Mapping and modelling the population and habitat of the roan antelope (Hippotragus equinus langheldi) in Ruma National Park, Kenya

Abstract

Wildlife-based tourism, which is Kenya’s second largest economic sector, is threatened by the risk of extinction of many wildlife species in the country. The endemic roan antelope (Hippotragus equinus langheldi) now only survives in Ruma National Park (RNP) where its population has been declining continuously since 1976. This thesis investigates the roan’s habitat use and selection, causes of population decline and population viability in RNP with the aim of recommending scientifically-based management interventions for population recovery and sustainable conservation. Roan movement patterns and habitat use were investigated using 4 home range estimation techniques. Habitat selection was studied at multi-spatial scales using compositional analysis, logistic regression, and information-theoretic (IT) and multi-model inference (MMI) techniques. Data for this study consisted of population estimates for roans and other grazers, Landsat images, soil maps, digital terrain data, rainfall records, snare distribution records, and roan ground tracking data. Identification of causes of population decline was carried out using both multivariate and univariate techniques. A generic population viability analysis (PVA) package was used to (i) estimate the likelihood of roan extinction under various management options; and (2) rank the management alternatives for roan population recovery. All 4 home range estimators are useful in characterizing different aspects of the roan home range, but overall the local convex hull method produced the most realistic home ranges. The three habitat selection methods yielded similar results but the IT techniques demonstrated superior qualities as they identified important habitat variables and produced the most accurate model predictions. MMI averaged models coupled with GIS data developed very informative habitat suitability and poaching risk maps. Analysis of habitat selection indicated different usage in seasons and spatial scales depending on water availability, habitat composition and burned status, and distribution of eco-geographical features. High adult mortality due to poaching with snares was identified as the main cause of roan population decline. Other important factors included habitat change and rainfall variability with its associated droughts and floods. PVA showed that under the current conditions, the roan population cannot persist more than 3 decades. Several anti-poaching and prioritized management interventions to curb poaching and promote population recovery are described.