This study applies a benchmarking framework to assess a 39-member ensemble of regional climate models which have dynamically downscaled CMIP6 models over the CORDEX Australasian region. Four modelling centres contributed regional climate models to this ensemble using three regional climate models (CCAM, WRF, and BARPA-R), and a total of five model configurations. Assessment is conducted over the Australian continent using a separation into four major climate zones over a 30 year historical climatological period (1985-2014). Rainfall and near-surface temperatures are compared against six benchmarks measuring mean state patterns, spatial and temporal variance, seasonal cycles, long-term trends and selected extreme indices. Benchmark thresholds are derived either from previous studies or comparison with the driving model ensemble. Major model biases vary between ensemble members and include dry biases in northern and southern Australia, winter wet biases and a persistent low bias in the winter diurnal temperature range across all the modelling centres. Daily variability at large length-scales is comparable in the driving global climate model and downscaled regional climate model length-scales, and long-term trends are largely determined by the driving global climate model. Overall, the ensemble was deemed to be fit for purpose for impact studies. Strengths and weaknesses of the systematic benchmarking framework used here are discussed.

Preparing for environmental risks requires estimating the frequencies of extreme events, often from data records that are too short to confirm them directly. This requires fitting a statistical distribution to the data. To improve precision, investigators often pool data from neighboring sites into single samples, referred to as “superstations,” before fitting. We demonstrate that this technique can introduce unexpected biases in typical situations, using wind and rainfall extremes as case studies. When the combined locations have even small differences in the underlying statistics, the regionalization approach gives a fit that may tend toward the highest levels suggested by any of the individual sites. This bias may be large or small compared to the sampling error, for realistic record lengths, depending on the distribution of the quantity analysed. The results of this analysis indicate that previous analyses could potentially have overestimated the likelihood of extreme events arising from natural weather variability.