Spatial connectivity in tree-level decision-support models using mathematical optimization and individual tree mapping

Abstract

Tree-level planning is gaining relevance supported by continuous advances in forest remote sensing methods. Spatial connectivity under spatial optimization methods applied to individual tree mapping data remains unexplored. This article presents a spatially explicit mathematical formulation that ensures the tree cuttings conforming a corridor can connect two points across a forest landscape. The created paths facilitate operational harvesting linking forest planning solutions to forest machinery operations. The model integrates production, economic and connectivity constraints using mixed integer programming as optimization method, while tree positions and attributes were mapped using airborne laser scanning. Model performance was tested towards different harvesting targets and for the assessment of trade-offs involving the connectivity of the solutions and the economic performance. The showcase area to test the model is pine forest located in Central Spain comprising more than 9000 detected trees. The results showed the mathematical formulation in the model is effective at creating paths connect proposed areas using optimized tree harvests. Global optimality was reached fast using mathematical programming for a complex and large combinatorial problem. The presented model is another step forward in the design of multi-objective tree-level planning models, enhancing the assimilation of planning solutions towards forest operations capable to maximize the use of individual tree mapping data.