Doppler orbitography uses the Doppler shift in a transmitted signal to determine the orbital parameters of satellites including range and range rate (or radial velocity). We describe two techniques for atmospheric-limited optical Doppler orbitography measurements of range rate. The unstabilised technique determines the Doppler shift directly from a heterodyne measurement of the returned optical signal. The stabilised technique aims to improve the precision of the first by suppressing atmospheric phase noise imprinted on the transmitted optical signal. We demonstrate the performance of each technique over a 2.2km horizontal link with a simulated in-line velocity Doppler shift at the far end. A horizontal link of this length has been estimated to exhibit nearly half the total integrated atmospheric turbulence of a vertical link to space. Without stabilisation of the atmospheric effects, we obtained an estimated range rate precision of 17μms-1 at 1s of integration. With active suppression of atmospheric phase noise, this is improved by three orders of magnitude to an estimated range rate precision of 9.0nms-1 at 1s of integration, and 1.1nms-1 when integrated over 60s. This represents four orders of magnitude improvement over the typical performance of operational ground to space X-Band systems in terms of range rate precision at the same integration time. The performance of this system is a promising proof of concept for coherent optical Doppler orbitography. There are many additional challenges associated with performing these techniques from ground to space that were not captured within the preliminary experiments presented here. In the future, we aim to progress towards a 10km horizontal link to replicate the expected atmospheric turbulence for a ground to space link.