Modelling and Analysis of Lightning-Induced Wildfires in Australia

Lightning-induced wildfires lead to significant loss of life and extensive property damage worldwide. This issue is especially critical in southeast Australia, where such wildfires account for 80-90% of the total area burned, emphasising the need for a comprehensive understanding of the contributing factors and mechanisms that drive these events. This study aims to investigate the complex interactions between climate, topography, lightning activity, and fire events in New South Wales (NSW), Australia. By analysing comprehensive datasets from 2017-2021, including ignition records, meteorological data, topographical information, and fuel characteristics, this research seeks to identify the key factors influencing lightning-attributed wildfires and predict the probability of lightning-caused fire occurrence. A Random Forest model is trained and tested to estimate the probability of fires caused by lightning strikes. Model performance was assessed through the Receiver Operating Characteristic, with an Area Under the Curve (AUC) around 0.7 in the validation datasets, indicating a good agreement between the estimated probabilities and the reported lightning-caused fires. The identified key factors that influence lightning fire ignitions include humidity, elevation, temperature, rainfall, soil moisture, and fuel moisture, highlighting the dominant influence of weather variables on wildfire ignitions. The preliminary results demonstrate a potential link between the geographic distribution of lightning-induced fires and the temperate climate zones, possibly due to the presence of dense vegetation and seasonal weather patterns. Our ongoing efforts focus on further refining the predictive model and conducting a more extensive analysis of the data to enhance our understanding of the dynamics of lightning-induced wildfires. Ultimately the study will provide insights for effective risk management and mitigation of lightning-caused wildfires in the regions prone to wildfires.