Simulation of sediment graph using hydrograph

Sediment graph is an important tool for soil and water resources management of watersheds. It is therefore important to find an appropriate procedure to simulate sediment graph data in different watersheds with inadequate and unreliable suspended sediment data. However, achieving simple procedures based on easily accessible data has been less considered. Hence, the present study was conducted in Galazchai Watershed in West-Azerbaijan Province, Iran, with an area of some 103 km2 to investigate the ability of synthetic sediment graph development by using hydrograph. Towards this attempt, a databank of 18 storm events occurred during autumn 2011 and spring 2014 was developed and corresponding total and direct hydrographs and sediment graphs were prepared. All total and direct sediment graph components were simulated using corresponding hydrograph components with the help of different bivariate regression models and the best performed model was ultimately determined by applying different criteria. The results showed that the simulation of different total sediment graph components was possible using hydrograph components except for base time and ordinates of 50 and 75 percent of peak. Direct sediment graph simulation results also verified the simulation ability of all components except for ordinates of 75 percent of peak. According to the results, simulation of temporal components with the lower relative errors had better performance than the other components. So that, for total sediment graph, the lowest relative estimation and validation errors for time to peak were 48.86% and 45.65%, respectively. Also, the best performed model developed for the direct sediment graph base time had the lowest relative estimation and validation errors of 23.03% and 21.75% and the highest coefficient of efficiency of 0.93.