Anthropogenic infrastructure shapes habitat selection of African wild dogs in a multi-use, privately owned landscape

This collection archives the analysis code and spatial data supporting the study of habitat selection by African wild dogs (Lycaon pictus) in a privately owned, multi-use landscape dominated by South Africa’s private wildlife industry. Using a landscape-of-fear framework, the work tests whether resident packs exhibit risk-avoidant or opportunity-driven responses to anthropogenic infrastructure (buildings, fences, public roads, and farm roads). Habitat selection was quantified at two spatial scales: second-order selection via resource selection functions (RSFs) and third-order selection via step selection functions (SSFs).The collection includes:Five R scripts used for data processing, model fitting, and visualizationFive infrastructure data layers compiled for the study (buildings, fences, tar public roads, dirt public roads, and farm roads)Four KML files of buffered Minimum Convex Polygons (MCPs) plus 4 km buffers used to define availability for second-order analyses for each of the study packs.One KML file of the combined MCPs, plus 4 km buffer, used to define the study region for infrastructure mapping.African wild dogs are an endangered species. Due to conservation concerns and restrictions on data sharing, the raw GPS location data used in this study is not included in this repository. The full GPS dataset is available upon request from the conservation NGO's that own the data, including the Waterberg Wild Dog Initiative (WWDI; www.waterbergwilddogs.org.za) and the Endangered Wildlife Trust (EWT; www.ewt.org). Restrictions apply to the use of these data, which were used under permission for this study.