Downscaled outputs of nine General Circulation Models (GCMs) and Regional Climate Models (RCMs) for local climate change per degree of global warming along with global warming projections and atmospheric pCO2 were used to construct probabilistic change scenarios by applying Monte Carlo Random Sampling (MCRS) techniques. A subset of these probabilistic climate change scenarios was used to study the individual influences of rainfall, temperature and pCO2 on wheat production at contrast locations in South Australia by using Agricultural Production Systems sIMulator (APSIM) -Wheat model. The individual effects of rainfall, temperature and atmospheric pCO2 on wheat production were isolated by looking at one variable each time while keeping the other two at the most likely scenarios.

Two sites: Keith and Minnipa, representing a wet site and a drier site respectively, were selected. Impacts on yield production, on failed sowing years due to the changed rainfall regime and on risks of wheat production were evaluated. Risk of wheat production was defined as the conditional probability of not exceeding the critical yield threshold, which was determined through economic analysis.

Research results show that rainfall played a significant role in wheat yield, in the failed sowing year and in risk level at both locations. However temperature and pCO2 have less impact compared with rainfall. Quite different influential patterns on wheat yield at these two locations were observed under these three driving forces.

Key word: climate change impact, wheat production, APSIM-Wheat Model, risk level