Applicability of a predictive distributional model for the distribution of the Asiatic black bear　(Ursus thibetanus japonicus) to a local population in the Tanzawa region of Japan

Abstract

A predictive distributional model of Asiatic black bears (Ursus thibetanus japonicus) is useful for the effective management of the species’ habitat. Applicability of a previously developed model to the Tanzawa local bear population was examined. Three different types of animal presence data were collected for validation of the model. The overall rate of correct classification was 67%?85%, indicating that the model can be extrapolated to other local populations with acceptable transferability. Accuracy of the model’s predictions differed depending on the data sampling design most likely as a result of bias in one of the sampling methods. The model tended to underestimate the actual presence of bears. We attribute this underestimation to 1) properties of the particular habitat used by the local Tanzawa population and 2) the model’s greater ability to predict the habitat of mountain bears than that of Satoyama bears.