A survival guide for assessing global fire risks to natural and human systems

Compared to tropical cyclones and floods, wildfires have historically been regarded as a secondary peril, which has contributed to the lack of a systemic perspective on global wildfire risk — an understanding increasingly required in disaster risk reduction. This gap is aggravated by the complex and interlinked climatic and anthropogenic drivers of wildfire occurrence, and the diverse and often difficult to interpret landscape of satellite-derived fire datasets, which complicates their use and may lead to inappropriate data choices. We address these challenges by providing an overview of the systemic pathways of fire impacts, available remote-sensing-derived fire products and modelling efforts. We highlight research gaps and provide recommendations on terminology and data choice. Burned area products are generally used to estimate the extent of burned land and exposed entities, but are characterised by relative high omission errors. Active fire locations constitute a sample taken at the time of satellite overpass, but can provide a proxy for fire intensity. Caution is warranted in intensity metric selection, as these can be biased by latitude, biome and missed measurements. Both data types have been used to derive event-based products. These offer key fire behaviour metrics, such as rate of spread and fire line intensity, yet limited operability results in irregular update cycles. Estimations of population’s exposure to fires should also take into account disperse effects from smoke (e.g. PM_2.5 concentrations ≥ 15 µg m⁻³). In parallel, ecological impacts are often associated with burn severity, while natural fire regimes and shifts thereof should be considered.