Evaluation of the USDA curve number method for agricultural catchments in the Australian semi-arid tropics

Abstract

The USDA-SCS curve number method (5 day antecedent rainfall version) was evaluated using rainfall and runoff data from four catchments at the Douglas-Daly Research Farm, 250 km S.W. of Darwin. These catchments were rotational cropped with maize and soybeans using either conventional or no-till technology. Optimized curve numbers (CNs) were calculated for each catchment and crop type for the various 5 day antecedent rainfall conditions (ARC) by using a grid-based statistical technique known as the Marquardt-Levenberg method. Significance levels (SL) of 0.05 or less were regarded as significant throughout. The curve number method (CNM) was able to describe significant levels of variation in observed rainfall and runoff data. By using optimized curve numbers, the CNM significantly explained up to 98% of the variation in data. Results showed that crop type and soil conservation bank spacing as individual factors had minor or no effect on optimized CNs, and hence on the ability of the CNM to predict runoff. Cropping technology and ARC, however, had a very strong influence on optimized CNs. Optimized CNs obtained in this study provide best estimates of CNs for use on commercial cropping catchments in the region.