Investigating habitat degradation of Ursus arctos using species distribution modelling and remote sensing in Zagros Mountains of Iran

Knowledge on habitat suitability of ecologically important species is central for their conservation management. The occurrence records of brown bear (Ursus arctos) in Kurdestan Province of Iran and environmental data were compiled to predict the ecological niche of this bear using Maxent species distribution algorithm. Satellite imagery of Landsat 5 and Landsat 8 for 1990 and 2017 was also employed to investigate human-induced land degradation in the suitable habitats of the species. The images were pre-processed and classified into seven land use and land cover (LULC) classes using support vector machine (SVM) algorithm. According to Maxent model, about 9% of the study area could be regarded as highly suitable habitat, followed by moderately suitable habitat (11.6%) and marginally suitable habitat (29.3%). The integration of Maxent potential habitat map and LULC changes revealed that more than 6997 ha of the suitable habitats of the species have been degraded due to different human activities mainly related to the conversion of natural lands and Daryan Dam construction.