Evaluation of WEPP for runoff and soil loss prediction at Gunnedah, NSW, Australia

Abstract

It is important to use historical data to test physically based runoff and soil erosion prediction models as well as the method to estimate model parameters. WEPP (Water Erosion Prediction Project) was validated for bare fallow and annual wheat treatments at Gunnedah, New South Wales, Australia. Wheat stubble was either burned or mulched. Climate, soil, management, and runoff and soil loss data were collected for the period 1980-87 for 3 bare fallow plots, and 1950-74 for 10 annual wheat plots. Three slope lengths from 21 to 62 m were established for the treatment with stubble burned. Slope steepness varied from 8% to 9% at the site. Effective saturated hydraulic conductivity and soil erodibility parameters were estimated from measured soil properties. No further calibration of these parameters was attempted in order to assess the true potential of the model for runoff and soil loss predictions. WEPP worked well for the bare fallow plots with prediction efficiency of 0.97 for event runoff and soil losses. WEPP generally over-predicted the runoff, and consequently, the soil loss for annual wheat treatments for the site. WEPP was able to predict the effect of slope length on sediment concentration and soil loss for the site. CLIGEN, which provides the continuous climate input to WEPP, was found to produce adequately the mean daily rainfall, but produced higher than expected peak rainfall intensity, resulting in higher runoff and soil loss for all treatments.