Automatic Calibration of Groundwater Simulation Model (MODFLOW) by Indeterministic SUFI-II Algorithm

Mathematical simulation of groundwater resource systems is one of the essential tools in managing these valuable resources and calibration of the groundwater simulation models is the time-consuming, and complicated step of these systems. Automated calibration, developed in recent years by researchers with different algorithms, is one of the effective methods to overcome these computational problems. On the other hand, lack of field data in terms of time and space and the hydrological and hydrogeological complexities leads to many uncertainties in the calibration results. The SUFI-II algorithm is an uncertainty-based automatic calibration method that is capable of calibration and uncertainty analysis of numerical simulation models. In this paper, for the first time, this algorithm is used to calibrate and analyze the uncertainty of hydrodynamic parameters (hydraulic conductivity and specific yield) of the MODFLOW model. The results of model implementation for the Ardabil plain groundwater model (Northwestern Iran), indicate an average of 62 percent of the observation data within the 95 percent confidence interval. Finally, the best intervals of parameters are determined for the hydraulic conductivity and specific yield by the proposed approach. Also, the calibration of the groundwater model has been carried out using PEST. According to the results, the root-mean-squared error (RMSE) value in this case (RMSE = 3.37) is higher than the SUFI-II method (RMSE = 1.86), which indicates better performance of the SUFI-II algorithm than the PEST model.