Evaluation of the habitat of striped hyaena (Hyaena hyaena) using aggregation method in the Shaho mountain Domain of Kermanshah province

Abstract

The striped hyena (Hyaena hyaena) is a global scale endangered species and has a high risk of local extinction in its population, therefore, investigation and evaluation of its habitat for covered areas seems necessary. This study was done to investigate the distribution status of this species in the Shaho Mountain domain in Kermanshah province. In this study, after collecting species presence points, habitat variables including slope direction, elevation, distance from rangelands, distance from agricultural land, distance from main road, residential density, ecotone, slope percentage and viewshed was identified and used in the analysis. In this regard, firstly, using single-class support vector machine models the habitat of the species was modeled. By confirming the validity of the model output through AUC criteria was used from the binary output of the model in order to provide quasi-absence sites 10 times the presense points within almost 5 km distance. Then maximum entropy models (MaxEnt), Back-Propagation (BP) Neural Network model (BP-ANN), and two-class support vector machine (SVM) were then used for modeling. The validity of which were calculated as 0.97, 0.97, 0.89 respectively. Then, all models were used in an aggregation scenario according to the weight obtained from model validation (AUC). Model sensitization was performed using random forest method. The results showed that the variables; viewshed, road distance and distance from the fields were the most important habitat variables affecting the striped hyena habitat in the region.