Habitat selection--Statistical methods

An innovative technique of evaluating resource selection for black rhino (Diceros bicornis) was used to assess the population utilization distribution (PUD) within a rhino sanctuary in Liwonde National Park, Malawi. The PUD enabled an evaluation of responses to habitat variables over a spatial gradient of resource selection. A Geographic Information System (GIS) was constructed using vegetation, browse availability, roads, rivers, water holes and satellite imagery. Linear models were developed to quantify habitat variables within the black rhino sanctuary and park. The sanctuary model was calibrated within a known core area (R^2=0.42, P<0.001), validated in a second area (R^2=0.56, P<0.001) within the sanctuary and, subsequently, used to predict potential black rhino habitat within the remaining sanctuary boundaries. The model for the entire Liwonde National Park predicted additional black rhino habitat (R^2=0.25, P<0.05). Population utilization distribution was found to be a powerful conservation tool for determining suitable black rhino habitat.