Recent investigations in complex terrain have found that remote sensing instrumentation commonly finds mean wind-speed differences when compared to cup anemometery. In many cases the difference is found to be an underestimation and varies from 2 to 9% depending on topology. We describe these differences in a theoretical sense for a five-beam sodar. An investigation is conducted on a New Zealand ridge with a five-beam sodar and three computational models, consisting of a potential flow model and two computational fluid dynamical simulations, OpenFOAM and the industry standard software WindSim. All models predict the difference to within 0.1-2.5%. A comparative assessment is made and it is found that, given the computing overheads, the potential flow model provides a good compromise in the prediction of mean wind-speed difference.