Modeling tree crown-mediated accumulation of downed woody debris in fire-maintained longleaf pine forests of the southeastern US from laser scanning data

Background Measurement of downed woody debris (DWD) across large areas is needed to realistically represent heterogenous fuel landscapes. Such measurements are key for improving predictions of fire behavior, effects, consumption, and emissions, and to develop and implement more effective forest management plans. However, DWD varies in amount, size, and structure, with its spatial distribution across the forest floor being driven by complex and interacting factors, such as forest productivity, tree growth dynamics, decomposition, disturbances, weather, and mortality events. Remote sensing, particularly airborne laser scanning (ALS), and ecological based knowledge can be leveraged to improve fuel characterization by estimating fuel properties at operational scales. In this study, we quantified DWD fuel loads using ALS data by linking expected rates of tree wood biomass (branch and bole biomass) annual deposition to surface fuel loads in fire-maintained woodlands dominated by longleaf pine. Terrestrial laser scanning (TLS) data were additionally used to develop crown proportion equations to estimate the branch biomass corresponding to fine (≤7.6 cm) and coarse (>7.6 cm) branch size components. We accounted for DWD accumulation by considering fire return intervals and decomposition rates by applying a spatially explicit implementation of the Olson fuel accumulation model, then accounting for fuel removal due to previous prescribed fire. Results The model provided unbiased estimates, suggesting that it can represent variability at broader spatial scales; however, we found a negligible relationship between fine woody debris estimates and independently collected field observations, which highlights the difficulty of both characterizing and assessing model performance of highly variable DWD loading at high spatial resolution. Many variables, such as consumption during previous fires, annual rates of branch deposition and decomposition, and the likely undersampling of the field data used for comparison contribute to the uncertainties of the estimates. Our analyses suggest that, a priori, the within-stand spatial distribution of DWD is not driven solely by the spatial distribution of live trees, and consumption during prescribed fire is likely patchier compared to constant and uniform rates considered in this study. Conclusions Further refinement and parameterization are required to model DWD and improve the characterization of within-stand fuel distribution; however, the methodology is a first step for integrating remote sensing into an ecological conceptual framework that provides DWD estimates to build heterogeneous fuel beds and may help fuel and fire managers more effectively address wildfire-related natural resource management challenges.