Enhancing CAFE DSS capabilities for multifunctional forest management that optimises different ecosystem services: a case study with a semiarid pine plantation

Process-based models (PBMs) are increasingly required to evaluate forest management under climate change, but their use in decision support is limited by site-specific parameterisation, which remains a manual, expert-dependent step performed outside the decision framework. We extended the CAFE Decision Support System, which couples the Biome-BGCMuSo model with multi-objective evolutionary optimisation, with (i) an autocalibration plugin that calibrates the model within the same optimisation environment used to generate management alternatives, and (ii) new indicators that respond directly to management and disturbance: a fuel-based ignition risk index and prospective indices of resistance and resilience to an imposed drought. The workflow was tested in a semi-arid Pinus halepensis plantation in eastern Spain, where a long-term thinning experiment provides independent field data. Calibrated against transpiration, soil water content and stored biomass, the selected parameter set achieved Nash-Sutcliffe efficiencies of 0.65, 0.73 and 0.95 and the same set was selected under every aggregation criterion tested. A global sensitivity analysis identified maintenance respiration, leaf nitrogen in Rubisco and boundary layer conductance as the dominant parameters, acting mainly through interactions. Across 100 Pareto-optimal solutions, a first thinning intensity of 60% or higher was observed in half of the cases (49 out of 100 solutions), followed by a second intervention around 12–13 years, consistent with the treatment and the duration of its effects documented in the field. Relative to the unmanaged control, the alternatives increased deep percolation 3.5-fold, reduced ignition risk by 24%, at the cost of a 48.9% reduction in standing biomass. Mean drought resistance and resilience did not differ from the control, but 37 alternatives improved on it in both. The study provides a replicable framework for coupling PBMs with multi-objective optimisation in forest planning, in which the case study serves as a test of the workflow rather than as a silvicultural recommendation.