شبیه‌سازی پیشروی آب‌شور در آبخوان‌های ساحلی (مطالعه موردی: سواحل جنوبی دریای خزر)

Introduction: The issue of seawater intrusion has become an environmental problem considering the increasing trend in groundwater extraction from coastal aquifers. Increased groundwater exploitation and lack of coastal aquifers management have caused seawater intrusion into coastal aquifer. The intrusion has led to the salinization of aquifers, causing many problems in the exploitation of water resources. This pumping has continually increased the risk of seawater intrusion and deterioration of freshwater quality in the sari-Neka aquifer. Seawater intrusion limits the usage of groundwater for agriculture, industry, and public water supply.
Materials and Methods: For the present study, Sari-Neka aquifer was selected. The study area is located in the southern shores of the Caspian Sea, in the northern part of Iran. The MODFLOW version 2000 is used to simulate a steady-state and transient groundwater flow system in Sari-Neka aquifer. To simulate solute transport, MT3DMS and SEAWAT are used. In MT3DMS, advection package, dispersion, and source/sink mixing packages are used. The numerical code MT3DMS does not consider the effect the density. Thus, SEAWAT­-­Variable Density Flow package was initialized. The necessary data for modeling of groundwater flow can be categorized into water resources data, meteorological data, hydrodynamic characterization, topography map, and geological information. To build the flow model, flow type (steady-state and transient state), initial conditions (groundwater level in September 2010 for the steady-state) and type of boundary conditions (general head boundary), flow package (LPF package), ‌temporal discretization (48 monthly stress periods from September 2010 to August 2014 in transient condition) and monthly time steps were assigned to the model. To prepare the flow and transport model grid, the study area was discredited horizontally into 3694 active square cells (500×500 m). The MT3DMS model was used to simulate the qualitative changes on the aquifer surface and the SEAWAT model to simulate the depth of the aquifer. Therefore, the conceptual model of solute transport was prepared by making the necessary changes in the conceptual flow model. September 2010 groundwater level data and TDS and Cl data are taken as the initial conditions in the flow and transport model, respectively. The Caspian Sea bordering the study area in the north is represented by a constant TDS concentration of 35000 mg/l and constant CL282.2 meq/l. In this model, we entered the water heads of the observation wells, hydraulic conductivity, storage coefficient, effective porosity, aquifer discharge, and aquifer recharge, porosity, Coefficient of molecular water diffusion, Longitudinal dispersivity, Horizontal transverse dispersivity, vertical transverse dispersivity.
Results and Discussion: The calibration of the flow model was carried out for both steady and transient conditions using the trial and error approach. Monthly groundwater levels of data from 14 observation wells were used for calibration purposes. Steady-state calibration for the flow model was performed by comparing the observed groundwater levels and calculated values of groundwater levels in September 2010. During calibration, hydraulic conductivity values were adjusted, until groundwater level values calculated by MODFLOW were matched the observed values within an acceptable level of accuracy (±1m). After steady-state calibration, the transient model was simulated for the four year period between September 2010 and August 2014 that was divided into 48 stress periods with monthly time steps. At the end of flow model calibration, the resulting hydraulic conductivity ranged from 5.3 to 21. 6 m/day, while the resulting specific yield values were from %3.4 to % 5.9. The validation flow model was simulated for the period between September 2010 and August 2014 (12 stress periods). The values of the correlation coefficient in the steady-state model, transient model and validation model in the flow model were obtained 0.99, 0.98, and 0.97, respectively. The results illustrate a good agreement between the observed and calculated groundwater levels. The transport model was calibrated using TDS and CL concentration data from September 2010 to August 2014 (8 stress periods) by adjusting parameters affecting the dispersion process. To confirm the accuracy of the model, TDS and CL concentration data from August 2014 to September 2015 were used for validation purposes. By considering the TDS and Cl concentration in September 2010 as the initial condition, the transient model was run. Transport model calibration was achieved through a trial-and-error. The values of the correlation coefficient in the transport model for TDS are obtained 0.83 and 0.87 in the transient model and validation model, respectively. The values of the correlation coefficient in the transport model for CL were obtained 0.82 and 0.86 in the transient model and validation model, respectively.
Conclusion: After the validation of transport model and assuming all the hydrogeologic conditions remain, a predictive 6-year simulation run using SEAWAT model indicates that further seawater intrusion into the coastal aquifers can occur in the study area.