Spatio-temporal optimization of tree removal to efficiently minimize crown fire potential

High-intensity wildfires have resulted in large financial, social, and environmental costs in the western U.S. This trend is not expected to decline soon, as there are millions of overstocked hectares at medium to high risk of catastrophic wildfires. Thinning is being widely used to restore different types of overstocked forest stands. Typically, thinning prescriptions are derived from average stand attributes and applied to landscapes containing a large number of stands. Stand-level thinning prescriptions have thus limitations when applied for reducing the risk of high-intensity wildfires. They use indicators of crown fire potential (e.g., canopy base height and canopy bulk density) that ignore variability of fuels within stands, location of individual cut- and leave-trees after treatments, and the temporal effects of these prescriptions for reducing crown fire potential over time. To address the limitations of current stand-level thinning prescriptions, a computerized approach to optimize individual tree removal and produce site-specific thinning prescriptions was designed. Based on stem maps and tree attributes derived from light detection and technology (LiDAR), the approach predicts individual tree growth over time, quantifies tree-level fuel connectivity, and estimates skidding costs for individual trees. The approach then selects the spatial combination of cut-trees that most efficiently reduces crown fire potential over time while ensuring cost efficiency of the thinning treatment