Search

Frequent, small wildfires can pose dangers to homes in the wildland–urban interface, but are not often included in wildfire hazard models. We assessed patterns of small wildfire occurrence probability in the Northeast region of the United States, focusing on (1) spatial and seasonal variations; (2) differences between small and large fires (size threshold of 4 ha); and (3) how predicted probabilities are influenced by inconsistent wildfire definitions in urbanised landscapes. We analysed fire incident report data from 2005 to 2017 to parameterise maximum entropy (MaxEnt) models based on land cover, topography, climatic water deficit, soil moisture and road density. Overall, wildfire occurrence was highest in areas with lower agricultural cover and with more low-density urban development (explaining 53.5 and 28.6% of variance, respectively, in our region-wide model), while larger fires were concentrated in areas with intermediate levels of development, higher climatic water deficit and more rugged topography. These patterns were largely consistent when we assessed models for individual states, but differences in wildfire reporting patterns led to differences in the effect of urban development on fire probability. Our results provide novel understanding of small wildfire patterns in the Northeast and demonstrate the need to more reliably quantify these hazards. We analysed fire incident report data to model wildfire occurrence probability across the Northeast region of the USA, in order to better understand how patterns and drivers of small fires (0.1–4 ha) differ from larger fires. Hazard from small fires is greater near low-density urban development. [ABSTRACT FROM AUTHOR]