A predictive spatial model for gray wolf (Canis lupus) denning sites in a human-dominated landscape in western Iran

摘要: Predictive models of species denning habitat are useful for understanding distribution core use areas and identifying with potential conflicts. Wolf den site selection in unmanaged dominance agriculture-lands human disturbance is not completely understood. We used a GIS multivariate model based on the Mahalanobis distance statistic 11 digital layers to evaluate gray wolf suitability Hamedan province (HP), western Iran. Results showed that >74 % dens occurred sites occupied 14 % (2,785 km2) study area. Based sensitivity analysis ROC, we found rangelands, elevation, roads, settlements streams were most predictive variables. Our very good overall performance according classification accuracy discrimination capacity. Denning elevated rugged terrains adjacent rangelands away from roads villages revealed HP highly influenced primarily by factors associated disturbance. The derived predicted map can be prioritize conservation, identify conflicts, implement adaptive management regions beyond wilderness protected areas.