The availability and quality of preferred forage changes seasonally (East, 1984), as does the availability of palatable water (Redfern, 2003). It could therefore be expected that herbivores may alter their ranges seasonally, ensuring that they can access the best forage as it becomes available. In the most extreme situations this can lead to a migratory existence (Fryxell, 1987), as discussed in Appendix 5. However, even animals that remain resident within one system throughout the year often have distinct or overlapping seasonal home ranges.
Fourteen adult zebra mares, seven from the Central Delta and seven from the Peripheral Delta, were fitted with GPS-capable collars (Vectronic GPS-Plus 4D), between July 2007 and April 2008, as described in section 3.2.1. For this analysis I used only the hourly GPS fixes.
A4.2.2. Home range estimation
The kernel-density-estimation technique (KDE) was used to calculate the dry season home range using the Animal Movement Extension (Hooge et al., 1999) for ArcView 3.2 (ESRI, 1999). KDE is a contouring method (Worton, 1989) and is considered superior to other home range estimations such as minimum convex polygons as it is less biased by distant points and is therefore less likely to include unused landscape (Hemson et al., 2005). KDE produces an utilisation distribution (UD); UDs quantify an animal’s relative use of space in terms of a probabilistic density function (van Winkle, 1975). The probability of an animal occurring at systematic locations through its home range is calculated as a function of relocation points, showing relative use of space throughout the home range: the height of the UD at a given location indicates the concentration of use at that location relative to the rest of the home range (White and Garrot, 1990). The benefits of using this method are that it removes psuedoreplication without presuming equal usage across the entire range. Psuedoreplicaiton occurs when an individual animal’s location is taken as the primary sampling unit (Hurlbert, 1984), as successive locations from the same individual are rarely independent (Otis and White, 1999). To remove this pseudoreplication, one resource selection value is needed per animal rather than per location. One method is to look at an animal’s use of space over time, for example a seasonal home range (Aebischer et al., 1993). However, with larger sampling unit, use may not be uniform over the entire area and so methods that presume equal usage of all areas within the home range may confuse ecological interpretation. Using the KDE approach negated these problems; provides one UD per zebra and providing a measure of relative use within the utilised area. Home range extent was taken as the 95% KDE isopleth.
Dry season ranges were calculated from GPS fixes recorded between April and October. Wet season ranges were calculated from GPS fixes between December and February. The cross-over months of March and November were not used due to the high climatic variation found in these months.Read more: Download PDF Document