Using Spatial Optimization to Create Dynamic Harvest Blocks from LiDAR-Based Small Interpretation Units

Abstract

Spatial and temporal differences in forest features occur on different scales as forest ecosystems 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 time and depend on the management objectives. This study presents a methodology for implementing these transient units, referred to as dynamic treatment units (DTU). LiDAR (Light Detecting and Ranging) data and field sample plots were used to estimate forest stand characteristics for 500-m2 pixels and compartments, and a set of models was developed to enable growth simulations. The DTUs were obtained by maximizing a utility function which aimed at maximizing the aggregation of harvest areas and the ending growing stock volume with even-flow cutting targets for three 10-year periods. Remote sensing techniques, modeling, simulation, and spatial optimization were combined with the aim of having an efficient methodology for assigning cutting treatments to forest stands and delineating compact harvest blocks. Pixel-based planning led to more accurate estimation of stand characteristics and more homogeneity inside the delineated harvest blocks while the compartment-based planning resulted in larger and higher area/perimeter ratio.