Seasonal and Interannual Variation in the Home Range and Habitat Selection of African Buffalo (Syncerus caffer) in the Klaserie Private Nature Reserve, South Africa

A Geographic Information System (GIS) was constructed of Klaserie Private Nature Reserve (KPNR), comprising buffalo herd locations, vegetation layers and water availability. Using this GIS, range behavior and habitat selection was examined at multiple scales, both geographically and temporally. We employed three methods of home range estimation: Minimum Convex Polygons (MCP), a kernel method and a local nearest-neighbor convex-hull construction (kNNCH). We found that for a dataset recording the locations of three major herds in the reserve over 5 years (1995-2000), with all three methods, the Southern Herd (SH) had the largest range, the focal study herd (FH) had the intermediate range and the Northern Herd (NH) had the smallest range; the k-NNCH method best described the ranges, as it accommodates user knowledge of known physical barriers, such as fences. An examination of 84 independent 10-day ranges of the focal herd across 9 years reveal that buffalo travel further and range wider in the dry season than the wet. Broad scale categories of vegetation types were used in habitat selection and showed that the three buffalo herds showed preference for a patch of Acacia shrubveld and patches of Combretum dominated short woodlands. A more detailed analysis of the focal herd (FH) using post-process remotely sensed vegetation data, ΣNDVI (Normalized Difference Vegetation Index) at a fine temporal interval failed to show preference for greener vegetation. Results of an analysis of habitat selection for distance-to-water revealed that buffalo are selecting for areas within 1km of water preferentially, within the wet season (for all herds) and across both seasons across all years (FH). Moreover, an isopleth analysis using the k-NNCH method showed a preference for riverine areas in both seasons. These findings suggest that buffalo are selecting for areas near water preferentially, but may range further in the dry season for higher quality food. The predictability of preferred habitat is not possible using ΣNDVI data, but we discuss the potential suitability of woody-dominated landscape in the dry season as preferred habitat, despite the apparent paradox.