Risk factors affecting the probability of damage by sika deer in plantation forests in Yamanashi Prefecture, Japan

摘要

Using a generalized linear mixed model approach, we determined the most important risk factors affecting the probability of damage by sika deer in a forest plantation in Japan. Candidate risk factors included tree species, stand age, peripheral dwarf bamboo community, topographical factors, snow depth, and human disturbance factors. Based on this model, we developed a risk map of forest damage. The model indicated that the most important risk factor was stand age, followed by tree species and maximum snow depth. Our predictive model has practical use due to its high classification accuracy (83.9%). To decrease damage from sika deer, an afforestation plan that incorporates these factors should be implemented. Because it is based on common, forest GIS data that have recently been compiled by several local governments in Japan, our modeling method of deer damage can easily be adapted to other areas.