Variability and drivers of burn severity in the northwestern Canadian boreal forest

• Published in: Ecosphere (Washington, D.C) Vol. 9; no. 2
• Main Authors: Whitman, Ellen, Parisien, Marc‐André, Thompson, Dan K., Hall, Ronald J., Skakun, Robert S., Flannigan, Mike D.
• Format: Journal Article
• Language: English
• Published: Washington John Wiley & Sons, Inc 01.02.2018
• Subjects: boreal forest; Boreal forests; burn severity; Carbon; Ecosystem assessment; Ecosystem management; Ecosystem recovery; Ecosystems; Extreme weather; fire severity; fire weather; Forest & brush fires; forest fire; Forests; fuel structure; Land management; Landscape; Mortality; Prescribed fire; Risk management; Soil erosion; Species composition; topoedaphic context; Trends; Variability; Vegetation; Weather; wildfire; Wildfires; Canada
• ISSN: 2150-8925; 2150-8925
• DOI: 10.1002/ecs2.2128

Burn severity (ecological impacts of fire on vegetation and soils) influences post‐fire stand structure and species composition. The spatial pattern of burn severity may compound the ecological impacts of fire through distances to seed sources and availability of bud banks and seedbeds. Land managers require spatial burn severity data to manage post‐fire risks, ecosystem recovery, and assess the outcomes of fires. This research seeks to characterize and explain variability in burn severity in the northwestern boreal forest. We assessed burn severity one year post‐fire in six large wildfires that burned in 2014. We measured burn severity using the Composite Burn Index, surface Burn Severity Index, Canopy Fire Severity Index, and percent overstory mortality, describing a range of surface and overstory fire effects. Burn severity was variable, ranging from unburned residuals to complete overstory mortality and intense combustion. We related field measurements to remotely sensed multispectral burn severity metrics of the differenced Normalized Burn Ratio (dNBR), the Relativized dNBR, and the Relativized Burn Ratio. Diagnostic models of burn severity using relativized metrics had lower errors and better (though not significantly so) fits to the field data. Spatial patterns of burn severity were consistent with those observed in other large fires in North America. Stand‐replacing patches were large, aggregated, and covered the largest proportion of the landscape. These patterns were not consistent across the four mapped burn severity field metrics, suggesting such metrics may be viewed as related, but complementary, as they depict different aspects of severity. Prognostic models indicated burn severity was explained by pre‐fire stand structure and composition, topoedaphic context, and fire weather at time of burning. Wetlands burned less severely than uplands, and open stands with high basal areas experienced lower burn severity in upland vegetation communities. This research offers an enhanced understanding of the relationship between ground observations and remotely sensed severity metrics, in conjunction with stand‐level drivers of burn severity. The diverse fuel complexes and extreme fire weather during the 2014 fire season produced the complex patterns and broad range of burn severity observed.