I know it'll be approximate, and I even doubt it will make any difference in what I intend to do - calculate bistatic Doppler signals from flight data and compare them with real signals in radio spectrograms.
That will of course be a question of geometry.
At very low and very high elevation angles between ground sensor and target the effect of height errors becomes negligible. The worst sensitivity to height errors will be for an elevation angle of around 35 degrees.
However the actual error will depend upon the distances involved. Clearly a 1000ft height error on a target at FL300 will have far less effect than a 1000ft height error on a target at 3000ft.