Using Simulated 3D Surface Fuelbeds and Terrestrial Laser Scan Data to Develop Inputs to Fire Behavior Models

Abstract

Understanding fine-scale variability in understory fuels is increasingly important as physics-based fire behavior models drive needs for higher-resolution data. Describing fuelbeds 3Dly is critical in determining vertical and horizontal distributions of fuel elements and the mass, especially in frequently burned pine ecosystems where fine-scale fuels arrangement drives fire intensity and resulting fire effects. Here, we describe research involving the use of highly resolved 3D models. We create fuelbeds using individual grass, litter, and pinecone models designed from field measurements. These fuel models are distributed throughout the fuelbed to replicate fuel distribution in rectified nadir photography taken for each plot. The simulated fuelbeds are converted into voxel arrays and biomass is estimated from calculated surface area between mesh vertices for each voxel. We compare field-based fuel depth and biomass with simulated estimates to demonstrate similarities and differences. Biomass distributions between simulated fuel beds and terrestrial laser scan data correlated well using Weibull shape parameters (r = 0.86). Our findings indicate that integration of field, simulated, and terrestrial laser scanner data will improve characterization of fuel mass, type, and spatial allocations that are important inputs to physics-based fire behavior models.