Classification of Fagus crenata forests and Quercus mongolica var. grosseserrata forests with regard to climatic conditions

We developed a statistical distribution model for Fagus crenata forests and Quercus mongolica var. grosseserrata forests to clarify the climate-distribution relations for the two dominating cool-temperate-forest types. The national standard grid systems at a resolution of ca. 1 km² were used as a framework for the analysis. Grid squares attributed as either F. crenata or Q. mongolica var. grosseserata forests were extracted from the entire vegetation data set, and the climatic data set was then combined with the extracted grid squares. Four climatic factors that presumably affect the distribution of the forests, i. e., warmth index, minimum temperature of the coldest month, winter (December-March) precipitation, and summer (May-September) precipitation, were derived from the climatic data set. The four climatic factors were then used to develop a classification tree model to devise classification rules from the relations between the forest distributions and climate conditions. The resulting classification tree model showed that the macro-scale distributions of F. crenata forests and Q. mongolica var. grosseserrata forests were accurately (ca. 90 percent classification accuracy) distinguished by the four climatic conditions. Fagus crenata forests were more likely to become established if the minimum temperature of the coldest month was over - 12.5 degrees, May-September precipitation was over 760mm, warmth index was under 74, and December-March precipitation was over 442 mm. In contrast, Q. mongolica var. grosseserrata forest showed a higher tolerance in terms of climatic conditions, but was suggested to be constrained by deep snow-cover. Accordingly, we concluded that a combination of the four climatic factors was able to distinguish the distribution of the two forest types accurately. This study also supported the view that Q. mongolica var. grosseserrata forests in Honshu are climatically stable, refuting the claim that these forests are always established secondarily after clear-cutting of F. crenata forests.