Climate models use broad definitions of vegetation-based biophysical parameters that may not represent the continentally-specific nature of Australian vegetation well. This paper explores the impact on the January simulation of climate over Australia of this common simplification. First, we map the Australian distribution of vegetation types onto the default classification used in one high-resolution climate model. Second, through a search of the literature we chose replacement values for the biophysical parameters to better reflect the properties of eucalypts. Third, we choose several sets of biophysical parameters and allow these to vary regionally. We assess the impact of these changes on simulated rainfall, temperature and latent heat flux over an ensemble of six different Januaries. We find that the model simulates rainfall and temperature reasonably well over Australia in January and that replacing the default parameter set with a single set of more appropriate values degrades model performance slightly. Allowing the biophysical parameters to vary regionally leads to some small improvements in the simulation of temperature and precipitation. We find large impacts on the simulated latent heat flux. Overall, the model is not substantially improved by careful selection of eucalypt parameters. We comment that given the shortage of observed data on eucalypts and the almost total absence of biophysical parameters for other Australian vegetation types, this lack of sensitivity to the biophysical parameters is reassuring since it implies that the lack of data is not presently seriously limiting to our particular modelling capability.