Sediment Estimation of the Mosul Dam Lake Using SWAT and Dual ANN Models

Managing sediment in river basins and waterways became a key point in water management. The continuous sediment deposition from the rivers and streams flowing to the reservoirs is today’s problem. Entering sediment from the Sweedy, Crnold, and Alsalam valleys located at the right bank to the reservoir of Mosul Dam in the north of Iraq has been estimated using the ANN model. This study aims to estimate the sediment values using only available precipitation data. The challenges and difficulties in this study were that the catchments of these valleys were ungauged with very limited available data, so the ANN models were used with the assistance of the SWAT using three different models in three main steps. In the first step, the flow, concentration, and load of sediments were calculated using the SWAT model for a period (1994-2018) using the available meteorological data. Then, as a second step, based on the available rainfall data and the flow resulting from the SWAT model, the ANN model was used to obtain the flow discharge. In the last step, two separate ANN models were used to estimate the sediment concentration and sediment load entering the dam reservoir from the right-side bank. The coefficient of determination (R^2) and Nash Sutcliff Efficiency (NSE) were used to evaluate the models. Finally, the result showed the ability of ANN to use two successive models (Dual ANN) to estimate the sediment concentration and sediment load for the valleys, as mentioned earlier.