A landscape-scale model of shrub and herbage dynamics in Central Australia, validated by satellite data

Abstract

Ecosystem models have an important integrative role in allowing available data to be used to predict the potential impacts of changing climate or disturbance patterns. No such model has previously existed for the rangelands of Central Australia. In this paper we develop an ecosystem model for arid chenopod shrublands south of Alice Springs which takes account of inherent landscape heterogeneity and temporal variability, but which uses simple empirical relationships because of the small amount of available data on which to base any such model. The model is focused on the dynamics of simple herbage and shrub biomass pools; further breakdown of species composition is not attempted. Herbage dynamics are most dependent on grazing effects, as measured by distance from permanent watering points and average paddock stocking rate, and on the erosional status of the soil. Shrub dynamics are most sensitive to erosional status. The model is tested against historical Landsat MSS imagery transformed by a plant cover index. Correspondence is good, except on highly erosional sites which appear more dynamic than predicted. Further work is needed to quantify rates of soil erosion and deposition at different grazing levels, and of shrub growth and dieback in arid climates. Accurate means of assessing plant biomass from satellite imagery also require further development.