Short-term predictive capabilities are crucial to protecting life and property. These capabilities help us to understand how fires are likely to spread and develop over time and inform planning and responses from communities and emergency services.
Predictive tools are often optimised for local conditions, which means their effectiveness can vary depending on where they are used. They also need ground-truth data to improve accuracy and modelling to take into account changing conditions.
Many organisations are investing in enhancing predictive tools to make them more accessible, more adaptable to local conditions, and to model more complex fire behaviours. Examples of these complex behaviours include the atmospheric conditions created by extreme fires and the extent to which ember spotting increases the severity of fires. The challenges of modelling such complex behaviours are accompanied by difficulties in accurately predicting fire spread overnight and sourcing accurate and timely ground-truth data.
Improving how tools predict fire behaviour, and how effectively organisations use those tools to respond to fires could help to reduce the impact of fires on our communities. As with any modelling problem, data is critical. Improving how we integrate data and modelling with on-the-ground operations will maximise the value of these tools and our responses.