Chemical and isotopic (87Sr/86Sr, 18O, D) constraints to the ...

PII S0016-7037(00)00618-9

Chemical and isotopic (87Sr/86Sr, d18O, dD) constraints to the formation processesof Red-Sea brines

M. C. PIERRET,1,* ,† N. CLAUER,1 D. BOSCH,2 G. BLANC,3 and C. FRANCE-LANORD4

1Centre de Geóchimie de la Surface/Ecole et Observatoire des Sciences de la Terre, 1 rue Blessig, 67084 Strasbourg Cedex, France2Laboratoire de Tectonophysique, Place E. Bataillon, 34095 Montpellier Cedex, France

3Departement de Geólogie et d’Oceánographie, Avenue des faculte´s, 33405 Talence Cedex, France4Centre de Recherches Pe´trographiques et Geóchimiques, BP 20, 54501 Vandoeuvre-les-Nancy Cedex, France

(Received March28, 2000;accepted in revised form October13, 2000)

Abstract—About twenty deeps filled with hot brines and/or metalliferous sediments, are located along theRed-Sea axis. These brines present a well-suited framework to study the hydrothermal activity in such a youngocean. The present study outlines the results of a geochemical approach combining major-, trace-element andisotopic (oxygen, hydrogen, strontium) analyses of brines in six of the deeps, to evaluate different processesof brine formation and to compare the evolution of each deep. Important heterogeneities in temperature,salinity, hydrographic structure and chemistry are recorded, each brine having its own characteristics. Theintensity of hydrothermal circulation varies among the deeps and ranges from being strong (Atlantis II andNereus) to weak (Port-Soudan) and even to negligible (Valdivia and Suakin) and it varies along the entireRed-Sea axis. These observations do not favour a unique formational model for all of the brines. For example,the brines of the Suakin deep appear to have been formed by an old sea water which dissolved evaporite beds,without significant fluid circulation and hydrothermal input, while others such as Atlantis II or Nereus Deepsappear to be dominated by hydrothermal influences. A striking feature is the absence of a relationship betweenthe position of the deeps along the axis and their evolutionary maturity.Copyright © 2001 Elsevier ScienceLtd

1. INTRODUCTION

The Red Sea is a young ocean which evolved from a conti-nental to an oceanic rift. The axial valley is characterised byheterogeneous spreading rates along the axis. It differs frommature oceanic ridges by having continuous oceanic crust in itsmeridional part (between 16° and 19°N), while the axial troughis discontinuous towards the north. Only a few isolated deepshave a basaltic floor north of 19°N (Bonatti, 1985; Le Quentrecand Sichler, 1991). These bathymetric depressions are charac-terised by the occurrence of hot brines and/or metalliferoussediments and represent unique sites in the study of ocean-ridgeformation. In this study, we examine the formation processes ofthe brines from six Red-Sea deeps.

Atlantis II Deep represents the largest deep of the Red Sea.It contains the warmest and the saltiest brines, as well as themost metal-enriched sediments. It is the most thoroughly stud-ied deep, partly because of the economic interest in its metaldeposits. Characterisation and formation of the brine layershave been studied since the 1960’s (Miller et al., 1966; Schoelland Hartmann, 1973; 1978; Hartmann, 1980; Blanc and An-schutz, 1995; Anschutz and Blanc, 1996; Hartmann et al.,1998a;b; Anschutz et al., 1999). Temperature and salinity havebeen recorded since 1965. The occurrence of mineralised brinesand sediments in this deep has generally been related to hy-drothermal activities and to leaching of the thick Mioceneevaporitic beds (Anschutz and Blanc, 1993a;b; 1995a;b; Blancet al., 1998; Degens and Ross, 1969; Pottorf and Barnes, 1983;

Zierenberg and Shanks, 1986). The available models proposethat sea water circulated and interacted with hot oceanic ba-salts, dissolved evaporitic salts during hydrothermal circula-tion, and due to their high density, the hydrothermal solutionswere trapped in the Atlantis-II depression.

The sediments and brines of the other deeps, such as those ofSuakin, Port-Soudan, Valdivia, Thetis and Nereus, were lessintensively sampled and studied and only a few comparisonshave been made among these different sites. Hydrothermal-fluid circulations through the evaporite deposits seem to be themain reason for the salty inputs into the deeps, inducing for-mation of brine layers at their bottoms. Alternatively, otherbasins such as those of Bannock and Tyro in the easternMediterranean Sea and that of Orca in the Gulf of Mexico,contain brines which do not appear to result from hydrothermalactivities. The origin of these basins relates to subsurface saltdissolution triggered by tectonic deformation (Camerlenghi,1990). Our study addresses such questions. Are the brines fromdifferent Red-Sea deeps formed the same way? Is the hydro-thermal activity always necessary, and does it have the sameinfluence in each deep?

In mature oceanic ridges, such as the East-Pacific Rise andthe Mid-Atlantic Ridge, the hydrothermal fluids have differentsalinities but never higher than 8% of NaCl (Edmond et al.,1982; Von Damm et al., 1985; Von Damm, 1988; 1995), andmineralizations occur as sulphide chimneys and mounds. Thedifference between these ridge mineralizations and the hydro-thermal mineralizations of Red-Sea deeps is largely due to theoccurrence of hot brine pools in the latter. The salty hydrother-mal fluids of the Red Sea are trapped in the deeps and theirmetals are precipitated as layers within the sediments. More-over, no hydrothermal fluids have been directly sampled in the

*Author to whom correspondence should be addressed.†Present address:Institut fur Nukleare Entsorgung, Postfach 3540,D-76021 Karlsruhe, Germany ([email protected]).

Pergamon

Geochimica et Cosmochimica Acta, Vol. 65, No. 8, pp. 1259–1275, 2001Copyright © 2001 Elsevier Science LtdPrinted in the USA. All rights reserved

0016-7037/01 $20.001 .00

1259

Red Sea, which implies that their chemical compositions andtemperatures could only be deduced. Ramboz et al. (1988)measured temperature and salinity within fluid inclusions ofbarite and anhydrite in the metaliferous sediments of the At-lantis II Deep and obtained a 400°C maximal temperature anda maximum salinity of 330. These values should represent goodapproximations for the hydrothermal fluids.

During the 1992 REDSED cruise aboard the R.V.MarionDufresne, sediments, waters and brines were for the first timecollected together in seven Red-Sea deeps (Suakin, Port-Soudan, Valdivia, Chain, Atlantis II, Thetis and Nereus; Fig.1). Analysis of these samples in the present study allowedcomparisons of the evolution of contemporaneous deeps. Phys-ical, chemical and isotopic determinations were made to char-acterise the brines and to deduce their origins and formationprocesses. The results address the following aims: (1) to eval-uate the different formation processes of the brines occurring ineach deep; (2) to highlight the differences among the deeps; (3)to quantify the impact of hydrothermal activity on each deep;and (4) to deduce the general evolution of these deeps relativelyto their location along the spreading axis.

2. GEOLOGIC SETTING AND PREVIOUS STUDIES

The Red Sea is a narrow oceanic depression, about 2000 kmlong, located between the Arabic and African plates. Its evo-

lution began with a rifting phase during Oligocene time (atabout 28–32 Ma), followed by a magmatic and spreadingphase. Sea-floor spreading with formation of oceanic crustlasted at least 7 Ma (Bohannon and Eittreim, 1991) with anexpansion rate of about 2 cm/yr. During the Miocene, highevaporation rates induced formation of thick evaporite deposits.These evaporites contributed to the formation of continentalbrines in the Gulf of Suez (Issar et al., 1971; Rosenthal et al.,1998) and to the brines of several Red-Sea depressions.

Discovery of the topographic depressions began in 1948when scientists aboard the Swedish research vessel O.V.Alba-tross reported salinity and temperature anomalies relative tonormal sea water at 21°109N and 38°099E (corresponding toDiscovery Deep). These anomalies were the first evidence forthe occurrence of differentiated layers in the brines of thedepressions. Seismic transects correlated the salinity and tem-perature anomalies with topographic deeps located along theaxis (Charnock, 1964; Miller, 1964). In 1965, several studiesrecorded temperatures of 56°C and salinities of 261 in Atlan-tis-II Deep. This deep was explored in detail during an inter-national expedition in 1966 in which a stratified brine body andthe first oceanic metalliferous deposits of economic interestwere described (Miller et al., 1966; Degens and Ross, 1969).Further investigations of the Red Sea resulted in the discoveryof about twenty deeps containing brine layers and/or metallif-erous sediments between 18° and 26°N, along the active riftzone (Backer and Schoell, 1972; Schoell, 1976).

Sea-floor spreading is often accompanied by hydrothermalactivity. In the Red Sea, sea water can interact with the oceanicbasalts and with the nearby sedimentary evaporites, blackshales enriched in metals, and biodetrital sediments. The denseand saline hydrothermal fluids also fill the bathymetric depres-sions. Thus, trapping combined with double diffusive convec-tion induces brine layering (Huppert and Turner, 1981; Mc-Dougall 1984a;b; Anschutz et al., 1998). Studies of fluidinclusions in anhydrite of Atlantis II sediments suggest a sup-ply of hot and salty fluids (Zierenberg and Shanks, 1983;Ramboz et al., 1988). After trapping, the brines evolved viadifferent processes. Brines may be enriched by chemical dis-solution, such as calcite or aragonite dissolution during settlingof biogenic particles in the water column. Moreover, differentchemical precipitations may take place into the brines (second-ary carbonates, silicates or Fe-oxides; Pottorf and Barnes,1983; Hartmann, 1985; Zierenberg and Shanks, 1986;Butuzova et al., 1990; Anschutz and Blanc, 1993a;b; Anschutzand Blanc, 1995a;b; Blanc et al., 1998).

Several studies have concluded, however, that not all Red-Sea brines were formed according to this model. For instance,Schoell and Faber (1978) on the basis of oxygen and hydrogenisotope ratios, concluded that the brines that were derived fromdeep waters at various times are of three different types: (1)isotopically unaltered present-day waters; (2) heavy palaeowa-ters from the last glacial maximum (18000 yr BP); and (3)lighter palaeowaters from the last climatic optimum (8000–5000 yr BP). Zierenberg and Shanks (1986) suggested that thebrine of Valdivia Deep formed by karstic dissolution of Mio-cene evaporites without hydrothermal input, that the brine ofSuakin deep did not record high-temperature water/rock inter-actions, and that the lower brine of Atlantis II Deep reflectsexchanges at approximately 255°C. Monnin and Ramboz

Fig. 1. Schematic map of the Red Sea with locations of the sevenstudied deeps and corresponding CTD-rosettes numbers.

1260 M. C. Pierret et al.

(1996) reported that anhydrite undersaturation seems to char-acterise brines with a hydrothermal signature. They suggestedthat those of the Atlantis II, Discovery and Suakin deepscontain a hydrothermal component, whereas the brine ofValdivia Deep in equilibrium with anhydrite results only fromdissolution of Miocene evaporites. As has been discussed byMonnin and Ramboz (1996) and Pierret et al. (2000) the originof the Suakin brine still remains a matter of debate. In sum-mary, the chemical and isotopic characteristics of the Red-Seabrines depend on parameters such as: (1) the supply of hydro-thermal fluids; (2) the nature of the interactions between seawater and either sedimentary deposits (evaporites, old sedi-ments) or basalts; (3) the contribution of dissolved biodetritalparticles; and (4) chemical precipitation from brines.

We analysed the waters from seven different deeps compris-ing from South to North, Suakin, Port-Soudan, Atlantis II,Chain, Valdivia, Thetis and Nereus (Fig. 1). The Suakin Deepis about 30 km long and 9 km wide with an average depth of2600 m; it consists of two subbasins (East and West) with thebottom formed, at least locally, by basalt. Two brine layerswere described in this southernmost deep of the Red Sea(Baumann et al., 1973). The Port-Soudan Deep is an elongatedgraben structure (2836 m maximum depth; Ba¨cker et al., 1975)located between Suakin and Atlantis II deeps (Fig. 1). It is asingle basin aligned in a NNW-SSE direction with a flat bottom(Guennoc and Thisse, 1982), and is filled with the thickestbrine layer known in the Red Sea. The Atlantis II Deep is thelargest brine-filled basin of the Red Sea with a surface area ofabout 52 km2 and a volume of about 4 km3 for the LowerConvective Layer (LCL) brine (Hartmann, 1998a;b; Anschutzand Blanc, 1996). The deep is a complicated graben structurelimited by antithetic tensional faults along its NE and SWborders (Ba¨cker et al., 1975). It consists of four subbasinsseparated by bathymetric highs that do not extend above the topof the brine pool. The Chain Deep located to the South ofAtlantis II, consists of three basins (Blanc and Anschutz, 1995),and is connected to Atlantis II at a depth of 1980 m (Anschutzet al., 1999). The Valdivia Deep is located about 20 km to theW-SW of Atlantis II on the western flank of the axial trough. Itconsists of several small and shallow bathymetric depressions.Its main basin was sampled. The Thetis Deep discovered in1972, 160 km northwest to Atlantis II Deep, is divided intoseveral subbasins. The main basin to the northwest is approx-imately 10 km long and 3 km wide with a depth of 1780 m. Thesediments of the Thetis Deep represent the second richest

metalliferous site in the Red Sea. The Nereus Deep situatedwithin the median valley, has an overall width of 12 km and alength of 40 km. It is limited in the SW and NE by steep faults(Fig. 1); Backer and Schoell, 1972; Ba¨cker et al., 1975; Ba¨cker,1976; Bignell and Ali, 1976; Scholten et al., 1991).

3. SAMPLING AND METHODS

Water and brine samples were collected in the seven deeps describedabove (Fig. 1). The results of different CTD-rosette characterizationsare summarised in Table 1.

The details of the sampling method have been described previously(Blanc and Anschutz, 1995; Anschutz and Blanc, 1996). A Bissett-Berman modified Olliver CTD-rosette continuously recorded temper-ature, conductivity and hydrostatic pressure. Water and brines werecollected in ascending Niskin bottles (12 L capacity). The salinitieswere obtained using a Goldberg optical refractometer. The pingerposition and the pressure recorded by the CTD-rosette allowed theevaluation of the sampling depth based on sound-velocity determina-tions for sea water and different brines. The values of sound velocitywere extrapolated from Matthew’s tables (Matthew, 1939). The con-ductivity data together with temperature and salinity allowed recon-struction of a continuous salinity record with linear functions specific toeach environment. The water samples were filtered (0.45mm pore size)after collection, and subsequently acidified with HNO3.

The elemental concentrations were determined by atomic absorption(Ca and Mg), emission spectrometry (Na and K), colorimetry (Si), ionicchromatography (Cl and SO4), ICP-AES (Mn, Sr, Fe, Rb), and ICP-MS(Li, Mo, Ba, U, Cr, Co, Cu, Zn and Pb). Reproducibility of the IAPSOsea-water standard varied between 5 and 10% for these elementalchemical determinations. The oxygen isotope compositions were ana-lysed by CO2 micro-equilibration after 24 h. Hydrogen was extractedby water reduction on uranium at 800°C (Friedman and Woodcock,1957). The D/H ratios were measured on a VG 602D-type massspectrometer and18O/16O ratios were determined using a triple collec-tor VG Optima mass spectrometer. ThedD andd18O relative to SMOW(Standard Mean Ocean Water) are given in per mil (‰) values. Theanalytical errors are62‰ and60.2‰ for thedD andd18O measure-ments, respectively. Corrections due to different salt concentrations andCO2 equilibration with H2O were applied according to Gat and Con-fiantini (1981). Strontium was extracted using a cation-exchange pro-cedure (type AG 503 12) according to Birck (1981) and was loadedwith a Ta activator on a single tungsten filament. The Sr isotopic datawere obtained on a VG Sector mass spectrometer. The NBS 987standard was routinely measured and for 10 determinations madeduring the course of the study, yielded an average87Sr/86Sr ratio of0.7102566 0.000014 (2s). Three CTD-rosette characterizations wereobtained in the Atlantis II Deep: B3 was located in the northernpassage, B4 in the eastern basin, and B6 in the west-southwesternpassage. Since they reflect identical stratification and chemical com-positions, in what follows, mean composition values are given as anaverage of the three CTD-rosettes of this deep.

Table 1. Hydrographic and CTD-rosette characteristics of the brines from the seven studied deeps.

CTD-rosette Depth (m) Type of brine Thickness (m) Temperature (°C) Salinity

Suakin B1 2800 Lower brine 22 23.5° 147.5Upper brine 29 23.3° 145.5

Port-Soudan B2 2600 1 brine 123 35.9° 214.5Chain B5 2050 1 brine 42 45.3° 270Atlantis II B3, B4, B6 2120 LCL 20503 bottom 66.1° 270

UCL1 29 55° 155UCL2 16 50° 118UCL3 7 46.5° 90

Valdivia B7 1673 1 brine 75 33.7° 242Thetis B8 1820 lack of brineNereus B9 2460 1 brine 15 29.9° 223.5Sea water 21.6° 40

1261Geochemical study of Red Sea brines

4. RESULTS

Temperature and salinity were determined in the sevendeeps. The variations during the last 20 yr were published byAnschutz et al. (1999). Four different stratified brine layerswere identified in the Atlantis II Deep; a Lower ConvectiveLayer (LCL) and three Upper Convective Layers (UCL1,UCL2, UCL3; Blanc and Anschutz, 1995). Suakin Deep con-tains two layers of brines, Thetis Deep is filled with Red-Seabottom water, and the four other deeps are characterised by onelayer of brine. All brines are warmer and saltier than the bottomwater (Table 1). The temperatures range from 23.5°C in Suakinto 66°C in the lower brine of Atlantis II and the salinities rangefrom 147.5 in Suakin to 270 in Atlantis II and Chain. TheSuakin brines differ from those of the other deeps by havingsignificantly lower temperatures and salinities. Each elementalprofile correlates with the temperature and the hydrographicstratification (Pierret, 1998).

Differences in pH, major (anions and cations) and traceelemental concentrations were observed among the differentbrines (Table 2). The most basic brine is that of Suakin Deep(pH 5 7.85) and the most acidic one corresponds to the LCL ofAtlantis II (pH 5 5). Nereus contains the brine which is themost concentrated in K, Ca, Sr, and Rb and least concentratedin Si, whereas the Atlantis II lower brine is the most enrichedin Si, Mn, Fe, Li, Ba and Zn. The lower brine of Suakin is theleast concentrated in Na, Cl, K, Fe (below detection limit) andLi. The brine which is the most depleted in Ca, Sr, Mn, Rb, Ba,Cu, and Zn, and the richest in Mg and SO4 is that of ValdiviaDeep (Table 2). Each brine is enriched in Na, Cl, Ca, Sr, Si, andRb relative to normal Red-Sea water, but the brines of AtlantisII, Chain and Nereus deeps have low concentrations of SO4 andthose of Atlantis II and Chain contain less Mg. These largedifferences suggest that the brines formed by different pro-cesses and that each deep operates as an isolated system allow-ing preservation of its chemical and physical characteristics.

Several bottom waters and brines were analysed for theirstable isotope compositions (oxygen and hydrogen; Table 3).The isotope signatures of the bottom water plot into two areas(Fig. 2) suggesting the presence of different water masses. Thedifferences may be partly explained by seasonal variations in

the exchange flow through the strait of Bab al Mandab con-necting the Red Sea to the Gulf of Aden (Thompson, 1939;Smeed, 1997), and by seasonal variations in temperature, cir-culation and current directions of the Red-Sea waters (Phillips,1966). The highestdD was found for the lower brine of Suakin(122‰) and the lowest for the lower brine of Atlantis II

Table 2. Average concentrations of the brines from the seven studied Red-Sea deeps.

Suakin PS Atlantis II Chain Valdivia Nereus Thetis

pH 7.85 6.43 5.21 6.48 6.21 7.43 8.1Alk (meq) 0.9 1.96 1.11 0.26 2.64 0.44 2.46Na (1) 2.48 3.79 4.88 4.81 4.13 3.74 0.54Cl (1) 2.76 4.04 5.37 5.17 4.59 4.35 0.64K (2) 38 49 69.9 66.8 52 80.2 12Ca (2) 56.1 33.52 146.3 143.3 25 237.7 12Rb (3) 8.2 6.7 25.3 25.1 3.1 31.1 1.5Mg (2) 72.2 64.6 32.8 36.8 95.3 80.4 63So4 (2) 33.3 44 6.7 8.5 72 7 33Fe (2) 0.007 0.107 1.4 0.02 0.13 0.06 ,Mn (2) 0.48 0.12 1.52 0.71 0.06 0.94 ,Li (3) 45.7 106 563.4 429 98.1 168.7 31Ba (3) 1.1 1.6 15.3 5.3 0.5 4.5 0.06Cu (3) 0.5 4.1 3.8 6.9 , 12 ,Zn (3) 2.6 4.7 41.6 9.9 , 18.4 ,

(1) in mol/1; (2) in mmol/1 and (3) inmmol/1. P.S - Port-Soudan Deep., 5 below detection limit.

Table 3. d18O and dD ratios (‰/SMOW) of various waters andbrines in the Suakin, Port-Soudan, Chain, Atlantis II, Valdivia, Thetisand Nereus deeps.

DeepType ofwater

Depth(m)

dD (‰)62‰

d18O (‰)60.2‰

Atlantis II LCL 2087 6 1LCL 2056 6 1.1LCL 2101 7 1.1LCL 2176 8 1.1LCL 2056 8 1.1LCL Average 13 1.8

Sea water 1944 12 1.9Sea water 1912 13 1.8

Chain B 2090 10 1.1Sea water 1905 10 1.7Sea water 1799 9 1.8

Suakin B1 2829 21 1.9B1 2808 20 1.9B2 1793 20 2B2 1778 16 2.3

Sea water 1730 7 2.1Port-Soudan B 2643 11 2.4

B 2540 12 2.5Sea water 2443 9 2.6Sea water 2366 8 2.7

Valdivia B 1600 9 2.2B 1540 9 2.2

Sea water 1520 11 2.2Thetis Sea water 1816 12


Nereus B 2448 11 3B 2443 10 3


(B - brine, B1 - lower brine in Suakin Deep, B25 upper brine inSuakin Deep and LCL5 Lower Convective Layer in Atlantis II Deep).


(16‰). Thed18O values range from13‰ for the Nereus brineto 11‰ for the lower brine of Atlantis II. Thus, Atlantis II hasthe lightest water in oxygen and in hydrogen (Fig. 2). Twopalaeowaters corresponding to the last glacial maximum(18000 yr BP), and to the last climatic optimum (8000–5000 yr

BP; Schoell and Faber, 1978) fall along the line fordD andd18O of present-day Red-Sea surface water, given by the rela-tion: dD 5 6 3 d18O (Craig, 1969; Fig. 2). The isotopesignatures of Red-Sea interstitial waters in Miocene to Pleisto-cene sediments, plot nearby (grey area in Fig. 2; Friedman andHardcastle, 1974; Lawrence, 1974). The figure outlines the factthat the isotope signatures cannot be explained by simplebinary mixing of two end-members. In fact, different originsand processes have to be considered: 1) the brines could be oldsea water which has (or has not) interacted with basalt and/orwith old sedimentary deposits; and 2) the isotopic fractionationmay have occurred during mineral precipitation and dissolu-tion. These processes are discussed below.

The 87Sr/86Sr ratios range from 0.709136 0.00002 (2s) inthe Thetis Deep filled with bottom sea water, to 0.7065660.00002 for the Nereus brine (Table 4). The Sr concentrationsrange from 0.09 mmol/L (Thetis) to 0.63 mmol/L (Nereus),which is in good agreement with the results obtained by Zier-enberg and Shanks (1986) on brines of the Atlantis II, Valdiviaand Suakin deeps. Observed87Sr/86Sr ratios plotted againstinverse Sr concentrations show a trend of low87Sr/86Sr value inSr enriched brines such as those of Nereus and Atlantis II, andhigh 87Sr/86Sr ratios at low Sr concentrations, such as thoseoccurring at Valdivia or Thetis (Fig. 3).

5. DISCUSSION

Previous studies on the brines of some Red-Sea deeps pointto varied geochemical properties, and the results obtained hereshow important heterogeneities in the chemistries and isotopiccompositions of the brines. Several aspects of brine evolutionwill be successively addressed in the discussion: 1) the forma-tion processes including the contribution of basaltic Sr; 2) theexistence of secondary mechanisms which may modify thebrines; and 3) the origin and characteristics of each studiedbrine. For deeps containing more than one brine layer, we wereespecially interested in the data for the lower pools, as they arethought to be the most illustrative of the origins of the brines.

5.1. Processes of Brine Formation: HydrothermalAlteration of Oceanic Crust and Dissolution ofEvaporites

The salinities of the studied Red-Sea brines (between 145and 270; Table1) cannot only be due to hydrothermal sea

Fig. 2.dD versusd18O diagram with isotopic compositions of bottomRed-Sea water and brines in the Atlantis II, Chain, Suakin, Port-Soudan, Valdivia and Nereus deeps. The equation of Craig, 1969relating surface Red-Sea water (dD 5 6 3 d18O) and the domain ofinterstitial waters in Miocene to Pleistocene Red-Sea sediments (1) iscalculated using the data of DSDP Leg 23 (Friedman and Hardcastle,1974; Lawrence, 1974). IW5 interstitial water; (2) Old sea-watersignature from Schoell and Faber (1978).

Table 4.87Sr/86Sr ratios and Sr concentrations (mmol/1) in the brines of the studied deeps. The two extreme cases are outlined in grey.

Deeps SampleDepth(m)

Temperature(°C)

[Sr](mmol/1)

87Sr/86Sr(62s in 1025)

Atlantis II B6-2 (LCL) 2158 66.4 0.479 0.706966 2Atlantis II B6-6 (UCL1) 2042 55.8 0.225 0.707476 2Atlantis II B4-7 (UCL 2) 2016 50 0.194 0.707696 2Atlantis II B6-8 (UCL 3) 2007 38.2 0.179 0.707826 1Chain B5-1 2090 45.4 0.468 0.707096 1Suakin B1-2 2829 23.5 0.189 0.708016 1Suakin B1-6 2778 23.3 0.185 0.709176 2Port-Soudan B2-1 2643 35.9 0.243 0.708506 2Valdivia B7-3 1600 33.7 0.167 0.708806 1Thetis B8-12 986 21.6 0.092 0.709136 2Nereus B9-1 2453 29.9 0.625 0.706566 2


water/basalt interactions and require halite (evaporite) dissolu-tion, as in the Salton-Sea hydrothermal system containing hy-persaline metalliferous brines (McKibben et al., 1988). It isproposed that the formation of the brines from the most studieddeep, Atlantis II, requires two main processes: hydrothermalalteration of the oceanic crust by sea water and chemicaldissolution of thick Miocene evaporites (NaCl, CaSO4) bycirculating sea water. We have attempted to quantify these twoprocesses using different geochemical tracers.

5.1.1. Elemental concentrations

When sea water interacts with a basaltic rock at high tem-perature during hydrothermal circulation, the physical, chemi-cal and isotopic characteristics of rock and waterboth change.Compared to the original sea water, hydrothermal solutions atmidocean ridges (MOR) have higher temperature and elevatedconcentrations of Li, K, Rb, Ba, Si, Ca, Sr, Mn, Fe, Zn, Cu,lower pH and SO4 and Mg concentrations that decrease to

almost zero (Edmond et al., 1982; Von Damm et al., 1985;1988; 1995). The chemical mobilisation of various elementsdepends on both the temperature and the depth of the circula-tion: SO4 is removed from sea water above 150°C (by anhy-drite precipitation and then sulfate reduction), the removal ofsea water Mg is due to smectite formation at low temperatures(,200°C) and chlorite crystallisation at higher temperatures.Below 1 km depths at 300 to 400°C, transformation of sulfur inrocks to H2S is accompanied by leaching of some metals (Fe,Zn, Pb, Cu). Li and other alkalis (K, Rb, Ba) are leached fromthe rocks at moderate to high temperatures (Honnorez et al.,1983; Rosenbauer et al., 1983; Alt, 1995; Hannington et al.,1995). In an active convection system, large volumes of seawater circulate at low temperature through a thin layer of rock,whereas only small amounts of sea water penetrate into thedeep rocks to react at higher temperature. In the Red Seasystem, thick evaporite layers constitute another type of rockwhere sea water can react and reach very high salinities duringcirculation. The main evaporite minerals are halite, anhydrite,gypsum and sometimes magnesite, so that Ca, Na, Cl, Mg orSO4 enrichments of brines relative to sea water may be due toevaporite leachings.

Specific chemical ratios can also be used to identify a hy-drothermal influence in Red Sea deeps, for instance the Fe/Mnratio is very different in hydrothermal fluids, hydrothermalbrines and non-hydrothermal brines (Table 5). Hydrothermalfluids are more enriched in Fe than Mn relative to sea water.The Zn/Cl ratio allows differentiation of Zn coming from seawater/evaporite dissolution, and Zn resulting from interactionwith basalt (Table 5). Hydrothermal high temperature circula-tion of fluids with Cl involves an increase in some metals likeZn in the fluids with a Cl concentrations close to that of seawater (high Zn/Cl), whereas evaporite leaching includes a largeincrease of Cl without clear Zn enrichment (low Zn/Cl).

The histograms in Figure 4 show that Atlantis II and Chaindeeps contain hydrothermal fluids which have interacted withthe oceanic crust at high temperatures (depletion in Mg andSO4, enrichment in Li, Ba, Rb, K, Fe, Mn, Zn, Cu, high Fe/Mnand Zn/Cl ratios). Nereus brine is depleted only in SO4 com-pared to sea water, but is enriched in Rb, K, L, Ba, Fe, Zn andCu, indicating a hydrothermal influence. The brines of theValdivia and Suakin deeps do not show clear hydrothermalorigins. The Fe/Mn ratio of Valdivia brine is high (of the sameorder of magnitude as that of hydrothermal fluids), however,

Fig. 3. 87Sr/86Sr ratios versus 1/[Sr] (Sr in ppm) of the differentstudied brines. Thetis Deep without brine can be considered to benormal deep sea water characterised by a87Sr/86Sr ratio of 0.70913 anda Sr content of 0.092 mmol/L. The Sr isotopic system can be simplifiedinto a two end-member model (section 5.1.2.). The positions of theend-member A (basaltic component) and of the two extreme end-members (B) have been reported (point B9 corresponding to Nereusbrine and B0 to Valdivia brine). The B end-member is essentiallycontroled by evaporites and is located along the dotted line.

Table 5. Fe/Mn and Zn/Cl chemical ratios of different brines from Red-Sea (this study).

Ratios Suakin Port-Soudan Valdivia Nereus Chain Atlantis II

Fe/Mn 0.0145 0.885 2.24 0.06 0.04 0.91Zn/Cl 0.93 1.24 1.15 4.22 1.9 7.59

Ratios HF(1) Bannock(2) Tyro (2) Orca(3) Salton-Sea(4)

Fe/Mn 2.45 0.004 0.0165 0.13 1.1Zn/Cl 370 nd nd nd 178

Also reported for comparison are the Fe/Mn and Zn/Cl ratios for an average hydrothermal fluid (HF), several Mediterranean brines (Bannock andTyro), a brine from Orca basin and a hydrothermal brine from Salton Sea. (1; Edmond and Von Damm, 1985; Von Damm, 1995), (2: Saager et al.,1993; De Lange et al., 1990; (3: Sackett et al., 1979), (4: McKibben and Williams, 1989; Thompson and Fournier, 1988). nd5 not determined.


Fig. 4. Histograms of pH (f), temperature (°C) (c) and some chemical elements (a,b, d, e, g, h, i, j) of brines of the Suakin,Port-Soudan, Atlantis II lower layer, Chain, Valdivia, Thetis and Nereus deeps. Thetis Deep which does not contain brines,represents generic deep Red-Sea water. The mean composition of hydrothermal fluid (from MOR) has been reported forcomparison, as well as the Fe/Mn ratios measured for the hydrothermal brine of the Salton Sea (McKibben and Williams,1989), and non hydrothermal brines of the Orca basin (Sackett et al., 1979), and Bannock and Tyro basins (Saager et al.,1993; De Lange et al., 1990).


this ratio is due to very low Mn concentrations (0.06 mmol/L),the lowest Mn content measured during the present study,combined with 0.13 mmol/L of Fe (Table 2). For comparison,Fe and Mn reach respectively 1.4 and 1.52 mmol/L, respec-tively 10 and 25 times more than the hydrothermal brine of theAtlantis II Deep. The Fe/Mn ratio of the Valdivia brine is notexplained by hydrothermal activity, but more probably byextreme Mn depletion of this deep. In view of the absence ofevidence for a hydrothermal component, the high salinities maybe simply due to submarine dissolution of exhumed evaporiteswithout deep circulation. Similar dissolution processes weresuggested for brine formation in the Orca Basin (Gulf ofMexico; Addy and Behrens, 1980), and in the Bannock andTyro basins (eastern Mediterranean Sea; Camerlenghi, 1990).The details of the processes in each deep are discussed below.

5.1.2. Sr isotopic compositions

The Sr in the brines may come from four major sources: seawater, dissolution of biogenic carbonates, marine evaporites,and basalts. The87Sr/86Sr isotopic ratios of sea water (0.7091–0.7092), biogenic carbonates (in equilibrium with sea water)and the Pliocene-to-Pleistocene sedimentary units (0.7089;Burke et al., 1982; Beets, 1992) are very similar. Sea water haslow Sr concentrations (0.09 mmol/L; 8 ppm). Old sedimentarydeposits, mainly consisting of biogenic carbonates, clay, andsilt (Supko et al., 1974), are less soluble than evaporites. TheMiocene evaporite units (mainly halite and anhydrite) have aconstant87Sr/86Sr ratio at 0.70894 (Zierenberg and Shanks,1986), but Sr concentrations that are highly variable (between10 and 2000 ppm; Whitmarsh, 1974; Zierenberg and Shanks,1986) because of variable proportions of anhydrite and halite.Basaltic Sr in hydrothermal fluids decreases the87Sr/86Sr ratiorelative to sea-water (e.g., Albare`de et al., 1981; Vidal andClauer, 1981; Dosso et al., 1991). The Red-Sea basalts have87Sr/86Sr ratios ranging from 0.70257 to 0.70308 (with anaverage of 0.7027) and an average Sr concentration of about140 ppm (Altherr et al., 1988; Eissen et al., 1989; Bosch, 1990).Variations of 87Sr/86Sr ratios between modern sea water(0.70913) and the evaporites (0.70894) are small. In addition,the Sr concentrations of evaporites (as large as 2000 ppm) aresignificantly higher than sea water (8–9 ppm). This is why themixing system may be simplified into a two-component mixingmodel. To calculate the contribution of basaltic Sr to eachbrine, the Sr isotopic system is simplified into this two end-

member model: one component (A) being the oceanic crust andthe second end-member (B) representing Sr from other sourcesessentially reflecting evaporite dissolution (Fig. 3). Thus, the Bend-member has a constant87Sr/86Sr ratio (the same as theevaporite units: 0.70894), but variable Sr concentrations due to(1) highly variable Sr concentrations in the evaporite units(Whitmarsh, 1974; Zierenberg and Shanks, 1986), (2) the vari-able thickness of the Miocene evaporites in the Red Sea (Stof-fers and Ku¨hn, 1974), and (3) the different conditions ofevaporite dissolution depending on the temperature and chem-ical composition of the migrating fluids. The location of thedifferent B end-members fluctuates between the B9 and B0positions (the B9 end-member representing the Nereus brineand the B0 end-member the Valdivia brine) along the dottedline in Figure 3. The contribution of the A and B end-memberscan be calculated with the following relation (Faure, 1986)(Table 6, Fig. 3, S5 sample):

[Sr]S z (87Sr/86Sr)S 5 a [Sr]A z (87Sr/86Sr)A 1 b [Sr]B

z (87Sr/86Sr)B (1)

wherea 1 b 5 1, a 5 basaltic Sr content andb 5 Sr contentof the B end-member.

The Nereus brine yields the highest contribution of basalticSr (a) at 13.7%. Thea values for the Atlantis II and Chainbrines range from 7 to 8%. This contribution is even smaller forthe Suakin (1.5%) and Port-Soudan (0.9%) brines and almostinsignificant for that of Valdivia (0.2%). We have also calcu-lated theaSW factor taking sea water as the end-member B(Table 6) as is the case for common sea water-basalt hydro-thermal systems at oceanic ridges. In this case,aSW is greaterthan in the previous case, but the evaporite Sr contribution hasnot been integrated into the budget calculation.

Zierenberg and Shanks (1986) calculated the contribution ofbasaltic Sr by isotope mass balance using only the isotopiccomposition without concentrations:

XA z [87Sr/86Sr]A 1 XB z [87Sr/86Sr]B 5 [87Sr/86Sr]S

with XA1XB 5 1 (2)

where XA represents the fraction of basaltic Sr, and XB repre-sents the fraction of evaporitic Sr in the brine. The authorsindicated that approximately 31% the brine Sr is derived frombasalts in the lower brine of Atlantis II. This amount is higherthan that obtained here because calculations that do not take

Table 6. Values of the different contributions in the Sr isotopic budget.

Brines [Sr] ppm 87Sr/86Sr [Sr] ppm End-MB a (%) asw (%)

Nereus 46.81 0.70658 32 13.7 33.6Atlantis II 34.97 0.70696 27 7 25.1Chain B 34.17 0.70709 25.1 7.9 24.6Port-Soudan 18.35 0.70850 17.2 0.9 13.2Suakin 14.92 0.70801 13 1.5 10.7Valdivia 12.40 0.70880 12.1 0.2 8.9Thetis 8.04 0.70913 / / /

The fourth column represents the Sr concentration of the variable B end-member corresponding to figure 3. The contribution of basaltic Sr (in %)is calculated according to the equation: [Sr]s z (87Sr/86Sr)s 5 a [Sr]A z (87Sr/86Sr)A 1 b [Sr]B z (87Sr/86Sr)B, wherea 1 b - 1. a 5 contribution ofthe A end-member (basalt) andb 5 contribution of the B end-member (variable). Theasw factor is calculated with sea water as the end-member B.See text for more explanation.


into account the concentrations of the different end-membersare less precise.

Volcanic glasses were observed in biodetrital Red Sea sed-iments (Schneider et al., 1976; Boger and Faure, 1976; Bogeret al., 1980) and in core sediments of the Suakin and Nereusdeeps (Ba¨cker et al., 1975; Jedwab et al., 1989; Bosch et al.,1994; Pierret, 1998). Alteration and leaching of the basalticdetritus may cause an increase of the contribution of the basal-tic Sr in the brines. It will be shown below that the brines aregenerally undersaturated with respect to amorphous silica (sat-uration indices of 0.13 and 0.10 in the brines of Suakin andNereus, Table 7). With a saturation index of 0.92 the brine ofAtlantis II Deep is the only one close to equilibrium, and novolcanic detritus was observed in its sediments. This means thatthe basaltic Sr of the Atlantis II brine is not supplied bydissolved volcanic particles.

5.1.3. The oxygen and hydrogen isotope compositions

The d18O anddD values of the brines during hydrothermalcirculation seem to depend on water/rock interactions. Thedifferences between hydrothermal-fluid (HF) and sea-water(SW) signatures (DHT-SW) due to fluid/basalt interaction at midoceanic ridge (MOR) are:DHT-SW(dD) from 0 to 14 andDHT-SW (d18O) from 0 to12.5‰ (Bowers and Taylor, 1985;Bowers, 1989; Bo¨hlke et al., 1994; Shanks et al., 1995). Duringmigration sea water may also interact with the thick Mioceneevaporites consisting mainly of halite and/or anhydrite withsome units of black shales rich in pyrite and sphalerite, and alsowith the Pliocene and Pleistocene sediments consisting of bio-genic and detrital materials (Supko et al., 1974). Fluid-sedimentinteractions cause isotopic evolution of water:D(dD) from 0 to23 andD(d18O) from 0 to12.5 ‰ (Bowers and Taylor, 1985;Bowers, 1989; Bo¨hlke et al., 1994; Shanks et al., 1995). Theoxygen and hydrogen signatures of the brines also stronglydepend on the signature of the original sea water. Schoell andFaber (1978) estimated the oxygen and hydrogen isotope com-positions of palaeoseawaters from the18O-variations in fora-minifera from Red-Sea sediments combined with the Craigrelation (dD 5 6 3 d18O; Fig. 2). Our data show that thebottom sea waters sampled in 1992 vary with the samplinglocation (Table 3 and Fig. 2). These wide temporal and spatialvariations are a source of uncertainty in the determination ofthe original oxygen and hydrogen isotope compositions of thebrines. However, hydrothermal activity started less than 20,000yr ago (Backer and Richter, 1973) and it is reasonable toassume that over 20,000 yr the isotope compositions of the

Red-Sea water have scattered around the Craig line, betweensea waters from two extreme climatic events (18000 yr BP and8000–5000 yr BP; Fig. 2). Except in the case of Suakin Deep,which has a very peculiar signature, the brine data plot in arelatively restricted area in thedD vs.d18O diagram (Fig. 2). Ahydrothermal origin implies an increase of thedD and d18Ovalues, and interactions between sea water and marine sedi-ments can explain a decrease ofdD. However, the oxygen andhydrogen isotope compositions can also be modified by frac-tionations during secondary processes that occur after brinesare trapped in the deeps.

5.2. Chemical Processes in the Brines

When a brine is trapped in a submarine depression, differentprocesses can take place. Mineral dissolution and precipitationmay occur, e.g., biogenic carbonate dissolution, clay precipita-tion, etc. Such processes may involve oxygen and hydrogenisotopic evolution. O’Neil et al. (1976) showed that the rates ofoxygen and hydrogen isotopic exchange may be more rapid insaline waters, which means that the isotopic compositions ofbrines may be modified during chemical reactions. Thefractionation factors for mineral-water (min-H2O) systems(amin-H2O 5 [1 1 (dmin/1000)]/[11 (dH2O/1000)]) area . 1for oxygen of clays and carbonates anda , 1 for hydrogen ofclays (Savin and Epstein, 1970; O’Neil et al., 1969). Therefore,chemical precipitation of clays and carbonates will tend todeplete water in18O and enrich water in D.

The saturation index (I) of several minerals has been studiedin the different available brines. It represents the ratio of theionic activity product (Q) of any mineral to its solubilityconstant (Ks): I5 Q/Ks. When I is equal or higher to 1, themineral can precipitate. Alternatively, the solution is undersatu-rated with respect to the mineral when I is smaller than 1, andthe mineral may dissolve. The EQPO program (Risacher,1992), based on Pitzer’s model (Pitzer, 1979; 1984; Felmy andWeare, 1986; Palaban and Pitzer, 1987; Greenberg and Moller,1989; Spencer et al., 1990), has been used to calculate activitycoefficients (gi) and activities (ai), and mineral-saturation indi-ces from chemical compositions and temperatures of the brines.The brines of Port-Soudan, Valdivia, Chain and Atlantis IIdeeps are clearly undersaturated with respect to calcite andaragonite, allowing dissolution of biogenic carbonates, whereaspreservation of skeletal carbonates is predicted in the Suakinand Nereus deeps (Table 7). Since the brine in the Atlantis IIDeep is undersaturated with respect to the biogenic carbonates,probably since the beginning of hydrothermal activity (An-

Table 7. Values of the saturation index of three carbonates (calcite, aragonite and dolomite), two sulphates (anhydrite and gypsum) and halite, quartzand amorphous silica in Red-Sea brines. NRSW5 Normal Red Sea Water.

I 5 Q/K NRSW Suakin Port-Soudan Valdivia Nereus Chain Atlantis II

Calcite 3.1 2.08 0.83 0.53 10 0.94 0.55Aragonite 2 3.49 0.54 0.34 6.47 0.61 0.59Dolomite 99.8 91.83 4.42 3.57 105.2 0.94 0.33Gypsum 0.26 0.61 0.7 0.91 0.86 0.88 0.84Anydrite .148 0.7 0.8 1.06 0.733 1.54 2.34Halite 4.13z 1023 0.1 0.33 0.49 0.41 0.89 0.94Quartz 0.148 2.75 12.68 5 1.85 3.07 10Amorph.Si 7.16z 1023 0.13 0.77 0.29 0.1 0.21 0.92


schutz and Blanc, 1993b), we modeled its isotopic evolutiondue to the dissolution of biogenic carbonate. Moreover, theavailable data on Atlantis II Deep are the most complete(history of the deep, volume and age of the brine). The biode-trital sedimentation rate is estimated to be 10 cm/1000 yr onaverage in the Red Sea and constitutes approximately 65% ofbiogenic carbonates (Ba¨cker, 1976; Blanc et al., 1998; Pierret,1998). The surface of the Atlantis II Deep is 60 km2 and thedeed was filled by a stable acidic brine about 15,000 yr ago.During 15,000 yr, 58.53 106 m3 of biogenic carbonates,representing Ncarb

Oxy 5 3.86 3 1012 mol of oxygen, may havebeen dissolved by 3.94 km3 of brines representing NBrine

Oxy 5262.43 1012 mol of oxygen. The dependence of the evolutionof d18O of the water on the quantities of dissolved calcite isgiven by:

d18Ofinal 5 d18Oinitial 3 [NBrineOxy /(NBrine

Oxy 1 NCaCO3Oxy ) NCaCO3

Oxy ]

1 d18OCaCO33 [NCaCO3Oxy /(NBrine

Oxy 1 NcarbOxy) NCaCO3

Oxy ]

5 d18Oinitial 1 18O2,

with

d18OCaCO35 113.9‰/SMOW (O’Neil et al., 1969)(3)

where18O2 is the variation ofd18O. The maximum value of18O2 is 10.2 ‰ if all the biogenic carbonates were dissolved inthe Atlantis II Deep. Chemical precipitation of carbonates frombrine (chemical sedimentation) would reduce this slight change(numerous secondary carbonates have been identified in theAtlantis II sediments; Hofman et al., 1998; Pierret, 1998). Thus,carbonates representing the most important mineral phase dis-solved by the brine (the siliceous species were well preservedin the Atlantis II sediments; Anschutz and Blanc, 1993), havea limited influence on the isotopic compositions of the brine inthe Atlantis II Deep. The oxygen and hydrogen isotope signa-tures are therefore mainly due to the characteristics of thefluid(s) forming the brines and secondary fractionations may beconsidered as having only a minor influence.

Monnin and Ramboz (1996) calculated the anhydrite-satu-ration index in different Red-Sea brines. Using chemical data ofthe brines collected in 1972, 1973 and 1977, the brine of SuakinDeep was found to be undersaturated (I5 0.3–0.4). Using1977 data, anhydrite is approximately in equilibrium (I5 0.93)with the brine of Valdivia Deep. In the case of the lower brineof Atlantis II, these authors showed that anhydrite is not per-manently in equilibrium with the brine; two periods of satura-tion (1966 and 1976) were followed by undersaturation periods.They proposed a relation between anhydrite saturation andhydrothermal activity. Since brines resulting only from evapo-rite dissolution must be saturation or nearby so with respect toanhydrite, they suggested that undersaturated brines contain ahydrothermal component. The 1992 data show that the Suakinbrine is always undersaturated, that the Valdivia brine is main-tained at equilibrium with CaSO4 and that the lower brine ofAtlantis II Deep is in a saturation period (Table 7). However,chemical evolution of a brine formed only by evaporite disso-lution can come under modifications of the equilibrium withanhydrite, such as sulfate reduction by bacteria and Ca precip-itation as secondary carbonates. Thus, the undersaturation of

the Suakin brine with respect to anhydrite is not necessarily anindication of hydrothermal activity.

5.3. Characteristics and Origins of the Different Red-SeaBrines

5.3.1.Thetis

Thetis is the only deep of this study without brine; it is filledwith generic Red-Sea bottom water. The waters of this deep(CTD-rosette B8) can, therefore, be considered to be the seawater end-member.

5.3.2. Atlantis II

The Mg and SO4 concentrations of the Atlantis II brine arelower than in sea water (Fig. 4e), although evaporites (mainlyconsisting in NaCl and CaSO4) were leached as indicated byelevated Na and Cl contents of the brine. Compared to seawater, the Mg and SO4 depletions are typical for hydrothermalfluids (Fig. 4e), and result from interactions between sea waterand the oceanic crust during hydrothermal circulation at tem-peratures above 250°C (cf. section 5.1). In addition, Li, Ba, Fe,Zn, Cu, Mn, K and Rb which characterize hydrothermal fluidsare the most concentrated in the brine of Atlantis II (Fig.4b,g,h,i). The Fe/Mn and Zn/Cl ratios are also characteristic ofa hydrothermal origin (Table 5 and Fig. 4d,j). This brine has thehighest temperature and salinity and, as is typical for a hydro-thermal origin it is also the most acidic (Fig. 4a,c,f). The stableisotope signatures (averaged18O 5 11.07‰ and dD 517.1‰) are close to thed18O and dD values (11.21 and17.4‰, respectively) obtained by Schoell and Faber (1978),and also the values (10.9 ‰ and16.25‰) obtained by Blancet al. (1995). The values clearly express the depletion in heavyisotopes of oxygen and hydrogen compared to present-daybottom sea water. Schoell and Faber (1978) explained thesestable isotopic characteristics by a supply of Red-Sea palaeo-water of the last climatic optimum (8000–5000 yr BP). Theauthors proposed that high-temperature exchanges betweenminerals and waters are improbable. Blanc et al. (1995) sug-gested that the brines resulted from mixing processes betweenRed-Sea palaeowaters of pluvial and warm periods, interstitialwaters expelled from evaporitic sediments, and hydrothermalinputs. Bowers (1989), Bo¨hlke et al. (1994) and Shanks et al.(1995) explained variations in oxygen and hydrogen isotopiccompositions as being due to water/basalt interactions duringhydrothermal circulation at MOR (section 5.1.3.) These frac-tionations systematically involve an increase ind18O of thefluid. In this case, the initial sea water which circulated beforefilling Atlantis II had a lowerd18O value than the present-daybrine. Only certain interstitial waters from Miocene deposits(which have variable values) have similar lowdD and d18Ovalues (Friedmann and Hardcastle, 1974; Lawrence, 1974). Thebrines may, therefore, result from mixing between sea waterand Miocene interstitial water which interacted with basaltsduring hydrothermal circulation. ThedD of the Atlantis II brineis lower than the lowest sea-waterdD from the last climaticoptimum (8000–5000 yr BP). Only interactions between seawater and old sediments can explain thedD measured signa-ture.

The 87Sr/86Sr ratios of the brines are lower than that of sea


water, which requires approximately a 7% addition of basalticSr (Table 6). This Sr isotopic composition (87Sr/86Sr 50.70696) is similar to those reported by Zierenberg and Shanks(1986) and Blanc et al. (1995) and can result from hydrother-mal alteration of basalt by sea water. The relatively low valueof the basaltic contribution was explained by the occurrence ofa large quantity of Sr provided by the leaching of Mioceneanhydrite, consistent with the high concentrations of Na and Clfrom leaching of evaporites. Evaporite dissolution can, there-fore, be envisaged during hydrothermal circulation. The disso-lution of biogenic carbonates in the acidic brine, the dissolutionof Miocene anhydrite and the hydrothermal alteration wouldresult in a high Ca concentration. Thus, the Atlantis II brinesconsist of a salty fluid reacting with basalt at high temperatureindicating significant hydrothermal activity.

5.3.3. Chain deep

The Chain Deep is adjacent and connected to Atlantis II to ata depth of 1980 m. This depth is above the LCL/UCL1 tran-sition zone in Atlantis II, and it does not allow a direct dis-charge of the lower brine from the top of Atlantis II into theChain basin. However, the isotopic signatures of the lowerAtlantis II and Chain brines are very close. Thed18O anddDvalues are of the same order of magnitude (Fig. 2 and Table 3)and both87Sr/86Sr ratios, 0.706966 0.00002 (Atlantis II) and0.70709 6 0.00001 (Chain) and Sr concentrations (0.48mmol/L and 0.47 mmol/L) are similar. Thus, in Figure 3, Chainand Atlantis II lower brine data are very close to one another.The salinity and K, Ca, Rb, Mg, SO4 concentrations are alsosimilar (Fig. 4a,b,e). The top of the LCL brine in Atlantis II andthe top of the Chain brine are at the same depth. Therefore, wesuggest the same origin for both: the brine in the Chain Deepcould have been derived from the lower brine of the Atlantis IIDeep. As direct communication above the seafloor cannotoccur because of the topography, it may have occurred beneaththe deeps along a fracture system. This hypothesis would alsoexplain why the top of the lower brine of Atlantis II Deep andthe top of the brine of Chain Deep are at the same depth.

The temperature and various elemental concentrations (Fe,Mn, Zn) in the brine of Chain Deep are lower than in AtlantisII Deep (Fig. 4h,i). Loss of heat may have occurred duringtransport. Also, some chemical elements could have been fixedin Atlantis II by chemical precipitation, or during transport.Iron is less mobile than Mn in solution, and the Fe/Mn ratio isactually lower in the Chain brine than in the Atlantis II lowerbrine (Fig. 4d). Thus, the chemical and isotopic parameterssuggest the existence of subsurface connections between theAtlantis II Deep and the Chain Deep with the consequence thatthe brine of the Chain Deep may have resulted from an outflowfrom the lower brine of Atlantis II but not by upper directdischarge.

5.3.4. Nereus deep

The Nereus brine has the lowest87Sr/86Sr ratio and thehighest Sr concentration of the studied brines. These parame-ters suggest the highest basaltic contribution (Table 6). TheSO4 concentration of the Nereus brine is lower than in seawater represented by the Thetis reference (Fig. 4e), but that of

Mg is higher. This distinction represents a significant differencerelative to the Atlantis II brine which has low concentrations ofSO4 and Mg relative to sea water. The enrichment in Mgrelative to sea water (Fig. 4e) may result from leaching ofevaporites. Actually, Mannheim (1974) identified Mg-rich lay-ers in the Miocene evaporite deposits. The concentrations of Li,Ba, K, Rb, Mn, Zn, and Cu, and the Zn/Cl ratio, support ahydrothermal influence but with a smaller input than that in theAtlantis II Deep (Fig. 4b,g,h,i,j). However, the contribution ofbasaltic Sr in the Nereus brine (13.7%) appears to be higherthan in the Atlantis II brine (7%). This difference can beexplained by the occurrence of basaltic glasses in the sedimentsof Nereus, as fragments of volcanic rocks were described in thesediments of this deep (Bignell and Ali, 1976; Bonatti et al.,1984; Jedwab et al., 1989; Bosch et al., 1994) but not in theAtlantis II Deep. This volcanic-glass may have been dissolvedby the brines and may consequently have increased indirectlythe basaltic component of the brine. Indeed, the Nereus brine isundersaturated with respect to amorphous silica (Table 7) andsuch materials may be dissolved by the brines and increase boththe Si concentration and the basaltic Sr concentration. A preciseevaluation of the contribution of volcanic glasses dissolution isdifficult, but Sr input from dissolution of volcanic fragmentscannot be neglected in the budget. The temperature of the brineis low (29.9°C; Table 1) and Nereus has the lowest thickness ofbrine (15 m, Table 1). The low temperature may be explainedby a significant loss of heat at the sea water-brine interface dueto the high surface/volume ratio of the brine.

The Sr isotope composition, depletion in SO4, and enrich-ment of some other elements (Li, Ba, K, Mn, Zn, Cu) incomparison with sea water suggest interactions between seawater and oceanic crust above 150°C. In this case, thedD andd18O values of the initial sea water increased and the inducedvariations are:D(dD) from 0 to 14‰, D(d18O) from 0 to12.5‰ (see section 6.1.c). The metal-enriched sediments ofNereus are Holocene (Bonatti et al., 1984), but it is impossibleto date the brine and to determine the age and isotope compo-sition of the initial sea water (cf. section 5.1.3.). Also, thehydrothermal circulations may be cyclical and initial sea watersof different ages with varied isotopic signatures may haveinteracted. However, thedD and d18O values of the Nereusbrine being higher than that of the bottom sea water (Fig. 2),may be explained by interactions between sea water and oce-anic crust.

The brine is saturated with respect to calcite and aragonite,thus thed18O value did not increase because of carbonatedissolution. The residence time of biogenic carbonates withinthe Nereus brine is short because of the sedimentation rate andno isotopic exchanges at equilibrium are possible. In addition,the sediments contain beds of biogenic carbonates and in someplaces, secondary carbonates such as siderite, rhodochrositeand manganosiderite in low quantities. Small amounts of sec-ondary clays were also identified in the Nereus sediments(Bignell and Ali, 1976; Bonatti et al., 1984; Jedwab et al.,1989). Precipitation of carbonates implies a decrease of thed18O values of water, while precipitation of clays implies adecrease of thed18O values and an increase of thedD values.However, such processes seem to have been negligible relativeto fractionation during sea water-basalt interactions. Thus, theoxygen and hydrogen signatures are probably mainly due to


interactions during migration of the hydrothermal fluids. Insummary, it may be assumed that the Nereus brine has beenformed from a fluid of hydrothermal origin with implied fixa-tion of SO4 at a lower intensity than in the Atlantis II Deep(lower Zn, Cu, Li, Ba concentrations in the Nereus brine).Some chemical elements could have been brought to the brineby dissolution of basaltic glass. Evaporites were dissolvedduring the migration of hydrothermal sea water resulting inhigh salinity and high Ca concentrations in the resulting brine(Fig. 4a,b). The high Ca concentration implies that the leachedevaporite probably contained a high proportion of anhydrite.

5.3.5.Port-Soudan deep

The Port-Soudan Deep is filled with a brine having a tem-perature of 35.9°C and a salinity of 214.5 (Table 1) implyingevaporite dissolution. The brine is enriched in Mg and SO4 withrespect to sea water (Fig. 4e). This means that it did not mainlyconsist in a high-temperature hydrothermal fluid, as was thecase for the Atlantis II and Nereus deeps. This brine is alsoenriched in Li, Ba, Zn, Cu, K, Fe and Mn relative to sea water(Fig. 4b,g,h,i), which is characteristic of a hydrothermal com-ponent. In addition, the Fe/Mn ratio is of the same order ofmagnitude as in Atlantis II (Fig. 4d). These data imply that thebrine of the Port-Soudan Deep did not only result from seawater dissolving evaporites. The contribution of basaltic Sr ison the order of 1% (Fig. 3 and Table 6). No basaltic fragmentswere identified in the sediment cores and the basaltic Sr in thebrine probably results from limited sea water-basalt exchangeduring sea water circulation. In this case, it can be envisagedthat the Port-Soudan brine is characterised by a limited hydro-thermal component with a limited sea water/basalt interaction.The fluids mainly supplying the Port-Soudan Deep probablydid not reach a high temperature, and therefore did not pene-trate the oceanic crust deeply.

5.3.6. Valdivia deep

The salinity of the brine in the Valdivia Deep (242) impliesdissolution of NaCl evaporites. We observed a clear enrichmentin SO4 and Mg relative to sea water (Fig. 4e), which excludesthe possibiliy that the brine was formed by a hydrothermal fluidcirculating in a basalt environment at high temperature. Theconcentrations of Zn and Cu are belows detection limits, thebrine was the least concentrated in Mn, Rb, Ba, Ca and Sr(Table 2), and Li and Fe concentrations are low. These ele-ments are typically enriched in hydrothermal fluids. The highFe/Mn ratio of Valdivia brine is due to its low Mn content(section 5.1.1.) but not hydrothermal activity. The content ofthe basaltic Sr (0.2%; Table 6) in the brine is negligible and the87Sr/86Sr ratio (0.708806 0.00001) is close to the Sr isotopiccomposition of evaporites (0.70894; Zierenberg and Shanks,1986). Thus, elemental composition and Sr isotopic data pre-clude significant contribution of a hydrothermal fluid. TheValdivia brine has the same oxygen and hydrogen isotopicratios as present-day bottom sea water and seems to be a recentwater without a clear hydrothermal input (Fig. 2). As for theNereus brine, the enrichment in Mg may result from leaching ofMg-rich evaporites. Thus, the evaporitic units dissolved in this

case, consisted primarily of halite and magnesite and lessanhydrite.

Two processes can explain the formation of Valdivia brine.Either it formed by sea water which dissolved an evaporiticcomponent during subsurface circulation, or as suggested byBacker et al. (1975), the position of the Valdivia Deep on theslope of the main trough, along a probable major transformfracture zone, induced salt-dissolution influenced by tectonicactivity. In both cases, the hydrothermal influence was limited.Nevertheless, the high salinity (S5 242) and the temperature(33.7°C) determined for this brine suggest shallow sea-watercirculation before entering the Valdivia Deep.

5.3.7. Suakin deep

The Suakin brine is the least salty and has the lowest tem-perature of the Red-Sea brines (Table 1, Fig. 4a,c). The pH isthe least acidic and close to that of sea water (Fig. 4f). The K,Fe, Li, and Si concentrations are the lowest, whereas the Mgconcentration is high (Fig. 4b,e,g,h). No sign of a clear hydro-thermal origin was found in the chemical compositions. TheSuakin brine has the most atypical oxygen and hydrogen iso-tope signatures of this study. Thed18O and dD values areclearly different from those of sea water (past and present-day)with dD distinctly enriched in deuterium (dD of about120‰)compared to the other brines. The brine’s isotopic data plots inthe field of the ancient interstitial waters (Fig. 2). We maytherefore adopt the idea proposed by Bath and Shackleton(1984), Friedman et al. (1988) and France-Lanord and Shepp-ard (1992), that the isotopic oxygen and hydrogen compositionsof the interstitial waters in the ancient Red-Sea sedimentsreflect the signature of bottom sea water variations trapped inthese sediments. Savin and Epstein (1970) and O’Neil andKharaka (1976) demonstrated that isotopic exchange betweenminerals and interstitial waters does not occur in the tempera-ture range of sedimentation and early diagenesis (T, 100°C).The brine is saturated with respect to calcite and aragonite, thusno oxygen variation (increase ofd18O) occurred by carbonatedissolution, as was the case for the Nereus brine. Alternatively,the isotopic signature of oxygen and hydrogen can be explainedif the brine was derived from an old water which dissolvedexhumed evaporites without any other influence. The brinecould have formed similarly to the Bannock or Tyro basinsbrines in the Mediterranean Sea (subsurface evaporite dissolu-tion). As is the case for the Suakin brine, the temperaturedifferences between brines and bottom sea water in the Medi-terranean basins are small (brine temperatures 1–2°C higherthan those of the bottom sea water; Boldrin and Rabitti, 1990).Two layers of brine are observed in the Suakin Deep, but it ispossible to obtain a stratification of the brine with a simplesubmarine dissolution of exhumed evaporites, as in the Ban-nock Basin which contains a double-layered brine (Bregant etal., 1990). Anschutz et al. (1999) showed that an abnormalsalinity can persist in such depressions over thousands of yearswithout additional input of salt. In fact, the occurrence oflayered brines in any deep cannot be considered to be a recordof past or present hydrothermal activities.

The Fe/Mn ratio is low (0.014) with values in the interval ofthe Bannock (0.004), Tyro (0.01) and Orca (0.13) brines (Fig.4d). In the absence of hydrothermal inputs, the properties of Fe


and Mn at the sea water/brine interface are dominated bydifferences in the redox potential; as the solubility of Mn inreduced environments is enhanced more than that of Fe. Saageret al. (1993) showed that the particulate/dissolved ratios areabout 1024 for Mn and 0.35 for Fe in the Bannock and Tyrobrines. Thus, the enrichment factors are about 2000 for Mn and10 for Fe relative to sea water. In the Suakin brine, the Feconcentration is very low (0.007 mmol/L) while the Mn con-centration (0.48 mmol/L) is typical of a salty basin withouthydrothermal activity.

As in the Mediterranean basins, the high level of stagnationprevented any mixing process other than molecular diffusion,while the degradation of organic matter led to the developmentof anoxic conditions (Bregant et al., 1990). Jorgensen (1983)showed that eutrophication stimulates sulphate reduction andaccelerates the formation of H2S. Thus, SO4 concentrations inbrines result from a balance between dissolution of evaporiticanhydrite and utilisation of SO4

22 during oxidation of organicmatter. The mechanism of hydrogen sulphide generation is wellknown in natural anoxic basins (Dryssen and Wedborg 1986;Anderson et al., 1988; Luther et al., 1990) and can be repre-sented, according to the Redfield et al.’s formulation (Redfieldet al., 1963), by:

2 CH2O 1 SO422N 2 HCO3

2 1 H2S (4)

In the presence of sulphate-reducing bacteria, sulphate ionsreplace oxygen as electron acceptors during oxidation of or-ganic matter. In the Tyro Basin, the reduced sulphur (H2S andR-SH) represent 5% of the sulphate concentration (Bregant etal., 1990). This is the reason why the SO4/Ca ratio differs fromthe anhydrite ratio and why the brine is not in equilibrium withanhydrite (major constituent in evaporite). As Mg-enrichedevaporitic layers have been identified (Mannheim, 1974), theMg enrichment may result from evaporite dissolution, as sug-gested for the Valdivia brine. The brine is not depleted in Mgor in SO4, which confirms the lack of high temperature basalt-sea water interactions.

The Sr isotopic composition implies a contribution of 1.5%of basaltic Sr (Fig. 3 and Table 6), appears to beconsistent witha hydrothermal influence. However, as in the case of the NereusDeep, basaltic glass was found in the sediment cores, and thebrine of Suakin is undersaturated with respect to amorphous Si(I 5 0.13; Table 7). Thus, the basaltic Sr may be due todissolution of these volcanic fragments. We propose that thebrines of the Suakin Deep have been formed by old sea waterwhich dissolved exhumed evaporites without hydrothermal cir-culation.

6. IMPACT OF HYDROTHERMAL ACTIVITY INRED-SEA DEEPS

The brines of the different deeps record various origins, andthe extent of hydrothermal influences in each of them differ:strong (Atlantis II and Nereus), weak (Port-Soudan) and neg-ligible (Valdivia and Suakin). The occurrence of hydrothermalfluids supplying deeps has clear consequences on the nature ofthe sediments deposited in these deeps.

The Atlantis II Deep has the most metal-enriched sedimentswith metal-sulphide units (Fe, Zn, Cu) and massive Fe/Mn-oxide beds which represent the only deposit of Fe, Cu, Ag and

Au of economic potential (Guennoc and Thisse, 1982; Mustafaet al., 1984; Nawab, 1984). The Nereus Deep contains Fe-Mnsediments (oxi-hydroxides, siderite, Mn-siderite, pyrite,sphalerite, chalcopyrite), and pyrite, hematite and Mn- andFe-carbonates were identified in Port-Soudan cores (Jedwab etal., 1989; Pierret, 1998). These sediments contain higher con-centrations of trace metals (Co, Cr, Zn, Cu) than genericRed-Sea sediments (Pierret, 1998). The chemical and mineral-ogical compositions of the sediments of Valdivia Deep areclose to normal biodetritic sediments of the Red Sea (Ba¨cker etal., 1975; Pierret, 1998). Suakin Deep differs from the others;its brine may have been formed without any hydrothermalcontribution. High seismic activity and many tectonic move-ments have been recorded in the southern part of the Red Seaand the area of the Suakin Deep (Makris et al., 1991). This deepcould have been created by a tectonic event which fractured thebottom of the sea and formed a depression. Evaporites mayhave been exhumed after the tectonic deformation and thedissolution of these evaporitic units by sea water could haveinduced brine formation. No major transform fault was ob-served in the Suakin region by Pautot (1983) who suggestedthat hydrothermally active deeps are preferentially located atthe junction between rift segments and transform faults. Thelack of evidence for a hydrothermal component in the Suakinbrine agrees well with the non hydrothermal origin proposedfor the framboı¨dal pyrite described in its sediments (Pierret etal., 2000).

Hydrothermal circulation depends on a network of faults,porosity of sediments, and the intensity of the geothermalgradient. The fluid does not necessarily reach the oceanic crustand the circulation may be shallow. Sea water and basalt mayinteract at different temperatures according to the migrationdepth of the sea water, strongly influencing the chemistry of thefluid (fixation of sulphate and/or Mg, enrichment in Li, Ba, andother metals). On the basis of Sr and Pb isotopic studies, Volkeret al. (1993) showed that the maturest oceanic crust (N-MORBtype) was located in the central part of the Red Sea, corre-sponding to the area of Atlantis II Deep. Eissen et al. (1989)proposed that short-lived magmatic chambers (100–1000 yr)operate along the Red-Sea rift. Bogdanov et al. (1997) studiedthe mineralogy and chemistry of sediment cores from thespreading centre in the western Wodlark Basin of the westernPacific Ocean. They showed that these centres are not synchro-nous along the spreading axis, each having a specific evolution.Similary, except for the Chain Deep which is connected withthe Atlantis II Deep, the different deeps in the Red Sea mayhave evolved independently.

7. CONCLUSIONS

The varied chemical and isotopic data obtained on brinesfrom different Red-Sea deeps suggest different mechanisms forbrine formation.

1. Some brines result from interactions with oceanic crust athigh temperatures during hydrothermal circulation. For in-stance, the brines of the Atlantis II and Chain deeps resultfrom hydrothermal fluids reacting with basaltic rocks at hightemperatures. It can be shown that the lower brine of theAtlantis II Deep fills the Chain Deep next to it. The brine ofNereus Deep is enriched in the same elements as the lower


brine of Atlantis II, has a lower87Sr/86Sr ratio and it isdepleted only in SO4. Fluids supplying this deep have in-teracted with the oceanic crust at high temperatures, but theintensity was probably lower than in the case of Atlantis IIDeep. These brines with basaltic Sr contributions between 7and 13.7%, are the only ones that unequivocally indicate ahydrothermal influence.

2. The brine of Port-Soudan Deep yields evidence for a slightbasalt interaction. Sea water probably circulated only super-ficially in the oceanic crust before discharging in this deep.

3. The brine of Valdivia Deep mainly results from evaporitedissolution (especially halite) by recent sea water.

4. The brines of Suakin Deep having lower temperatures,salinities, and Fe concentrations differ significantly from theother brines. The oxygen and hydrogen isotopic signaturesare comparable to those of old interstitial waters. SuakinDeep may have formed after tectonic activity, and the brinesseem to have resulted from old sea water which dissolvedexhumed evaporitic units without any hydrothermal contri-bution.

5. In summary, the formation of the Red-Sea brines involvedthree major processes: leaching of Miocene evaporites, con-vective fluid circulation, and interaction with basaltic crustand old sedimentary units. Varied combinations of theseprocesses allow distinction of the brines in the studieddeeps. This conclusion emphasizes the fact that hydrother-mal activity is not identical along the whole Red-Sea axis; itdoes not favour a unique formational model for the brines.A striking feature of the brines is the lack of a relationshipbetween the position of the deeps along the axis and theirevolutionary maturity. Suakin is the closest deep to thecontinuous oceanic crust (in the southern part of the RedSea), but its brines probably formed without hydrothermalactivity.

6. The different deeps appear to have evolved independently,except for Chain Deep which is connected with Atlantis IIDeep. The highest hydrothermal activity seems to have beenin the area of the Atlantis II Deep.

Acknowledgments—We would like to thank sincerely, Dr. G. Faure(The Ohio State University, USA), Dr. S. D. Scott (University ofToronto, Canada), an anonymous reviewer and the Editor in charge ofthe script for their very thoughtful comments and remarks and also fortheir highly improving help in the English presentation. We are alsograteful to the chief scientist of the O.V. Marion-Dufresne, to itscaptain and to its crew for completion of the 1992 REDSED expedition.We thank F. Gauthier-Lafaye for discussions about oxygen and hydro-gen isotopic results and F. Risacher for help and avaibility duringapplication of the EQP0 program. F. Chabaux is especially thanked forhis improving suggestion during the preparation of the script. Manythanks are also due to Robert Rouault and Jean Samuel (Centre deGeochimie de la Surface, Strasbourg) and to Etienne Devantibault(Laboratoire de Tectonophysique, Montpellier) for analytical assis-tance during the course of the study. This work was funded by IFRTP-TAAF and the French Marine Geosciences Committee INSU-95/1.

Associate editor:R. H. Byrne

REFERENCES

Addy S. K. and Behrens E. W. (1980) Time of accumulation ofhypersaline anoxic brine in Orca basin (Gulf of Mexico).Mar. Geol.37, 241–252.

Albarede F., Michard A., Minster J. F., and Michard G. (1981)87Sr/

86Sr ratios in hydrothermal waters and deposits from the East PacificRise at 21°N.Earth Planet. Sci. Lett.55, 229–6.

Alt J. C. (1995) Subseafloor processes in Mid-Ocean Ridge hydrother-mal systems. InSeafloor hydrothermal systems: Physical, chemical,biological and geological interactions(ed. Geophysical Monographby the American Geophysical Union) pp. 85–114.

Altherr R., Henjes-Kunst F., Puchelt H., and Baumann A. (1988)Volcanic activity in the Red Sea axial trough-evidence for a largemantle diapir ?Tectonophysics150,121–133.

Anderson L. G., Dryssen D., and Hall P. O. (1988) On the sulfurchemistry of a super-anoxic fjord, Framvaren, South Norway.Mar.Chem.23, 283–293.

Anschutz P. and Blanc G. (1993a) L’histoire se´dimentologique de lafosse Atlantis II (mer Rouge). Les apports de la micropaleóntologie.C.R. Acad. Sci., Paris se´rie II 317,1303–1308.

Anschutz P. and Blanc G. (1993b) Le rapport NaCl/eau des bouesmineralisee de la fosse Atlantis II (mer Rouge). calcul de la teneur enhalite des se´diments et implication sur la paleótemperature du milieu.C. R. Acad. Sci., Paris se´rie II. 317,1595–1600.

Anschutz P. and Blanc G. (1995a) Chemical mass balances in metal-liferous deposits from the Atlantis II Deep, Red Sea.Geochim.Cosmochim. Acta59, 4205–4218.

Anschutz P., Blanc G., and Stille P. (1995b) Origin of fluids and theevolution of the Atlantis II deep hydrothermal system, Red Sea:Strontium isotopic study.Geochim. Cosmochim. Acta59, 4799–4808.

Anschutz P. and Blanc G. (1996) Heat and salt fluxes in the Atlantis IIDeep (Red Sea).Earth Planet. Sci. Lett.142,147–159.

Anschutz P., Blanc G., Chatin F., Geiller M., and Pierret M. C. (1999)Hydrographic change during 20 years in the brine-filled basins of theRed Sea.Deep Sea Res.46, 1779–1792.

Anschutz P., Turner J. S., and Blanc G. (1998) The development oflayering, the fluxes through double-diffusive interfaces, and thelocation of hydrothermal sources of brines in the Atlantis II Deep;Red Sea.J. Geophys. Res. C.103,27,809–27,819.

Backer H. (1976) Fazies und Chemische Zusammensetzung rezenterAusfallungen aus Mineralquellen im Roten Meer.Geol. Jahrbuch.D17, 151–172.

Backer H. and Richter H. (1973) Die rezente hydrothermal-sedimenta¨reLagesta¨tte Atlantis II Tief im Roten Meer.Geol. Rundsch.62,697–741.

Backer H. and Schoell M. (1972) New deeps with brines and metal-liferous sediments in the Red Sea.Nat. Phys. Sci.240,153–158.

Backer H., Lange K., and Richter M. (1975) Morphology of the RedSea central Graben between Subair Islands and Abul Kizan.Geol.Jahrbuhr.D13, 79–123.

Bath A. and Shackleton N. (1984) Oxygen and hydrogen istope studiesin sqeezed pore-waters, Deep Sea Drilling Project Leg 74, Hole525B: Evidence for Mid-Miocene ocean isotopic change.Initialsreports of the DSDP.74, 697–699.

Baumann A., Richter H., and Schoell M. (1973) Suakin Deep: Brinesand hydrothermal sediments in the deepest part of the Red Sea.Geol.Rundsch.62, 684–697.

Beets C. J. (1992) Calibration of late Cainozoic marine strontiumisotope variations and its chronostratigraphic and geochemical ap-plications. Thesis, Vrije Univeriteit, Amsterdam. 133 p.

Bignell R. D. and Ali S. S. (1976) Geochemistry and stratigraphy ofNereus Deep, Red Sea.Geol. Jahrbuhr.D17, 173–186.

Blanc G. and Anschutz P. (1995) New stratification in the hydrothermalbrine system of the Atlantis II Deep, Red Sea.Geology23,543–546.

Blanc G., Boule`gue J., and Michard A. (1995) Isotope compositions ofthe Red Sea hydrothermal end-member.C.R. Acade´mie des Sciences.320,1187–1193.

Blanc G., Anschutz P., and Pierret M. C. (1998) Metalliferous sedi-mentation in the Atlantis II Deep: A geochemical insight. InSedi-mentation and Tectonics of Rift Basins: Red Sea-Gulf of Aden.(ed.B. H. Purcer and D. W. J. Bosence) pp. 510–524. Chapman et Hall.

Bogdanov Y. A., Gurvich E. G., Lisitzin A. P., Serova V. V., andGorbunova Z. N. (1997) Sediments of the active rift zone in thewestern Woodlark basin and the development of hydrothermal andvolcanic activity.Mar. Geol.142,143–170.

Boger P. D. and Faure G. (1976) Systematic variations of sialic and


volcanic detritus in piston cores from the Red Sea.Geochim. Cos-mochim. Acta40, 731–742.

Boger P. D., Boger J. L., and Faure G. (1980) Systematic variation of87Sr/86Sr ratios, Sr compositions, selected major-oxide concentra-tions, and mineral abundance in piston cores from the Red Sea.Chem. Geol.29, 13–38.

Bohannon R. G. and Eitreim S. L. (1991) Tectonic development ofpassive continental margins of the southern and central Red Sea witha comparison to Wilkes Land, Antartica.Tectonophysics198,129–154.

Boldrin A. and Rabitti S. (1990) Hydrography of the brines in theBannock and Tyro anoxic basins (eastern Mediterranean)Mar.Chem.31, 21–33.

Bonatti E., Colantoni P., Lucchini F., Rossi P. L., Taviani M., andWhite J. (1984) Chemical and stable aspects of the Nereus Deep(Red Sea) metal-enriched sedimentation.Mem. Soc. Geol.27, 59–72.

Bonatti E. (1985) Punctiform initiation of seafloor spreading in the RedSea during transition from a continental to an oceanic rift.Nature316,33–37.

Bosch D. (1990) Evolution geóchimique initiale et prećose d’un rift:Systematique isotopique Pb, Sr, Nd du diapir mantellique de Zabar-gad, de son encaissant gneissique et de son hydrothermalisme: Com-paraison avec les se´diments me´talliferes et les MORB de la MerRouge. Conse´quences geódynamiques et me´tallogeniques. The`se.Montpellier II. 231 p.

Bosch D., Lancelot J., and Boulegue J. (1994) Sr, Nd and Pb isotopeconstraints on the formation of the metalliferous sediments in theNereus Deep, Red Sea.Earth Planet. Sci. Lett.123,299–315.

Bohlke J. K., and Shanks III W. C. (1994) Stable isotope study ofhydrothermal vents at Escanaba Trough: Observed and calculatedeffects of sediments-seawater interaction. InHydrothermal, and Bi-ologic Studies at Escanaba Trough, Gorda Ridge, Offshore NorthernCalifornia (eds. J. L. Morton, R. A. Zierenberg, and C. A. Reiss)U.S.G.S. Bull.2022,223–239.

Bowers T. S. and Taylor H. P. (1985) An integrated chemical andstable isotopic model of the origin of mid-ocean ridge hot springsystems.J. Geophys. Res.90, 12,583–12,606.

Bowers T. S. (1989) Stable isotope signature of water-rock interactionin mid-ocean ridge hydrothermal systems: Sulfur, oxygen, and hy-drogen.J. Geophys. Res.94, 5775–5786.

Bregant D., Catalano G., Civitarese G., and Luchetta A. (1990) Somechemical characteristics of the brines in Babbock and Tyro Basins:Salinity, sulfur compounds, Ca21, F2, pH, As, PO4

32, SiO2, NH3.Mar. Chem.31, 35–62.

Burke W. H., Denison R. E., Hetterington E. A., Koepnick R. B.,Nelson H. F., and Otto J. B. (1982) Variation of sea water87Sr/86Srthroughout Phanerozoic time.Geol.10, 516–519.

Butuzova G. Y., Drits V. A., Morozov A. A., and Gorschkov A. I.(1990) Processes of formation of iron-manganese oxyhydroxides inthe Atlantis II and Thetis Deep of the Red Sea.Spec. Publ. Int. Ass.Sed.11, 57–72.

Camerlenghi A. (1990) Anoxic basins of the eastern Mediterranean:Geological framework.Mar. Chem.31, 1–19.

Charnock H. (1964) Anomalous bottom water in the Red Sea.Nature203,591.

Craig H. (1969) Geochemistry and origin of Red Sea brines. InHotbrines and recent heavy metal deposits in the Red Sea(eds. E. T.Degens and D. A. Ross) pp. 208–242. Springer-Verlag, New York.

Degens E. T. and Ross D. A. (1969) Hot brines and recent heavy metaldeposits in the Red Sea. Springer Verlag, New York., 600 p.

De Lange G. J., Middelburg J.J., Van Der Weijden C. H., CatalanoC. H., Luther G. W. III, Hydes D. J., Woittiez J. R. W., andKlinkhammer G. P. (1990) Composition of anoxic hypersaline brinesin the Tyro and Bannock Basin, eastern Mediterranean.Mar. Chem.31, 63–88.

Dosso L., Hanan B. B., Bougault H., Schilling J. G., and Joron J. L.(1991) Sr-Nd-Pb geochemical morphology between 10° and 17°N onthe Mid-Atlantic Ridge: A near MORB signature.Earth Planet. Sci.Lett. 106,29–43.

Dryssen D. and Wedborg M. (1986) Titration of dilphides and thiols innatural waters.Anal. Chim. Acta180,473–479.

Edmond J. M., Von Damm K. L., McDuff R. E., and Measures C. I.(1982) Chemistry of hot springs on the East Pacific Rise and theireffluent dispersal.Nature297,187–191.

Eissen J. P., Juteau T., Joron J. L., Dupre´ B., Humler E., andAl’Mukhamedov A. (1989) Petrology and geochemistry of basaltsfrom the Red Sea axial rift at 18° North.J. Petrol.30, 791–839.

Faure G. (1986) Principles of isotope geology. Second Edition. JohnWiley and Sons. U.S.A. New-York, 589 p.

Felmy A. R. and Weare J. H. (1986) The prediction of borate mineralequilibria in natural waters: Application to Searles Lake, California.Geochim. Cosmochim. Acta50, 2771–2783.

France-Lanord C. and Sheppard S. M. F. (1992) Hydrogen isotopecomposition of pore waters and interlayer water in sediment from theCentral Western Pacific, leg 129.Proc. O.D.P., Sci. Res.129,295–302.

Friedman I. and Woodcock A. H. (1957). Determination of deuterium-hydrogen ratios in Hawaiian waters.Tellus9, 553–556.

Friedman I. and Hardcastle K. (1974) Deuterium in interstitial watersfrom the Red Sea cores.Init. Rep. DSDP.23, 969–970.

Friedman I. and Hardcastle K. (1988) Deuterium in interstitial waterfrom deep-sea cores.J. Geophys. Res.93, 8249–8263.

Gat J. R. and Confiantini R. (1981) Stable isotope hydrology. Deute-rium and 18O in the water cycle. InInternational Atomic EnergyAgencyVienna.

Greenberg J. P. and Moller N. (1989) The prediction of mineralsolubilities in natural waters: A chemical equilibrium model fot theNa-K-Ca-Cl-SO4-H2O system to high concentration from 0 to250°C.Geochim. Cosmochim. Acta53, 2503–2518.

Guennoc P. and Thisse Y. (1982) Gene`se de l’ouverture de la merRouge et des mine´ralisations des fosses axiales.Doc. Bourr. Rech.Geol. Min. 51, 89. p.

Hannigton M. D., Jonasson I. R., Herzig P. M., and Petersen S. (1995)Physical and chemical processes of seafloor mineralization at Mid-Ocean Ridges. InSeafloor Hydrothermal Systems; Physical, Chem-ical, Biological and geological interactions.Geophysical Mono-graph91, 115–157. Copyright 1995 by the American GeophysicalUnion.

Hartmann M. (1980) Atlantis II deep geothermal brine system. Hydro-graphic situation in 1977 and changes since 1965.Deep Sea Res.27A, 161–171.

Hartmann M. (1988) Atlantis II deep geothermal brine system. Chem-ical processes between hydrothermal brines and Red Sea deep water.Mar. Geol.64, 157–177.

Hartmann M., Scholten J. C., Stoffers P., and Wehner F. (1998a)Hydrographic structure of brine-filled deeps in the Red Sea: Newresults from the Shaban, Kebrit, Atlantis II and Discovery Deep.Mar. Geol.144,311–330.

Hartmann M., Scholten J. C., Stoffers P. (1998b) Hydrographic struc-ture of brine-filled deeps in the Red Sea: Correction of Atlantis IItemperatures.Mar. Geol.144,331–332.

Hofman P., Schwark L., Brachert T., Badaut D., Rivie`re M., and PurcerB. H. (1998) Sedimentation, organic geochemitry and diagenesis ofcores from the axail zone of the southern Red Sea: Relationships torift dynamics and climate. InSedimentation and tectonics of riftbasins: Red Sea-Gulf of Aden.(eds. B. H. Prucer and D. W. J.Bosence). chap. G2, pp. 479–504. Chapman et Hall.

Honnorez J., Laverne C., Hubberten H., Emmermann R., and Muehlen-bachs K. (1983) Alteration processes in layer 2 basalts from DSDPHole 504B, Costa Rica Rift. Init. Rep. DSDP69, 509–546.

Huppert E. H. and Turner J. S. (1981) Double-diffusive convection.J. Fluid Mech.106,299–329.

Issar A., Rosenthal E., Eckstein Y., and Bogoch R. (1971) Formationwaters, hot springs and mineralization phenomena along the easternshore of the Gulf of Suez.Bull. Int. Ass. Sci. Hydrol.16, 25–44.

Jedwab J., Blanc G., and Boule`gue J. (1989) Vanadiferous mineralsfrom the Nereus Deep, Red Sea.Terra Nova1, 188–194.

Jorgensen B. B. (1983) Processes at the sediment-water interface. InThe major biogeochemical cycles and their interactions(eds. B.Bolin and R. B. Cook), p. 477–509.

Lawrence J. R. (1974) Stable oxygen and carbon isotope variations inthe pore waters, carbonates and silicates, sites 225 and 228, Red Sea.Init. Rep. DSDP.23, 939–942.


Le Quentrec M. F. and Sichler B. (1991) 3-D inversion of deep towmagnetic data on the Atlantis II Deep (Red Sea): Hydrothermal andgeodynamic interpretation.Tectonophys.198,421–439.

Luther III G. W., Catalano G., De Lange G. J., and Woittiez J. R. W.(1990) Reduced sulfur in the hypersaline anoxic basins of the Med-iterranean Sea.Mar. Chem.31, 137–152.

Manheim F. T. and Siems D. E. (1974) Chemical analyses of Red Seasediments.Init. Rep. D.S.D.P.23, 923–938.

Matthews D. J. (1939) Tables of velocity of sound in pure water andseawater in echo sounding and sound ranging. Hodrogr. Dep., Ad-miralty, London.

McDougall T. J. (1984a) Fluid dynamic implications for massivesulfide deposits of hot saline fluid flowing into a submarine depres-sion from below.Deep-Sea Res.31, 145–170.

McDougall T. J. (1984b) Convective processes caused by a dense, hotsaline source flowing into a submarine depression from above.Deep-Sea Res.31, 1287–1309.

McKibben M. A., Andes J. P., and Williams A. E. (1988) Active oreformation at a brine interface in metamorphosed deltaic lacustrinesediments: The Salton Sea geothermal system, California.Econ.Geol.83, 511–523.

McKibben M. A. and Williams A. E. (1989) Metal speciation andsolubility in saline hydrothermal fluids: An empirical approach basedon geothermal brine data.Econ. Geol.84, 1996–2007.

Miller A. R. (1964) Highest salinity in the world ocean?Nature203,590.

Miller A. R., Densmore C. D., Degens E. T., Hathamay J. C., ManheimF. T., McFarlin P. F., Pocklington H., and Jokela A. (1966) Hotbrines and recent iron deposits in deeps of the Red Sea.Geochim.Cosmochim. Acta42, 1103–1115.

Monnin C. and Ramboz C. (1996) The anhydrite saturation index of theponded brines and sediment pore waters of the Red Sea deeps.Chem.Geol.127,141–159.

Mustafa H. E. Z., Nawab Z., Horn R., and Le Lann F. (1984) Economicinterest of hydrothermal deposits. The Atlantis II project. Proceed-ings of 2nd international seminar on offshore mineral resources.Brest.1984,509–539.

Nawab Z. (1984) Red Sea mining, a new area.Deep Sea Res.31,813–824.

O’Neil J. R. and Kharaka Y. K. (1976) Hydrogen and oxygen isotopeexchange reactions between clay minerals and water.Geochim.Cosmochim. Acta40, 241–246.

O’Neil J. R., Clayton R. N. and Mayeda T. K. (1969) Oxygen isotopefractionation in divalent metal carbonates.J. Chem. Phys.51,5547–5558.

Palaban R. and Pitzer K. S. (1987) Thermodynamics of concentratedelectrolyte mixtures and the prediction of mineral solubilities to hightemperature for mixtures in the system Na-K-Mg-Cl-SO4-OH-H2O.Geochim. Cosmochim. Acta51, 2429–2443.

Pautot G. (1983) Les fosses de la mer Rouge: Approche geómor-phologique d’un stade initial d’ouverture oceánique reálisee al’aidedu Seabeam.Oceanologica Acta6, 235–244.

Phillips O. M. (1966) On turbulent convective currents and circulationof the Red Sea.Deep Sea Res.13, 1148–1160.

Pierret M. C. (1998) Les saumures et les se´diments de sept fosses del’axe de la Mer Rouge (19–23°N). Etude mine´ralogique,geochimique et isotopique. Roˆle de l’hydrothermalisme. The`se deDoctorat. Universite´ Louis Pasteur, Strasbourg. 383 p.

Pierret M. C., Blanc G., and Clauer N. (2000). Sur l’origine de la pyriteframboıdale dans les se´diments de la fosse Suakin (Mer Rouge).Compt. Rend. l’Acad. Sci.330,1–8

Pitzer K. T. (1979) Theory: Ion interaction approach. InActivity coef-ficient in electrolyte solutions(ed. R. M. Pytkowicz.) Vol. 1, pp.157–208. CRC Press, United States.

Pitzer K. S. (1984) Ionic fluids.J. Phys. Chem.88, 2689–2697.Pottorf R. J. and Barnes H. L. (1983) Mineralogy, geochemistry and ore

genesis of hydrothermal sediments from the Atlantis II Deep, RedSea.Econ. Geol. Mon.5, 198–223.

Ramboz C., Oudin E., and Thisse Y. (1988) Geyser-type discharge inAtlantis II deep, Red Sea: Evidence of boiling from fluid inclusionsin epigenetic anhydrite.Can. Mineral.26, 765–786.

Redfield A. C., Ketchum B. H., and Richards F. A. (1963) The

influence of organisms on the composition of sea water. InThe Sea(ed. M. N. Hill) Chap. 2, pp. 26–77. New York. Wiley-Interscience.

Risacher F. (1992) Geóchimie des lacs sale´s et crouˆtes de sel del’altiplano bolivien.Sci. Geolog. 45, 135–219.

Rosenbauer R. J., Bischoff J. L., and Radtke A. S. (1983) Hydrothermalalteration of graywacke and basalt by 4 molal NaCl.Econ. Geol.78,1701–1710.

Rosenthal A., Jones B. F., and Weinberger G. (1998) The chemicalevolution of Kurnub Group paleowater in the Sinai-Negev prov-ince-a mass balance approach.Appl. Geochem.22, 1–17.

Saager P. M., Schijf J., and De Baar H. J. W. (1993) Trace-metaldistribution in sea water and anoxic brines in the eastern Mediter-ranean Sea.Geochim. Cosmochim. Acta57, 1419–1432.

Sackett W. M., Brooks J. M., Bernard B. B., Schwab C. R., Chung H.,and Parker R. A. (1979) A carbon inventory for Orca basin brinesand sediments.Earth Planet. Sci. Lett.44, 73–81.

Savin S. M. and Epstein S. (1970) The oxygen and hydrogen isotopegeochemistry of clay minerals.Geochim. Cosmochim. Acta31, 25–42.

Schneider W., Koch G., and Lange G. (1976) Vulkanische Gla¨ser insedimenten des Roten Meeres.Geolog. Jahr.D17, 107–134.

Schoell M. and Stahl W. (1972) The carbon isotopic composition andthe concentration of the dissolved inorganic carbon in the Atlantis IIdeep brines (Red Sea).Earth Planet. Sci. Lett.15, 206–211.

Schoell M. and Hartmann M. (1973) Detailed temperature structure ofthe hot brines in the Atlantis II deep area (Red Sea).Mar. Geol.14,1–14.

Schoell M. (1974) Valdivia VA01 Rotes Meer-Golf Von Aden. Hy-drography II1 III, Daten. Bundesanstalt fu¨r Bodenforschung. Han-nover. 1063 pp.

Schoell M. and Faber E. (1978) New isotopic evidence for the origin ofRed Sea brine.Nature275,436–437.

Schoell M. and Hartmann M. (1978) Changing hydrothermal activity inthe Atlantis II deep geothermal system.Nature274,784–785.

Scholten J. C., Stoffers P., Walters P., and Plu¨ger W. (1991) Evidencefor episodic hydrothermal activity in the Red Sea from the compo-sition and formation of hydrothermal sediments, Thetis Deep.Tec-tonophys.190,109–117.

Shanks III W. C., Bo¨hlke J. K., and Seal II R. R. (1995) Stable isotopesin Mid-Ocean Ridge hydrothermal systems:interactions between flu-ids, minerals, and organisms. InSeafloor Hydrothermal Systems:Physical, Chemical, Biological, and Geological interactions.(ed.S. E. Humphris, R. A. Zierenberg, L. S. Mullineaux, and R. E.Thomson) pp.194–221. American Geophysical Union.

Smeed D. (1997) Seasonal variation of the flow in the strait of bab alMandab.Oceanolog. Acta20, 773–781.

Spencer R. J., Moller N. and Weare J. H. (1990) The prediction ofmineral solubilities in natural waters: A chemical equilibrium modelfor the Na-K-Ca-Mg-Cl-SO4-H2O system at temperature below25°C.Geochim. Cosmichim. Acta54, 575–590.

Stoffers P. and Ku¨hn R. (1974) Red Sea evaporites: A petrographic andgeochemical study.Init. Rep. D.S.D.P.23, 821–847.

Supko P. R., Stoffers P., and Coplen T. B. (1974) Petrography andgeochemistry of Red Sea dolomite.Init. Rep. D.S.D.P.23,867–878.

Thompson E. F. (1939) Chemical and physical investigations. Theexchange of water between the Red Sea and Gulf of Aden over the“sill.” John Murray Expedition, 1933–34. Sci. Rep.2, 83–103.

Thompson J. M. and Fournier R. O. (1988) Chemistry and Geother-mometry of Brine Produced from the Salton Sea Scientific DrillHole, Imperial Valley, California.J. Geophys. Res.13, 165–173.

Vidal P. and Clauer N. (1981) Pb and Sr isotopic systematics of somebasalts and sulfides from the East Pacific Rise at 21°N (projectRITA). Earth Planet. Sci. Lett.55, 237–246.

Volker F., McCulloch M. T., and Alther R. (1993) Submarine basaltsfrom the Red Sea: New Pb, Sr and Nd isotopic data.Geophys. Res.Lett. 20, 927–930.

Von Damm K. L., Edmond J. M., Measures C. I., and Grant B. (1985)Chemistry of submarine hydrothermal solutions at Guyamas basin,Gulf of California.Geochim. Cosmochim. Acta49, 2221–2237.

Von Damm K. L. (1988) Systematics of postulated on submarinehydrothermal solution chemistry.J. Geophys. Res.93, 4551–4561.

Von Damm K. L. (1995) Controls on the chemistry and tem-


poral variability of seafloor hydrothermal fluids. InSeafloorHydrothermal Systems: Physical, Chemical, Biological,and Geological Interactions(eds. S. E. Humphris, R. A.Zierenberg, L. S. Mullineaux, and R. E. Thomson).Geophysical Mongraph91, 222–247. American GeophysicalUnion.

Zierenberg R. A. and Shanks III W. C. (1983) Mineralogy and geo-chemistry of epigenetic features in metalliferous sediment, AtlantisII deep, Red Sea.Econ. Geol.78, 57–72.

Zierenberg R. A. and Schanks III W. C. (1986) Isotopic contraints onthe origin of the Atlantis II, Suakin and Valdivia brines, Red Sea.Geochim. Cosmochim. Acta50, 2205–2214.


Chemical and isotopic (87Sr/86Sr, 18O, D) constraints to the ...

Documents