Australia is a warm and dry continent that has been inhabited for around 60 thousand years. Despite this lengthy period of habitation, but because of the nature of the Australian landscape, there are relatively few bioclimatic records of the continent’s palaeoclimate, particularly from drier regions and that span the entire period of human occupation. Climate modelling presents an opportunity to fill this gap, but to date there have been few attempts to do so. This PhD investigates three periods of the palaeoclimate of the Australian region: the mid-Holocene (6000 years before 1950), 12 ka (kilo-annum before 1950), and the Last Glacial Maximum (21,000 years before 1950). The climate modelling is performed with the Community Earth System Model (CESM) and then downscaled with the Weather Research and Forecasting (WRF) model. The results of the simulations are compared against a baseline simulation for the pre-industrial (1850 common era). The simulation results are also compared, where possible, to existing reconstructions of the palaeoclimate from bioclimatic evidence.For the mid-Holocene climate of Australia and the equatorial tropics of the Indonesian–Australian monsoon region the finer-resolution WRF simulations reduce the bias between the model and bioclimatic data results for the mean temperature of the warm month, mean annual precipitation, and moisture availability index available in the bioclimatic proxy data set. The model results show that temperatures over southern Australia at the mid-Holocene and pre-industrial period were similar, and temperatures were slightly warmer during the mid-Holocene over northern Australia and into the tropics, compared to the pre-industrial period. During the mid-Holocene precipitation was generally reduced over northern Australia and in the Indonesian–Australian monsoon region, particularly during summertime. The mid-Holocene simulations show that the use of finer-resolution models can improve our understanding of the palaeoclimate.At the end of the deglaciation existing evidence indicates that conditions over Australia were becoming warmer, but there is an inconsistent picture of the monsoon and its impacts. The use of a finer-resolution model to simulate the palaeoclimate at 12 ka shows that the monsoon was active during this period, albeit in a less dominant form than the pre-industrial period. Temperatures were generally 0.25–2 °C cooler than the pre-industrial period in the south of Australia and the eastern tropics, however over the central and western equatorial tropics of Australia annual temperatures were 0.25–2 °C warmer, most notably in the austral spring.The third set of simulations centre on the Last Glacial Maximum when the temperatures over Australia were cooler and precipitation was generally reduced, although there were some locations of increased precipitation. The anomalies in tropical precipitation are explained by dynamical processes from changes in water vapour convergence. The simulations show that the Indonesian–Australian monsoon was active during this period, albeit with reduced precipitation. The east coast of Australia experienced increased precipitation resulting from increases in East Coast Lows.These findings are the first from the use of downscaled climate models to investigate the palaeoclimate over Australia at high spatial resolution. This is of particular importance over many inland parts of Australia where preservation of bioclimatic evidence is poor or non-existent, and therefore climate modelling is the only method by which this region’s past climate can be described. The high-resolution simulations quantify more precisely the changes in climate and identify the meteorological features that caused these changes, for example, the broadening of the monsoon trough, intensification of the Carpentaria heat low, meridional gradient of precipitation over the Lake Eyre Basin, and the zonal gradient of Tasmanian precipitation. The elucidation of these climatic features cannot be obtained from bioclimatic evidence or coarser-resolution models.Aboriginal populations were generally stable until the onset of the Holocene, with populations retreating to refugia in times of unhospitable climatic conditions. Existing interpretations of these climates have insufficient spatial or temporal resolution to accurately describe the effects on Aboriginal populations. The high-resolution simulations presented in this thesis greatly enhance our understanding of the climate faced by early Aboriginal populations, highlighting the improved value of using models such as WRF to simulate the palaeoclimate. Recommendations for future research priorities are made that could build on the results of this thesis. This work makes a significant contribution to the understanding of Australia’s palaeoclimate and provides a valuable resource that may be used for the interpretation of archaeological discoveries.