Bias correction of sediment rating curve and estimation of watershed erosion rate

The discharge-sediment rating curve is usually fitted by a power function that transforms the discharge into the sediment yield or concentration. To derive the power function, the discharge and sediment data are taken logarithm so that the power function can be fitted in the form of a linear function. Although the linear function derived by the least squares method is statistically unbiased, the detransformation of the linear function to the power function yields a systematic bias due to the conversion of residuals. Thus, this study applied two bias correction methods to solve the problem of biasedness. Discharge and sediment data from two basins, Wu River and Kaoping River in Taiwan, were used to test the bias correction methods. Results show that the bias correction methods really improve the sediment estimation in the study basins. The watershed erosion rate (different from the soil erosion rate) regarding each subbasin in the study area was calculated by the corrected rating curve. The watershed erosion rates in Wu River basin are from 0.3 to 5.0 mm/year, and those in Kaoping River basin are from 0.7 to 5.3 mm/year. Moreover, parameters of the discharge-sediment rating curve were analyzed, and the link between parameters and the riverbed gradient and erosion power was discussed.