Comparison of Techniques to Determine Eastern Grey Kangaroo Home Range

Abstract

The ellipse, modified harmonic mean, and fourier transform models were used to examine the home range of adult eastern grey kangaroos (Macropus giganteus) in the Sandy Creek Reserve, Bago State Forest, New South Wales (NSW), Australia. Locations of kangaroos were determined by sighting numbercollared or radio-collared individuals. The ellipse model was used to determine diurnal home range area, orientation, and geographical mean center. Those females and males occupying the primary study area had similar home range areas (9 = 23.2 and 20.5 ha for M and F, respectively). Orientation and geographical center of the range varied with both time of day and season due to movements related to sheltering behavior. In winter, diurnal shelter was not sought and forage was poor; activity centers shifted further into the open pasture than in spring. Home range area of females was significantly reduced when accompanied by young. A modified harmonic mean model was used to determine home range areas and to assess longer-term shifts in activity centers for widely ranging males. The ability of this method to discriminate multiple activity centers resolved long-term movement patterns of radio-tracked males, which were related to transient associations with female groups during the spring-summer mating period. The fourier transform model was used to determine home range areas and boundaries. These areas were much smaller than those determined by the other 2 methods, and the boundaries were often discontinuous. J. WILDL. MANAGE. 51(4):921-930 Movement patterns of macropodids are largely unknown. Some early studies used distances between capture sites and subsequent resightings as indices of mobility (Frith 1964, Ealey 1967, Bailey 1971). More recently, repeated observations of known individuals have led to the estimation of home range areas (Kaufmann 1974, Inns 1980, Johnson 1980, Croft 1982, Jarman and Taylor 1982, Priddel 1984). We studied home range areas of eastern grey kangaroos using 3 estimation techniques: ellipse, harmonic mean, and fourier transform models. The ellipse model (Jennrich and Turner 1969) was chosen because it is statistically robust compared to the popularly used convex polygon technique (Mohr and Stumpf 1966, Van Winkle 1975, MacDonald et al. 1980). Ellipses are recommended for comparisons of area, location, and orientation (Van Winkle et al. 1973, Koeppl et al. 1975, Hawes 1977, Inglis et al. 1979). However, the model is parametric, and therefore should only be used on data that are normally distributed; also, the shape of the home range index obtained will be restricted to an ellipse regardless of its true boundaries. To overcome the latter 2 limitations, 2 nonparametric techniques were used. The harmonic mean model (Dixon and Chapman 1980) places less emphasis on the boundaries and concentrates on the differential use of space within the home range. This involves the calculation of the center of activity or the geographic location of the point of greatest activity. More than 1 center of activity is allowed because the data do not necessarily come from a normal distribution. A contour or isopleth can then be drawn to include a predetermined proportion of data points. The fourier transform model (Anderson 1982) applies a smoothing function to a bivariate frequency distribution of data points and calculates the area under the resulting distribution. We wish to especially thank M. Ellis for his assistance in all aspects of this project. Numerous volunteers assisted in kangaroo captures with R. Williams and D. Read providing notable support. We thank the members of the NSW For. Comm., stationed at Batlow and Tumut, for their cooperation and encouragement. R. Hill and J. Taylor helped prepare the figures and B. J. Fox commented on the manuscript. We thank P. R. Krausman for valuable criticisms of an earlier draft. This research was supported by Aust. Res. Grants Scheme, Grant D17815394.