Abstract
Behavioral classification and observational error correction methods are often needed when analyzing mammal behavior using GPS. The switching state-space model (SSSM) has been proposed for this problem. However, various positioning intervals have been used to attempt to balance the research purposes and battery life of the GPS. Thus, we may come to different conclusions regarding the appropriate positioning intervals using the same analysis. In this study, we used Asiatic black bear (Ursus thibetanus) behavior data. We generated various time interval positioning data from the real positioning data, to verify the effect of the interval on the SSSM estimates. By combining the SSSM with activity sensors, we can classify the behavior into transiting, resting, and foraging, using 5- to 120-minute intervals. Moreover, time-spatial scales (which recognize the core area) increased as the positioning interval increased. We compared the relative moving distances for the different positioning intervals. Our analysis of the habitat selection performed better using short positioning intervals of 30 minutes or less. We expect that new knowledge will be obtained by applying SSSM with short interval positioning data (e.g., at 5- to 30-minute intervals).
