Sediment Rating Curve Parameters Relationship with Watershed Characteristics in the Semiarid River Watersheds

Abstract

In order to effective management of water resources, an accurate estimate of the water transporting load is necessary. But no individual method has been shown to consistently provide accurate load estimates among different water sampling sites. We evaluated the accuracy of several load estimation methods by simulation of the relationship between river discharge and sediment yield, across a broad range of sampling and environmental conditions. Sediment rating curves (SRCs) are generally used to estimate the suspended sediment load of the rivers and drainage watersheds. Since the regression equations of the SRCs are obtained by logarithmic retransformation and have a little independent variable in this equation, they can be used to estimate the true sediment load of the rivers. One of the modification methods to compensate this bias is developing various correction factors. For optimization of SRCs and to evaluate the bias correction factors in Kale Shour and Kashaf Rud watersheds, seven hydrometric stations of this region were selected with the easy-to-find variables in the upstream watershed of stations. Investigation of the accuracy index (ratio of estimated sediment yield to observed sediment yield) and the precision index (the estimated coefficient of variation) of different bias correction factors of Food and Agriculture Organization coefficient (FAO) and three regression models, the quasi-maximum likelihood estimator (QMLE), smearing estimator, and the 7-parameter minimum-variance unbiased estimator (MVUE) with the least significant difference test showed that FAO coefficient increases the estimated error in all of the stations. Applications of MVUE in linear and mean load SRCs have not statistically meaningful effects. The QMLE and the smearing factors increased the estimated error in mean load SRCs, but that does not have any effect on linear SRCs estimation.