The use of crop yield prediction as a tool for land evaluation studies in Northern Australia

Abstract

Adequate prediction of crop yield is essential in quantitative land evaluation studies. The AUSCANE simulation model, local expert opinion and historical records were used to predict crop yield and ccs (a measure of sugar content in the cane stalk) for sugarcane production for land mapping units at two different scales of investigation in the Herbert River district of northern Australia. Simulations were undertaken for a period of 25 years. The model performed adequately on clay soil types; however, it generally overestimated yield and ccs values on sandy soil types and underestimated values on loamy soil types. Yield estimates were influenced by mean daily solar radiation and changes in the soil water balance. Results from the yield prediction process have played a significant role in the development of a tripartite system of land evaluation based on crop yield prediction, expert systems for biophysical and economic land suitability and risk analysis.