Simulation and Assessment of Groundwater for Domestic and Irrigation Uses

The alluvial fan of Mandali located between latitude 30 ̊45’00” N longitude 45 ̊30’00” E in east of Diyala Governorate, Iraq. Thirty-five wells were identified in the study area with average depth of 84 m and estimated area of 21550 ha. A three-dimensional conceptual model was prepared by using GMS program. From wells cross sections, four geological layers have been identified. The hydraulic conductivity of these layers was calculated for steady state condition, where the water levels for nine wells distributed over the study area were observed at same time. Afterward, PEST facility in the GMS was used to estimate the aquifer hydraulic characteristics. Other characteristics such as storage coefficient and specific yield have been determined from one year field observations that were collected by General Authority of Groundwater, Diyala Governorate. Also, the observations were used for calibration of unsteady state model. Then wells were hypothetically redistributed and increased to 103 wells, assuming a distance of 1500 m between the wells, a well productivity rate of were 7 l/s, annual rainfall rate was used for recharging. Three different wells operating times were suggested and these 6, 12, and 18 hr/day with total discharge of 150, 300, 450 m/day and maximum drawdown of 7, 11, and 20 m respectively. For water quality assessment, the collected groundwater samples were analysed at the laboratory. Results showed that the TDS in all wells was ranged from 1000-3000 mg/l but TDS in well number 18 was exceeded 3000 mg/l which indicate that the groundwater in this well is not recommended to be used for irrigation. According to Iraqi standard for drink (IQS 2009), it can be used for drinking if saline treatment units were provided.


Introduction
Water being one of the necessities for life, the human through history has striven to locate and develop it. Over 90% of freshwater available at any time on the earth lies under the land surface [1]. Groundwater reverse surface water, is available in some quantity almost everywhere that man can settle in, is more stable in periods of drought, and has many other advantages over surface water. Groundwater can be used for drinking, irrigation, industries and municipalities.
The lack of surface water due to climatic change and construction of dams on the water sources in the riparian countries increases the importance of groundwater as an alternative source. Therefore, it is requiring knowledge and study on groundwater at the above areas in order to control and organize usage, as well as to define the quality and indicate the suitability for domestic, irrigation and industrial purposes [2]. Lately, software has been developed as hydraulics modelling to simulate groundwater flow, which gives a better idea on of the nature of groundwater usage [3].
The Mandali city located between latitude 30˚45'00" N and longitude 45˚30'00" E, at the eastern of Diyala Governorate, near the Iranian border as shown in Figure 1, the city suffered from a lack of surface water sources. The population and agriculture in this city have been influenced by surface water deficit and started using groundwater as an alternative source. Many wells have been driller randomly in these areas, also the use of groundwater is not ideal, so it is necessary to study the characteristics of groundwater in this city and created conceptual model by using GMS software [4].

Figure 1. Exact location of the study area (Mandali City)
The area of the study is determined by the border of the alluvial fan, which includes both the center cities of Mandali and Qazaniyah as showing in Figure 2.

Figure 2. Mandali alluvial fan
The objectives of the present research are to constructed a conceptual model for the Mandali alluvial fan, then determine the hydraulic properties of aquifers of the study area by using (GIS & GMS) software [5]. Then define the best distribution of wells. However, it also includes qualitative evaluation for the groundwater for various uses such as domestic, irrigation and others.

Research Methodology
The research methodology consists of collecting data for 31 wells drilled at the study area, represented by the location of wells, their depths, and groundwater levels, and the physical and chemical characteristics of groundwater and these were presented using GIS program including the alluvial fan boundaries borrowed from the satellite maps. The data was exported as shapefile in the GMS program. From wells cross sections the longitudinal section of the soil can be create then building MODFLOW, the surface of MODFLOW was editing with (DIM) map for study area. The conductivity of soil layers was determined at the steady state and the storage coefficient and Specific Yield would be determined at the unsteady state by using PEST package. The following steps demonstrate the process of creating a conceptual model step by step:  Insert the wells data as tables in Excel program, this tables includes the coordinates of the wells and all the geometry data for wells. then import this data in GIS program  Import the geological maps of the study area then make a georeferenced for this map and match them with the satellite map in GIS software.
 Determination and drawing the border of the study area.
 From available report of wells using cross section to prepare tables as text file including the layers of soil and thickness.
 Open a new project in the GMS program with a definition of the coordinate system and units.
 Import all tables, map and shapefiles that was created in previous steps in GMS program.
 Created borehole data from table in step 4 then create cross-sections between boreholes.
 From digital elevation map (DEM) for study area can be created surface grid (TIN).
 Edit surface of cross-section by snapping with surface net (TIN).
 Generating solid data from cross-section for study area (generating aquifer).  Create conceptual model then from it can be generating boundary layer, sink and source layer, recharge layer, and observation layer.
 From GIS data can be define the layers in conceptual model.
 The establishment of a three-dimensional grid covering all solid formation of the study area, the distance between line in X and Y axis are 100 m and create five layers in Z axis.  Create MODFLOW model then active steady state in zone of model type and estimate parameter in PEST package.
 It is necessary to inactivate the cells outside the model domain in three-dimensional grid.
 Interpolating the data in solid model and conceptual model with MODFLOW model, then enter the initial estimated value for hydraulic conductivity of soil layers with determine range (min & max).
 Define starting head from record of water table to observation wells was taken at same time.
 Run the MODFLOW model and after several cycles the hydraulic conductivity value will be stabilized and the simulated water table for the observation wells are close to the recorded values.

Data of the Study Area
The following data were gathered and used as inputs for the simulation model.

Geological Information for the Study Area
The geological formations in the study area are Muqdadiyha (Lower Bakhtiari). It consists of sandstone, silt and Claystone beds. This type of formation is for confined aquifer. It is a visible on the surface in the high northeastern of alluvial fan and extend into the lower layers of the study area [6,7]. Above the Muqdadiyha layer, the alluvial fan was form and it considered one of the modern formations resulting from the flow of water in the flood seasons of Wadi Harran. The fan was composed of several layers formed from coarse materials such as gravel and sand deposits in high discharges of the flood in the lower layers, and then the fine sediments above that layer to form a higher layer with low permeability in the low discharge as shown in Figure 4 below.

Climatological Information for the Study Area
Rain is one of the most important sources of recharge for groundwater in dry areas where there are no water bodies or other sources of surface water. The climate of the study area is generally hot and dry in summer, cold and rainy in winter, the rainfall is between 100-300 mm per year, as shown in the

Topography of the Study Area
The study area is located between a mountainous area on the eastern side near the Iraqi-Iranian border and Al-Naft valley to the west. The level of the region is sloping from the northeast to the southwest, the average elevation of the ground surface ranges from 300 meter above the sea level at the east to 50 meter above the sea level at the west as shown in Figure5.

Wells Data
The data of the wells can be dividing divided into two parts: firstly, the geometric data, including the coordinates, depth, water level, diameter and cross section of the wells shown Table 2. Secondly, groundwater quality data of the wells as shown in Table 3. The first part of data was used for studying the hydraulic properties of the aquifers and established a conceptual model using the GMS software and redistribution of wells in the study area. While the second part was used for assessing the water quality. Note; X, Y Coordinates in decimal degree, Q =Productivity of wells, and dia. = diameter of wells.
The cross section of the wells used to create boreholes, then can be drawn longitudinal sections and soil layers (solid data), shown in Figure 6.

Figure 6. Cross section of Borehole
From the geological map and the data available on the study area, several configurations of the soil layers were defined. The formation of Muqdadiyha is the oldest form, non-permeable layer as mentioned earlier. The layer of gravel mixed formed, above it with some material like sand and clay and above the layer of gravel formed a layer of clay with a little of gravel and sand too. Above these layers formed a mixture of sand, gravel, salts and clay at surface of earth as shown in Figure 7.
In the next step the establishment of a three-dimensional grid required covering all the solid shape. The interval distance between two lines is 100m in X-axis, and Y-axis. The vertical interval was 50 m for each layer in the grid (Zaxis) [8]. The number of layers in grid is five layers started from zero level to 350m above sea level.

Figure 7. Soil layers at the study area
From GIS data, the layers of the conceptual model can be created by including boundary conditions, sources and sinks data, recharge zone and observation wells data. Then grid is constructed, and the data was initializing to be used with MODFLOW, before the conceptual model was converted to a grid-based numerical model. For giving stratigraphy for MODFLOW, models, which consider the main objective of using solid models, a grid-independent of the layer elevations were define, which used for directly, re-create the MODFLOW grid geometry after the grid resolution changed. As well as solid models of stratigraphy can be easily, create by the use of GMS "horizons approach" [9].

Defining Layers of the Conceptual Model
Boundaries layers: It is very important to define the boundary conditions of the study area, for creating a model that simulates the reality, and since the study area is local, most of the borders are of the type of constant head except the northeast it is a non-permeable border of the water. The region does not have rivers or drainage effect on the level groundwater, shown in Figure 8.
Sources and Sinks layer: That layer will be including all sources of recharge and discharges of groundwater in study area. Since the area is devoid of surface water resources, all the inhabitants of this city depend on the groundwater. The specifications of wells of study area were mention in Table 3, except the operating (pumping) data, which will be estimate for each situation separately.
Recharge layer: This layer represents the amount of net rainfall that feeds the groundwater, as usually rainwater goes in the form of surface runoff or evaporation. It was calculated as a daily average of the year at steady state condition by taking the annual rainfall rate (250) mm divided by the number of days of the year multiplying by deep percolation percentage which ranges from (5-20) % of the amount of rainfall [10].
Recharge = (250/ (365×1000))×15%=0.0001 m/day After completing all above data, the next steps are creation MODFLOW model from solid data then interpolated MODFLOW with all layer in conceptual model, the readings of the observation wells, were recorded after stopping pumping to obtain the best static water level for the purpose of increasing accuracy during the calibration process of the model.

Estimating Properties of Layers and Running MODFLOW
Before starting run of MODFLOW, soil specifications should be introduced as default values according to soil layer type, then by giving minimum and maximum value for conductivity of aquifer at steady state accordant to range of hydraulic conductivity values for geological materials (After Todd, 1980), assuming the storage coefficient and specific No flow yield equal to zero at steady state. The starting head must be defined for each layer of MODFLOW model see Figure 7. Then, from the PEST package provided by the program, the conductivity of layers values can be estimated and calibration of model at same time. Now MODFLOW model can be running, and after several cycles the values of the hydraulic conductivity can be estimate, the error ratio in the initial values will be reduce until it reaches the optimal values, the program will be stopping, and then these values used to define the soil properties in the study area as showing in table 4. For check calibration are showing in Table 5 and Figure 9. Table 4. Hydraulic conductivity of soil layers (k)  The error was represented by the colored bar. When the color bar is green that's means bar lies entirely within the target, but when the error is less than 200%, means that the bar is outside the target, so the bar is draw in yellow. As well as when the error is greater than 200%, the bar will be drawing in red. In this case, the color bar is green. The red triangle represents the dry cells and the blue triangle represents the flooded cells.
From the results obtained by the MODFLOW model, the hydraulic conductivity of Layer two (clay layer) was found to be of low value (3m/day). That layer represents a perched aquifer in the local area. The water flows in the horizontal and vertical direction above the clay layer, at the end of the alluvial fan the clay layer will be close to the ground surface, that leading to flow of water above the surface. From field observations, the groundwater flows at the fan's border in form of springs. The gravel layer below the clay layer it is a semi-confined layer and it is the main storage layer of groundwater in Mandali alluvial fan.

Determine the Storage Coefficient and Specific Yield at Unsteady State
After determination of the hydraulic conductivity values of the soil layers, the values of storage coefficient and specific yield can calculate in the same way, as the conductivity was calculated, with the need for data for one of the observation wells at lies. The management of transient information from a different source would be create a transient simulation, including pumping wells data, Table 8, recharge data shown in Figure 10, and water table of observation wells showing in Table 6. The first step in an unsteady state it is to convert the MODFLOW model from a steady to transient form, then import the transient data like rainfall, pumping periods, flow rates of wells in study area and stress period shown in Tables 7 and 8.   After determining the values of the storage coefficients and the specific yields, and the calibration the model was done, the results of the unsteady state of the simulation model can be compared with observation records, as shown in Table 9.  Figure 11 shows the differences between water table in observation well measured in the field and level of water obtained by the conceptual model and the difference is within the permissible limits.

Run-time Scenarios and Distribution of Wells
After running the MODFLOW model in a normal state as shown in Figure 12, red triangles in the high regions are representing dry cells while blue triangles at the low end of the alluvial fan are representing the flooded cell. In fact, dry cells are only found in the top layer of the highlands, and submerged areas are the natural springs of water. These springs were formed because of a high hydraulic gradient of water, and because of the second layer (clay layer) which has low hydraulic conductivity, and then the flow above this layer will be in two directions, horizontal flow and vertical flow. The third layer (gravel layer) is the main storage for groundwater in the study area.

Figure 12. Drawdown in water table after 6 hrs.
After determining the hydraulic specifications of the soil layers in the study area, the wells were hypothetically increased and different spacing between these wells were tested. To determine the optimum distribution with acceptable drawdown of water table, 103 wells with spacing of, 1500 m, are used in this scenario. Assuming the productivity of the well was 7 liters per second, or about 600 cubic meters per day. Three daily operating times were proposed and these are 6, 12 and 18 hours, with annual rainfall rate of (0.0001 m/day), the discharges from wells for the three scenarios were 150, 300, and 450 m 3 /day respectively. The proposed operation of these wells was for one year only.
 Scenario 1: The operation time was 6 hours with discharge 150 m 3 /day, Figure 13 shows that the drop in the water table was between 1m and 7m, and the maximum drawdown was in the middle of the study area.
 Scenario 2: The operation time was 12 hours with discharge of 300m 3 /day, Figure 14 shows that the drop in the water level was between 1m and 11.5 m, and the maximum drawdown of water table was in the middle of the study area too.
 Scenario 3: The operation time was 18 hours with discharge of 450m 3 /day, Figure 15 shows that the drop in the water level was between 2.5m and 20m, and the maximum drawdown of water table was in the middle of the study area too.

Assessment of Groundwater Quality
It is essential to study the physical and chemical characteristics of groundwater in the study area before recommend it for drinking or irrigation usage. It was well known that in the movement of groundwater it may pass through regions where impurities will be dissolved and deteriorating its quality [11,12]. There are many classifications to assess water quality, but the simplest of these classifications is depending on the value of total dissolved solids (TDS) on the basis of as shown in Table 10. According to this classification, the groundwater of the wells in the study area in Table 3 is slightly saline water except for well (W18) [13,14]. It is a moderately saline water show Figure 16.

Conclusions
In this research, Groundwater Modeling System (GMS) software was used to build three-dimensional conceptual model and management of groundwater usage in the Alluvial fan of Mandali aquifer after the determining the hydraulic conductivity, storage coefficient and specific yield. Three scenarios with different daily operation time (6, 12 and 18) were tested for feasible wells distribution based on minimum drawdown. The wells is hypothetically increased to 103 over the study area with spacing of 1500 m. From GMS software simulation results for the study area, the following conclusions can be drawn:  For daily operation of 6 and 12 hours, the maximum drawdowns were found to be 7m, and 11.5 m, respectively.
 For daily operation of 18 hours, the maximum drawdown was 20 m, so this operation scenario will be recommended for operating the proposed number of wells and meet the demand in the study area.
 For all wells excluding well number 18, the range of TDS was from 1000 to 3000 mg/l. However, for well number 18, the TDS was found more than 3000 mg/l which indicate that the groundwater in this well is not recommended to be used for irrigation but according to Iraqi stander for drink (IQS 2009) it can be used for drinking after advanced treatment.

Conflicts of Interest
The authors declare no conflict of interest.