Background The 2023 Quebec wildfires in Canada attracted international attention, affecting 4.5 M ha of boreal forest.

Aims We investigated if causal factors could be identified that explained spatial variation in the wildfire burn severity of the Quebec 2023 wildfires.

Methods We conducted a series of Boosted Regression Tree (BRT) models to investigate if the spatial distribution of the within-fire burn severity index could be attributed to climatic, vegetation and topographic variables.

Key results The most important model variables were Topographic Wetness Index (TWI) (relative contribution 40%), Fire Weather Index (FWI) (relative contribution 23%), Topographic Position Index (TPI) (relative contribution 14%), forest age (relative contribution 14%) and vegetation type (relative contribution 9%). Higher burn severities were associated with larger FWI values, smaller TWI and TPI values, and forest age circa 20–40 years.

Conclusions The model variables influence burn severity by the moisture content and flammability of plant biomass. Forest management is implicated indirectly through the accumulated impacts of harvesting which have skewered age class distributions toward younger forests.

Implications Given projected worsening fire weather conditions, it is prudent to consider how forest management can be modified by, among other things, protecting and restoring depleted older forest age classes within the forest management zone.