Adaptation of a Spatially-Explicit Individual- Based Forest Dynamics Model SORTIE-ND to Conifer-Broadleaved Mixed Stands in the University of Tokyo Hokkaido Forest

Abstract

We calibrated a spatially-explicit individual-based forest dynamics model SORTIE-ND to conifer-broadleaved mixed stands in the Uni- versity of Tokyo Hokkaido Forest, and evaluated its accuracy of re- constructing past stand structure development under a single-tree selection system. We also modified the model so that it would be applicable under various geographic conditions by incorporating vari- ables that were derived from the Digital Elevation Model (DEM). We used inventory data collected from 12 permanent plots to estimate the model parameters. The analysis involved 23 tree species and species groups. The model consisted of five submodels: light, growth, mortality, snag dynamics, and recruitment. After the calibration, we evaluated model accuracy on a stand scale. The results of the evaluation showed that SORTIE-ND fairly accurately reconstructed the past development of basal area, diameter class distribution, and species composition. It was indicated that SORTIE-ND can be flexi- bly adapted to the regional characteristics of the study site, and has a large potential for exploring single-tree selection regimes that allow for the conservation of stand structure in conifer-broadleaved mixed stands.