Uncertainty Analysis of the Water Level of the Karun and Dez Dam Reservoirs Due to the Inflow Hydrograph Uncertainty

The Real-time flood control operation of a reservoir system can greatly reduce human and financial losses. In this study, a model for real-time flood control operations in reservoirs under the occurrence of floods in April 2019 and the uncertainty of hydrographs of inflows is presented. The presented model includes three modules: the flood that occurred in April 2019, Monte Carlo-HEC-ResSim simulation, and uncertainty analysis. The considered uncertainty factor is the hydrograph of the input to the reservoirs and side flows to the river. A Monte Carlo-HEC-ResSim simulation was performed according to the real flood of April 2018. In order to quantify the uncertainty of the HEC-ResSim model in the simulation of the water level of the dam reservoir, two 95 Percent confidence band factors (P-factor) and the band width factor index (d-factor) were used. Based on the results of the Monte Carlo simulation, the uncertainty in the water level of Dez, Karun and Gotvand dams due to the uncertainty of the hydrograph of the inflow was 0.037, 0.107 and 0.034 Percent, respectively. Therefore, the highest and lowest uncertainty in the water level due to the uncertainty of the inflow hydrograph is related to the Karun 1 and Upper Gotvand dams, respectively. In addition, the uncertainty bandwidth of the HEC-ResSim model in simulating the water level in Dez, Karun 1, and Upper Gotvand reservoirs was 0.151, 0.407, and 0.808, respectively. Considering that the maximum allowable amount of the accepted bandwidth factor is equal to one, the uncertainty bandwidth obtained in all the parameters (including the input and output values from the reservoirs, the volume of the reservoirs and the control points downstream of the dams) is accepted. This indicates the low uncertainty of the HEC-ResSim model in reservoir exploitation operations. The 95ppu values of the observational data in the 95 Percent confidence band for the water level of the reservoirs in the three studied dams were 100 Percent. A high 95ppu value indicates that the model has a strong physical and theoretical basis. For other parameters, the 95ppu values were low due to the low uncertainty of the parameters.