Assessment of Plant Functional Types on Burn Severity in Interior Alaska, 2017-2020

Wildfires are increasing in both frequency and severity across boreal forests, making remote observations of burn severity paramount for above and belowground fuel monitoring, modeling, and risk management. However, traditional burn severity metrics such as the Normalized Burn Ratio (NBR) are often unilaterally applied across burn scars, despite evidence suggesting burn severity may vary with vegetation structure and composition. We present data to calculate a new, scalable integrative vegetation-fire index (IntFire) that incorporates plant functional type (PFT) specific sensitivities to burn for improving severity assessments. The IntFire index was developed using ninety-four nonzero composite burn index (CBI) observations measured across four upland and lowland boreal forest fires (total burned area: 909 square kilometers) in central Alaska. We mapped pre-fire PFT cover using 3 meter resolution PlanetScope imagery, synthsesizing very high-resolution WorldView-2 and NASA G-LiHT imagery to inform fine-scale classifications. Linear and logarithmic regression models were used to correlate each PFT with its best performing burn severity metric then integrated into IntFire. We demonstrate the scalability of IntFire by applying the optimal regression models to both fine and meso-scale optical products (i.e., PlanetScope and Harmonized Landsat/Sentinel, HLS) using derived and spectrally unmixed PFT maps.