Wildfires are one of the major natural hazards facing the Australian continent. Chen (2004) rated wildfires as the third largest cause of building damage in Australia during the 20th Century. Most of this damage was due to a few extreme wildfire events. For a vast country like Australia with its sparse network of weather observation sites and short temporal length of records, it is important to employ a range of modelling techniques that involve both observed and modelled data in order to produce fire hazard and risk assessment products with utility. This paper details the use of statistical modelling of both observations and climate model simulations (downscaled gridded reanalysis information) in the development of extreme fire weather potential maps.
Indicators of fire weather potential such as the McArthur Fire Forest Danger Index (FFDI) are widely used by fire management agencies to assess fire weather conditions and issue public warnings. FFDI is regularly calculated at weather stations using measurements of weather variables and fuel information. As it has been shown that relatively few extreme events cause most of the impacts, the ability to derive the spatial distribution of the return period of extreme FFDI values contributes important information to the understanding of how potential risk is distributed across the continent.
The long-term tendency of FFDI has been assessed by calculating the return period of its extreme values from point-based observational data using extreme value distributions. The spatial distribution of the return period of FFDI was obtained by applying an advanced spatial interpolation algorithm to the measurements made at weather stations. As an illustration, maps of 50 and 100-year return period of FFDI under current climate conditions are presented (based on both observations and reanalysis climate model output).
