Towards global prediction of fire risk in a changing climate

The identification of areas susceptible to fire is critical for planners, managers, and decision makers in developing effective mitigation strategies. At Climate X we are producing risk estimates to help businesses and communities mitigate and adapt for climate change related losses. Climate X provides risk scores and expected financial losses from a range of physical hazards.  The risks posed by wildfire are increasing in many regions and especially within the wildland–urban interface. We developed a machine learning model to predict changes in fire risk at 90m that can be applied globally. The approach combines meteorological drivers of fire weather utilising CORDEX regional climate models with local fire susceptibility modelling trained from Earth observation records of fire occurrence.  By 2100, we find an average 7% increase in fire risk across the US and western Europe under the rcp8.5 scenario. Our results demonstrate how the combined application of machine learning, climate and Earth observation data can provide time sensitive assessments of fire risk at global scale.