Prediction of Escaped Powerline‐Caused Fires in Alberta, Canada

Abstract

Powerlines have increasingly become a source of wildfire ignitions, contributing to more escaped and catastrophic fires during extreme fire weather. Temporary power shutoffs are implemented to reduce this risk, yet decisions on their implementations are often based on limited scientific evidence. Focusing on the boreal forests in Alberta, Canada, we developed an integrated framework to assess the likelihood of Escaped powerline‐caused fires (EPF) (EPFs) from occurrence and early spread perspectives. Using records of powerline‐caused fires from 1990 to 2023, we integrated MaxEnt and machine‐learning models to predict occurrence and escape probability from fire weather, fuels, and topography at the regional scale, overcoming the reliance on detailed powerline operational parameters. Although the frequency of powerline‐caused fires has increased in recent decades, most fires remain small and are contained during initial attack (IA). Fire weather emerged as the primary drivers of both fire occurrence probability and spread. Higher fire weather substantially increases the probability of powerline‐caused fire occurrence, while fires in coniferous forests are more likely to escape initial attacks. Under extreme fire weather (Fire Weather Index (FWI) >30), the EPF likelihood is more than three times higher than under moderate conditions (5 ≤ FWI <11), with the significant increases concentrated in southwestern Alberta.