A SIMPLE-MODEL FOR PREDICTING HERBAGE PRODUCTION FROM RAINFALL IN RANGELANDS AND ITS CALIBRATION USING REMOTELY-SENSED DATA

This paper describes the development of a model for estimating herbage production from rainfall and evapotranspiration in rangelands and its calibration using a vegetation cover index derived from remotely-sens ed data. The model operates on a time unit of 1 month and has a maximu m of three calibration parameters: a water use efficiency term, a cove r depletion rate and a term describing the percentage of tree and shru b cover present. The model can represent variability at the whole land scape level and is well-suited to calibration. Parameter values derive d when the model is inverted show variations which can logically be as sociated with differences in land condition and grazing pressure. Thes e parameter values make it possible to model differences in herbage pr oduction in a particular area under different rainfall conditions and in response to changes in land condition or the level of grazing.