Using spatial optimization to create dynamic harvest blocks from LiDAR-based small interpretation units

摘要

Spatial and temporal differences in forest features occur on different scales as forestecosystems evolve. Due to the increased capacity of remote sensing methods to detect these differences,forest planning may now consider forest compartments as transient units which may change in timeand depend on the management objectives. This study presents a methodology for implementingthese transient units, referred to as dynamic treatment units (DTU). LiDAR (Light Detecting andRanging) data and field sample plots were used to estimate forest stand characteristics for 500-m2pixels and compartments, and a set of models was developed to enable growth simulations. The DTUswere obtained by maximizing a utility function which aimed at maximizing the aggregation ofharvest areas and the ending growing stock volume with even-flow cutting targets for three 10-yearperiods. Remote sensing techniques, modeling, simulation, and spatial optimization were combinedwith the aim of having an efficient methodology for assigning cutting treatments to forest standsand delineating compact harvest blocks. Pixel-based planning led to more accurate estimationof stand characteristics and more homogeneity inside the delineated harvest blocks while thecompartment-based planning resulted in larger and higher area/perimeter ratio.