Predicting the distribution of Eastern Grey Kangaroos by remote sensing assessment of food resources.

Abstract

This study demonstrates how the distribution of animals can be described using
remotely sensed data at a scale in the order of square kilometers. Kangaroo distribution
has been monitored at regional scales using aerial surveys and detailed field study. This
study attempts to fill the gap between local and regional scales by using Landsat derived
vegetation characteristics to provide animal distribution details at local scale. Field surveys
of Eastern Grey kangaroos and vegetation biomass were undertaken at the Warrumbungle
National Park, New South Wales, Australia. The distribution and abundance of kangaroos
and plant biomass were compared with remotely sensed vegetation characteristics taken
from Landsat TM imagery. The distribution of green, short (< 5cm) blade grass biomass
(the preferred kangaroo food resource) was patchy and positively correlated with kangaroo
density and Landsat spectral bands 1, 2, 3 and a principal component combination of bands
1-7 (excluding band 6). Total population density was positively correlated with blade
grass biomass and Landsat band 3. The dispersion of kangaroos within habitats was
patchy, even though the Landsat image defined habitats as being homogeneous. This study
clearly demonstrates the value of Landsat data to environmental management in the past
and present.