WATER RESOURCES RESEARCH, VOL. 35, NO. 6, PAGES 1877-1894, JUNE 1999

Geochemical and boron, strontium, and oxygen isotopic constraints on the origin of the salinity in groundwater from the Mediterranean coast of Israel

Avner Vengosh,1 Arthur J. Spivack,2 Yohanan Artzi, 3 and Avner Ayalon4

Abstract. In order to identify the origin of the salinityand formation of saline plumesin the centralpart of the Mediterraneancoastal aquifer of Israel, we determinedthe elementaland boron, strontium,and oxygenisotopic compositions of fresh and brackish groundwater(C1 up to 1500 mg/L). We distinguishbetween two key anthropogenic sources:(1) sewageeffluents used for irrigationwith high Na/C1,SO4/C1, and B/C1ratios andlow Br/C1 ratios relative to seawaterratios, low ;5•B values (0-10%o) and high ;5•80 values(>-4%0); and (2) importedwater from the Sea of Galilee that is artificially rechargedto theaquifer with high Br/C1 (3 x 10-3) and;5•80 values (-1%o) anda low 87Sr/86Srratio of 0.70753.The brackish groundwater from the salineplumes have relatively low Na/C1ratios (0.5-0.8) and high Ca/Mg, Mg/C1, and Ca/(SO4 q- HCO3) (> l) ratios relativeto seawaterratios; marine SO4/C1 and Br/C1 ratios; ;5•B values of 24.8-49.9%0; 18 87 86 ;50 of -2.95%0 to -4.73%0; and Sr/ Sr ratios of 0.708275-0.708532.The composition of most of the investigatedgroundwater from the saline plumesdiffers from those of the 87 86 anthropogenicsources, imported water, fresh uncontaminatedgroundwater ( Sr/ Sr of 0.70866,;5•B of 20-30%o),and saline water from the adjacentEocene aquitard. Only in ß 18 areas of artificial rechargedoes local groundwaterhave high Br/C1 and ;50 values that are typicalto the Sea of Galilee. The linear relationshipsbetween chloride and most of theions, including B andSr, the relatively high ;5•B (>30%0) and low ;5•80 (<-4%0) values,and the chemicalsignature of the salineplumes (e.g., marine Br/C1and SO4/C1 ratios),suggest that (1) mixingprocesses control the chemicalcomposition of the brackish water within the aquifer, and (2) the salinepostulated end-member has a chemical compositionthat resembles modified seawater with a marineand higher ;5•B values, and a 87Sr/86Srratio of <0.7083.We proposethat most of thesalinization phenomena and the formationof salineplumes in the inner parts of the coastalaquifer are derivedfrom upconingof underlyingnatural salinewater bodiesand enhancedby overexploitationand draw-downof the overlyingfresh groundwater.

1. Introduction rechargeof salinewater) or nonpointsources (e.g., agriculture return flows,irrigation with sewageeffluent) derived directly Salinizationis one the most widespreadprocesses that de- from anthropogeniccontamination. Salinization can alsooccur gradeswater quality and endangersfuture water exploitation. becauseof "natural processes"such as seawaterintrusion and In manyareas, particularly in arid and semiaridzones, ground- saline-waterflow from adjacentor underlyingaquifers [e.g., water salinizationlimits the supplyof potablefresh water. This Maslia and Prowell,1990]. problemis intensifiedin coastalaquifers where human activi- The Mediterraneancoastal aquifer (Figure 1) is one of the ties (e.g.,water exploitation,agriculture, reuse of wastewater) importantwater resourcesin Israel, supplying20% of the na- result in acceleratingwater-quality deterioration. Monitoring tional water consumption.The salinityof groundwaterin this and identifyingthe origin of the salinityare crucialfor water aquifer has increasedduring the last few decades,and saline managementand remediation.Yet the variety of salinization plumeshave formed in its central region. Severalexplanations sources,particularly in unconfinedaquifers, makes this task to the riseof salinityin the aquiferinclude (1) recyclingof salts difficult.Groundwater salinization can resultfrom either point from irrigationof local groundwaterand/or evapotranspiration sources(e.g., leakageof industrialand domesticwaste water, of meteoricwater; (2) contaminationby irrigationand/or leak- age of wastewater;(3) rechargeand irrigationwith imported waterfrom the Seaof Galilee;(4) flowof salinewater from the •HydrologicalService, Jerusalem, Israel. 2Centerfor MarineSciences, University of NorthCarolina at Wil- adjacentEocene aquifer in the east;and (5) upconingof saline mington. water from underlyingsources [Magaritz et al., 1990;Mercado, 3jacob Blaustein,Institute for Desert Research,The Water Re- 1985; Ronen et al., 1987; Rosenthalet al., 1992; Vengoshand sourceResearch Center, Ben Gurion University of the Negev, Sede Rosenthal,1994; Vengoshand Ben Zvi, 1994]. Boqer, Israel. The purposeof the presentstudy is to establishthe mecha- 4GeologicalSurvey, Jerusalem, Israel. nism(s) for groundwater salinization and to identify the Copyright1999 by the American GeophysicalUnion. sourcesof salinity by utilizing an array of geochemicaland Paper number 1999WR900024. isotopictracers (boron, strontium, and oxygen).Each of these 0043-1397/99/1999WR900024509.00 isotopictracers has previouslybeen used separatelyto delin-

1877 1878 VENGOSH ET AL.: GEOCHEMICAL AND ISOTOPIC CONSTRAINTS

A

•0• Tel•viv/ ,• /' / Jerusalem .• Ashdod/ / '

B

Figure 1. (a) A generalmap of the Mediterraneancoastal aquifer in Israel and (b) the studyarea with the salinespots. Wells locationsare marked by solid circles.Note the locationsof the salinespots and chloride concentrations(in milligramsper liter) of the most salinewells in that area. eate fluid origins and chemical modifications,for example, the aquifer is not confined.In most parts of the aquifer, the oxygenfor tracing the source and nature of the recharged unsaturatedzone is made of calcareoussandstone (kurkar), fluids[e.g., Gat, 1974, 1981],boron for identifyingthe impacts loam (hamra), and heavysoils. In its westernpart the unsat- of anthropogenicsources [e.g., Vengosh et al., 1994],and stron- urated zone consistsof sandstonethat is reflectedby the high tium for evaluatingwater-rock interactions [e.g., Banner et al., qualityof the local groundwater(<50 mg C1/L). In most areas 1989; Chaudhuri et al., 1987; Fritz et al., 1987; Johnson and the aquifer overliesthick and relativelyimpermeable units of DePaolo,1997a, b]. Yet integrationof isotopicdata for a va- the SaqiyeGroup. The SaqiyeGroup (Figure 2) is composed riety of elementsprovides a better diagnostictool for deter- of (1) Oligoceneand early Miocene marls of the Bet Guvrin mining the origin of the salinity. Formation;(2) middleMiocene marlsof the Ziqim Formation; (3) upperMiocene evaporite units (representing the Messinian 2. Hydrogeological Setting event);and (4) Pliocene-agedunits of alternatingthick marine The coastalaquifer lies along the Mediterranean coast of shalesand limestonesandstone layers of the Yafo Formation. Israel (Figure 1) and is composedof Pliocene-Pleistocenecal- Along the easternmargin the aquifer restsupon the Eocene careoussandstone (Figure 2) and layersof clays[Garish and (ShefelaGroup) aquitardin the southerncoast (Figure 2). Friedman, 1969]. Its thicknessvaries from 200 m in the west The annualvolume of groundwaterpumped from the aqui- along the coastlineto a few meters in the eastern margins, fer, startingfrom the 1950s,has been 400 x 106to 450 X 106 20-25 km east from the seashore.In the westernmargins the m3/yr,about one fifth of thetotal water consumption in Israel aquifer is dividedinto severalconfined aquifer systems(suba- [WaterCommissioner, 1995]. Overexploitation beyond the nat- quifers) with different piezometriclevels. In its central and uralreplenishment (---340 x 106m3/year) caused a continuous easternparts there are no continuouspartitioning layers and drop in piezometriclevels between the 1950sand late 1980s. VENGOSHET AL.: GEOCHEMICALAND ISOTOPICCONSTRAINTS 1879

W

CoastalPlain aquifer

dunes Mediterranean

te•rrestrial

marine rocks •lays J Calcareoussandstone •_ _• • (Kurkar• __=--•••?' [•-:•-- m••------••neclaysandshales I Age Group Formatiol• Hydrogeology ______----- YafoFormation Kurkar coastalaquifer • ' SaqiyeGroup QuarternaryIPliestocene 5 10 15 Yafo aquiclude

Km from coast ß Mavqi'im (evaporites) Ziqim

i 12;[Miøcene Bet-Guvrin aquitard

øligøceneMaresha

Adulam

Figure 2. Schematichydrogeological cross section of the coastalaquifer and lithohydrologicalunits.

Asa result, hydrologic depressions formedin the central areas, and are used for domesticuse and irrigation over the coastal changingthe natural east-to-westflow regime in the aquifer. plain.In addition,a largevolume (--•20 x 106m3/year during Since the early 1990s,pumping has been reduced and water the lastfive years) has been recharged into the the aquifer levelshave consequently increased. Nevertheless, groundwater sincethe mid 1960sthrough Lake Azriqa'm (an artificiallake salinization,which began during the 1950s,has continuedde- in an old quarry), northwestof Be'er Toviyya; Lake Merar, spitethe restorationof the water levelsin the last 10 years.In southof Giva't Brener;and rechargewells in the Giva't Brener the Be'er Toviyyaregion (Figure 1) the parallel increasein area. water level and salinity has continued since the late 1960s [Vengoshand Ben Zvi, 1994]. As a result, salineplumes and 3.2. Analytical Techniques spotshave developed in the innerparts of the aquiferand in its Elementalanalyses were carried out in the analyticai labo- easternmargins (Figure 1). The locationsof mostof the saline ratory .of the HydrologicalService in Jerusalem.Boron con- plumesare associatedwith the hydrologicdepressions in the centrationswere determinedwith a modified spectrophoto- central part of the aquifer. Severalsaline spotshave existed metrictechnique using the reagentAzomethine H [Kiss,1988]. sincethe 1930s[Artzi, 1999.; Vengosh et al., 1996]. Bromideconcentrations were-determined by flow injectionion analyzer(QuickChem 8000• [Vengosh and Pankaratov, 1998]. 3. Methods Boronisotopic compositions in sampleswere determinedby negativethermal ionization mass spectrometry (NTIMS) [Ven- 3.1. Sample Description goshet al., 1989;Eisenhut et al., 1996]. Boron was separated The selectedresearch area representssaline plumes from from natural sampleswith a boron-selectiveresin, Amberlite the centraland easternparts of the aquifer.From 1992to 1995 IRA-743[Kiss, 1988], eluted wi•h 1 N HC1,mixed with a solu- we collectedgroundwater from pumpingwells in areasof high tion of MgC12and Ba(OH)2, loadedonto Re singlefilaments, salinity (Figure 1). We also collectedfresh uncontaminated and analyzedwith a reversepolarity, "dynamic"collector, the groundwaters(<100 mgC1/L) from the western part of the Finnigan MAT-261 massspectrometer, at the University of aquifer. Regensburg,Germany. Details of the analyticaland massspec- In orderto characterize thegeochemical composition ofthe trometryprocedures are givenby Eisenhutet al. [1996].Some anthropogenicsources that may affect the salinizationpro- sampleswere analyzedwith a VG-336 massspectrometer at cesses,we investigatedimported water from the National Wa- the Universityof North Carolina, Wilmington, in which a B- ter Carrier, originatingfrom the Sea of Galilee, as well as free seawater matrix was added to enhance ionization. The treated sewageeffluents. The treated sewageeffluents, used mode of filament loadingand massspectrometry procedures for irrigationover the aquifer in the central coast,were sam- were strictlyrepeated in samplesand standardswere used in pled from open reservoirs.The importedwaters of the Sea of order to minimize the variability of mass spectrometer- Galilee are transportedthrough the National Water Carrier inducedisotopic discrimination. A standarddeviation of up to 1880 VENGOSH ET AL.: GEOCHEMICAL AND ISOTOPIC CONSTRAINTS

2%o was determinedby repeat analysesof National Instituteof Standardsand Technology (NIST) SRM-951 standard.Isotope ratiosare reported as per mil deviation(•B) in the •B/•øB ratios relative to the measured standard, NBS SRM 951: /5•B= [(•B/•øB)samp•e/(•B/•øB)NBs951-- ]-] X ]_000 The mean of the absolute•B/•øB ratios of NIST SRM-951 replicatesanalyzed with the sampleswas 3.9935 ñ 0.008 (at Regensburg)and 4.015 ñ 0.005 (at Wilmington).Duplicate massspectrometry analyses were carried out for someof the sampleswith externalprecision below 2%o. Moreover,several sampleswere cross-checkedby analyzingthem at Regensburg and Wilmington.Their ratioswere identical,within the 2%o precision. Strontiumwas separatedby cation exchangechromatogra- phy usingstandard techniques. Isotope ratios were determined usinga two-detectordynamic collection routine on a VG-3361 massspectrometer at the Universityof North Carolina,Wil- mington.We usedtungsten filaments. Prior to collectionof Sr ratios,85Rb was monitored by Daly detection to insurethat the 87Rbisobaric interference was negligible. All measured87Sr/ 86Srresults were corrected to an 86Sr/88Srratio of 0.1194using exponentialcorrection. Through this procedure, seawater 87Sr/ 86Sryielded a ratioof 0.70923.Fractionation corrected ratios were normalizedto the measuredseawater ratio assumingthat the modern seawater ratio is 0.709199. Oxygen isotope ratio measurementswere made on a VG SIRA-II massspectrometer at the GeologicalSurvey of Israel. Resultsare givenin per mil valueswith respectto the standard meanocean water (SMOW) standard(•80) [Craig,1961]. The •80 valueof thewater was determined after equilibration with CO2, by shakingwith 2 mL of water for 4-6 hoursat 25øC [Epsteinand Mayeda, 1953]. Analyticalreproducibility of du- plicatesmeasured on different dayswas better than 0.1%o.

4. Results

4.1. Anthropogenic Sources Treated wastewaters are characterizedby a salinityrange of 300 to 550 mg C1/L, and a wide range of ion ratios (Table 1). High Na/C1,SO4/C1, B/C1, K/C1 and low Br/C1ratios relative to the marine ratios are typical of waste water from the Dan Region [Vengoshand Keren, 1996] (Figure 3). Waste water fromopen reservoirs has /5•80 values of -3.45%• to 0.92%• (Table 1). Importedwater from the Sea of Galilee has a sa- linityof 220rng C1/L, a highBr/C1 ratio (3 x 10-3), a /5•80 valueof -1%•, a/5•B of 22.8%•, and a low 87Sr/86Srratio of 0.70753.

4.2. Brackish Groundwater The brackishgroundwater in the investigatedarea (Table 2) is depletedin Na+ andK +, andenriched in Mg2+ andCa 2+ relativeto diluted seawater with a similarsalinity (Figure 4). In addition, we found small but consistentdifferences between the different saline spots:(1) in groundwaterfrom Yavne, Hazor, and Be'er Toviyyathere is a linear correlationbetween C1- and Ca2•-, whereas in Giva't Brenerthe Ca2+ variationis scattered;(2) in groundwaterfrom Yavneand Be'er Toviyya the SO42- contentis depleted,whereas in Giva'tBrener it is enrichedrelative to the seawaterSO4/C1 ratio; (3) in ground- water from Yavne the Na/C1 and SO4/C1ratios decrease,and Ca/(HCO3 + SO4) increaseswith salinity,while in Giva't VENGOSH ET AL.: GEOCHEMICAL AND ISOTOPIC CONSTRAINTS 1881

Chemical compositionof sewageeffluent and Sea of Galilee

1.2 [] Dan Region ß I ß [] DanRegion

1.0

o [] [][]

sea water ratio

0.8 sea water ratio Sea of Galilee [] []

0.7 Chloride(mg/l) , Chlor,ide(mg/l) 200 300 400 500 600 200 300 400 500 600 0.005 0.010

[] 0.00,• 0.008 [] Dan Region Sea of Galilee

.002 'ß ! 0.006

0.00: [] [] sea water ratio 0.004 [] [] mm mm m mm mm mm mm mm m mm • []

0.001 DanRegio•n [] [][] 0.002 Sea of Galilee [] mmm m•m• sea water ratio 0.00œ [] Chloride (mg/l) 0.000 , Chloride(lng/l) . 200 300 400 500 600 200 300 400 500 600 0.12 0.004 ßß D•anRegion 0.1C [] [] [] Dan Region ß ß 0.003 o.o• Sea of Galilee [] 0.0• seawater ratio 0.002

ø-ø4'Sea of Galilee 0.001

0.02 sea water ratio []

0.0C , , 0.000 ! , 200 300 400 500 600 200 300 400 500 600 Chloride (mg/l) Chloride (mg/l)

I wastewater from the Dan Region [] wastewater from openreservoirs ß Sea of Galilee

Figure3. Differentions to chloride(molar) ratios versus chloride concentration (in milligramsper liter) of sewageeffluents from the Dan Region Reclamation Plant and open reservoirs in the coastal plain of Israel.Note the high Na/C1,K/CI, SO4/C1,B/CI, F/C1 and the low Br/CI ratios of wastewaterrelative to the marine ratios.

Brener theseratios do not showany trends;(4) mostof the Gedera Gan Mordechai,and Revadim)were measuredfor brackishwater has a marineBr/CI ratio,while some samples boronisotopes both in 1992and 1995(Table 3). The from Giva't Brenerand Hazor havehigh Br/C1ratios. resultsshow a range of 0.3%0 to 10%o difference,which is The /5•B valuesof thebrackish groundwater range from attributedto the naturalvariability of the boronisotopic ratio. 24,8%0to 49.9%0,and 87Sr/86Sr ratios range from 0.708275 to 0.708532(Table 3). Freshwater from the westernpart of the aquifer(CI < 100mg/L) has lower •i•iB values (21.2-32.4%o 5. Discussion [Vengoshet al., 1994])and higher 878r/868¾ ratios (0.708609- The continuedrise of salinitywith time (Figure 5) reflectsan 0.708663;Table 3). increasingfraction of a salinesource that is degradingthe It shouldbe noted that most of theisotopic results are reported groundwaterquality. Assumingthat the saline sourcehas a for samplescollected at differenttimes. For example,several high salt contentrelative to originallocal groundwater, which wells(Hazor A, Yavne3, G. BrenerLevinson, Gedera Moa'za, hasa chlorideconcentration of lessthan 100 mg/L,based on 1882 VENGOSH ET AL.' GEOCHEMICAL AND ISOTOPIC CONSTRAINTS

oo o.o. o. oo o. o. o. o. o o o o oo. o.o. . . .o. o.o.o.o.o.o.o. oo

..... oo. o. o...... o.o. ß ß o. ß ß . o.

o

oo• . • • • ...... ,

o

. • . o o o . • ß . • • • ...... •• ß ß?•. ..I I • --I ...... ••.•••I I I I I I I I o o

o

.,• VENGOSH ET AL.: GEOCHEMICAL AND ISOTOPIC CONSTRAINTS 1883

o 0 0 0 0 0 0 0 0 0 •'n-• .•oooooooo•••

o

o 1884 VENGOSH ET AL.: GEOCHEMICAL AND ISOTOPIC CONSTRAINTS

0 10 20 30 400 10 20 30

0.07

0.06

0.05

0.04

0.03

0.02

0.01 CI 0.00

ß o

Y o

o ß o / o / o

CI CI

! i 10 20 30 400 1• 2;

Saline plumes: ß Yavne ß Be'erToviyya o Giva't Brener [] Gedera a Revadim [] Hazor

Figure4. Chlorideversus dissolved ions concentrations (mmole/L) in groundwaterin the studyarea as comparedto seawaterratio (solidline). Note the differencesamong the salinespots. The dashedline Y representchemical variations observed during sampling for several hours of well Gan Darom (see in Table2). Notealso the linear relationships between Ca andC1 in groundwaterfrom Be'er Toviyya (BT) andHazor (H).

historicaldata, a small fraction of salinewater would dominate 5.1. The Chemical and Isotopic Characteristicsof the the chemicaland isotopiccompositions of the contaminated Potential Saline Sources groundwaterin the salineplumes. The followingsections char- acterizethe differenttypes of the potentialsaline sources and Knownpotential saline sources in the Mediterraneancoastal comparethem to thecomposition of thebrackish groundwater. aquifersof Israel includelateral salinewater flowsfrom the VENGOSH ET AL.: GEOCHEMICAL AND ISOTOPIC CONSTRAINTS 1885

Table 3. Boronand Strontium Isotopic Compositions of InvestigatedGroundwater From the CentralCoastal Aquifer of Israel

B/C1 Source ID Date 8•B 87Sr/86Sr SE Sr C1 Sr/C1 Sr/Ca B (x 1000)

Mediterranean sea water 39.0 0.709200 .-. 9.59 22000 0.18 0.00038 5.3 0.8

Be'er ToviyyaSaline Plume Be'er Toviyya5 12612402 July9, 1995 -.. 0.708338 3.91E-05 1.70 763 0.90 0.0031 ...... Be'er Toviyya5 12612402 July 13, 1992 46.2 ...... 404 ...... 0.18 1.48 Be'er Toviyya7 12612302 June 13, 1995 ... 0.708275 3.80E-05 1.40 587 1.92 0.0028 ...... Be'er Toviyya7 12612302 July 1, 1993 40.3 ...... 544 ...... 0.24 1.45 Kefar Varburg D 12512302 July 6, 1995 "' 0.708384 1.70E-05 1.50 644 1.88 0.0029 ...... Kefar Varburg D 12512302 July 1, 1993 41.4 ...... 257 ...... 0.14 1.8 Kefar VarburgA 12512401 July 12, 1992 45.6 ...... 457 ...... 0.26 1.88 Be'er Toviyya3 12612403 July 13, 1992 47.6 ...... 715 ...... 0.26 1.17 Be'er Toviyya6 .-- July 1, 1993 49.9 ...... 711 ...... 0.32 1.48

Hazor Hazor A Kibutz 13212302 Aug. 1, 1995 35.3 0.708495 9.00E-05 1.70 489 2.80 0.0032 0.29 1.94' Hazor A Kibutz 13212302 July 19, 1992 31.7 ...... 508 ...... 0.16 1.03 Hazor kibutzB 13112301 July20, 1992 32.6 ...... 303 ...... 0.20 2.13 Hazav 1 13312602 Aug. 16, 1992 34.9 ...... 333 ...... 0.20 2.01 Ashdod5 12711701 July 16, 1992 31.9 ...... 410 ...... 0.10 0.76 Ashdod6 12811902 July 16, 1992 33.7 ...... 325 ...... 0.33 3.33 Ashdod2 12911801 July 16, 1992 37.5 ...... 345 ...... 0.39 3.71 Ashdod10 13012001 July 16, 1992 36.6 ...... 333 ...... 0.42 4.14 Gan Hadarom 1 13412102 Aug. 31, 1995 ... 0.708473 1.40E-05 2.20 520 3.41 0.0045 ...... Gan Hadarom 1 13412102 June 6, 1993 38.0 ...... 705 ...... 0.19 0.88 Gan Hadarom 1 13412102 June 6, 1993 30.6 ...... 576 ...... 0.18 1.02 Gan Hadarom 1 13412102 June 6, 1993 32.1 ...... 640 ...... 0.19 0.97

Giva't Brener G. Brenner Levinson 14013104 Oct. 18, 1995 37.2 0.708532 3.41E-05 1.60 414 3.11 0.0023 0.18 1.43' G. BrennerLevinson 14013104 July 17, 1995 28.8 ...... 412 ...... 0.19 1.52 G. BrennerSiman Tov 14013102 July 17, 1995 38.3 ...... 343 ...... 0.19 1.77 G. BrennerBerkovitz 14213001 July 17, 1995 48.4 ...... 405 ...... 0.12 0.96 Giva't BrennerA 14113103 July 17, 1995 29.6 ...... 328 ...... 0.13 1.25 G. BrennerMeshtefet 14113104 July 17, 1995 24.8 ...... 218 ...... 0.14 2.05 KevuzatShiler C 14213002 July 17, 1995 48.5 ...... 251 ...... 0.10 1.35 KevuzatShiler B 14112902 July 17, 1995 44.9 ...... 258 ...... 0.13 1.67

Gedera GederaMoa'za 13512901 Aug. 31, 1995 39.9 0.708425 2.90E-05 1.50 479 2.52 0.0026 0.20 1.37' GederaMoa'za 13512901 May 24, 1993 35.1 ...... 463 ...... 0.25 1.77 GederaMoa'za 13512901 May 23, 1993 35.1 ...... 471 ...... 0.24 1.67 Gedera Gan Mordechai 13512902 July 6, 1994 34.4 0.708393 3.40E-05 2.50 772 2.61 0.0031 0.32 1.36' Gedera Gan Mordechai 13512902 May 23, 1993 43.2 ...... 731 ...... 0.28 1.26

Yavne Yavne 3 13812701 Aug. 30, 1995 .-- 0.708345 1.60E-05 1.30 407 2.57 0.002 ...... Yavne 3 13812701 July 1, 1992 34.6 ...... 363 ...... 0.17 1.54 Yavne 3 13812701 Aug. 1, 1995 43.6 ...... 407 ...... 0.16 1.29' Yavne 2 13812401 July 1, 1992 32.9 ...... 405 ...... 0.11 0.89 Yavne C 13512203 July 1, 1992 35.6 ...... 221 ...... 0.18 2.64 Yavne A 13512301 July 1, 1992 32.9 ...... 227 ...... 0.21 2.96 Yavne 13 13712401 July 1, 1992 30.2 ...... 135 ...... 0.12 2.94 Yavne 10 13712601 July 1, 1992 34.4 ...... 221 ...... 0.20 2.94

Revadim Revadimkibuutz 13113201 Aug. 1, 1995 38.5 0.708473 6.70E-05 2.80 1352 1.67 0.0039 0.62 1.50' Revadimkibuutz 13113201 July 15, 1994 38.2 ...... 1409 ...... 0.67 1.56

Freshwater Ashdod19 13211504 July 3, 1995 '" 0.708663 3.40E-05 0.40 29 11.11 0.0035 ...... Ashdod 18 13211503 June 3, 1995 "- 0.708609 3.80E-05 0.50 47 8.57 0.0027 ...... Ashdod18 13211503 July 16, 1992 29.8 ...... 45 ...... 0.13 9.18 Ashdod 14 1311502 June 13, 1995 ... 0.708660 1.60E-05 0.40 32 10.07 0.0031 ...... Ashdod17 13211502 July 16, 1992 30.5 ...... 0.01 48 ...... 0.10 6.83 Imported Water (Sea of Galileo Azriqam reservoir* Oct. 18, 1995 22.8 0.707528 1.40E-05 0.60 223 2.17 0.0028 0.10 1.47 Azriqa'm well 12812201 Oct. 18, 1995 ... 0.708350 1.95E-05 0.85 207 3.31 0.0033 ...... Giv'at Brenner6 14012701 Aug. 30, 1995 ... 0.708149 3.19E-05 0.80 221 2.92 0.0021 ...... Isotopicresults are reportedin 811Bvalues as normalized to the standardNBS SRM-951. *Analysesby MS VG-336 at Wilmington,North Carolina. 1886 VENGOSHET AL.: GEOCHEMICALAND ISOTOPICCONSTRAINTS

16OO

• Revadim ß Be'erToviyya 1200

'800 Israel limit

400

0 5o 55 60 65 70 75 80 85 90 95 year

6oo

[] Hazor Israel limit 500 • Yavne o Giva't Brener 400 = Gedera

300

200 EPA/WHO Limit

100

i ß ! ß i ß ! ß ! ß ! ß i ß i ß i 50 55 60 65 70 75 80 85 90 95 year Figure5. Variationof chlorideconcentrations withtime of selectedwells from the saline plumes in the centerof theaquifer. Note the comparisons to drinking-water limit regulations of chloride concentrations. adjacenteastern Eocence aquitard, anthropogenic salt flux SO4/C1ratios. Artzi [1999]shows that a salinespring near fromthe surface via irrigation with wastewater over the aqui- Hulda,located east of the researcharea, along the eastern fer, and artificialrecharge and irrigationof importedwater marginof the coastalaquifer, has geochemical characteristics fromthe Seaof Galilee.In addition,recycling of freshwater thatare typical of typeII. Theonly available geochemical data throughpumping and irrigation may also result in long-term of salinewater of typeI indicatehigh B/C1 ratios, 87Sr/86Sr accumulationof saltsin a phreaticaquifer. The chemicaland ratios of 0.708102-0.708132,and 8•B valuesof 38-48.5%0 isotopiccompositions of thesesources are summarizedin Ta- (A. Vengoshand A. Starinsky,unpublished data, 1995). ble 4 and Figure 6. Uncontaminatedfresh water (C1 < 100 mg/L) from the Salinewater from the Eocence aquitard has a widerange of westernpart of the aquiferhas high Na/C1 (>1) and Br/C1 salinity,up to 5000 mg C1/L [Rosenthalet al., 1992;Nissim, (>1.5X 10-3) values[Vengosh and Pankaratov, 1998], 8280 1991;Vengosh and Rosenthal,1994; Y. Livshitz,The influence valuesof about-5%0 [Gat, 1981],low 15•Bvalues of 21.2- of naturaland artificial factors on the chemical composition of 32.4%0,[Vengosh et al., 1994],and a 87sr/a6srratio of 0.70866. groundwaterin the north-western Negev and southern part of Domesticwastewaters from the Dan RegionReclamation the coastalplain, Ph.D. thesis in preparation,1999]. Recent Projectand some wastewaters (i.e., with a chlorinityof---300 studies[Artzi, 1999; Y. Livshitz,Ph.D. thesis in preparation, mg/L) from open reservoirsin the centralcoast of Israel are 1999]indicate two major types of salinegroundwater: type I, characterizedby a uniqueanthropogenic chemical signature. shallowgroundwater associated with fluvial sedimentscharac- HighNa/C1 (1.1), SO4/C1, and B/C1 (5 x 10-3) aswell as low terizedby high Na/C1 (1.2), SO4/C1, and B/C1 ratios; and type Br/C1ratios (5 x 10-4) areattributed to applicationsof NaC1 II, deepsaline groundwater characterized by marine Na/C1 and saltand boron-enriched detergents [Vengosh et al., 1994;Ven- VENGOSH ET AL.: GEOCHEMICAL AND ISOTOPIC CONSTRAINTS 1887

Table 4. The Chemicaland IsotopicCompositions of PotentialSaline Sources and Major SalinePlumes From the Coastal Aquifer Source C1,mg/L Na/C1 SO4/C1 Br/C1,x103 B/C1,x103 a180,%0 allB, %0 87Sr/86Sr

Seawater* 440 0.86 0.05 1.5 0.8 -5 39 0.70923 Wastewater 300 1.1 0.09 0.5 2 to 5 -3 to 1 0 to 10 .-. Sea of Galilee 220 0.78 0.06 3.0 1 -1 23 0.70753 Eocene saline water Nahal Oz 2 1300 1.23 0.12 1.4 8.7 .-. 43 0.70813 Hulda spring 1300 0.86 0.05 1.5 1.8 ...... Saline Plumes Revadim 1350 0.66 0.05 1.5 1.5 -4.2 38 0.708473 Gedera 772 0.54 0.05 1.6 1.8 - 4.0 43 0.708393 Be'er Toviyya 763 0.78 0.05 1.4 1.8 -4.5 46 0.708338 Hazor 516 0.76 0.05 1.5 1.7 - 4.0 35 0.708495

*Fifty times diluted.

goshand Pankaratov,1998]. The variationsof the different ion to about 20-30%0, since the original sewagesignature is ratioswith chlorinity(Figure 3) suggestthat the fluids in the 0-10%o[Vengosh etal., 1994].In contrast,sea water (/5•B = open reservoirsare a mixture of domesticsewage effluents 39%0)would be modifiedto a highervalue, up to 60%0,which with a salinityof about 300 mg/L (i.e., typical of the sewage is significantlydifferent from that of sewage-contaminated from the Dan RegionReclamation Project) and localbrackish groundwater.The influenceof the aquifermatrix on the 8?Sr/ groundwaterswith higherchlorinity and lower Na/C1, SO4/C1, 868r ratios of dissolvedSr 2 + is more difficult to determine since and B/C1 ratios that were added to the reservoirs.Although Sr2+ canbe derivedfrom (1) dissolutionof Pleistocenecalcite this addition reduces these ratios, the anthropogenicgeo- matrixwith a 87Sr/86Srratio of >0.7090and (2) ion-exchange chemicalsignature is still distinguishable(at least for several with clayminerals [Johnson and DePaolo,1997a, b]. parameterslike SO4/C1,Br/C1, B/C1 ratios) from the marine The effect of water-rockinteractions on the Sr isotopicsys- ratios(Figure 3). The effluentsare also enriched in •80 (rel- tematicsis recordedin two groundwatersamples that repre- ative to local groundwater)because of the evaporationin the sentthe spreadingof rechargedwater in the aquifer,originat- open reservoirs.Vengosh et al., [1994]have shownthat sewage ingin the Seaof Galilee.The C1content, Br/C1 ratio, and/5•80 effluentsare characterizedby /5•B valuesin the rangeof values of these wells (Azriqa'm Sochnotnear the recharge 0-10%o. Consequently,the assemblageof high Na/C1,SO4/C1, basinof Azriqa'm and Giva't Brener 6; Table 2) are almost andB/C1 ratios; low Br/C1ratios; low /5•B values;and high identicalto that of the Sea of Galilee, which reflectsa long- /5•80values are typicalof domesticwastewater in Israel,in term anddirect recharge of the importedwater into the aquifer particular wastewaterthat is being used for irrigation in the (i.e., minimumdilution or mixingwith local groundwater).In central part of the coastalaquifer. contrast,the 87sr/a6srratios of the twowells (0.708350 and Imported waters from the Sea of Galilee are characterized 0.708149,respectively) are significantlyhigher than that of the byhigh Br/C1 (3 x 10-3) and/5•80values (-1%o), whichare rechargewater (0.707528).The Sr contentin thesesamples is related to the origin of brinesflowing to the Sea of Galilee also higher, suggestinga considerableaddition of matrix- [Stalinsky,1974], and evaporationprocesses over the lake [Gat, derivedSr with a higher87sr/a6Sr ratio. Consequently, it seems 1974, 1981],respectively. The high Br/C1ratio differsfrom the that dissolutionof the Pleistocenematrix with a 87Sr/86Srratio typical low Br/C1 characteristicof domesticwastewater. The of >0.7090 can modifythe originalSr isotopiccomposition of 87sr/a6Srratio of the Sea of Galilee is 0.707520. externalfluids in the aquifer.

5.2. Water-Rock Interactions 5.3. Elemental Systematics There are significantchemical and isotopicdifferences be- The chemicalvariations reflect two major patterns: (1) linear tween the potential saline sources(Table 4). Nevertheless, relationshipsbetween C1- andNa+, Mg2+, Ca2+, SO42-,and these geochemicalcharacteristics may be considerablymodi- Br- in brackishgroundwater from the saline areas of Be'er fied because of water-rock interactions. Mechanisms that can Toviyya,Yavne, Gedera, and Hazor and (2) nonlinearrela- modifythe originalchemical and isotopic compositions are (1) tionshipswhich are typicalof the Giva't Brener area (Figure base-exchangereactions with clay mineralsthat affect Na +, 4). While the linear relationshipsbetween C1- and the other Ca2+, and Sr 2+ andhence the Na/C1, Ca/(SO4 + HCO3),and ions indicate mixing processeswith a distinguishablesaline 87Sr/86Srratios; (2) adsorptiononto clay minerals, which af- end-member,the scattervariations suggest water-rock modifi- fects B, K +, and /•B; and (3) carbonatedissolution- cations.The chemicalcomposition of the brackishgroundwa- precipitation,which affects Ca 2+, Sr2+, andHCO3- ionsand ter primarily reflects the composition of the saline end- 87Sr/86Srratios. Consequently, while/•80, Cl-, andBr/C1 are member,particularly for the ion ratioswhich are lesssensitive consideredto be conservativetracers in the aquifersystem, the to dilutionwith low-TDS (total dissolvedsolids) fresh water. other geochemicaland isotopicparameters used in this study The variationsof Na/C1,SO4/C1, Br/C1, and Ca/(HCO3 + SO4) (/•B, 87Sr/86Srratios) may be influencedby water-rock inter- ratios suggesta saline sourcewith a marine Br/C1 ratio, low actions. Na/C1 (-0.5) and K/C1, and high Ca/Mg, Mg/C1, and Ca/ The effectof boron adsorptionon the isotopicvariations is (HCO 3 q- SO4) ratiosrelative to seawater. limited to about 20%0 [Spivacket al., 1987], and this would In brackish water from Revadim area, on the eastern side of increasethe/•B valuesof sewage-contaminatedgroundwater the aquifer,and in the Be'er Toviyyasaline plume, Na/C1 ratios 1888 VENGOSH ET AL.: GEOCHEMICAL AND ISOTOPIC CONSTRAINTS

Na/C1ratio salineplume,s 3• Seaof Galilee.•© ß I?.(•?.<(t [] .... , .... , .... , .... ! ..... 0.4 0.6 0.8 1.0 1.2 1.4

SO4 /C1 ratio Sources: salineplumes [• Eocence- Type I • Eocence-Type II .seawater 0.05 0.15 [• wastewater salineplumes Br/C1ratio (xOOO)

I;';'i';';1 .... []•!• • Seaof.o,Galilee 0.0 1.0 2.0 3.0

salineplumes B/C1 ratio (xOOO)

c/ e_•c& ...... ••, . t,%,%,:,:,:,! 0 2 4 ; ' • •o salineplumes b•80 (%,)

10. fold diluted• seawater l?;,;,i,;,>;.,;,;,;,.;,;,• .:.•O,••&';ß -5 -4 -3 -2 -1 0

&O&.•x.* {5"B,saline(%,)plumes I •q> seawater

0 10 20 30 40 50 87Sr/86Srratios I saline•plumesI Sea of Galilee seawater I .© 0.7074 0.7078 0.7082 0.7086 0.7090 0.7094 Figure6. Valuesof someof the geochemicaland isotopic parameters of the potentialsaline sources as comparedwith those of themajor saline plumes. Tenfold diluted seawater refers to mixingbetween seawater and freshwater with/5•80 • -5%,.

arehigher (0.66 and 0.80, respectively), but this water also has ratiosin the brackishgroundwater are differentfrom thosein a Ca-chloridecomposition (i.e., Ca/(HCO3+ SO4)>1). The freshwater (high Br/C1), the Seaof Galilee(high Br/C1), and Cl-richwater at Yavne has SO4/C1ratios that are lower than wastewater(low Br/C1, high SO4/C1), we arguethat the chem- thatof seawater (<0.05), whilein the otherareas the Cl-rich ical datarule out (1) recyclingof saltsfrom the irrigationof water has a marine SO4/C1ratio. localgroundwater and/or evapotranspiration of meteoric water The chemicalcomposition of the brackishwater differs from (i.e.,a long-termsalinization from recycling of freshwater), (2) thoseof uncontaminatedfresh water (i.e., low chlorinityof contaminationby irrigationand/or leakage of wastewater, and <100 mg/L,which has high Na/C1 and Br/C!), domestic sewage (3) rechargeand irrigation of waterfrom the Seaof Galilee. effluentsused for irrigation(high Na/C1, SO4/C1, B/C1, and low The chemicaldata cannot,however, rule out the input of saline Br/C1),imported water of theSea of Galilee(high Br/C1), and waterfrom the Eocenceaquitard of typeII with marineNa/C1, bothtypes of salinewater from the Eocence aquitard (Table 4, SO4/C1,and Br/C1 ratios (e.g., Hulda spring;Table 4). Never- Figure6). Whilethe low Na/C1 and high Ca-chloride signals of theless,the abovementioned linear correlationsbetween C1 the brackishgroundwater can be changedby base-exchangeand other ions suggestthat mixingprocess control the water reactions,the SO4/C1and Br/C1are not influencedby these chemistrysystematics which is not consistentwith base- processes:and are considered conservative tracers. Since these exchangereactions. VENGOSH ET AL.: GEOCHEMICAL AND ISOTOPIC CONSTRAINTS 1889

The Ca-chloride,low-Na/C1 ratio compositionof the brack- Brackishground water ish groundwaterresembles the compositionof modified sea water and hence suggeststhat the saline plumes are derived (this study) from natural salinewaters that originatedfrom seawater and becamedepleted in Na+ andenriched in Ca2+ andMg 2+. Two mechanismsare known to explain these modifications.The first is evaporationof sea water followed by halite precipita- tion, combined with dolomitization. Since sea water has a Na/C1ratio lessthan unity, halite precipitationfurther reduces this ratio in the residualbrine. Accordingly,the Na/C1ratio of 0.66 suggestsan ---20-fold evaporatedsea water (data from 0' McCafieryet al. [1987]). Conversly,the Br/C1 ratio increases -6 -5 -4 -3 -2 -1 0 1 during precipitation of halite from sea water, and thus one wouldexpect corresponding high Br/C1 ratios of > 1.5 x 10-3 [Carpenter,1978; Statinsky, 1974]. The Br/C1ratios of most of freshground water the investigatedbrackish water, however,are marine (Figure (Gat and Dansgaard,1972) 6). Nevertheless,hypersaline brines with a similar chemistry (e.g., low Na/C1, high Ca, relatively low Br/C1 ratios) were found in sabkhaenvironments in the Bardawil lagoon of the Northern Sinai [Levy, 1977]. In addition, salinewater with a N 6- chloride content of up to 30,000 mg/L and similar chemistry was discoveredalong the western part of the coastalaquifer [Vengoshet al., 1991]. Alternatively,the relativedepletion of Na + andenrichments of Ca2+ andMg 2+ maybe the resultof base-exchangereac- 0 1 tions of the original seawater with clay minerals.Accordingly, 8180 (%,SMOW) the water/reactiveclay mineral ratio must be low in order to have a net affect on the residual brine. In either case, it is clear that the compositionof most of the brackish groundwater Seaof •alilee differs completely from those of the anthropogenicsources ß Yavne and thus eliminatesthem as the major sourcesof salinity. waste water An anthropogeniccontribution is apparent in severalwells ß Be'er Toviyya in the investigatedarea. In somewells (e.g., Hafetz Hayim C o Giva't Brener andHafetz Hayim Kibbutz)chemical composition reflects con- [] Gedera taminationfrom sewageeffluents, while in other wells (e.g., 6 Revadim [] Hazor Giva't Brener A, 5, and 6, and Azriqam well; Table 2), which are located near the recharge sites,the impact of the Sea of ß o Galilee is recognizable.High Na/C1, SO4/C1,B/C1, and low Br/C1 ratios provide evidence for sewage contamination 00e•0 [] whereasthe impact of the imported Sea of Galilee is mainly -5 , identifiedby high Br/C1 ratios and 8180 values. ß i ß , ß , ß , ß i ß i ß , 0 200 400 600 800 1000 1200 1400 5.4. Oxygen Isotope Systematics Chloride (mg/1) Regionaluncontaminated groundwater has 8180 values be- Figure7. Histogramsof 8180 valuesand chlorideversus tween -5.5%0 and -4.0%0, with a mean of -4.7%0 (Figure 8•80 of groundwaterand sewageeffluents from openreser- 7). The 8180values of groundwaterssampled in the present voirsin the coastalplain of Israel.The 8•80 valuesof fresh studyvary between -4.8%0 to -2.4%0 (one sampleyielded a uncontaminatedgroundwater, sampled during early 1970s,are valueof -0.4%0; Figure7). Most of the brackishwater (22 out takenfrom Gat andDansgaard [1972]; 8180 values below the of 33 samples),particularly those sampleswith high salinity, dash line (<-4%0) in the lower figure are in the range of has8180 values of <-4%0. Thisisotopic range generally over- uncontaminatedgroundwater and thus representgroundwater laps with those of uncontaminatedgroundwater reported by samplesthat are not influencedby anthropogenic(wastewater Gat [1974, 1981].The other wellswith high 8180 values and Sea of Galilee) fluids. (>-4%0; Figure 7) are locatedspecifically in Giva't Brener and the Azriqa'm area and have alsohigh Br/C1ratios (> 1.5 x 10-3),reflecting local recharge from the Seaof Galilee. solvedions would increasesignificantly as a result of ---10% The 8180 values of the brackish water differ from those of mixingsea water (i.e., C1 --- 2000 mg/L),while 8180 would sewageeffluents in openreservoirs (8180 = -3.5-1%o) and increaseby only ---0.5%0. the Sea of Galilee (-1%o). Consequently,the oxygenisotope data confirmthat theseanthropogenic fluids are not the saline 5.5. Boron Isotope Systematics sourcefor the underlyinggroundwater. Mixing with seawater In the coastal aquifer systemthere are several sourcesof or a brine derived from sea water, as inferred from the chem- boron that can affect the isotopiccomposition of the investi- icaldata, would not significantlymodify the originallow 8•80 gated brackishgroundwater [Vengosh et al., 1994]: (1) fresh value of the fresh groundwater.For example, the total dis- groundwaterwith 8XlBvalues of 21.2-32.4%o,(2) modernsea 1890 VENGOSH ET AL.: GEOCHEMICAL AND ISOTOPIC CONSTRAINTS

0.06

,aitewaterBe'er Tovifya .

0.04 ß Be'er Toviyya o Giva't Brener [] Gedera 0.02 [] Hazor a Yavne •x Revadim seawater ratio . waste water 0.00 0 40 Chloride (mmole/1)

50

• 40

r.• 30 - -- ater •21 20 • tt•• ' ;S•ofGalilee

lO • • t.•groundsewage-contaminated water o waste water 0 50 100' 150 1/Boron (1/mmole) Figure8. Chlorideconcentration versus boron and 1/boron versus /5•B values of groundwaterfrom the studyarea, as compared to seawaterratios, wastewater, and fresh groundwater. The straightlines represent mixingbetween freshwater and saline water with/5•B values of 39%o(line a) and50%0 (line b) andbetween freshwaterand wastewater(line c). waterwith/5•B valuesof 39%0,(3) a salinesource or sources spondto mixingbetween a salinesource with a seawater-like withunknown i5•B values,(4) wastewaterwith an anthropo- /5•Bsignature and fresh groundwater. Similarly, the/5•B vari- genicisotopic signal of 0-10%o, (5) boronderived from dis- ations of brackishwater from the Be'er Toviyya area corre- solutionof marinecalcite matrix with a 15•Bvalue of 20-25%0, spondto mixingwith a salinewater with a high/5•B value and (6) residualboron, derived from adsorptionequilibrium (>50%0; Figure8). betweendissolved boron and exchangeableboron on claymin- The boron isotopecomposition of the brackishwater is eralsin theaquifer, relatively enriched in •B. significantlydifferent from those of anthropogenicsources, The boron-chloriderelationship in mostof the investigated suchas sewageeffluents (/5•B - 0-10%o) or sewage- groundwaters(Figure 8) suggeststhat boron,like mostof the contaminatedgroundwater (10-25%o [Vengoshet al., 1994]), dissolvedions, is nearlyconservative in the aquifer.The brack- and thusfurther rule out salinizationfrom leakageor irrigation ishgroundwaters have B/C1 molar ratios (1.5 x 10-3) thatare withsewage effluents. The high/5•Bvalues are similar, how- approximatelytwice that of seawater (0.8 x 10-3) and ever,to thoseof salinewater associatedwith saltwaterintrusion in valuesof 25-49.9%0(Figure 8, Table3). On the basisof these the westernpart of the aquifer[Vengosh et al., 1994].Conse- /5•B variations,two types of brackishwater are identified:(1) quently,the &rib values suggest that the brackish water is derived waterwith high&•IB values (>40%o), typically of the saline from a marinesource with a/5•B valueof 39%0or higher. plumesof Be'erToviyya and Shieller(see locations in Figure Therelatively high B/C1 ratios and/5•B values of thebrack- 1), and (2) waterfrom the othersaline plumes in the central ishwater are not consistentwith simplemixing between fresh aquiferand from the eastern part of theaquifer with a water and unmodifiedMediterranean seawater.However, high rangeof 25-40%0.The/5•B valuesof thelatter group corre- B/C1with a marine&•B signaturecan be generatedduring VENGOSH ET AL.: GEOCHEMICAL AND ISOTOPIC CONSTRAINTS 1891 evaporationof seawaterand precipitationof halite [Vengoshet chemicalcriteria) interactedwith carbonaterocks older than al., 1992], which is consistentwith the first mechanismsug- early-middleMiocene carbonates;(3) the postulatedsaline gestedearlier for the formationof the salinewater. In contrast, water did not originatefrom modernPleistocene-age seawater; simple adsorptionof boron from seawater(i.e., the second and (4) recyclingof freshwaterthrough pumping and irrigation mechanismof modificationof seawaterby base-exchangere- is not the main cause for salinization.Consequently, the Sr actions)would removeisotopically light B from solutionand isotopicdata confirman "allochthonous"saline source with an wouldresult in low B/C1and high/5•B values,which contra- isotopiccomposition of older sedimentaryrocks. dicts our data. A 87Sr/86Srratio of <0.7083corresponds to an early-middle Miocene or older seawaterisotope composition. In principle, 5.6. Strontium Isotope Systematics this may indicate that the saline sourceis derived from early- In the coastal aquifer systemthere are several sourcesof middle Miocene or older seawateror that younger seawater Sr2+ that canaffect the Sr isotopiccomposition of the inves- interactedwith carbonaterocks of that age and the Sr isotope tigatedbrackish groundwater: (1) freshwater with 87Sr/86Srcomposition of the water was reset by this interaction.On the ratios>0.70860, (2) modernseawater with a 87Sr/86Srratio of basisof its relativelyhigh Sr/C1ratio (Figure 9a; Table 3), it is 0.70920,(3) salinewater with unknown 87Sr/86Sr ratios, (4) Sr clear that the Sr isotoperatio is the result of water-rock inter- derived from dissolution of Pleistocene calcite matrix with a action, and it thus does not constrain the time of the intrusion 878r/86Srratio >0.7090,and (5) "exchangeableSr" derived of the fossilseawater. The combinationof high Ca/Mg ratios from ion-exchangereactions with clay minerals. andthe low 87Sr/86Sr ratio of <0.7083of thepostulated saline Rainwatersover Israel have a 87Sr/86Srrange of 0.7078- end-memberindicates that the originalseawater was modified 0.7092, reflectingmixing of Senonianto Eocene dust and sea during dolomitizationreactions with early-middle Miocene or spraydissolved in therainwater [Henit et al., 1993].The 87Sr/ older carbonates.According to this approach,the 87Sr/86Sr 868rratio of freshgroundwater (C1 < 100mg/1) is lowerthan constrainsthe minimum age of the carbonaterocks with which that of the Pleistocenecalcite matrix -0.7092 [Starinskyet al., evaporated seawater interacted. During dolomitization,the 1980]and thus probably reflects the meteoric87Sr/86Sr ratio. originalhigh 87Sr/86Sr ratio of thesea water would be reduced The 87Sr/86Srratio of the exchangeableSr in the coastalaqui- if the carbonate rocks are older [Sassand Starinsky,1979; fer is not known.Johnson and DePaolo[1997b] argued that the Starinskyet al., 1983;Stein et al., 1997].Thus the 878r/86Sr ratio isotopicratio of the exchangeableSr reflectsthe composition of theresidual Ca-chloride brine is controlledby (1) theS7Sr/ of water with which it has interactedin the past. Since fresh 86Srratio andage of the carbonaterock, (2) the fractionof groundwaterin the coastal aquifer is generated within the Ca2+ (andSr 2+) derivedfrom the rock(i.e., degree of inter- aquiferregion, the expected87sr/g6sr of exchangeableSr is action[Sass and Starinsky, 1979]), and (3) the initial87Sr/86Sr between0.70860 (the freshwaterratio) and 0.7092(the Pleis- ratio of the evaporatedseawater. The early-middleMiocene Sr tocenecalcite cement ratio). isotopic signal can be a result of any combinationof these As shownbefore, the Sr isotopicratio of rechargedwater processes,but indicatesthat the original seawatercould not from the Sea of Galilee is modified as a result of water-rock have interactedwith younger carbonaterocks. interactionsin the aquifer.While C1, /5•80, and Br/C1 of two groundwatersamples (Azriqa'm and Giva't Brener6 wells)in the vicinityof rechargeareas are almostidentical to thoseof 6. Conclusions: Synthesis and Mechanism of theSea of Galilee,the 87Sr/86Srratios are significantly higher, Salinization reflectingthe addition of Srwith a higher87Sr/86Sr ratio. This Figure 10 summarizesthe two principalprocesses that affect impliesthat Sr isotopicexchange is rapid, at least in the time the geochemicalcomposition of the investigatedgroundwater: frame of artificialrecharge of the Sea of Galilee to the coastal (1) rechargeof water importedfrom the Seaof Galilee,which aquiferduring the last30 years.The Sr/Caratio of the brackish isreflected by high/5•80 and Br/C1 ratios in groundwaterin the water (-0.003) is higherthan that of typicalcalcite (-0.0005); vicinity of the recharge areas, and (2) mixing with a Ca- thusSr 2 + wasnot derived from simple dissolution of thecalcite enrichedsaline water that has a high/5•B(>40%0) and a low cement.Instead, calcite solution/recrystallizationwould result 87Sr/a6Srratio (<0.7083). The correlationbetween /5•B and in relative enrichment of dissolved Sr2+ over Ca 2+ due to the 87Sr/86Srratios may also reflect mixing between dissolved B low Sr distributioncoefficient (0.05) in calcite[Katz et al., 1972]. and Sr, derived from dissolution of the carbonate matrix, and Although,strontium (Figure 9), as well as calcium(Figure a high-/5•B,low-87Sr/86Sr saline source. On the basisof the 4), vary linearly with chloridein the brackishwaters, the Sr geochemicaland isotopic constraints,we suggestthat fossil isotopesystematics indicate that Sr is not derivdedfrom con- seawaterhas interacted with early-middle Miocene or older servativemixing of two water types.If Sr variationswere de- rocksand hasbeen dilutedconsiderably (> 10 times)because terminedby conservativemixing, the 87Sr/86Srratio should of mixing with the local groundwater in the Pliocene- varylinearly with 1/Sr.This is not the case(Figure 9). Most of Pleistocenecoastal aquifer. The marine sedimentaryrecord the brackishwater samples,particularly those with high salinity (Figure 2) along the Mediterraneancoast reflects numerous (e.g.,Beer Toviyya), have 878r/86Sr ratios scattering between inland seawater invasions. Remnants of seawater could be en- 0.70825 and 0.70855. This is most likely due to the addition of trappedon the bottom,below (the SaqiyeGroup) and adjacent Sr resultingfrom carbonate solution/precipitation with a 87Sr/ (Avdat Group; Figure 1) to the coastalaquifer. 86Srin thisrange or slightlyhigher. This is alsoconsistent with The entrapped seawatercould enter the coastal aquifer by the relativelyhigh Sr/C1ratios. two mechanisms:(1) lateral flow from the adjacentEocene On the basis of these observations, several conclusionscan aquitardin the eastand (2) verticalupconing from underlying be drawn with respectto the origin of the saline component: salinewater sourcesat the bottom of the aquifer [e.g.,Maslia (1) thereis salinizationby an externalsource with a 87Sr/86Srand Prowell,1990]. Table 4 showsthat the chemicalcomposi- ratio of <0.7083; (2) the fossilseawater (based on other geo- tion of both typesof salinewater from the Eoceneaquitard are 1892 VENGOSH ET AL.: GEOCHEMICAL AND ISOTOPIC CONSTRAINTS

0.035

[] I • 0.030 • 0.025

[] o • 0.020 [] [] '4..• 0.015 o& ¸ *-• 0.010 sea water ratios

0.005 XSeaof Galilee_ ......

0.000 10 20 30 40 Chloride (mmole/1)

0.7088

0.7086 o A& [] 0.7084 [] ß ß ß ß Be'er Toviyya ß o Giva't Brener 0.7082 o fresh water [] Gedera 0.7080 possible [] Hazor 5•1 øcesses: ,• Revadim 0.7078 ß Yavne ß rechargewater 0.7076 ß Sea of Galilee 0.7074 .... 0 100 200 300 1/Sr (1/mmole) Figure9. Chlorideversus strontium concentrations, and i/strontium versus 87Sr/86Sr ratios of groundwater in the studyarea. Note the enrichmentof Sr2+ in the brackishgroundwater over seawater Sr/C1 ratios. The differentprocesses represent (1) mixingwith low Sr freshwater,(2) mixingwith seawater with 878r/86Sr ratios of 0.7092,(3) mixingwith saline water with 875r/86Sr ratios of <0.7083,(4) water-rockmodification in which dissolvedSr and/orexchangeable Srwith a high87Sr/86Sr ratios modifies the original Sr isotopicratio, and (5) mixingwith rechargeof the Sea of Galilee. different from that of the saline plumes in the aquifer. One tablishedthrough faults, local structures,and unconformities, may argue,however, that base-exchangereactions may modify asevidenced in the easternpart of the aquifer [Rosenthalet al., the original marine-like compositionof the Eocence saline 1992; Weinbergerand Rosenthal,1994]. The evidencethat sev- water. Nevertheless,Artzi [1999] showedthat the chemical eral salinespots with high salinityexisted since the 1930s[Ven- compositionof saline groundwaterfrom the coastal aquifer, goshet al., 1996] suggeststhat this salinizationprocess is a located along the Eocene-Pliestocenecontact at the eastern natural phenomenon.It shouldbe noted that salinizationre- margin of the coastalaquifer, is identicalto that of the Eocene sulting from substantialquantities of saline water that flow salinewater. Thus the Eocenegeochemical signature is clearly upward becauseof a fault systemthat breachesimpermeable recognizedin the adjacent groundwaterwithin the coastal unitswas also demonstratedin the upper Floridan aquifer in aquifer,which is different from the compositionof the inves- Georgia [Hanshawet al., 1965; Krauseand Randolph, 1989; tigated saline plumes. Maslia and Prowell,1990]. We proposethe followingconceptual model for the salin- The flow of the underlyingfossil brine to the aquifer was ization mechanismin the Mediterranean coastal aquifer of enhanced,however, because of the reductionof the hydraulic Israel. Hydraulic connectionsbetween underlyingunits con- pressureand formation of deep hydrologicaldepressions, re- tainingpressurized saline water and the surficialaquifer units sultingfrom extensivepumping of freshwaterover the central enabledupward flow of the salinewater and salinizationof the parts of the aquifer. The fact that salinizationhas increased overlyingfreshwater. The hydraulicconnections between the despitethe rise in water levelsduring the last decadeindicates overlyingaquifer, the low-permeabilitymarine shales(Yafo that the continuedupward flux of salinewater is still greater Formation;Figure 2), and the lower unitswere probablyes- than the down flow and horizontalrecharge of freshwater. VENGOSH ET AL.: GEOCHEMICAL AND ISOTOPIC CONSTRAINTS 1893

-1.00 Sea of Galilee

-2.00 ß Be'er Toviyya wa?e water o Giva't Brener •o • Gedera o [] Hazor "•'-3.00 o ß Yavne /x Revadim • waste water -4.00

-5.00 ! 0.000 0.001 0.002 0.003 Br/C1 ratio

5O C5alineend-member?

• 40 sea water

carbonate matrix ......

20 (Pleistocene)[ 0.7082 0.7084 0.7086 0.7088 0.7090 0.7092 8, Sr / Sr ratios Figure10. Variationsof Br/CIversus 8180 and 81lB versus 878r/86Sr ratios. Note (1) the influenceof the rechargewater from the Seaof Galilee(high Br/CI and i•180)on localgroundwater; (2) the wastewater signature(low Br/CI, high i•18 O, low i•11 B), and (3) the correlationbetween i• 11 B and 878r/a6srratios, indicatingmixing between fresh groundwater and a salinesource with high•11B and low 878r/a6srratios. Alternatively,the later reflectmixin[ betweendissolved B and Sr, derivedfrom dissolutionof the carbonate matrixand a high-•11Band 10w-a7sr •868r saline source.

The inferred existenceof deep saline water has water re- Acknowledgments. We appreciateand thank the thoroughreview sourceimplications. Seawater has been consideredas the ma- of T. Bullen(USGS, Menlo Park) and J. K. Bohlke(USGS, Reston), who significantlyimproved the earlierversion of the paper.We thank jor raw materialfor desalinizationin Israel.However, the large I. Pankratov (HydrologicalService), R. Nissan, and E. Rosenberg volumeof salinegroundwater that thiswork pointsto may be (Hebrew University) for their analyticalwork. Some of the boron a better sourcefor desalinization,as its use may be more isotopeanalyses were carriedout at the laboratoryof K. G. Heumann economicaland have a reducedenvironmental impact. More- (Universityof Regensburg,Germany) and we thank him for his gen- over, pumpingof underlyingsaline water may reducethe salt erous hospitality.A. Starinsky(Hebrew University), I. Gavrieli, Y. Yecheli (GeologicalSurvey), and E. Adar (Sede Boqer) helped in flux to the aquifer and may reduce the rate of salinization ideasand comments.Y. Artzi wassupported by the Water Commission process.Most of the pumping wells in the coastal aquifer of Israel.The projectwas partly supported by the InternationalAtomic penetratedto the upper and middle sectionsof the saturated Energy Agency, Vienna. zone. Sincethe inferred salinereservoir occupies the deepest part of the aquifer, its impact on the overlyingfreshwater References aquiferhas been limited. Thus a new deep drillingprogram is Artzi, Y., Salinization mechanismsof ground water in the coastal required to illuminate the extent and distributionof the saline aquifer of Israel: The Gan-Yavne Gedera Region, M.S. thesis,Ben reservoir.We suggestthat this salinegroundwater should also Gurion Univ. of the Negev, Beer Sheva,Israel, 1999. Banner, J. L., G. J. Wasserburg,P. F. Dobson,A. B. Carpenter, and be consideredas an important source for desalinizationin C. H. Moore, Isotopic and trace elementsconstrain on the origin order to solvethe severelack of potablewater in the semiarid and evolution of saline ground water from central Missouri, area of the Middle East. Geochim. Cosmochim.Acta, 53, 383-398, 1989. 1894 VENGOSH ET AL.: GEOCHEMICAL AND ISOTOPIC CONSTRAINTS

Carpenter,A. B., Origin and chemicalevolution of brinesin sedimen- case study from the Coastal Plain aquifer of Israel, J. Contam. tary basins,Okla. Geol. Surv.Circ., 79, 60-77, 1978. Hydrol., 11, 149-171, 1992. Chaudhuri,S., V. Broedel, and N. Clauer, Strontiumisotopic evolution Sass,E., and A. Starinsky,Behaviour of strontiumis subsurfacecal- of oil-field waters from carbonatereservoir rocks in Bindley field, cium chloride brines: Southern Israel and Dead Sea Rift Valley, central Kansas, USA, Geochim. Cosmochim.Acta, 51, 45-53, 1987. Geochim. Cosmochim.Acta, 43, 885-895, 1979. Craig, H., Isotopicvariations in meteoricwaters, Science, 133, 1702- Spivack,A. J., M. R. Palmer, and J. M. Edmond. The sedimentary 1703, 1961. cycleof the boron isotopes,Geochim. Cosmochim. Acta, 51, 1939- Eisenhut, S., K. G. Heumann, and A. Vengosh, Determination of 1950, 1987. boron isotopicvariations in aquaticsystems with negativethermal Starinsky,A., Relation betweenCa-chloride brines and sedimentary ionizationmass spectrometry as a tracer for anthropogenicinflu- rocks in Israel, Ph.D. thesis, 200 pp., Hebrew Univ., Jerusalem, ences,Fresenius J. Anal. Chem., 354, 903-909, 1996. Israel, 1974. Epstein,S., and T. K. Mayeda,Variations of •80 of watersfrom Starinsky,A., M. Bielsky,B. Lazar, E. Wakshal,and G. Steinitz,Ma- natural sources,Geochim. Cosmochim.Acta, 4, 213-224, 1953. rine 87Sr/86Srratios from the Jurassicto Pleistocene:Evidence from Fritz, B., N. Clauer,and M. Kam, Strontiumisotope data and geo- groundwatersin Israel, Earth Plant. Sci. Lett., 47, 75-80, 1980. chemical calculationsas indicators for origin of saline waters in Starinsky,A., M. Bielsky,B. Lazar, G. Steinitz, and M. Raab, Stron- crystallinerocks, in Saline Water and Gasesin CrystallineRocks, tium isotopeevidence on the historyof oilfield brines,Mediterra- nean Coastal Plain, Israel, Geochim. Cosmochim.Acta, 47, 687-695, editedby P. Fritz and S. R. Frape, Geol.Assoc. Can. Spec.Pap., 33, 1983. 121-126, 1987. Stein, M., A. Starinsky,A. Katz, S. L. Goldstein,M. Machlus,and A. Gat, J. R., Local variabilityof the isotopecomposition of groundwater, Schramm,Strontium isotopic,chemical, and sedimentologicalevi- in IsotopeTechniques in GroundwaterHydrology 1974, vol. II, pp. dence for the evolution of Lake Lisan and the Dead Sea, Geochim. 51-60, Int. At. Agency,Vienna, Austria, 1974. Cosmochim.Acta, 61, 3975-3992, 1997. Gat, J. R., Groundwater,in StableIsotope Hydrology: Deuterium and Vengosh,A., and A. Ben Zvi, Formationof a saltplume in the Coastal Oxygen-18in the WaterCycle, Tech. Rep. $er., no. 210, pp. 223-240, Plain Aquifer of Israel: The Be'er Toviyyaregion, J. Hydrol.,160, Int. At. Agency,Vienna, Austria, 1981. 21-52, 1994. Gat, J. R., and W. Dansgaard,Stable isotope survey of the freshwater Vengosh,A., and R. Keren, Chemicalmodifications of groundwater occurrencesin Israel and the Northern Jordan Rift Valley, J. Hy- contaminatedby rechargeof sewageeffluent, J. Contain.Hydrol., 23, drol., 16, 177-212, 1972. 347-360, 1996. Gavish,E., and G. M. Friedman,Progressive diagenesis in Quaternary Vengosh, A., and I. Pankaratov, Chloride/bromide and chloride/ to Late Tertiary carbonatesediments: Sequence and time scale,J. fluorideratios of domesticsewage effluents and associatedcontam- Sediment.Petrol., 39, 980-1006, 1969. inated groundwater, Ground Water,36, 815-824, 1998. Hanshaw, B. B., W. Back, M. Rubin, and R. L. Wait, Relation of Vengosh,A., and E. Rosenthal,Saline groundwater in Israel: Its bear- carbon-14 concentrations to saline water contamination of coastal ing on the water crisisin the country,J. Hydrol.,156, 389-430, 1994. aquifers,Water Resour. Res., 1, 109-114, 1965. Vengosh,A., A. R. Chivas,and M. T. McCulloch, Direct determina- Herut, B., A. Starinsky,and A. Katz, Stontium in rainwater from tion of boron and chlorine isotopesin geologicalmaterials by neg- Israel: Sources,isotopes and chemistry,Earth Planet.Sci. Lett., 120, ative thermal-ionizationmass spectrometry, Chem. Geol., 79, 333- 77-84, 1993. 343, 1989. Johnson,T. M., and D. J. DePaolo, Rapid exchangeeffects on isotope Vengosh,A., A. Starinsky,A. Melloul, M. Fink, and S. Erlich, Salin- ratios in groundwater systems,1, Development of a transport- ization of the coastalaquifer water by Ca-chloridesolutions at the dissolution-exchangemodel, WaterResour. Res., 33, 187-195, 1997a.. interfacezone, alongthe CoastalPlain of Israel, Hydrol.Rep. 20/91, Johnson,T. M., and D. J. DePaolo, Rapid exchangeeffects on isotope 35 pp., Hydrol. Serv., Jerusalem,Israel, 1991. ratiosin groundwatersystems, 2, Flow investigationusing Sr isotope Vengosh,A., A. Starinsky,Y. Kolodny,A. R. Chivas,and M. Raab, ratios, Water Resour.Res., 33, 197-209, 1997b. Boron isotopevariations during fractional evaporation of seawater: Katz, A., E. Sass,A. Starinsky,and H. D. Holland, Strontiumbehav- New constraintson the marine vs. nonmarinedebate, Geology,20, iour in the aragonite-calcitetransformation: An experimentalstudy 799- 802, 1992. at 40-980, Geochim. Cosmochim.Acta, 36, 481-496, 1972. Vengosh,A., K. G. Heumann, S. Juraske, and R. Kasher, Boron Kiss,E., Ion-exchangeseparation and spectrophotometricdetermina- isotopeapplication for tracingsources of contaminationin ground- tion of boron in geologicalmaterials, Anal. Chim.Acta, 121, 243- water, Environ. Sci. Technol.,28, 1968-1974, 1994. 256, 1988. Vengosh,A., Y. Artzi, I. Zirlin, I. Pankratov,H. Harpaz, E. Rosenthal, Krause,R. E., and R. B. Randolph,Hydrology of the Floridan aquifer andA. Ayalon,Monitoring water qualityin the CoastalPlain aquifer systemin southernGeorgia and adjucentparts of Florida and South of Israel (Givat Brener, Gedera, Yavne) (in Hebrew), Hydrol.Rep. Carolina, U.S. Geol. Surv.Prof. Pap. 1403-D, 65 pp., 1989. 4/96, 34 pp. Hydrol. Serv., Jerusalem,Israel, 1996. Levy,Y., The origin and evolutionof brinesin coastalsabkhas, north- Water Commissioner,Hydrological Service annual report: Develop- ern Sinai, J. Sediment.Petrol., 47, 451-462, 1977. ment,exploitation situation of water resourcesin Israel (in Hebrew), Magaritz, M., A. Amiel, D. Ronen, and M. C. Wells, Distributionof Jerusalem,Israel, 1995. metals in a polluted aquifer: A comparisonof aquifer suspended Weinberger,G., and E. Rosenthal,The fault pattern in the northern material to fine sedimentsof the adjacentenvironment, J. Contain. Negev and southerncoastal plain of Israel and its hydrogeological implicationsfor groundwaterflow in the Judea Group aquifer,J. Hydrol., 5, 333-347, 1990. Maslia, M. L., and D.C. Prowell, Effects of faults on fluid flow and Hydrol.,155, 103-124, 1994. chloride contaminationin a carbonateaquifer system,J. Hydrol., Y. Artzi, JacobBlaustein for Desert Research, The Water Resource 115, 1-49, 1990. ResearchCenter, Ben Gurion University of the Negev, Sede Boqer McCaffrey, M. A., B. Lazr, and H. D. Holland, The evaporationpath 84990, Israel. of seawater and the coprecipitationof Br- and K+ with halite, J. A. Ayalon, GeologicalSurvey, 30 Malchei Israel Street, Jerusalem, Sediment.Petrol., 57, 928-937, 1987. Israel. Mercado,A., The use of hydrogeochemicalpatterns in carbonatesand A. J. Spivack,Center for Marine Sciences,University of North and sandstoneaquifers to identify intrusionand flushingof saline Carolinaat Wilmington,601 S. CollegeRd., Wilmington,NC 28403-3297. water, Groundwater, 23, 635-646, 1985. A. Vengosh,Hydrological Service, P.O. Box 6381,Jerusalem 91063, Nissim,I., Characterizationof a hydrologicalsystem in arid zone:The Israel. ([email protected]) EoceneAvdat Group chalk, central and northernNegev, Israel (in Hebrew, English abstract), M.S. thesis, Hebrew Univ., Rehovot, Israel, 1991. Ronen, D., M. Magaritz, E. Almon, and A. Amiel, Anthropogenic anoxification("Eutrophication") of the water table regionof a deep phreatic aquifer, WaterResour. Res., 23, 1554-1560, 1987. Rosenthal,E., A. Vinokurov, D. Ronen, M. Magaritz, and S. Mosh- (ReceivedSeptember 30, 1998;revised January 20, 1999; kovitz. Anthropogenicallyinduced salinizationof groundwater:A acceptedJanuary 21, 1999.)