Groundwater resource management based on three-dimensional flow modeling of Wadi Baysh aquifer system: southwestern Saudi Arabia

Abstract

Regional groundwater modeling is a valuable tool for ensuring the sustainable development and management of groundwater resources as it allows the assessment of the aquifer behavior under different types of hydrological stresses. The objective of this study was to use 3D flow modeling to assess the best options for managing groundwater in the Wadi Baysh basin in Saudi Arabia. The model for the Wadi Baysh aquifer system was based on groundwater monitoring data for 10-year period from 2007 to 2016. The model was tested with different hydrological stresses until the year 2050. The model behavior shows that groundwater levels in Wadi Baysh aquifer show no signs of depletion at the current pumping rates. The impact of the water stored in the dam reservoir is clearly seen and the zone of influence is very well demarcated. At this stage, groundwater use in the Wadi Baysh aquifer system is sustainable and no decline in groundwater levels is observed.

Introduction

Due to the benefits associated with the various economic activities such as fisheries, tourism and recreation, more than 40% of the world’s population resides within 100 km of the coast. Access and availability of clean water play a vital role in the socio-economic development of coastal zones (Neumann et al. 2015). However, the availability of freshwater in coastal areas can be easily influenced by climate change and human activities (Harley et al. 2006; Stefanova et al. 2015). This mainly includes saline water intrusion as a result of the rise in sea levels or due to the effects of over-abstraction of groundwater (Taylor et al. 2013). Consequently, developing an understanding of the hydrogeological processes and the response of the aquifer to the various components of the hydrological cycle becomes a prerequisite for proper groundwater management in these areas. This is especially the case in coastal regions that have an arid climate where groundwater is often the only source of freshwater.

Jazan region in Saudi Arabia is one such example of a coastal area where groundwater from the Wadi Baysh aquifers system constitutes an important source of potable water. Consequently, it is important to determine the hydrological response of the aquifer system due to increasing groundwater abstraction and decreasing rainfall associated with climate change.

The main objective of this study was to carry out the 3D numerical simulation of the groundwater flow of Wadi Baysh aquifer system to understand the behavior of the groundwater flow, the impact of various hydrological stresses on the aquifer system and the impact of flooding on the groundwater levels and quality. The objective was achieved by first developing a conceptual model for the aquifer system using the available hydrogeological information and then preparing a 3D numerical model for predicting the future groundwater scenario based on different hydrological stresses for building an effective and sustainable water management plan.

The prepared model is based on finite-difference estimation for solving partial differential formulas of groundwater movement (Senthilkumar and Elango 2004). The flow equation, with the specification of water flux or head condition or together based on the boundary conditions of the aquifer and the situations of primary water head, constitutes mathematical aspects of groundwater movement.

Study area description

Wadi Baysh basin is an important part of Jazan province and is located in the Tihama—Asir region, which is a narrow coastal strip along the Red Sea. The study area is located between longitudes 42° 24′ to 43° 27′ E and latitudes 17° 05′ to 18° 03′ N. The catchment area of Wadi Baysh basin is about 7309 km². The basin (115 km in length) gets its recharge from Asir Mountain with the drainage eventually joining the Red Sea, as shown in Fig. 1. The catchment area of Wadi Baysh is about 0.4% of the total area of Saudi Arabia. The area receives annual rainfall ranging from 100 to 380 mm with estimated annual rainfall volume ranging from 0.73 to 2.8 Billion Cubic Meters (BCM).

Fig. 1

Location map and DEM of Wadi Baysh basin

Wadi Baysh basin shows different relief characteristics with elevation ranging from 0 m above mean sea level (amsl) at the Red Sea Coast to 2980 m amsl along the water divide in the Asir Mountain. Wadi Baysh basin is typical of many basins in the southwestern part of the Kingdom which is occasionally flooded resulting in the destruction of infrastructure and leading to a significant loss of water to the Red Sea. To control flood events, a dam was constructed in the upper part of the Wadi Baysh basin.

Geology and hydrogeology

Geologically more than 50% of Wadi Baysh basin consists of the Precambrian basement rocks which form a part of Arabian Shield. Figure 2 shows the geological map of Wadi Baysh. A small sedimentary succession of rocks outcrops in the center of the study area and includes the Cambrian-Ordovician Wajid sandstone, the Jurassic sandstone of the Khums Formation and the Jurassic Limestone of the Hanifa Formation.

Fig. 2

Geology of Wadi Baysh Basin

The flood basalts and igneous intrusions of Gabbro and Diabase and ferruginous laterite and silcrete deposits below the flood basalts represent the rocks from the Tertiary period.

The lower part of the Wadi Baysh basin is occupied by the unconsolidated deposits of the Quaternary period. These unconsolidated deposits occur in the Wadi beds and also as extensive sheets covering much of the coastal plain. The alluvial layers are thin, ranging from a few meters to more than 80 m. The main Quaternary deposits include the terrace deposits consisting of inactive deposits of gravel, sand and silt capped by a veneer of lag gravel; the Quaternary alluvium consisting of gravel sand and silt in modern stream channels; and Quaternary reworked alluvium comprising mainly of sand and silt adjacent to modern stream channels. Structurally, Wadi Baysh basin is dissected by many major and minor structures which play a significant role in enhancing groundwater recharge as shown in Fig. 2.

Groundwater is the main source for irrigation in the Wadi Baysh basin. The main aquifer of Wadi Baysh basin is the Quaternary unconsolidated deposits which contain groundwater in the unconfined conditions. There are hundreds of drilled wells, most of them fully penetrating the Quaternary aquifer. This aquifer is composed of coarse sand and gravel with clay intercalations. Its thickness varies between a few meters in the Wadi deposits to about 100 meters in the coastal areas. Figure 3 shows the depth to groundwater level in the Quaternary alluvial deposits which form the main aquifer unit of the region. The map has been prepared from water level observations in 100 wells monitored in 2015 and 2016. It ranges from 2 to 52 m below ground level for 2015 (Fig. 3a) and from 1 to 57 m for 2016 (Fig. 3b).

Fig. 3

a Depth to groundwater table map of the study area for 2015; b depth to groundwater table map of the study area for 2016. Values are in meters below ground level

Model conceptualization

The basic step involving any groundwater flow modeling is the preparation of a conceptual model that contains the most important hydrogeological characteristics of the system.

The numerical modeling program MODFLOW (McDonald and Harbaugh 1988) was used in this study for simulating groundwater movement in the Wadi Baysh aquifer. Wadi Baysh basin covers an area of about 7309 km² based upon the Universal Transverse Mercator (UTM) of World Geodetic System (WGS) 1984. The model domain for the Wadi Baysh basin was divided into 168 rows and 174 columns, with a single vertical layer in the regional model. Multiple aquifer units with varied topography conditions and limited abstraction of water for drinking and agricultural characterize the study area. The region of interest, which is the downstream portion of the basin (Dam area), has an area of 1404 km² and was modeled in detail with a grid pattern of 100 × 100 m. The remaining study area was divided into rows and columns with a grid size of 500 × 500 m (Fig. 4).

Fig. 4

Discretization and Boundary Conditions of the Study area

This simulation of the Wadi Baysh aquifer system is based on the detailed study of geology, well lithology, piezometric levels and hydrological cross sections. The aquifer in the study area consists of unconsolidated gravels and sands of Quaternary age intercalated with some clay. The groundwater is found in unconfined condition and the aquifer has an area of about 1404 km². The groundwater follows the topography from the north and northeast to southwest.

Boundary conditions

The hydrologic boundaries of the modeled Wadi Baysh basin are represented in Fig. 4. Since the entire Wadi Baysh catchment has been included in the model domain, no flow boundaries were assigned to the model as the horizontal flows from outside the catchment has been neglected. The south-west boundary of the domain is restricted by the Red Sea and is considered to be a constant head boundary.

The elevation of the top of the aquifer was derived from the Shuttle Radar Topography Mission (SRTM) data of 30-m resolution. Data were downloaded and processed in ARC GIS 10.3, and 100,000 data points were selected and transferred to the model domain. The elevation values ranged from 2966 to 2 m (Fig. 5) and were used as input to the model as the top elevation data. The bottom elevation data were mainly derived from the lithological descriptions of the well drillings and field visits. The modeled aquifer is considered to be a single layer under unconfined conditions.

Fig. 5

Digital Elevation Model, a used as the model’s top layer and b 3D disposition of the study area

Input data of the model

The initial heads of groundwater for Wadi Baysh basin were measured in 2007. Based on intensive studies and analysis of the rainfall events, hydrographs and water level measurements, it is considered that the groundwater head values of January 2007 adequately represent the baseline spatial distribution of potentiometric heads in the Wadi Baysh aquifer.

Aquifer dimensions

The aquifer geometry comprises determining the top, bottom slices and thickness of the unconfined aquifer. Based on the available data from SRTM and the lithological descriptions of the well drilling-logs, the elevations of the top and bottom of the aquifer were determined. The 3D disposition of the study area is depicted in Fig. 5.

Aquifer parameters

The aquifer parameters, which include transmissivity and specific yield, were calculated from the interpretation of data from 12 aquifer tests in the study area (12 pumping and 12 recovery data) as shown in Table 1. Most of the pumping test data were available only for the Quaternary formation, and data pertaining to the Precambrian basement rocks were not available. Hence, standard textbook values for the basement rocks were used in the model (Todd 2005).

Table 1 Pumping test results

Groundwater sources and sinks

Groundwater withdrawal

Groundwater from the Wadi Baysh basin is extracted for a variety of uses including domestic and agriculture. Based on the inventory of 102 pumping wells during the field visits, it was calculated that the aquifer is being pumped at the rate of 60 m³/h. The average pumping duration is 6 h per day, three times a week. The total groundwater abstraction in the area is therefore about 360 m³/day and 1080 m³/week. The pumping wells (n = 102) were distributed in the model domain using the well package, and are shown in Fig. 6. Many indications based upon the land use, climate and soil characteristics show that the inhabitants of the area use groundwater only during the dry summer season (when there is no rainfall).

Fig. 6

Spatial distribution of the pumping wells (n = 102) in the study area

Groundwater recharge

Groundwater recharge of Wadi Baysh basin aquifer is controlled by geological, topographical and land cover aspects. Based on the intensive studies of surface hydrology and groundwater monitoring, it is likely that rainfall events and flash floods are the main sources of aquifer recharge. Rainfall amounts of more than 20 mm have a pronounced effect on the groundwater levels and aquifer recharge. The infiltration capacity of geological formations present in the Wadi Baysh basin ranges from 0–12%. Table 2 shows the rainfall infiltration factor used in modeling for groundwater recharge calculation (Chatterjee and Ray 2014) which is based on the equation:

Table 2 Rainfall vs. infiltration factor used in groundwater recharge calculations

$${\text{Recharge }}\left( {{\text{m}}^{3} } \right) \, = {\text{ Area }}\left( {{\text{km}}^{2} } \right) \, \times {\text{ Infiltration factor }}\left( \% \right) \, \times {\text{ rainfall intensity }}\left( {\text{m}} \right)$$

(1)

The recharge values are calculated based on the infiltration factor, geological formation and intensity of rainfall. The rate of recharge from the main channel to the aquifer is estimated based on the difference between the channels head and the groundwater head. Water heads of the main channels were estimated depending on the head in the adjacent wells. Three recharge zones were demarcated in the study area which includes (a) an area comprising of Quaternary sediments, (b) an area underlain by Precambrian basement rocks that cover most of the northern part of the basin and (c) a small area underlain by Cenozoic rocks in the eastern and central portion of the study area (Fig. 7).

Fig. 7

Recharge zones of the study area (blue colure for Quaternary, green color for Permian and Jurassic and White color for Precambrian)

Dam location, main channel and lake in the model domain

The location of the dam was put in the model domain using the well package, and the main channels in the model domain were incorporated using the river package of MODFLOW. The height and base of the channel were input to the model using the SRTM elevation data. A lake on the upstream side of the dam was also incorporated in the model to assess the impact of the dam flooding. The location of the dam in the model has been shown in Fig. 8.

Fig. 8

Model domain, main channels and Dam location of the study area

Mathematical and controlling formulas and Model program

MODFLOW determines groundwater flow through a porous medium by numerically solving partial differential equations developed from Darcy’s Law (Darcy 1856) and the principle of conservation of mass (McDonald and Harbaugh 1988) as cited by Alsharhan et al. (2001). The following equation is solved for each cell in the model domain:

$$\frac{\delta }{\delta x}K_{xx} + \left\{ {\frac{\delta }{\delta x}K_{yy} } \right\}\frac{\delta h}{\delta y} + \left\{ {\frac{\delta }{\gamma z}K_{zz} } \right\}\frac{\delta h}{\delta z} - \left\{ {W = S_{\text{s}} } \right\}\frac{\delta h}{\delta t},$$

(2)

where K_xx, K_yy and K_zz represent the hydraulic conductivity in x, y and z directions, respectively, along Cartesian co-ordinate axes, which are supposed to align with the main direction of hydraulic conductivities; h represents the hydraulic head; W is the volumetric flux per unit volume and represents sinks and/or sources; Ss represents the specific storage of the porous material; and t represents time. MODFLOW simulates the changes in the hydraulic heads of the aquifer system (hydraulic heads) to the specified hydrologic stress (sources and sinks).

The groundwater simulated in the Wadi aquifer model is derived from: the infiltration of water in recharge areas; return flow; and the infiltration of water along the main channels during rainfall events. The principal source of discharge from the model domain is by groundwater abstraction by pumping wells. At each finite-difference grid cell node, the model simulated the hydraulic heads and mass water balances of the aquifer system.

Model calibration

The process of model calibration strategy was used in such a manner so that the well-known strictures were varied as little as possible, and the poorly defined or unknown factors were varied to the maximum to get the best conformity between computed and observed field data. Model calibration under steady-state condition was carried out to reduce the variation amongst the modeled and the actual groundwater level conditions. Groundwater head data recorded during Jan 2007 in 22 wells distributed spatially throughout the study area were used for calibration under steady-state conditions.

The hydraulic conductivity values were the most poorly defined parameter in the model as the data were available from only 12 pumping tests which were carried out in the study area. In addition, the hydraulic conductivity values obtained from pumping tests were for the Quaternary sediments and no data were available for the Precambrian basement rocks which cover almost 90% of the study area. This necessitated other methods of estimating aquifer parameters in these areas.

This was done by assessing the lithological variations in the water-bearing formations of the study area by analyzing borehole lithology from accessible large diameter open wells.

On the basis of this analysis, the values of hydraulic conductivity were varied up to 10% of the obtained values from pumping tests in the quaternary formations, so that the computed groundwater heads mimic the observed heads (Fig. 9a, b).

Fig. 9

Comparison of computed and observed groundwater head for 31 days (a) and 120 days (b)

The outputs clearly indicate that there is a perfect and excellent match between the computed and observed groundwater levels in most of the wells of the study basin. Through numerous trial runs, the (RMSE) Root Mean Square Error and the Mean Error (ME) were minimized in model-runs using the method reported by Bredehoeft 2005.

Transient-state simulations using the model were carried out for the period January 2007 to December 2016 for a period of 10 years with monthly stress-periods (120), with each stress-period consisting of 30 time-steps. The calibration of the model under transient conditions was obtained by numerous repeated assessment runs using a trial-and-error process to get the best possible match between the simulated and measured field groundwater levels, both spatially and temporally.

In the transient model calibration, the incorporated hydraulic conductivity values were altered from those calibrated by the model under steady-state condition. On the basis of the best match amongst the observed and computed groundwater levels from 22 observation stations spread throughout the study area during the period from January 2014 to December 2016, the transient models were considered to be suitably calibrated. The sensitivity of the groundwater model was checked by changing the input parameters. The model stricture of concern was tested for a series of values and the changes in the model were determined by recording the response in the RMS error of the groundwater level difference between the modeled and actual measured data.

Model budget

Water budget components for the Wadi Baysh aquifer model were calculated using the program ZONEBUDGET (Harbaugh 1990) for the steady-state and unsteady periods. The water mass budget for the Wadi aquifer system is represented in Table 3.

Table 3 Mass budget for the Wadi aquifer system

Simulation results

Under transient-state conditions, the model was run for a period of 10 years from January 2007 to December 2016 which indicated that simulated and observed heads had a good agreement. Figure 10a, b clearly shows that the computed groundwater head during June and December closely matches the observed groundwater heads. A detailed observation of the computed groundwater head of the aquifer clearly indicates that the highest groundwater heads are observed in the northern side of the study area and that the lowest heads are observed in the southern part of the basin, which forms the common evidence of the elevation (Fig. 10a, b).

Fig. 10

Comparison of computed and observed GW head during June 2015 (a) and simulated groundwater head during December 2015 (b)

The regional water level trend is from the north to the south. The groundwater movement paths for the month of June 2015 are given in Figs. 14. Almost all the wells show respectable matching between the measured and computed groundwater heads. The Root Mean square (RMS) error ranges from 8.96 to 9.45 m and the normalized root mean square error ranges from 5.2 to 5.4%. The correlation coefficient for the model ranges from 0.988 to 0.989. The observation from the RMS and normalized RMS error clearly indicates that the simulated groundwater heads are similar to those measured in the field.

Furthermore, the results of this simulation suggest that groundwater use from Wadi aquifer system is currently sustainable. The flow vectors clearly indicate the groundwater flow direction. Groundwater flow in the Wadi aquifer system follows the topography and flows from northeast to the southwest following the channels.

Impact of the dam

The calibrated model was used to understand the impact of the dam on the groundwater resources of the Wadi Aquifer system. The groundwater flow contours (Fig. 11) clearly indicate the period of flooding and non-flooding.

Fig. 11

Impact of the dam flooding on groundwater head (thicker contour line infers the groundwater head with flooding in the dam and lighter/thinner contour line infers groundwater head without flooding in the dam

A closer study of the contours at elevations of 200 and 180 m apparently shows how they overlay each other. The thicker contour is groundwater head with the flooding in the dam and lighter/thinner contour line is groundwater head without flooding. The thicker 200-m contour line is above the thinner 200-m line. This clearly indicates the buildup of the groundwater head in the upstream side of the dam and excess area is covered by the thicker 200 m contour line in the upstream side.

In the case of the 180-m contour line, which is located on the downstream side of the dam, the situation is reversed. The thinner 180-m contour line is above the thicker 180-m contour line. This situation shows the movement of the 180-m thicker contour line, i.e., the groundwater head with the flooding in the dam occupies more area in the downstream side. The buildup of the groundwater head (180-m contour) in the downstream side of the dam is a clear indication of increased recharge to the system. Increase in groundwater head is observed (Fig. 11) both upstream and downstream of the dam, although the level of increase is higher on the downstream side of this structure.

Model forecast

The hydrologic behavior of the aquifer to varying water inputs and outputs from the system was simulated to ensure that proper management of groundwater resources in the Wadi Baysh basin. Different modeling scenarios were generated for a forecasting period starting from 2017 to 2050. All the input data of hydrology, hydrogeology, meteorology and basin characteristics were entered to the model up to 2050. Forecasting scenarios were planned for the development of optimum management policy.

Normal rainfall situation

This model scenario is simulated to forecast the regional groundwater heads of Wadi Baysh basin till the year of 2050 based upon the measured rainfall data and the current water levels during the project period (2015–2016). The modeled groundwater heads of the year 2030 show that there is no more change in the water heads throughout much of the study area compared with the current situation of 2016 as shown in Fig. 12. This suggests that the water balance of the Wadi Baysh aquifer is stable under the current rainfall regime and with current groundwater use. The model output indicates that there are no significant changes in groundwater heads between 2030 and 2050 and that the potentiometric head contours for these two periods of time are comparable.

Fig. 12

Predicted groundwater head in 2030 under the business as usual scenario

Assessing the impact of flooding

The scenario where flooding of the dam takes place at four-yearly intervals until 2030 was also modeled. The rainfall data of 2015 and 2016 were used for this model prediction analysis. In this scenario, groundwater levels increased both in the downstream and upstream sides of the dam site, and the resulting pattern of groundwater flow is shown in Figs. 13 and 14. However, the groundwater level reaches the same level which is observed during the normal rainfall season, within a year (Fig. 13).

Fig. 13

Predicted groundwater head with flooding once in 4 years

Fig. 14

Impact of flooding in Dam on the downstream side—Change/mismatching in contour lines indicating the impact of recharge

The effect of increasing groundwater abstraction

The effect of increasing groundwater abstraction from the aquifer was also simulated. The present pumping wells (n = 102) were distributed in the model domain using the well package. Additional wells were placed at suitable locations within the basin to match the expected increase in groundwater demand.

In scenario 1 of these model runs, 200 wells were included in the model, and in scenario 2500 (Fig. 15a) wells were added. The impacts of pumping on potentiometric heads in these two scenarios were then investigated for the years 2030 and 2050.

Fig. 15

a Increase in pumping wells (n = 500). b Impact of increase in pumping in 2030

Scenario 1 (200 wells) shows that the groundwater head declines by 0.02 to 0.11 m/year. In scenario 2 (500 wells), the largest declines in potentiometric heads (Fig. 15b) take place in the central and western portions of the study area in 2030. The movement of the 200–250-m head elevation contour in the central portion of the model domain indicates the extent of the decline with an increase in pumping.

As the rate of pumping increases, groundwater use from the aquifer system becomes less sustainable with large declining potentiometric head trends occurring in the central and western portions of the model domain. However, there is no change in the declining trend beyond 2030 and, it is steady and ranges from 0.08 to 0.25 m/year until the year 2050.

Impact of decreasing recharge on potentiometric heads

The simulation was used to predict the impact of the decrease in recharge in the Wadi Aquifer system. The impact on the groundwater head during the year 2050 (Fig. 16a) clearly shows that regions downstream of the dam, central portion and western portion are affected by the declining groundwater heads due to no recharge.

Fig. 16

a Impact of groundwater head in 2050 with and without recharge. b impact on groundwater head with recharge and without recharge with an increase in pumping rate (n = 500) a in 2050

Further, there is a mismatch between the groundwater head with recharge and without recharge in the regions downstream of the dam site (Fig. 16a). A shift in the potentiometric contour lines from 175 to 25 m is clearly observed. Further, the modeled scenarios were applied for predicting the groundwater heads with an increase in pumping and no groundwater recharge for the year 2050 (Fig. 16b). The predicted model clearly shows a drastic decline in groundwater head in the central, western regions and downstream of the dam site.

Effect of recharge from the dam

The model was also used to assess the potential changes in recharge that would take place if there was no dam in the region. Figure 17 clearly shows the difference in the potentiometric heads between the presence and absence of the dam and its contribution to groundwater recharge. Blue contour lines (Fig. 17) are the groundwater head values if the dam exists and black contour lines represent the groundwater head values if the dam does not exist. This figure shows that the presence of the dam has a significant impact on recharge in the area. For example, the area covered by the blue contour line (200—225 m) expands and indicates the presence of a large groundwater mound downstream of the dam. Furthermore, the area covered by the blue contour lines (175 m onwards) also expands clearly showing the effects of increased recharge with the head in the dam maintained at 64 m. The benefit to downstream regions is also observed by the shifting of contours of groundwater head. In the absence of the dam, the increase in groundwater abstraction structures would have a drastic impact on the Wadi aquifer system.

Fig. 17

Groundwater head with dam recharge and without dam recharge in

Water quality—solute modeling

Simulations were also carried out with the calibrated model to determine the likely impact of the flooding in the dam on the Total Dissolved Solids (TDS) content of groundwater. The concentration of TDS in the study area ranges from 305 to 8125 mg/l and all the sampling locations are situated in the southern portion of the study area. Changes in the TDS content of groundwater were simulated using the 2015 TDS concentrations as the initial concentrations in the model. The methodology adopted for TDS simulation was according to Vetrimurugan et al. (2017). The simulated TDS concentrations are represented in Figs. 18a, b. The results of the simulation show that the TDS concentrations decline from about 2000 to 1000 mg/L in the western portion of the model domain near the dam after the flooding event due to the effects of flooding events. The predicted TDS concentration until 2030 is also similar to the concentration observed in Fig. 18b. After 2030, there is no change in the groundwater quality due to recharge.

Fig. 18

TDS concentration in Wadi Baysh basin a before flooding and b after flooding

Summary and conclusions

• 3D finite-difference flow and solute transport models were applied for simulating the groundwater heads in the Wadi Baysh aquifer system for a time period of 10 years (2007–2016) to study and manage the behavior of the aquifer.

• The modeled scenarios show that Wadi Baysh aquifer system is characterized by steady-state conditions under the current situations. The groundwater movement in the area is from northeast to the southwest and closely follows the regional topography.

• The main channels of Wadi Baysh have a considerable contribution to the steady-state condition of the aquifer system.

• Flooding events in the dam will increase groundwater recharge to the aquifer and create a large groundwater mound near the structure.

• The model mass budget shows that river contributes to 61,000 m³ of water to the aquifer. The modeling indicates that seepage from the main channels of the Wadi system is the most significant source of recharge to the aquifer and that the current level of groundwater discharge is in a condition of steady-state equilibrium with the recharge.

• Simulations indicate that the TDS content of groundwater near the dam will decrease from about 2000 mg/L to 1000 mg/L due to the effects of recharge from the dam

• Simulations that include an additional 500 pumping wells in the central and western portions of the study area indicate that the potentiometric head in the aquifer will decline at a rate of 0.08–0.25 m/year.

• A prolonged drought (i.e., no rainfall recharge on the aquifer system) with increased pumping will cause large declines in potentiometric heads in the central and western parts of the study area, but areas downstream of the dam will be minimally affected.

• To manage the Wadi aquifer system safely in the future, the increase in groundwater abstraction structure needs to be restricted to less than 200 additional wells so the decline in groundwater head is minimized and is maintained at a decline-rate of between 0.02 and 0.11 m/year (i.e. the overall decline in heads over a decade would be 0.2 to 1.1 m). With one or two flooding events in a decade, the groundwater decline would be arrested and the groundwater heads in the Wadi aquifer system would return to the present situation.