Spatial distribution of cattle in arid rangelands as detected by patterns of change in vegetation cover.
Abstract
Patterns of change in vegetation cover, as measured by multitemporal remote sensing using the LANDSAT Multispectral Scanner, were used to model domestic cattle distributions in the arid rangelands of central Australia. Model assumptions that paddock shape and waterpoint layout have no effect on rate of cover change (and thus cattle distribution), other than through changes in the area of a paddock and the individual landscape types within it, at each distance from water seemed justified. Model assumptions that rate of cover change in a particular landscape type at a given distance from water is unaffected by the mix of other landscape types around it also seemed justified. Land degradation appeared to have only a limited effect on the distribution of cattle compared with differences in grazing preference for individual landscape types and the number of cattle present in these rangelands. Generalized cattle distribution models can be derived for broad landscape types on the basis of distance from water and grazing preference. These models can be applied irrespective of the number of cattle present. The models have potential for incorporation into decision support systems for paddock design and waterpoint layout. Such systems could ensure the effective and sustainable utilization of arid rangeland vegetation by cattle.