Simulation of Groundwater Balance Using Integrated Surface and Groundwater SWAT-MODFLOW-NWT Model (Case Study: Mahabad Plain)

 Surface and groundwater conjunctively interact at different spatial or temporal scales within a plain. In many plain, surface and groundwater resources are used in combination in agriculture. Therefore, it is important to accurately predict the components of groundwater and surface water balance. Despite the rapid expansion of numerical models over the past two decades, there is still a need for comprehensive and integrated assessment of surface and groundwater components. In particular, the interconnection of both surface and groundwater models is important to connect both surface and groundwater, especially the water balance in the unsaturated root zone. In this study the effect of water recharge due to deep percolation from simultaneous supply of irrigation water from surface and groundwater sources, and rainfall from the SWAT model were used to simulate groundwater balance using the combined MDOFLOW-NWT model.

 In this study, the effect of recharge values obtained from the SWAT model was analyzed to simulate the fluctuation of water table, and groundwater balance components using the integrated model of MODFLOW-NWT model in the Mahabad plain. One of the important steps in quantifying the impact of irrigation management, and the change in land-use on the surface and groundwater balance was the simulated recharge due to the deep percolation of rainfall and irrigation water. This was done by the SWAT model, and was used as the boundary condition to the MODFLOW-NWT model. Calibration and validation of groundwater model were also done by trial-and-error and automatic PEST methods. The simulation period was performed for 10 years from the hydrological year of 2009-2010 to 2018-2019, from which 6 and 4 years were used as the period for calibration and validation were from 2009-2010 to 2014-2015 and 2015-2016 to 2018-2019, respectively. Groundwater balance components are naturally different for different years. Therefore, the study was conducted for dry, wet, and normal years. Hydraulic conductivity and specific yield were the used as initial calibration parameters in the MODFLOW-NWT model.

 The results showed a higher hydraulic conductivity and specific yield values for the aquifer was in the southern, central, and northeastern areas of the plain, and the lowest values were in the northern and near the outlet of the plain. After the calibration process, the results showed that an average, 9% of rainfall, and 36% of irrigated water percolate to the aquifer. These observations were confirmed based on a satisfactory and acceptable estimate of the water table level of the model for both calibration and validation periods. The statistical RMSE criteria for calibration and validation periods were 0.35 and 0.34 m, respectively. Also, the results of R2 and NSE criteria were estimated as 0.94 and 0.91 for the calibration period, and 0.93 and 0.89 for the validation period, which indicates that the model was properly calibrated and was well able to simulate groundwater level. The groundwater hydrographs developed from piezometers’ readings, show that the recharge values estimated by the SWAT model, considering the change in land use and irrigation management across the plain, were able to properly simulate groundwater level across the aquifer. Specifically, the studies showed a continuous drop in groundwater level created in the southern and southwestern regions of the aquifer (piezometers of Fakhrighah, Gorg tapeh, and Serah Haji Khosh) due to the presence of high-consumption crops such as apple and alfalfa, and the higher number of operation wells.

The results of this study showed that the recharge values obtained from the calibrated SWAT model was crucial parameters for proper simulation of groundwater, and can significantly improve the model results. The results of the main components of the groundwater balance for different years showed that the amount of recharge due to the infiltration of rainfall, and irrigation were different for each year. Also, interactions between surface and groundwater resources vary from about 30 to 50 million cubic meters between years, indicating a significant interaction between the water resources. In general, the SWAT-MODFLOW-NWT model can be used as a practical tool for proper management of surface and groundwater resources under different management scenarios.