Trends in extreme indices in 60-year simulations over Australia.

We investigate indices of extreme temperature and precipitation in simulations of the Australian climate over 60 years (1950-2009). Three different physical configurations of the Weather Research and Forecasting (WRF) modelling system are selected from a large ensemble to generate as much independent information as possible while spanning the uncertainty range found in the full ensemble. Boundary conditions from NCEP/NCAR Reanalysis 1 are used to drive WRF simulations at 50 km spatial resolution.

The unprecedented length of the regional climate simulations allows the study of long-term trends in the recent climate. In particular trends in various precipitation and temperature extreme indices are calculated for both model output and observational datasets. The model ability to reproduce the trends in the climate indices and their spatial patterns is examined. Model output is also compared with indices calculated from the driving global reanalysis to assess the potential benefits of dynamical downscaling for the study of changes in climate extremes.

We find that the RCMs compare well with observations in terms of the inter-annual to inter-decadal variability of different climate extremes at continental scales. The RCMs show good agreement with observation for maximum temperature extremes, whereas for minimum temperature ones the RCMs show a stronger warming trend. However, the differences among observational datasets suggest that the uncertainty in observations may also play a role in this disagreement. At regional scales, partly contrasting results are obtained depending on the area and the index examined. Although no substantial benefit is obtained for precipitation at this spatial resolution, especially in the East Coast and the Northwest, the RCMs are able to improve the spatial patterns of recent changes in extreme temperature provided by the large-scale driving reanalysis.