Optimizing high-dimensional forestry for wood production and carbon sinks

Our model for optimizing stand-level wood production and carbon sinks includes individual-tree models for forest growth, an advanced model for soil carbon, detailed wood production economy, and an intertemporal objective for the value of wood production and carbon sinks/emissions. Carbon stocks include aboveground biomass and carbon in forest soil and in wood products. Optimization of the management regime, rotations and thinning timing, and type and intensity are carried out by reinforcement learning. Including the social price of carbon causes a regime switch from continuous cover forestry to clear-cuts, postponed lighter thinning, and a longer rotation, and, with a high carbon price, to solutions with pure clear-cuts or solutions utilizing stands as pure carbon stocks. Carbon price has profound effects on stand values, and the bare land value may well exceed the value before a clear-cut. The total average carbon stock is maximized with harvest and a long rotation instead of "no harvesting". Bioenergy, carbon capture, and storage (BECCS) always increases the value of wood production but not necessarily the value of carbon sinks. With BECCS, increasing carbon stocks in trees and forest soil remains optimal.