The present study examines the response of groundwater systems to expected changes in the Mediterranean Sea (rise of <1cm/yr) and Dead Sea levels (decline of ∼1 m/yr). A fast response is observed in the Dead Sea coastal aquifer, exhibited both in the drop of the water levels and in the location of the fresh-saline water interface. No such effect is yet observed in the Mediterranean coastal aquifer, as expected. Numerical simulations, using the FeFlow software, show that the effect of global sea level rise depends on the coastal topography next to the shoreline. A slope of 2.5‰ is expected to yield a shift of the interface by 400 m, after a rise of 1m (∼100 years), whereas a vertical slope will yield no shift. Reduced recharge due to climate change or overexploitation of groundwater also enhances the inland shift of the interface.
 Subsurface seawater intrusion is an extremely important factor in management of coastal aquifers [e.g., Van Dam, 1999]. Its rate is usually determined by the rate of increase of groundwater salinity [Melloul and Zeitoun, 1999]. A direct estimation of the rate of the intrusion can be conducted by the analysis of radiocarbon and tritium isotopes in monitoring wells [Yechieli et al., 2001; Sivan et al., 2005].
 Sea levels have changed as a result of alterations in climatic conditions during the geological history. Currently, there is a common hypothesis which claims that the climate is largely influenced by human activity leading to global warming. As a result, polar icebergs and on-land glaciers are melting and raising the sea level. A sea level rise of ∼0.5 m by the year 2100 was suggested by Warrick et al. .
 The objective of this study was to understand the possible effects of sea level rise on coastal aquifers, taking into account variations in the hydrogeological configuration and precipitation due to climate change. Because of the small magnitude of the changes in ocean levels, which makes it difficult to actually track its effect on groundwater in the field within a reasonable monitoring time window, the field conditions of the Dead Sea hydrological system were also examined. The unique, rapidly changed system of the Dead Sea provides data that can be extrapolated to other coastal systems, such as the Mediterranean, provided their differences are taken into account.
2. Site Characterization
 The present study was conducted in two different coastal aquifers, one adjacent to the Mediterranean Sea and the other adjacent to the Dead Sea. The Mediterranean aquifer is a good representative of many coastal aquifers near the ocean around the world, while the Dead Sea provides a unique opportunity to study the response of groundwater to sea level changes in the field. Both aquifers consist of several subaquifers, separated by less permeable layers. A description of the two aquifers is provided in sections 2.1 and 2.2.
2.1. Israeli Mediterranean Coastal Aquifer
 The Israeli Mediterranean coastal aquifer extends for more than 120 km along the eastern part of the Mediterranean Sea (Figure 1). Its thickness decreases eastward from ∼200 m near the coastline to a few tens of meters and less near the eastern boundary (Figure 2). The aquifer, belonging to the Kurkar Group, consists of interlayered sandstone, calcareous sandstone, siltstone, and red loam, which alternate with continental and marine clays of Pleistocene age, overlying impervious marine clays of the Saqiye Group of Pliocene age [Issar, 1968]. East of the shoreline, up to a distance of 5–8 km, clay interlayers subdivide the aquifer into four subaquifers. Some of the lower subaquifers are confined, while the upper ones are phreatic [Nativ and Weisbrod, 1994]. The connection of the lower subaquifers to the sea is debatable; while some claim that they are connected [Bear and Kapuler, 1981], others say that they are not [Kolton, 1988; Kafri and Goldman, 2006]. The precipitation over most of the coastal aquifer is ∼600 mm/yr, and the recharge coefficient was taken to be about 0.3 [Gvirtzman, 2002]. Thus, the recharge to most parts of the aquifer is ∼200 mm/yr.
 Basic information on the fresh-saline water interface is available from hundreds of wells, from TDEM studies [Goldman et al., 1989] and from chemical and hydrological monitoring [Vengosh et al., 1991; Yechieli et al., 1997; Hydrometric Department, Israeli Hydrological Service, unpublished report, 1995]. In several areas, the fresh-saline water interface has moved inland over a distance of more than 1 km as a result of overpumping during the last few decades [Melloul and Zeitoun, 1999]. However, since the array of monitoring wells is not sufficient because of scarcity of boreholes in many parts of the aquifer, especially at different distances from the shoreline, the actual extent of penetration is not known from the entire aquifer.
2.2. Dead Sea Coastal Aquifer
 The Dead Sea is situated in the deepest part of the Dead Sea Rift System. It is a terminal lake with no outflow of surface or subsurface water. The Dead Sea groundwater system consists of two main aquifers (Figure 3): the Upper Cretaceous Judea Group Aquifer and the Quaternary alluvial coastal aquifer [Arad and Michaeli, 1967; Yechieli et al., 1995]. The present study is focused on the coastal aquifer, which consists mainly of clastic sediments, such as gravel, sand, and clay deposited as fan deltas and lacustrine sediments such as clays, aragonite, gypsum, and salts. The alternations between gravel and clay subdivide the aquifer into several subaquifers that differ in their groundwater level and chemical composition. The transmissivity of the aquifer was estimated in an interference pumping test in Arugot Wadi to be 1500 m2/d [Wollman et al., 2003], yielding a hydraulic conductivity value of 30–100 m/d, depending on the specific thickness of the aquifer, which is not known. This aquifer is bounded by normal faults, which set Cretaceous carbonate rocks of the Judea Group against Quaternary alluvial and lacustrine sediments. The recharge of the aquifer is mainly through lateral flow from the Judea Group aquifer, which is replenished in the highlands 10–30 km to the west and by flash floods. Direct recharge is negligible because of the arid climate and high evaporation in the Dead Sea region. Because of the scarcity of observation boreholes penetrated into the deeper subaquifers, the monitoring of the water levels and the interface location was conducted mainly in the upper subaquifer.
 The Dead Sea salinity and density are 340 g/L and 1.24 kg/L, respectively [Lensky et al., 2005]. The extremely high density of the Dead Sea induces a very shallow interface between the fresh water and the brine. According to the Ghyben-Herzberg approximation, the depth of the interface is 4.35 times that of the groundwater head above sea level compared to 40 times in the ocean [Yechieli, 2000].
 The situation in the Dead Sea is different from that near the Mediterranean because of the rapid drop of its level by 1 m/yr due to its negative water balance. This drop is a result of the considerable exploitation and diversion of water upstream from the Sea of Galilee and the Jordan and the Jarmuch rivers since the 1960s and is due to the Dead Sea brine pumping of the Israeli and Jordanian Dead Sea industries [Yechieli et al., 1998; Lensky et al., 2005]. Fluctuation of the Dead Sea occurred also in the Pleistocene and Holocene times [Bartov et al., 2002; Bookman et al., 2004], which probably affected the groundwater system [Yechieli et al., 2009], but these fluctuations are not within the scope of this work.
 The Dead Sea and the adjoining groundwater system are hydraulically interconnected, as expressed in a relatively fast (a few days) groundwater level response to level changes of the Dead Sea [Yechieli et al., 1995]. The drop in groundwater level decreases with increasing distance from the shoreline inland [Yechieli et al., 2009]. Preliminary measurements imply that in some locations near the shoreline, the drop is quite similar to that of the Dead Sea, while reaching a dynamic steady state [Kiro et al., 2008], whereas in other locations the groundwater level drop is significantly smaller than that of the Dead Sea [Yechieli et al., 1995]. The fresh-saline water interface is also responding to the drop of the sea level and the eastward shift of the Dead Sea shoreline, resulting in the fast flushing of the Dead Sea coastal aquifer. Hydrological modeling of the Dead Sea system shows that circulation of saline water continues even during the present fast decrease of the Dead Sea level [Kiro et al., 2008].
 The present study examines the response of groundwater systems to changes in sea levels, using both field data and hydrological simulations.
3.1. Field Studies
 The field studies were conducted in the Dead Sea coastal aquifer, where significant changes are expected to occur within a short time because of the rapid drop of the lake's level. Water levels were measured in several exploration boreholes located at different distances from the shoreline using a manual water level meter.
 The location of the fresh-saline water interface was measured with an in situ device of electrical conductivity (EC) which measured the EC in the depth interval between the water table and the bottom of the boreholes. Usually, the EC values correlate well with the salinity values [Yechieli, 2000]. The situation is, however, more complicated in the Dead Sea area. At very high salinities, higher than half that of the Dead Sea brines, there is no correlation between EC and salinity [Yechieli, 2000]. Therefore, the location of the fresh-saline water interface can only be inferred from measurement of EC values below 160 mmho/cm, according to calibration curves given by Yechieli  which provided the relationship between EC and salinity. The EC profiles were conducted for 6 years in order to examine the response of the interface to change in Dead Sea levels. The conductance of EC profiles was problematic in some of the monitoring boreholes since it requires the insertion of a large device into the 2 inch pipe. Therefore, in many wells, only monitoring of water levels is available.
 Field studies were also conducted in the Mediterranean aquifer, both measurement of water levels and location of the interface. However, because of anthropogenic effects, such as pumping and artificial recharge, the actual effect of seawater rise cannot be determined in the field.
3.2. Numerical Setup of the Mediterranean Aquifer
 The response of groundwater to a consistent rise of Mediterranean Sea level was modeled by FeFlow, a finite element simulator [Diersch and Kolditz, 2002] that solves the coupled variable density groundwater flow and solute transport. Flow boundary conditions were set as follows: no flow on the right, bottom, and left boundaries, and recharge of 200 mm/yr on the top ground surface boundary and prescribed hydraulic head of the sea pressure on the seafloor. Salinity is prescribed on the seafloor as 0.035 g/g and 0 g/g on the top boundary. The numerical mesh consists of 25,907 nodes and 48,653 triangular elements. The horizontal and vertical hydraulic conductivities are 30 and 3 m/d, respectively; the porosity is 0.1; the specific storativity is 10−4 m−1; and the longitudinal and vertical dispersivity coefficients are 10 m and 1 m, respectively. The seawater level rises at a rate of 1 cm/yr for 100 years. Steady state simulations were run for two possible topographic configurations, which probably represent the extent of situations in most coastal aquifers: (1) low angle coastal slope of about 2.5‰ and (2) a vertical coastal cliff of 30 m.
 Additional simulations included the following: (1) two different recharge values, 0.2 m/yr representing average present conditions and a lower rate of 0.1 m/yr; (2) a pumping value of 5 × 106 m3/yr per km width, which was represented as conducted from a single point, located about 3 km from the shore (see explanation in section 4), and (3) subdivision to lower confined and upper phreatic subaquifers by a confining clay layer subdividing the left portion of the aquifer into two subaquifers.
3.3. Numerical Setup of the Dead Sea Aquifer
 The eastern boundary of the model is a specified saltwater head boundary representing the bottom of the Dead Sea. The initial head of this boundary is set to +70 m at the shore, which is assumed to be the thickness of the aquifer at this point. This is a representative thickness of several subaquifers in the alluvial coastal aquifer of the Dead Sea area. The specific density of the seawater is 1.24 with a concentration of 340 g/L total dissolved solids. The western boundary is a specified flow boundary with a total inflow of 3.2 m3/d per meter width of fresh water. This flow rate was used in order to produce a hydraulic head gradient of about 0.005 m/m at the aquifer top, which is similar to that measured in the field. The numerical mesh consists of 6617 nodes and 12,181 triangular elements. A steady state simulation was performed in order to generate the initial flow and concentration distribution at time zero. A transient simulation was then performed in which the Dead Sea level declined at a rate of 1 m/yr for 20 years, a total 20 m decline.
4. Results and Discussion
4.1. Mediterranean Coastal Aquifer
 While the levels of the Mediterranean Sea show a minute rise of ∼10–20 cm in the last 20 years [Rosen, 2004], groundwater monitoring in the vicinity of the shoreline in the last 20 years does not show a measurable consistent change in the water levels. Moreover, no significant change was observed in the location of the fresh-saline water interface that can be related to sea level rise (V. Friedman, personal communication, 2009). Indeed, such effects were not expected because of seasonal variations, coupled with overpumping, which mask the expected small effect of sea level rise. Moreover, the effects of tidal fluctuations on the fresh-saline water interface, which is intensified artificially in monitoring boreholes [Shalev et al., 2009], is an additional process that masks the possible detection of changes.
 Numerical modeling was conducted in order to forecast the expected hydrological response in the next 100 years because of an expected ∼1 m rise in sea level (Figures 4 and 5). In general, the sea level rise results in a similar change in water levels. Since the specific conditions along the Mediterranean coast vary, the simulations were conducted for two different coastal and two different climatic conditions. Simulations show that for the case of a steep topography, simulated as a vertical cliff (Figure 4b), the fresh-saline water interface will not move significantly (by only 1 m). For a milder slope of 2.5‰, the interface will move inland to a distance of 400 m, which coincides with the shift of the shoreline (Figure 4c). Therefore, the actual distance depends on the slope of the topography, whereby a shallower slope would yield a farther inland shift of the shoreline and the location of the interface. In the Israeli coastal aquifer, where the topography near the sea is usually about 1–2%, the interface will move 50–100 m inland.
 The possible climate change, namely, a decrease in precipitation and thus in aquifer recharge, was also simulated. Since the expected change in recharge is not known, the present study examined an extreme reduction of 50% in aquifer recharge. This, by itself, will lower the water table and will significantly move the interface inland to a distance of about 1200 m (Figure 4c). A sea level rise following such conditions will result in a farther inland interface shift.
 Additional simulations were conducted in order to examine the prevailing case in many coastal aquifers, where overpumping exists, causing upconing of seawater. Most of the simulations thus include a pumping well of 5 × 106 m3/yr (for a strip of 1 km width) at a distance of 3 km from the shoreline. This value, which is quite extreme for a normal pumping regime, was taken in order to test the situation of extreme cases. This was done despite the fact that the average pumping in the Israel coastal aquifer is some 2 × 106 m3/yr per kilometer-wide strip, perpendicular to the flow direction. As a matter of fact, more intensive pumping is carried out in places, causing upconing of seawater. Moreover, increase in pumping may occur in the future because of the water shortage of this region. In the case of a sea level rise, while pumping still occurs, some change in the geometry of the interface was found, but the location of the interface toe remained at the same place (Figure 5a). Lower values of pumping yield (e.g., 2–3 × 106 m3/yr) have smaller influence on the location of the interface.
 Simulations were also conducted in order to test the actual configuration of most coastal aquifers which are subdivided to subaquifers (Figure 5b). The simulations in the present study were conducted for the relatively simple case of only two subaquifers, where the upper one is phreatic and the lower one is confined. The configuration of the confined fresh-saline water interface agrees with the analytical solution of Van Dam , which calculated the distance of the interface toe from the shoreline to be L = −αkD2/2q0, where α is the relative density difference (ρs − ρf)/ρf = 0.027; k is the hydraulic conductivity (30 m/d); D is the aquifer thickness (200 m); and q0 is the groundwater discharge, which is calculated by the product of the recharge (0.2 m/yr) and the length of the recharge zone (20000 m) to be 4000 m2/yr (11 m2/d). This calculation yield a value of L = 1473 m, which agrees with our results.
 This is quite similar to the situation of the two upper subaquifers in the Israeli coastal aquifer (see Figure 2). The lower subaquifers, whose connection to the sea is still debatable, were not considered in these simulations. In these simulations, the phreatic aquifer behaves in a manner quite similar to the previous simulation, as expected. On the other hand, in the confined aquifer, only a very small change in the location of the interface is exhibited.
 The effect of pumping is also examined for the two subaquifer configurations where pumping is executed from the lower subaquifer. The simulations show a significant change in this configuration, whereby the interface in the lower subaquifer moves inland (Figure 5c). Similar to the case of one homogeneous aquifer, the sea level rise will shift the interface farther inland, but the interface toe is expected to remain at the same place. The reason for this is that with pumping, the well is the only way fresh water can discharge out of the lower subaquifer. Therefore, the interface between the fresh and saline water must be at the well itself.
 These results show that the expected sea level rise by itself will not change considerably the location of the fresh-saline water interface in the Israeli coastal aquifer because of its relative steep coastal topography. On the other hand, a significant reduction of recharge due to climate change will cause an inland shift of the fresh-saline water interface. These results agree with those of Sherif and Singh  which show that huge flat deltas with shallow topography in the oceans are expected to be more vulnerable to significant seawater intrusion due to sea level rise. Low lying deltaic areas, such as in the Netherlands, are also expected to be affected by sea level rise [Oude Essink, 2001].
4.2. Dead Sea Coastal Aquifer
 The coastal aquifer of the Dead Sea consists, as described before, of clastic material. In general, this material is finer and of lower permeability close to the shoreline and becoming coarser and more permeable with distance from the shoreline. Exceptions are the stream valleys, such as the Arugot stream, that consist of highly permeable alluvial fan sediments even near the shoreline.
 The response of groundwater system to the Dead Sea level drop was examined by both numerical simulations (Figure 6) and field monitoring. The Dead Sea level drop is shown to significantly affect the position of the fresh-saline water interface (Figure 6b), which is consistent with the field observation. The change in groundwater heads (Figure 7) is a function of the permeability of the aquifer and distance from the sea. Boreholes adjacent to the shoreline exhibited water levels drop close to that of the Dead Sea levels (Figure 7) [Kiro et al., 2008]. Along the Arugot and Darga alluvial fans, for example, where the sediments are mostly gravel of high permeability, the groundwater level drop was similar to that of the Dead Sea level up to a distance of 700 m inland (Figure 7a). In the Zeelim boreholes, which are located in the alluvial fan of Wadi Zeelim at much larger distance from the shoreline (∼4 km), the drop in water level is significantly lower than the drop of the Dead Sea levels (Figure 7d).
 Boreholes outside the mainstream valleys have shown, in general, smaller water level drops as compared to that of the Dead Sea (i.e., EG-3a and the Tureibe boreholes; Figures 7b and 7e). This is explained by their relatively lower permeability being located outside the main alluvial fans. It should be noted that the distance from the monitoring boreholes to the shoreline increase with time because of the retreat of the shoreline, also affecting the hydraulic gradient.
 The drop of the Dead Sea level, and the eastward movement of the Dead Sea shoreline, also affects the location of the fresh-saline water interface (Figures 8 and 9). The decline of the Dead Sea is expected to result in a drop in the location of the interface, which is to be exhibited in monitoring boreholes by a lowering in EC profiles with time (Figure 8). The EC profiles in borehole near the Dead Sea, at a distance of ∼70 m, indeed show a significant change in the last 3–5 years (Figure 8a). On the other hand, no significant change in the interface location was observed in Darga borehole, which is located much farther (about 700 m) from the shoreline in the alluvial fan of Wadi Darga (Figure 8b).
 The fast decline of the fresh-saline water interface is resulting in a fast rate of flushing in some parts of the Dead Sea coastal area [Kafri et al., 1997; Kiro et al., 2008]. This is exhibited by the decrease in EC values at given depths in some of the boreholes (Figure 9). This process is probably controlled by the hydraulic properties of the sediments near the Dead Sea, whereby the gravel in the Arugot alluvial fan allows a more rapid flushing, while in other locations a slower flushing is expected.
5. Summary and Conclusion
 The response of two coastal aquifers, namely, the Mediterranean and the Dead Sea coastal aquifers, to sea (base) level variations is preliminarily analyzed here. Both are typical coastal aquifers, but they differ in being subjected to a future slow Mediterranean Sea level rise (1 m/yr) in the case of the first aquifer and to a rapid (1 m/yr), already occurring decline of the Dead Sea in the case of the latter. The Mediterranean aquifer is a good representative of many coastal aquifers near the ocean around the world, while the Dead Sea provides a unique opportunity to study the response of groundwater to sea level changes in the field. The applicability of the unique Dead Sea system to other coastal aquifers is more complicated, but it is worth the effort since it is probably one of the only places in the world where actual changes in groundwater system due to changes in base level can be monitored in the field.
 The fast response of the Dead Sea coastal aquifer to change in the base level implies that the response will also be fast in the Mediterranean coastal aquifer since the hydraulic conductivity in both aquifers is quite similar. However, because of the slow rate of change of the base level, the changes in the groundwater levels and interface locations are not expected to be noticed here. The response of the groundwater in the Mediterranean aquifer is expected to be fast, almost immediate in the present change rate of the sea levels.
 The simulations of the Mediterranean coastal aquifer yield the following results: In the case of a steep coastal topography, simulated as a cliff, the shoreline and thus the fresh-saline water interface are not expected to shift inland. A considerable inland shift of the shoreline, accompanied by an inland shift of the interface, is expected in the case of a low-angle, almost flat coastal topography. Reduced recharge due to climate change or overexploitation of groundwater also enhances the inland shift of the interface.
 In the case of the Dead Sea coastal aquifer, the response to the current rapid Dead Sea level decline is already observed and monitored. The amount of groundwater level drop in relation to that of the Dead Sea is controlled by the aquifer's permeability and the distance from the shoreline. Groundwater level drop is greater and is detected more inland in high-permeability alluvial fan areas as compared to low-permeability areas. Regarding the fresh-saline water interface, its drop is observed by using repeated EC profiling in boreholes. In addition, a continuous freshening of the overlying portion of the aquifer as well as the vadose zone takes place.
 We thank Haim Hemo for the field measurements in the Dead Sea area and Vladimir Friedman from the Hydrological Service of Israel for providing us the data of the Mediterranean coastal aquifer. Ran Gabai and Batsheva Cohen are thanked for their help with the data and the figures. We thank the Hydrological Service of Israel for providing the Dead Sea level and some of the Tureibe levels.