Evaluating an individual‑based model's ability to reproduce fine‑scale spatial structure in boreal mixed forests

Abstract

Process-based forest models are increasingly used to guide management, but few are validated against fine-scale spatial patterns that emerge from neighborhood interactions. We tested whether the spatially explicit individualbased model SORTIE-ND, which simulates growth, mortality, and recruitment as functions of neighbourhood interactions among individual trees, can reproduce observed fine-scale structure in boreal mixedwoods. Using long-term data from the Lake Duparquet Research and Teaching Forest (Qúebec) station, we initialized simulations from transect plots representing younger post-fire stands and compared simulated outcomes to independent hectare plots of similar ages along a 249-year chronosequence. The spatial structure was quantified with inhomogeneous L-functions for univariate and bivariate patterns, and model performance was assessed by comparing observed curves to simulation envelopes. SORTIE-ND reproduced fine-scale patterns for balsam fir and trembling aspen, showed partial agreement for white spruce, and failed to match the observed clustering of paper birch. Cross-species patterns were captured for fir–aspen but not for pairs involving white spruce. These results indicate that SORTIE-ND can approximate fine-scale spatial patterns for dominant species in boreal mixedwoods, but limitations remain where key processes (e.g., vegetative propagation, substrate dependence) are under-represented. We discuss implications for stand- to landscape-scale management and recommend model extensions and more independent validation to improve generality.