Predicting the spatial distribution of the abundance of Siebold’s beech in a montane cool-temperate region based on environmental factors

Using a geographic information system (GIS), our goal was to predict the potential distribution of Siebold’s beech (Fagus crenata Blume) in a montane cool-temperate region at a fine spatial resolution based on topographical features. The study was conducted in Akashibayama National Forest in the village of Kamikawa, Niigata Prefecture, central Japan. Species composition was investigated in 28 sample plots selected in the study area. A digital elevation model (DEM) was created, and topographical, hydrological, and light factors were calculated using the DEM. Then, the relationship between species composition and these environmental factors was examined using tree-based multivariate regression to derive regression trees. The species composition for the six major species selected, which included Siebold’s beech, was used as the response variable, and environmental factors were used as explanatory variables. For the derived tree-based regression model, the shaded relief, slope, specific catchment area, and curvature were selected as explanatory variables. The model classified natural vegetation into six forest types, and the result was consistent with the moisture preferences of these major species. The model was applied to the GIS to predict and map the species composition of the major species, especially the relative basal area of Siebold’s beech.