Real-time monitoring of thinning intensity using harvester data and remote sensing-based single-tree maps

Abstract

Achieving the recommended basal area and stem density of remaining trees in cut-to-length thinning operations can be a challenging task for the harvester operator. In recent years, thinning intensities have often violated recommended limits in the Nordic countries, and operators have expressed their desire for decision support and reporting assistance in thinnings. Here, we introduce a new concept named R-TIM (Real-Time Thinning Intensity Monitor) for the real-time monitoring of thinning intensity. R-TIM leverages a remote sensing-based single-tree level digital twin of the stand, maintained up to date during harvesting through matching harvested trees with their counterparts in the map. The properties of standing tree stock are calculated from the digital twin and the properties of removed stock from harvester measurements. Using simulations and field data from a test site in southern Finland, our tree matching algorithm attained an accuracy of over 85% for mean crane-tip positioning errors of 1 m or less. In the same positioning error range, the mean error in R-TIM-reported basal area of standing stock was below 0.5 m2/ha. The crane-tip positioning error estimated in the real harvesting was 0.71 ± 0.03 m, which is representative of modern crane-tip positioning accuracy and signals the feasibility of R-TIM for operational use.