Evaluation of the efficiency of using surface water simulation results to improve the accuracy of groundwater simulation in Silakhor shallow aquifer located in Lorestan province (Iran)

This paper evaluates the efficiency of using surface water simulation results to improve the accuracy of groundwater simulation due to improving the accuracy of its input boundary conditions. The three most important data sets of boundary conditions in groundwater flow simulation are groundwater recharge, evapotranspiration rates and their regional distributions that little information is usually available about them. Moreover, the incorrect definition of these values can lead to uncertain groundwater modeling that is not applicable for groundwater resources management. In the most of previous studies done in Iran, percentage of the average regional precipitation and the local pan evaporation data (available from local weather stations) are used as the regional groundwater recharge and evapotranspiration rates, respectively but they have high uncertainty in their quantities and spatial distributions. In this research in order to solve this problem, the values of groundwater recharge and evapotranspiration rates and their regional distributions, obtained from SWAT modeling results, were used as the related input boundary conditions data in groundwater flow simulation using MODFLOW model. SWAT model is a comprehensive watershed model that calculates these values in each Hydrologic Response Unit (HRU). The study aquifer in this research was Silakhor shallow aquifer with the area of 590 km2 located in Lorestan province of Iran. The annual water table depth in different locations of this aquifer had been reported from 0.75 to 40 meter. Using this approach for improving boundary conditions in groundwater simulation and its evaluation was carried out for first time in Iran. The SWAT model was run for a period of 7 years from 2002 to 2008. After calibration and verification of the model using hydrodynamic data of Silakhor and Tire-Doroud gauging stations, the values of groundwater recharge and evapotranspiration rates in each HRU were extracted. The MODFLOW model was run for the period of 9 months of 2009 using these simulated boundary conditions and the other required information. For comparison of the calculated values with the observed values of water table depths in the 20 piezometers for the period of 9 months of 2009, the Root Mean Square Error (RMSE) and Mean Absolute Errors (MAE) were obtained 1m and less than 1m, respectively. In continuous of this research, the values of 5, 10 and 15 percent of regional average precipitation and the evaporation data in Silakhor pan evaporation gauging station were used as usual boundary condition for groundwater simulation using the MODFLOW model. For comparison of the simulated values of water tables depths obtained from this usual approach with the observed ones, the both of RMSE and MAE values were calculated more than 1 meter. Therefore this research showed that using integrated SWAT-MODFLOW models was more applicable in the groundwater simulation in the study aquifer.