Societal demands for forest biodiversity and ecosystem services (BES) are growing and diversifying, which necessitates careful decision-making in forest management. Optimisation methods can support the decision-making process and resolve trade-offs between various BES objectives, and are successfully applied for forest management at the stand and landscape levels. However, there is an increasing interest in optimising management planning at an even finer resolution: the individual-tree level. This systematic review summarises the studies that optimise individual-tree decisions in forest management, taking individual-tree data as input and prescribing a management decision for every tree as the output. Tree-level management planning directly incorporates relevant tree attributes into the planning process - rather than relying on aggregated proxies - and complements developments in precision forestry, remote sensing and autonomous forest machines. We identified 47 relevant studies, which use diverse optimisation techniques such as heuristic algorithms, mathematical programming and machine learning. Several management targets and constraints (e.g., economic value, biodiversity and the structural features of the forest) have been addressed in the studies. Rich information about individual trees is available, although the attributes typically gathered during field inventory, like species, tree height and diameter at breast height, are still the most commonly used in decision-making. Identified directions for future research are to integrate natural disturbance risk predisposition, link tree-level optimisation with management plans at larger spatial scales and develop the real-world implementation of the individual-tree decisions.
Perry et al. (Sun,) studied this question.