measurement noise / tolerance, so it does not really matter how you define the ground speed in such a case.
For assessing the accuracy of GPS reported ground speed, the definition suddenly is important. Not in practical terms, because I am sure few people will care about the 0.16% difference, but for standardization/ certification ambiguity is a big thing.
Thanks for the feedback so far.