An evaluation of the weather research and forecasting model's ability in simulating fire weather for the Southwest of Western Australia

The southwest of Western Australia (SWWA) is prone to bushfires, and these have significant social, environmental and economic impacts. One of the major influences on fire weather is climate, and it is therefore important to understand current and future changes in fire weather in relation to climate change. Global climate models (GCMs) can be used to investigate current and future changes in fire weather, however, their coarse resolution (100 to 250 km) limits their applicability at the regional scale. Regional climate models (RCMs) are used to dynamically downscale GCMs to a regional scale (1 to 10 km). This study evaluates a RCM, the Weather Research and Forecast Model (WRF), in its ability to simulate fire weather over the period 1981 to 2010 in the SWWA . Fire weather was quantified using the McArthur Forest Fire Danger Index (FFDI), which is the current operational index used for fire danger warnings in Australia. FFDI was computed from both WRF and observational data, and results show that WRF captured the observed FFDI trend, albeit with a slight overestimation. Errors in WRF derived FFDI were mainly caused by WRF’s underestimation of relative humidity, which caused the FFDI to be overestimated, particularly along coastal regions. Fire Danger Ratings (FDRs), which are fire risk categories derived from the FFDI, were also examined, and it was shown that WRF was able to simulate low-risk FDRs with greater skill in comparison to high-risk FDRs. Although WRF performed poorly in simulating high-risk FDRs, these categories are a rare and unusual occurrence of the upper-distribution, and the majority of the FFDI distribution was well-represented by WRF. This study shows that overall, WRF was a useful tool for simulating fire weather over the SWWA