Reliability of sediment rating curves for a deciduous forest watershed in Iran

The suspended sediment load in a river is an indicator of its health. Yet, the direct measurement of suspended sediment load is difficult. Most assessments of suspended sediment load are therefore estimated from relationships developed between suspended sediment concentration and discharge, as sediment rating curves. However, these are not able to provide sufficiently accurate results. The present study aimed to assess the accuracy of the sediment rating curve in the educational forest watershed of Tarbiat Modares University in northern Iran. Water and suspended sediment sampling were conducted via depth-integration at daily and storm-wise scales, and the samples were analysed through decantation procedure. Towards this attempt, various sediment rating curves were examined for different data sets classified on the basis of various governing conditions. The relationships between study variables were assessed by applying a bivariate regression method and descriptive statistics. The results of fitting rating curves on suspended sediment and water discharge, with respective averages of 1.256 +/- 2.835 g L-1 and 0.556 +/- 0.444 m(3) s(-1), verified that no sediment rating curve could be reliably used for the study area. Scrutinizing the results also verified that the data stratification and increasing time resolution led to better model performance, though their accuracy has not been statistically and logically tested.