Andrew Dowdy ^* Andrew King

• The University of Melbourne, Parkville, Australia

This study examines several methods and new ideas for climate analysis, including expanded ensembles, that combine model projections from different greenhouse gas emissions pathways and different time periods. These methods are tested on Australian projections data previously made available based on outputs from the Energy Sector for Climate Information (ESCI) project that included all available dynamical downscaling approaches with bias correction designed with attention to detail on extremes. The expanded ensemble method provides larger sample sizes to help enhance confidence, with results showing that the projected changes per degree of global warming have relatively small differences when calculated using two different emission pathways and different time periods, with smaller differences than variations between individual models in the ensemble. Results include maps of mean values and extremes for temperature and rainfall metrics, as well as for compound events associated with dangerous bushfire weather conditions, providing new insights on climate change in Australia. The results also show that extremely dangerous fire conditions such as those of the Black Summer 2019/2020 and of Black Saturday in February 2009 are currently still very rare, but that climate change has already increased the chance of their occurrence, as well as larger increases projected in the future for higher amounts of greenhouse gas emissions. New analysis is also presented for changes in rainfall-based metrics associated with agriculture and biogeography such as Goyder's Line, discussed in relation to the use of climate analogues for adaptation decision making.