Estimation of pasture growth rate in the south west of Western Australia from AVHRR NDVI and climate data

Abstract

Regular estimates of regional pasture growth rate (PGR) would prove invaluable for wool producers in the agricultural zone of Western Australia (WA) wishing to adopt efficient techniques of managing the supply of annual pasture. Data from the AVHRR sensor are routinely processed to provide bi-monthly NDVI (normalised difference vegetation index) composite images of the agricultural zone of WA. We used a time series of NDVI imagery for 1995 to 1998 together with coincident climate layers in a light use efficiency (LUE) model to generate PGR maps for WA. We only operated the model for areas of the agricultural zone where pasture constituted more than 60% of the land cover. The regional estimates were compared with whole-farm averages made at three farms using pasture exclosures positioned across the main land management units (LMUs) on each property. Significant relationships were obtained between predicted and observed PGR for the three farms Kojonup, Porongurup and Williams of R 2 0.67, 0.75 and 0.70, respectively. A validation study in 1999 at five farms used a potential productivity index for LMUs to scale the regional estimates down to paddock level. The relationship between predicted and observed PGR, for data averaged by LMU, had an R 2 of 0.69. Given that little information on regional PGR currently exists, estimates generated by such LUE models from in-line AVHRR processing made available on the internet or by subscription can greatly improve the basis for feed management across wool growing enterprises in the south west of WA. This approach could readily be applied to other highly seasonal grazing systems around the world.