Are estimates of anthropogenic and natural influences on Australia's extreme 2010-2012 rainfall model-dependent?

Australia experienced much above average rainfall in association with strong, extended La Nina conditions during 2010-2012. Was the heavy Australian rainfall influenced by La Nina conditions and/or anthropogenic greenhouse gases? We investigate the relative contributions of anthropogenic climate change and natural climatic variability to the likelihood of heavy Australian rainfall using three distinct model datasets. Area-average rainfall anomalies for model simulations with natural forcings only were compared to simulations with both anthropogenic and natural forcings using 16 models participating in the Coupled Model Inter-comparison Project Phase 5. Using fraction of attributable risk to compare the likelihood of unusual rainfall between the parallel experiments, we find attribution statements are uncertain, with FAR values sensitive to the attribution parameters considered, including thresholds, regions and seasons. When heavy rainfall probabilities were next investigated in ensembles of two atmospheric general circulation models, run with and without anthropogenically-induced sea surface temperature changes, results were model-dependent. Overall, the attribution of seasonal-scale heavy Australia rainfall to a particular cause is likely more complicated than for temperature extremes. As estimates of the greenhouse gas attributable change in rainfall risk may depend on the model datasets considered, it is also useful to consider model outputs from several datasets and using various estimates of counterfactual surface conditions to establish robust attribution statements for extreme rainfall events. In contrast, comparing the likelihoods of heavy rainfall during simulated La Nina years with El Nino/neutral years reveals a substantial La Nina influence on Australian rainfall during 2010-2012 that is robust to changes in the attribution framework.