Superheating in the Red Sea? The heat-mass balance of the Atlantis II Deep revisited

Earth and Planetary Science Letters, 97 (1990) 190-210 Elsevier Science Publishers B.V., Amsterdam - Printed in The Netherlands

190

[XLeP]

Superheating in the Red Sea? The heat-mass balance of the Atlantis II Deep revisited C. R a m b o z a n d M. D a n i s * Centre de Recherches P~trographiques et G~ochimiques, B.P. 20, 54501 Vandoeuvre-l~s-Nancy (France) Received August 3, 1989; revised version accepted November 9, 1989 Atlantis II is the only hydrothermally active of the five Central Red Sea deeps, hot new brine being supplied by a geyser spring. It is filled with two stratified anoxic brine layers, namely, the Lower Conductive Layer (LCL) at the bottom and the Upper Conductive Layer (UCL) above it. The other deeps were filled by brine overspill. Hydrological and geochemical data on the Atlantis II Deep show that, between 1965 and 1977, only negligible amounts of lower brine have spilled over towards Chain A Deep and that, in contrast, the lower brine component of the UCL has increased significantly. The heat-mass balance of the Atlantis II Deep lower brine between 1966 to 1977 is considered. Mass balance calculations are based on the published bathymetry of the Deep and on the measured rising rate of the interface between the UCL and the LCL, including a possible loss of mass out of the system. Heat balance equations involve (1) the conductive flow of heat at the bottom of the Deep, (2) the heat lost by conduction through the lower interface, (3) the flow of heat due to hot brine advection and (4) the heat lost by lower brine advection out of the LCL. The equation of Pitzer et al. (1984) [57], extrapolated above 300°C, is first used to calculate the cp's of the hydrothermal brine: it predicts moderately increasing cp's with temperature. Heat-mass balance equations, solved with this cl, model, yield a spring temperature of 461 ° C for the time interval 1966-1977. This temperature is unrealistic as it is = 70 ° C higher than 390 o C, the boiling temperature of the lower brine at 220 bar. Besides, it is shown that none of the parameters in the calculation, other than Cp, can attain values as to decrease 7", below 390 ° C. The modified equation of Born predicts a near exponential increase of the hydrothermal brine cp's above 300 o C. Spring temperatures of 342°C or 353°C are calculated with this cr model, depending whether the mass lost towards the UCL is taken into account or not. These temperatures compare favorably with the mean temperature of 330 ° C obtained for the trapping of fluid inclusions in epigenetic anhydrites from the discharge zone sediments [21]. Hence it is concluded that (1) the changes in the mass and temperature of the Atlantis II Deep lower brine between 1965 and 1977 cannot be interpreted in terms of a realistic spring temperature, unless the hydrothermal fluid has presented very high heat transport properties above 300 o C; (2) this can be either in the stable region below the boiling point, as predicted by the equation of Born, or in the metastable region above 390 ° C: superheated liquids indeed present heat capacities which increase exponentially with temperature. Two periods in the recent hydrothermal activity of the Atlantis I1 Deep are distinguished: (1) From 1965 to 1976, the Deep has received an excess heat supply, e.g., compatible with the influx of superheated liquids. The spring flow rate, including the mass of lower brine expelled towards the UCL [29], was around 367 kg s t. (2) The hydrothermal activity then decreased from 1976 to 1979. The influx of fluids with abnormal heat transport properties was greatly reduced and the mean spring flow rate probably dropped to = 280 kg s-1.

1. Introduction The sediments at the bottom of the Atlantis II Deep form one of the largest known submarine deposits of Zn, Cu, Ag and Au [1]. They have

Contribution C.R.P.G. No. 000. * Present address: Universit6 de Bordeaux 1, Institut de Technologie A, G6nie M~canique, 33405 Talence Cedex. 0012-821X/90/$03.50

© 1990 Elsevier Science Publishers B.V.

been forming there for the past 11,000 years. They have been preserved because Atlantis II Deep, like the other Central Red Sea deeps (Discovery, Valdivia, Chain A and Chain B), is filled with reducing stratified brines which separate the underlying sulfides from the overlying oxidizing oceanic waters of the Red Sea. In contrast, most sulfides deposited in the open ocean are later dispersed, dissolved or oxidized [2]. Stratified hot brine pools are one way that aids the formation

S U P E R H E A T I N G IN T H E RED SEA?

191

and preservation of economic base metal sulfide deposits in oceanic environments. Geochemical studies have established that, at 1 3 ° N and 2 1 ° N East Pacific Rise (EPR) and in the five Central Red Sea deeps, metals are deposited from hydrothermal fluids enriched in Ca and depleted in Mg, 02 and SO4 relative to seawater [3-5]. Along the EPR, the mineralizing fluid is modified seawater having acquired its chemical and isotopic characteristics by interacting with basalts at high temperature [6-8]. Temperatures of the discharging brine of up to 358 ° C at 21 ° N EPR have been measured directly [9]. In contrast, the conditions of hydrothermal feeding of the hot brine pools from the Central Red Sea, and particularly that of the Atlantis II

TABLE

Deep, are still controversial in spite of 30 years of oceanographic studies. Although there is no doubt that the brines in the Atlantis II Deep partly originate from the Miocene evaporites deposited on the flanks of the rift [10-12], it is still debated whether their reducing and metalliferous potentials result from the interaction with newly-formed basalt [13,14] or with organic-rich shales within or above the evaporites [15]. The temperature of brine discharge at the bottom of the Atlantis II Deep has never been measured directly. Mean temperatures have been indirectly estimated to be in the range 1 1 5 - 2 5 0 ° C , based on mineral equilibria, geochemistry of the brines and heat-mass balance calculations (see the reviews in [14,16] and Table 1). These temperatures are much lower than those

1

Changes in the temperature and volume of tlae lower and upper brines of the Atlantis I1 Deep from 1965 to 1980, and their interpretation in terms of the temperature and flow rate of the hydrothermal spring feeding the Deep Reference

System

At

AT

[36]

UCL + LCL

11-65

0.34

AD

Changes ~

Me

D°

15-20

36 ~

0

T~

q~ -

X-66 [50]

UCL + LCL S = 49.3 d Xu=47

[48]

[39]

X-66 11-71

d, x 1 = 3 3

0 1.4

0

0

0

> 114

0

0

0

< 360

UCL:

+ 1.3

0

0

0

0

> 104

LCL:

+ 0.6

1.4

a

UCL + LCL S = 49.3 d x . = 4 7 d; xa = 33 d UCL + LCL S = 42

U C L : + 1.3 LCL: +0.6

3.1

IV-71 III-72

0.75

0.3

0

0

0

210

0.19

LCL

V-65

0.48

0.22

0

0

0

215

0.28

S = 43.5

XI-77

0

0

0

+

140

0.56

0

0

230

+

280

0.56

0.22

20 ~

150

0

490

0.35

x u + x I = 61.9 [23]

x 1 = 69 [49]

LCL

IV-65

UCL:

+ 1.3

S = 40

XI-77

LCL:

+0.6

x u = 30; x 1

=

78

This

LCL

II-65

0.40

0.22

20 e

168 ':

0

353

0.33

work

S = 43-43.4

XI-77

0.40

0.22

20 e

168 ':

*

342

0.39

x 1 = 63.3-64.6 A t = time length; A T = heating rate ( ° C y r - 1 ) ; A D = rising rate of the lower interface (m yr 1); ~,~ ( ¢ , o ) = conductive input (output) of heat (HFU); M e = mass of fluid expelled out of the system; T~ and q~ = calculated temperature ( ° C ) and rate o f f l o w (103 k g s 1) o f the inflowing brine respectively; x u and x I = t h i c k n e s s ( m ) o f the upper and lower brines respectively; S = surface o f the system ( k m 2 ) . ~.bx Heat lost by pure conduction through the lower interface, the thickness of which is taken as 5 m , 1.5 m a n d 1 m respectively; d a f t e r [33]; e a f t e r [36]; + = considering that half the mass of the injected brine is lost through the lower interface; * = lower brine lost towards the UCL after [30].

192

derived from recent mineralogical a n d / o r fluid inclusion data on high-temperature sulfide-sulfate assemblages in epigenetic veins cross-cutting the sediment of the southwest basin of the Atlantis II Deep [16-22]. Oudin et al. [20] showed that these veins are equivalent to the high-temperature hydrothermal assemblages at 2 1 ° N EPR. Fluid inclusions in the epigenetic barites and anhydrites demonstrate that the crystals were precipitated directly from boiling hydrothermal solutions at temperatures up to 390-403° C [20,21]. It is worth noting that Kaplan et al. [15] measured isotopic fractionation between interstitial sulfates and total sulfides, from the same stratigraphic level of the sediments, in the range of 1.011-1.017, compatible with an isotopic temperature of between 400 o and 550°C. Monin and Plakhin also proposed a discharge temperature of 490 o C, based on heat-mass balance calculations (Table 1). The recent directly observed fluctuations of the temperature and volume of the lower brine provide one of the best ways of averaging the heat transfer over the Atlantis II Deep, and hence to calculate mean hydrothermal discharge temperatures [22]. The many generations of fluid inclusions in anhydrite and barite, with different temperatures and salinities [20,21], are compatible with the presently observed temperature and salinity fluctuations of the lower brine [23]. Geochemical studies of the various sulfide-bearing levels demonstrate that the characteristics of the hydrothermal spring (temperature, chemistry) have been relatively uniform over the past 15,000 years [14]. This paper reconsiders the temperature of the Atlantis II Deep hydrothermal spring using heatmass balance calculations. It explains why previous calculations yielded low temperatures, compared with recent thermometric information derived from mineralogical, isotopic or fluid inclusion data in epigenetic veins. Data on the structure and hydrology of Atlantis II, Discovery, Chain A and Chain B Deeps over the past 15 years provide the bases for reconstructing the processes accounting for the geothermal activity in the Atlantis II Deep, they are therefore summarized below. Revised equations for the heat-mass balance of the Atlantis II Deep are proposed. The equations are solved in terms of the temperature and flow rate of brine discharge in the Atlantis II Deep.

c . R A M B O Z A N D M. D A N I S

2. The hot brine pools in the Atlantis II Deep area

The Central Red Sea Deeps (Discovery, Atlantis II, Valdivia, Chain A and Chain B) have in common a stratified structure characterized by well mixed layers with, both horizontally and vertically, nearly constant composition and temperature, alternating with interfaces with large temperature gradients [24]. Only the structures of Atlantis II, Discovery, Chain A and Chain B Deeps are discussed below (Figs. 1-3; Tables 1 and 2). The Valdivia Deep is unrelated to the Atlantis II Deep because, unlike other Central Red Sea Deeps, it is formed by dissolution of Miocene evaporites [25,26].

2.1. Physical interpretation of the deeps The physical interpretation of the observations on the stratified hot brine pools from Central Red Sea as established by Turner [22,27] are summarized below. Stratified fluid layers with sharp interfaces have been obtained in the laboratory on heating uniformly a tank from below under conditions of high heating rate and low salinity gradients. Their formation is a direct consequence of the diffusivity of salt in water being about two orders of magnitude smaller than that of heat. The stability of the heat-salt double diffusive layers depends on the ratio D of the separate contributions of salinity and temperature to the density difference across the interface. As D is experimentally increased from 1 to 7, the stratified layers become increasingly stable, with little dependence on the heating rate. The transfer of heat and salt through the sharp interfaces dominantly occurs by molecular diffusion at increasing D ratios, in contrast to convective mixing in the layer. For D ratios between 2 and 7, the flux of heat through the interface is observed to be a hundred times larger than that of salt [27]. The Central Red Sea stratified systems are rather different from the experimentally studied ones. They have formed under conditions of low heating rate and comparatively high salinity gradients. Only the lower interface of the Atlantis II Deep, being so sharp (about 1 m thick [23]), can be considered as equivalent to the interfaces studied experimentally, and interpreted as a diffusive interface separating two strongly convective layers [27]. However, the characteristic D ratio of the lower interface of the

SUPERHEATING IN THE RED SEA?

193

38*04

38°02

i

I

I-

I

' ~. ~ - ' ~ " 'N

38°06

[

I

I

I

"[ depth (m) 1950 -t

21°26I

2049.4

depth (m) 1950 2000 : ~ . . ~

150 North

2049.4 2050 . . . . . . . . . . . 20

~\\\\\\\\\\\\\\\\\\~

30

40

50

60

2100 2150

West Basin

r<°O>l i

i

20

40

30

50

60 z I

k

AtlantisIT~ 21"22 i

r

depth (m)

21°20 - 2100 1

~'"'.

-

21501- Southwest Basin , 20 30 40

I

; ,\

~1 ,'~) I r"~~ depth (m) ~l 1950,

T (°C] ' :~0 60

') '\

Chain

A ....

t K\\Xk'NN\\N-'~.

21"18 I/

~ lk\\\\\\\\\\\\"~m~

~. ,--,I

~1 k N \ X X X N X \ \ X \ N N N N N ' ~ ~

"

'~

J "" t

....

2049.4 .............

z u bc

"

/

- t ,t - e,

2100 •

~-

2150 Ea^, Bas,n

r

~\\\\\\\\\\\\\\\\\'~

J/ '

v/It, , Chain B

'~\\\\\\\\\\\\\\\\\'~

// |I

',\\\\\\\\\\\\\\\\\'~ ' - - -" ~ . N N \ \ \ \ \ \ x ~ t

l.

20

21°18~0 II I I' ' •

~,.

I D I s c o v a r y.~NNNN\NNN\NNN\\NN~,

N•

in

....

/

I

2Rm '

~ , I

I

I

30

\

{ T (°C) 4o

50

00

~--~top of the LCL 1 .--'limit of the UCL / "

I

1

/

Fig. 1. Simplified bathymetnc map of the brine pools in Atlantis II Deep area (after [40]) and characteristic temperature-depth profiles of the four basins of the Atlantis II Deep (after [24]: stations 50, 61, 75, 89, April 1971). Depths are corrected for higher sound velocity in brines [65].

Atlantis II Deep has remained = 17 between 1965 and 1977; this is in excess of the D values of the most stable diffusive interfaces investigated so far in the laboratory (up to 12). MacDougall [28] has recently developed a model showing that diffusive interfaces with characteristic D ratios beyond 20 are stable. The qualitative conclusions concerning the Central Red Sea stratified pools, derived from

the experimental studies [22], are therefore fully justified. The lower and upper interfaces in the Atlantis II Deep (see Fig. 2 and Table 2) are quite stable because the heating rate of the Deep is moderate and the salinity gradient across the interface is high. The two interfaces in the Discovery Deep are even more stable due to smaller temperature

194

C R A M B O Z A N D M. D A N I S

(November 1977)

ATLANTIS i

i

I

I

A

1 9 7 5 1 ~

,•E

~_ 2000

~

~.

2°25 I

U.I.

U.I.

'~_ L.I.

L.I.

(_3 2050 2048.0

2048.0

2075 I

4'

3O

I

s'

5I

8o

1~)

T (°C)

2020

i ',..._

2067

30

(February 1972)

DISCOVERY

2060

215

15

Salinity (wt %)

~

U.I.-

U.I.

L.I

~'~---.~

2057

L.I. ""--7

~21oo 2180

i

2220

2260

,

,

444

2300 20

,

44.6

,

,

B

44.8

I 30

I 40

J

I

5

50

10

T (*C) CHAIN

1940

I

I

r

i

t

15 20 Salin~y (wt.%)

30

(April 1971) i

04-1971

i

v .c: 1980

C ~ 2049.4

o

I

25

A

2049,4

-)

2060

20

I

30

I

40 T (*C)

I

50

60

/

I

10

I

15 /0 Salinity (wt %)

I

25

30

U.l.=upper interface L.l.=lower interface Fig. 2. Temperature-depth (left) and salinity-depth profiles (fight) in Central Red Sea Deeps: Atlantis II (A), Discovery (B), Chain A and B (C). The temperature profiles, corrected for higher sound velocity in brines [65], are drawn after [24]. Salinity profiles are tentatively reconstructed after chlorosity, salinity and chlorinity data (Atlantis II Deep [23], Discovery Deep [37], Chain A Deep [31], respectively).

SUPERHEATING 1N THE RED SEA?

195

TABLE 2 Temperature-depth structure of Atlantis II and Discovery Deeps in February 1972 (after [23,24]) and in October 1966 (numbers in brackets, after [22]). Only the 1972 depth data are corrected for higher sound velocity in brines [65] Atlantis II Deep (southwest basin) depth (m) Normal seawater

X

T

1992

Discovery Deep depth (m)

A T/A x

< 23

X

1987

r1992 (1984)

T

A T/A x

< 23

1987 (1986)

Transition layer

2019 (2009) UCL

]"

\ Lower interface LCL

27 (25) 23 (28)

0.9 (0.9)

> 2050 (2042)

0.3 ( 0 . 4 )

48 (37)

50.8 (44)

3

2049 (2037) 2050 (2042)

2035 (2023)

(4)

35.75 (36)

2038 (2027) 1 (5)

9 (2.4) 59.8 (56)

T = temperature ( ° C); X = thickness of the layer (m); A T / A x (lower) convective layer.

differences across them (Table 2). The composition and temperature are maintained constant by convection in the two layers of the Atlantis II Deep. The transfer of salt and heat through the sharp lower interface of the Atlantis II Deep proceeds mainly by diffusion. The former flux can be neglected as compared to the latter, on account of the low diffusivity of salt relative to that of heat. Brewer (in [29]) interpreted an increase of 2.2%o of the salinity of the upper brine from 1966 to 1971 in terms of a mean net flux of salt through the lower interface of around 6 10 -8 g cm -2 s -1. Between 1966 and 1977, on account of their temperatures and chlorinities, both the U C L and the transition layer with Red Sea Deep water (Fig. 2) have always presented a temperature excess of -- 8-12 ° C, as compared to conservative mixing of seawater with LCL and UCL, respectively [30]. This provides evidence for large conductive heat fluxes at the top and bottom interfaces of the UCL.

2.2. Contrasting structure and origin of the deeps Atlantis II, Discovery, Chain A and Chain B Deeps [31] have many dissimilarities despite the fact that they contain a convective layer in contact with the sediment i.e., the Lower Convective Layer ( = LCL), with a similar chlorinity (Table 2): (1) The temperature of the LCL is different from one Deep to another, and is highest in the Atlantis II Deep (Fig. 2). In addition, the temper-

2057 (2042) > 2057 (2042)

22 (15)

0.4 ( 0 . 6 ) 44.8 (45)

= temperature gradient across the interface; U C L (LCL) = upper

ature gradient across both the upper and lower interfaces is steeper in Atlantis II than in Discovery Deep (Table 2), due to thinner interfaces and the higher temperature of the convective layers in the former Deep. (2) The LCL is found at greater depth in Discovery and Chain B Deeps than in Atlantis II and Chain A Deeps (Fig. 2). (3) The temperature of,the LCL in the southwest basin of the Atlantis II Deep remains constant down to the surface of the sediment whereas a 50 m thick cooler layer is identified at the bottom of the Discovery Deep [24,32] (Fig. 2B). (4) The Atlantis II Deep Upper Convective Layer ( = UCL) is 7 times thicker than that in Discovery Deep and is found at greater depth (Table 2; Fig. 2) [23,33]. (5) Many authors have noted that the temperature of the Atlantis II Deep brines changes over periods of a few minutes to a day [34-36] due to small scale convection currents [29]. The structure of the Atlantis II Deep is also highly changeable over the longer term: the temperature of the U C L and of the LCL has evolved continuously over 15 years and the volume of the LCL has increased simultaneously (Fig. 3A). From 1972 to 1977, Chain A Deep has presented a more limited activity, marked by an increase of 1.4°C in the temperature of the lower brine layer (Fig. 3C). The m a x i m u m temperature of the LCL in Chain B Deep is only known from April 1971 to February

196

C

A

2048.0

rLc, 8,7,'~,""

~ T (°cl

62r

59, 4

X(m)j

T

t2048

5~2-

T (°C)

46~44.7

...

55.92

x,,

.'=490

2050.5

/ 2050.5

,"

46 ~/ / /11/ 4z /I/ 42

2050

gli

~

~

-12057

38

r c,

52.25

510

, ,~..5,.05

481

44~7

2049

2049.4~' ~.5

"~

E""

I49.87 ~ ~.

/" /"

~

44.7-4 =

"2055

/

5(3

DANIS

TLCL - 44.79

413

,'

52

M.

. . . .

42

54

AND

X(m)

DISCOVERY

44 ~ . 44~72

56L,

RAMBOZ

36

36 ~ . . . . . . . ,

/

,',887

//.

,

1965

,

-~,

,

EEl

-- " 3 5 . 6 ,

,

,

.

.

.

.

1970

1975

1970

1975

.

48.39

/

I

2050

TUCL

} T (°C)

744"26 /

v/

ATLANTIS

41.2

II

40 _..,. I

1965

I

I

I

l_

I

1970

I

I

I

I _.._1

I

I

I

1

1975

1965

Fig. 3. Temperature of the lower brine (TLcL: circles), of the upper brine (TucL: triangles) and depth of the lower interface (X: squares) in Atlantis II (A), Discovery (B) and Chain A and B Deeps (C) from 1965 to 1979. The ticks on the triangles correspond to the maximum temperature measured in the convective layer (data after [23, 24, 34, 37, 43, 49, 50, 66]).

1972: it has remained constant during that period (Fig. 3C). Similarly, Discovery Deep has been essentially inactive, as the temperature of both convective layers has remained unchanged. The volume of its LCL has slightly decreased over the same period (Fig. 3B). The Preussag data summarized in Fig. 3B do not support the statement by Bubnov et al. [37] that the lower interface in the Discovery Deep has deepened by 18 m from October 1966 to June 1976. The difference between the structure of the Atlantis II Deep and that of the other deeps (Figs. 2 and 3; Table 2) points to more active convection in the former Deep. The measured very high heat flows of up to 60 H F U (2.5 W m -2) in the southwest basin of the Atlantis II Deep [36], and the high spring flow rate implied by the hydrological data over 15 years [23] confirm that the Atlantis II Deep is presently an active submarine geothermal system [39]. Pugh [32,38] and Turner [22] first suggested that the layering in the Atlantis II Deep could not be maintained by conductive heat transfer through the sediment and that the large temperature increase of the LCL had resulted from convection. Schoell and Hartmann [31] showed that there is no direct connection between the pools of lower brine in Discovery, Chain A

and Atlantis II Deeps. Backer et al. [25] suggested the possibility of a subsurface connection between the lower brines in these Deeps. However, the similar chemical composition of the lower brines in Discovery, Chain A, Chain B and Atlantis II Deeps [31,40], the measured very high heat flows in the Atlantis II Deep as compared to those in the Discovery Deep [36], and the cooler layer at the bottom of the latter Deep suggest that the lower brine in the Atlantis II Deep may have spilled over into the Discovery and Chain Deeps (Fig. 1) [22,32,36]. According to this interpretation, the recent temperature evolution of the four Deeps (Fig. 3) suggests that some Atlantis II Deep lower brine has flowed over towards Chain A Deep between 1972 and 1977, but none towards Chain B or Discovery Deeps. The observed 5 m increase in the thickness of the Chain A lower brine between 1965 and 1977 [24], however, represents a negligible fraction of the volume increase of the lower brine in the Atlantis II Deep.

2.3. Localized hydrothermal feeding of A tlantis H Deep Erickson and Simmons [36] proposed that the supply of heat at the bottom of the Atlantis II Deep was uniform but, because of his laboratory

197

S U P E R H E A T I N G IN T H E R E D SEA?

experiments, Turner [22] ruled out this possibility. If the lower brine layer in the Atlantis II Deep had formed by uniform heating, it should be thinner than actually observed, given the low heating rate. Turner [22] suggested that both heating from below and local injection of brine into the bottom of the Deep was the cause of the observed layering. The most direct evidence for a local and still active influx of hot brine in the southwest basin is provided by the continuous survey of the temperature-depth structure of the Atlantis II Deep from 1965 to 1980. The Preussag group has studied in detail the horizontal and lateral variations of the brine temperature of the Atlantis II Deep from 1965 to 1977 [23,31,43,44]. The hydrology of the Deep can thus be reconstructed [39]. Schoell and H a r t m a n n [31] were the first to demonstrate that presently, the Atlantis II Deep contains a single upper layer of brine with a homogeneous composition and temperature, overlying four basins with lower brine, characterized by regionally different temperature structures (Fig. 1). Currents in the lower brine pool are restricted by topographic highs which reach beyond the top of he lower brine. A specific feature of the temperature profiles in the lower brine of the southwest basin is a transition zone, below the lower interface, which is a few meters thick and characterized by temperatures higher and more scattered than at greater depth [31,44] (Fig. 1). This high-temperature zone is probably due to the injection of hot brine directly into the bottom of the southwest basin as predicted by Turner [22]. It rises as a plume above the vents, spreads out below the interface and then mixes turbulently with the lower brine. The mixing process is complete, as shown by the negligible scatter of the temperatures below the high temperature zone down to the surface of the sediment [311. In 1966, Erickson and Simmons [361 measured large temperature gradients in the sediment at the bottom of the southwest basin. The perturbation of the temperature resulting from brine injection apparently has dissipated more slowly in the sediment, due to the more sluggish rate of conductive heat transfer through solids. The west basin is supplied by new hot brine at the top of the lower convective layer, as shown by the attenuated transition zone on the temperature profiles from this

basin in 1971 and 1977 [24,44]. Both the maxim u m temperature in the transition zone and the mean temperature in the underlying layer are lower than in the southwest basin due to the greater distance from the brine vents. Finally, the temperatures of the lower brine below the transition zone are more scattered than in the southwest basin, indicating a lower degree of mixing of the injected hot brine. Temperature profiles in the lower brine from the east and north basins present large temperature gradients (Fig. 1). In the east basin and in an isolated depression in the west basin, a sublayer of lower brine in contact with the sediment is characterized by a constant temperature 2 - 3 ° C lower than that in the overlying lower brine. This structure was interpreted to result from the recent influx of hotter newly-formed brine coming from the west part of the Deep and trapping relics of the older and colder brine at the bottom of isolated basins [31]. Constant temperature profiles in the colder layer [31] and in the underlying sediments of the east basin [36] prove that no heat was transiently lost or gained through the bottom of this basin. In the north basin, the temperature of the lower brine typically decreases with depth and the maximum temperature is systematically lower than the mean value measured in the other basins (Fig. 4). These features are the result of the very restricted communication between the north basin and the lower brine in the west basin.

'LT ('C) 62

S°uthwestbasin~

:12

~'&

North basin 58

! 56

time 1966

I

119681

I

1970

I

I

1972

I

119741

119761

11978 I

I

~._

Fig. 4. E v o l u t i o n of the t e m p e r a t u r e of the lower brine with time in three b a s i n s of the A t l a n t i s II D e e p (1 = [35]; 2 = [31]; 3 = [44]; 4 = [49]). Ticks as in Fig. 3.

198

In summary, evidences for a localized influx of hot brine in the southwest basin based on the mineralogy and geochemistry of the sediments at the bottom of the Atlantis II Deep are the following five points: (1) Epigenetic magnetite-sulfate-sulfide-bearing veins and vugs typical of the southwest basin are directly precipitated from the hydrothermal brine at and below ~ 4 0 0 ° C [21]. (2) Both the mineralogy of the veins and the specific geochemistry of the sediment in the southwest basin (particularly the covariations of the metals including Ag and Zn, with S [45]) indicate that the reduced venting brine in the Atlantis II Deep supplies the metals and sulfur. This conclusion is in agreement with the observation by Hartmann [23] that the increase in the reducing conditions and heavy metal supply has accompanied the increase in the temperature of the lower brine from 1966 to 1971. (3) The hydrothermal-chemical sediments in the southwest basin are recrystallized and metamorphosed where the hot brines are venting, with a related reduction of iron in oxides and silicates. A temperature anomaly is thus marked on a larger scale by the transformation of goethite to hematite in the west and southwest basins [41,46] and, on a more restricted scale, by the transformation of hematite to magnetite in the southwest basin [47]. Similarly, dioctahedral clays become trioctahedral [16,19]. (4) The hydrothermal-chemical sedimentation at the interface between the lower and upper brines is enhanced in the southwest basin because hot brine upwelling is more active immediately above the brine vents [47]. (5) The highly brecciated southwest basin sediments, with numerous pieces of basaltic glass, provide evidence for intense magmatic activity and bottom instability accompanying brine venting [41], in good agreement with the inferred episodic and catastrophic spilling over of brine from the Atlantis II Deep into the near-by deeps to the south. Finally, because the sediments in Chain A, Chain B and Discovery Deeps are a marginal facies of Atlantis II and are characteristically differentiated from those in the southwest basin of the Atlantis II Deep [41], no brine is considered to be venting at the bottom of these three deeps.

C. R A M B O Z A N D M. D A N I S

2.4. Geyser-type actiuity Ross [48] first suggested that the feeding of the Atlantis II Deep was of geyser-type, given the high flow rate of brine discharge calculated over 52 months from October 1966 to February 1972. Based on more precise bathymetric measurements using a bathysonde [23], a precise estimate of 0.28 m 3 s-1 for the flow rate of the brine over a period of about 12 years was calculated from the increase in the volume of the lower brine: this value is about 20 times the flow rate of the Old Faithful Geyser in Yellowstone National Park. The changes in the temperatures of the UCL and of the LCL in the Atlantis II Deep over 15 years (Fig. 3A) confirm that the hydrothermal activity in that Deep is discontinuous and changing yearly [23,44,48-50]. Mineralogy provides additional evidence that the hydrothermal activity in the Atlantis II Deep is intermittent: (1) Sval'nov et al. [51] described barite crusts on basalts from the Central Sill of the Atlantis II Deep which presented rythmic variations of the isotopic composition of oxygen and sulfur. (2) Some epigenetic anhydrite and barite crystals from the southwest basin are polygenic, being formed from brine pulses with a progressively increasing salinity [21]. These crystals contain liquid-rich and vapor-rich inclusions which prove that the discharging spring from which they were formed was boiling on the sea floor [21]. According to White [52], this is evidence for geyser-type activity. 2.5. Other fluid flows in the Atlantis H stratified brine system Highly changeable chemical gradients in interstitial waters from the sediments show that the flux of pore water to the LCL is small [53]. There are, however, abundant geochemical evidences that both interfaces characterizing the Atlantis II stratified brine pool (i.e., the lower interfaces and the transition layer with Red Sea deep water: Fig. 2) have been transiently disrupted. (1) For the lower interface of the Atlantis II Deep, this can be inferred from the following facts. A millimetre- to centimetre-thick level of limonite enriched in lepidocrocite exists in the amorphous horizon on top of the sediment [41]. Lepidocrocite is an iron oxyhydroxide, whose for-

199

S U P E R H E A T I N G IN T H E R E D S E A ?

mation appears to be related to the rapid oxidation of Fe n [42]. The formation of this mineral in the Atlantis II Deep is probably associated with a transient influx of Fen-rich lower brine into the upper brine. Lepidocrocite was subsequently sedimented all over the Deep. H a r t m a n n [30] also noted that the upper brine contains an excess of dissolved sulfate as compared to the theoretical value based on conservative mixing, which has continuously increased from 1966 to 1977. This excess sulfate originates from the dissolution of particulate anhydrite which was supplied by the lower brine through the lower interface [21,26]. Turner [22] has shown that, in double diffusive systems, "molecular diffusion across the lower interface is, by far, more dominant than turbulent mixing". Particulate anhydrite must have been supplied to the upper brine by a transient disruption of the lower interface by the hot brine plume. Preliminary calculations by Urvois and Watremez (in [54]) indeed confirm that such a process is realistic. (2) Hartmann [23] described that = 5 m thick constant-composition constant-salinity sublayers formed within the U C L and the transition layer of the southwest basin, which resulted from the mixing of increasing amounts of LCL or UCL, respectively, with seawater. (3) In the southwest basin, --- 200 m above the top of the UCL, Merlivat et al. [55] detected large amounts of 3He of deep origin, suggesting that, in 1983, the flow rate of the geyser spring may have been transiently high enough for the hydrothermal brine to vent directly in seawater. It is likely that the sedimentation of the lepidocrocite level all over the Deep has corresponded, a few thousand years ago, to a major catastrophic disruption of the stratified brine system, implying a lower brine loss of the same order of magnitude as that required to fill the Discovery Deep. Concerning the past 12 years however, geochemical data show that the LCL component of the U C L has increased by around 1.8 wt.% between 1966 to 1977 [30]. This allows to evaluate that = 28% of the mass brought by the geyser spring over that period has been lost due to a transient disruption of the lower interface.

3. The heat-mass balance of Atlantis II Deep

3.1. Reference system In the Atlantis II Deep, the transition layer between the U C L and the normal Red Sea water is very thick (25-31 m: Table 2) and characterized by highly changeable temperature profiles [23]. The heat transfer through such a complex upper interface cannot be predicted by referring to a standard model. A precise estimate of this parameter would require detailed knowledge of the changes in the temperature structure of the layer with time. As this information is not available, the lower brine pool is chosen as the reference system for the heat-mass balance calculation in preference to a system including both the U C L and the LCL (Table 1). The lower interface of the Atlantis II Deep has been very sharp over the past 15 years, therefore heat exchanges across it have been essentially conductive. The diffusive loss of mass through the lower interface will be further neglected for the following reasons: (1) Mass transfer through heat-salt diffusive interfaces involve salt diffusion, excluding any diffusion of water. The diffusive loss of salt on top of the lower brine does not significantly modify the density of the lower brine and, therefore, it can be neglected in the mass balance equation. (2) Salt fluxes through double diffusive interfaces are minor compared with diffusive heat fluxes (section 2.1), as confirmed by the near constant salinity of the lower brine between 1965 and 1977 [23].

3.2. Equations The heat-mass balance calculation applies to a time interval At = ( t f - - t o ) (see Table 3 for explanation of symbols). The LCL is modelled as a prismatic box with an area S and a height x (Fig. 5A). The heat-mass balance of the Atlantis II Deep lower brine over the time interval (t 0, tf) can be expressed as a function of three unknown parameters (Fig. 5A and 5B): M e = the mass of fluid expelled from the Atlantis II Deep during At; T~= the temperature of the hydrothermal brine; M s = the mass of hydrothermal brine discharged during At into the LCL.

2OO

C. R A M B O Z A N D M. D A N I S

TABLE 3 Notation

Cp CpO

Hi (Qi)

specific heat of the fluid (J kg 1 K - l ) mean specific heat of the lower brine at 220 bar in the temperature range 5 6 - 6 2 ° C (taken to be constant) enthalpy (quantity of heat) of the lower brine in the Atlantis II Deep at time t i (at time to: Q0,

/4o) LCL Me M o (Mr) M~

lower convective layer of the Atlantis II Deep mass of lower brine expelled from the Atlantis II Deep mass of the lower brine pool at time t o (tt) mass of hydrothermal fluid added to the lower brine during the time interval (to, /f) mean mass flow rate of the spring at the bottom of the Atlantis II Deep during the time interval

(t0, tr).

ti

t o (tf)

re To (Tf) T~f

T~

mean surface area of the lower interface during the time interval (t 0, tr) time at which some lower brine is expelled from the Atlantis II Deep unspecified time initial (final) time temperature of the lower brine expelled from the Atlantis II Deep temperature of the lower brine at time t o (/f) reference temperature mean temperature of the spring at the bottom of the Atlantis II Deep during the time interval

(to, t0 upper convective layer of the Atlantis II Deep volume of the lower brine at time t i thickness of the interface between the upper brine Xli and the lower brine of the Atlantis II Deep (Fig. 5) depth of the interface between the upper brine 21i and the lower brine of the Atlantis II Deep variation of the enthalpy (of the quantity of heat) AHo- f ( A Q 0 - r ) of the lower brine during the time interval (t 0, te) aHo (aQe) enthalpy (quantity of heat) transported out of the lower brine pool by the mass flux M e during the time interval (to, tf) AQ~ (AQ °) conductive input (output) of heat through the bottom (the top) of the lower brine pool during the time interval (to, tf) heat brought by the geyser spring to the lower brine during the time interval (to, tr) At time interval (t 0, tf) temperature difference across the lower interface ATli(t) of the Atlantis II Deep at time t mean thermal conductivity of the lower interface in the time interval (t 0, if) ,}~ (,% conductive heat flux at the bottom (at the top) of the lower brine mean density of the lower brine in the range P0 5 6 - 6 2 ° C (taken to be constant) UCL

During the time interval At, all the mass produced by the spring (i.e. Ms) accounts for (a) the observed increase in the lower brine mass (i.e. M r - M0), and (b) the mass of brine having flowed over towards the U C L (i.e. Me; Fig. 5B): Me-

M 0 + M e = M s

(1)

We shall assume that, during the time interval (t o, tf), only one catastrophic expulsion of brine from the Atlantis II Deep has occurred at time t = t e. Let (M0, To) and ( M r , T f ) be the mass and the temperature of the system at time t = t o and t = tf respectively (Fig. 5A and 5B), let T~ be the temperature of the fluid expelled at time t = te, and Tre f be a reference temperature. As the evolution of the system is isobaric, the variations of the enthalpy of the system in the time interval (t 0, t f ) can be expressed as follows: A H 0 f = H f - Ho= Q r - Q o To =

Note that equation (2) implies that the heat budget only depends on the initial and final states of the system, and is independent, in particular, of whether the spring regime was steady or discontinuous. During the time interval (to, f f ) , the system exchanges heat by four processes (Fig. 5C): (1) Conductive input of heat through the sediments: AQic=

ftrsdpi at

(3)

to

(2) Heat lost at the top of the system by conduction:

A Q ° = - ftti'SdP°

dt=-ftlfXS(ATli(t)/Xli)dt

(4) where h is the thermal conductivity of the lower interface and ATli(t)/Xli is the temperature gradient across the lower interface of the Deep. (3) Heat produced by the geyser spring: T

AQ~ = + f ~M~cp(T) dT

(5)

v Tref

(4) Heat transported out of the lower brine layer by the mass Me:

T AQ~ = - f °M~cp(T) dT d Tref

(6)

SUPERHEATING IN THE RED SEA?

201

x

® s

rn~

of th~ L C L

I

to

tt

)

t=te

t=t o

t=tf

.I.o

a,o

~

(~

~ c°

.1'.1:, ,t,.,,t..t.,

A

A

Ts , qs

A

,pm¢

~

Lc L

"Is, qs ~

A

[--]

t**,a,~,

"Is, qs u e L

Fig. 5. The heat-mass b a l a n c e of the lower brine pool in the A t l a n t i s II Deep. A. Schematic g e o m e t r i c model for the Deep. B, C. Exchanges of mass and heat t a k e n in account in the calculations, i n c l u d i n g the possibility of the e x p u l s i o n of a mass of lower brine Me out of the pool at time t e.

The conductive exchanges through the lateral walls of the box are negligible because the lower brine reservoir in the Atlantis II Deep is about three orders of magnitude longer or wider than thick. The heat balance equation for the system is:

T

3.3. Parameters

,-T0

~IMrcp(T) dT-JTr~rMoep(T) dT

T.

] at

+ f ~M,cp(T) dT- f ~Mecp(T) dT ~'Yrcl

J

yref

(8)

J Te

Combining equations (2) to (6):

T

ff'M,e (T) dT+ f: Moc,(T)dT + f ~M~Cp(T)dT

Q r - Q0 = AQ~ + AQ~' + AQ~ + AOe

= ftltraf~c at-ftltf~ka[AZli(t)/xli

Replacing T~f by T~ in equation (7), it follows:

(7)

-z,, S, V. Hartmann [23] determined to within +0.5 m the depth of the lower interface, zli, in 1965, 1971, 1972 and 1977; the values are given in Fig. 3. The surface area of the lower interface (S) from 1965 to 1980 has been estimated from new planimetric measurements of the

202

C R A M B O Z A N D M. D A N I S

bathymetric map by Backer and Richter [41]. Hence, the volume of the lower brine (V) was determined, taking into account the changing depth of the lower interface (Fig. 3A) [23]. For example, in 1965, S and V were estimated to have been 43 km 2 and 2.7 km 3 respectively, the lower interface being at a depth of - 2 0 5 0 . 5 m. These volumes are 10% lower than those published by H a r t m a n n [23]. - - x~. The transition between the U C L and the LCL takes place over a vertical distance of 1 + 0.2 m [23]. As the fluctuations of the thickness of the lower interface with time are not k n o w n precisely, Xl~ is taken to be constant from 1965 to 1980. - - ¢',!. In 1966, Erickson and Simmons [36] estimated that the heat flux through the sediment was in the range 0.63-0.84 W m 2 ( 1 5 - 2 0 H F U ) , by interpreting the temperature profiles measured at that date in the sediments all over the Deep. Obviously, the measured thermal gradients were transient, as discontinuous convection is the dominant process in the Atlantis II Deep. The conductive influx of heat to the Atlantis II Deep has therefore probably changed with time. It is worth noting that Erickson and Simmons [36] interpreted a composite temperature profile of all the 1966 thermal data in terms of the cooling of a 10 m thick layer of sediment from 74 ° to 5 6 ° C in 6 - 7 years. In the absence of data, the conductive influx of heat to the Deep ~ is assumed to be constant in time and space over the Deep and equal to 0.67 W m 2 (16 HFU). - - p. As the salinity of the L C L has not significantly departed from 26 wt.% eq. NaC1 from 1965 to 1980 [23,56], both the lower brine and the hydrothermal fluid are modelled in terms of a 6 M NaC1 solution. ( T r - To) has not exceeded 1 0 ° C from 1965 to 1980 therefore, in the temperature interval (To, Tf), p is considered constant and equal to P0 = 1.18154 g cm -3 [57]. It follows:

M o = ooVo

(9)

Mf = PoVf

(10)

where V0 and Vf are the volume of the L C L at t o and tf respectively. ~.. In the temperature range (T0, Tf), ~ is considered to be constant and equal to the thermal conductivity of a 6 M NaC1 solution at 59 o C, -

-

i.e. 0.625 W m 1 K 1 based on an extrapolation of the data compiled by Phillips et al. [58]. - - T~. The mass of lower brine expelled out of the L C L in the time interval 1965-1977 is not known precisely nor can its temperature be constrained in the range ( T0, T~). This is because the expelled brine can have been either a fraction of lower brine at a temperature comprised between To and Tf, or the pure hydrothermal brine at temperature T~ or any mixture of these two endmembers at a temperature comprised between To and T~ (sections 2.2 and 2.5). In the following calculations, only the loss of a lower brine component at a temperature between To and Tf is taken into account and the loss of a fraction of the hydrothermal fluid is excluded. The latter simplification however, has no effect on the calculated spring temperature and simply induces an underestimation of the actual spring flow rate. Additionally, brine expulsion from the Atlantis II Deep reservoir is arbitrarily assumed to have occurred at a time when the temperature of the L C L was T~= ( T r + T0)/2. The influence of such an hypothesis on T~ is evaluated hereafter. cp(T). In the temperature range ( T0, Tf), the heat capacity of the lower brine can be taken to be constant and equal to Cpo = 3.2044 × 103 J kg I K ~ [57]. In the temperature range 200-300 ° C, the cp's of pure water and of moderately saline solutions increase by -- 10% [57,59]. Calculations by Pitzer et al. [57] and Helgeson [59] predict that the specific heat of 3 - 6 M H 2 0 - N a C 1 solutions under pressure also increases moderately between 200 and 300 ° C (Fig. 6). In the following calculations, two models have been used to calculate the Cp'S of the hydrothermal brine taken as a 6 M H 2 0 - N a C I solution: (1) The values predicted by the equation of Pitzer et al. [57] below 300 ° C have been fitted by a least square method and extrapolated to higher temperatures. Fig. 6 shows that the cp's thus extrapolated increase by ~ 15% in the range 100-350°C. (2) The Cp'S of pure water near its critical point can be calculated using the equation of Born [60]. This equation, with additional terms to take into account the compressibility of the fluid [61], and with the equation of Bradley and Pitzer [62] to model the dielectric constant of pure water, have been used to predict the Cp'S of the hydrothermal -

-

SUPERHEATING

1@ 31

IN THE

RED

203

SEA?

Cp(Z.K-I.K9-1)

I

H20

I I I

BORN E q u a t i o n (I]H20

]~i

(i.e. cp(T)), equation (8) combined with equations (9) and (10) becomes:

I~

NaCI solutiorq (6Mdall P = 200 bars

a{ter

Bradley

CpoPo(Vr+ V0)(Tr- V0)/2

/I /

- f(:+~,)/2M~cp (T) dT

and Pit .... 197~)

1

PITZER

ET

= d P ~ S ( t f - to) - X S ~,'/ ( A T, i ( t ) / x ,i ) at

R

5 2.5

T (oC) 1 1~

151~

2B~I

25~

3EI~

35~]

(11)

Equations (1) and (11) express the heat-mass balance of the Atlantis II Deep as a function of the

4110

three unknown parameters T,, M~, M e, given that the temperature of the expelled lower brine is arbitrarily taken to be (Tf+ T0)/2. These simplified equations have been solved using the data indicated above, including the two possible models for the heat capacities of the hydrothermal brine. The mean temperature of the geyser spring from 1965 to 1977 calculated for each of the two Cp models is shown in Fig. 7 as a function of a variable loss of mass out of the system. Fig. 8 shows the spring temperatures calculated over

Fig. 6. Calculated isobaric heat capacities of a 6 M HzO-NaC1 solution under pressure as a function of temperature, using the equation of Pitzer et al. [57] extrapolated above 300° C, and the equation of Born [60], modified after Wood et al. [61] (calculations from Quint, Grolier and Coxham, unpublished).

brine at 200 bar (Quint, Grolier and Coxham, unpublished). Fig. 6 shows that the Cp'S thus predicted for the brine increase exponentially above 300 o C. On account of the above approximations, and depending on the model chosen for the cp's

TEMPERRTURE CRLCULRTED FOR THE SPRING Ts(oC) (325 - Cp R F T E R EXTRRPOLRTION 575OF P I T Z E R ET R L .

[563

i¢/

Cp

375

OF

L ~ / Ts=461°C


a25

325

"%

"///A

/ .... 275

/////,.

,

275 3

~ .'4 .5 (~ FLOW RRTE

THE MODEL MODIFIED

/ * ~ :..S::...O'57.. f

/

///h

AFTER BORN

qs

(103

%, @

Kg/S)

Fig. 7. Temperature calculated for the spring (T,) as a function of the mass flow rate (q~), using the data summarized in Fig. 3 and, to calculate the cp's, either (A) the equation of Pitzer et al. [57] extrapolated above 300 ° C, or (B) the modified equation of Born [60,61] (see Fig. 6). Numbers above the curves indicate variable thickness of the lower interface (xli = 0.5, 1, 1.5 m). The minimum spring flow rate indicated by arrow number 1 ( = 280 kg s 1), is imposed by the observed increase in the volume of the lower brine (see Fig. 3A). Lower flow rates are therefore excluded (hatched area). The flow rate indicated by arrow number 2 (358 kg s 1) includes the mass of lower brine lost towards the UCL between 1965 and 1977 [30]. In case the fluid expelled out of the pool was at the temperature of the bulk reservoir, T, is then read on the ~ curves. If the expelled fluid was at temperature Z~, the calculated spring temperature remains equal to the maximum value on the T~ curve, whatever the spring flow rate is above 280 kg s 1 (see text). The spring temperatures in the light dotted area are also excluded as they are above the boiling temperature of the hydrothermal brine on the sea floor [21]. Fig. 7 shows that the changes in the temperature and mass of the lower brine from 1965 to 1977 cannot be interpreted in terms of a realistic spring temperature unless the hydrothermal brine has presented very high cp values above 300 o C.

204

C. RAMBOZ AND M. DAN[S

X N a C ] =22_+_1 klTX 15

Level

1085

values -

320°C

10

m a x i m u m values -- 3 9 0 ° C

0150

200

25[3

300

350

400

450

TH(oC) Fig. 8. Histogram of homogenization temperatures to liquid of the fluid inclusions with a composition in the range 22_+ 1 wt.% eq. NaC1 in epigenetic anhydrites from southwest basin sediments [21]. Mean homogenization temperatures around 320°C indicate average trapping temperatures on the sea floor of 333 ° C, on account of the pressure correction [63].

shorter periods of time from 1965 to 1980, assuming no loss of mass out of the lower brine pool. 4. Discussion

The precision of the calculated Ts and q~ depends on the accuracy of the physical parameters of the Deep, on the magnitude of the brine overspill, and on the model chosen to calculate the cp's of the hydrothermal brine, to be discussed below.

4.1. The period 1965-1977 The homogeneous temperature-depth data set of Preussag provides a unique way to average the temperature and flow rate of brine discharge in the Atlantis II Deep from 1965 to 1977. The minimum flow rate of brine discharge, corresponding to no brine leakage from the reservoir, has been 280 kg s - I (Fig. 7). The uncertainty in the calculated flow rate due to errors in the Deep planimetry, is around + 50 kg s l. This is however a minimum value as it does not take into account the uncertainty of the map itself, which is quite difficult to evaluate. Discharge flow rates have been all the greater as more brine has flowed over out of the Deep. For the period 1965-1977, the hydrological and geochemical data summarized in sections 2.2 and

2.5 point to the following facts: (a) brine overspill towards the near-by deeps has been negligible; (b) the lower brine component of the UCL has increased by 1.8 wt.% [30]. Considering that the volume of the U C L was ~ 1.48 km 3 in 1977 (based on bathymetric and hydrological data, [25] and [23]), such a lower brine loss implies an additional flow rate of 78.40 kg s - l , i.e., a total spring flow rate of 358.4 kg s 1 for that period. Calculations show that, from 1965 to 1977, the heat lost from the top of the lower brine (¢,o) has been about 7.03 W m -z (xl~ = 1) and the heat flux required to increase the temperature of the lower brine pool from 56 ° to 61.5°C has been about 3.34 W m - : . As the conductive input of heat at the bottom of the Deep has been taken constant and equal to 0.67 W m 2, ,i,o is the major term of the heat-mass balance equations [22]. A flux of that magnitude is about that required to heat the U C L by ~ 2 2 ° C in the time interval 1965-1977, neglecting conductive heat loss on top of the UCL (to be compared with the measured temperature excess of 8 - 1 2 ° C observed for the UCL, relative to conservative mixing [30]). Previous heat-mass balance calculations have neglected the heat lost by conduction from top of the chosen system. This explains why the temperature of the spring in the Atlantis II Deep has been frequently underestimated at values below 215 ° C (Table 1).

S U P E R H E A T I N G IN T H E R E D SEA?

4.2. Calculated spring temperature: sensitivity of the model --cp. The calculated spring temperature strongly depends on the model chosen to calculate the cp's. With our extrapolation of the equation of Pitzer et al. [57] giving values of cp(T) between 3240 and 7100 J kg -1 K -1 in the temperature range 56-460 ° C (Fig. 6), a maximum spring temperature of 461°C is required to account for the observed changes in the mass and temperature of the lower brine (Fig. 7A). Taking into consideration a heat output through the top interface of about 6.3 W m -2, Monin and Plakhin [49] obtained quite a comparable temperature (490°C; Table 1). The temperature of the spring feeding the Atlantis II Deep, however, cannot have exceeded 3 9 0 ° C because, at a pressure of 220 bar, this is about the temperature of unmixing at its composition [21]. Several factors, other than the choice of another model for the Cp'S (Fig. 7B), can be considered to calculate a realistic temperature of brine discharge from the heat-mass budget of the Atlantis II Deep (i.e., ~ ~< 390 ° C): (1) This could be the loss out of the LCL of a mass of lower brine equivalent to at least 40% of the total mass of hydrothermal brine brought by the spring (Fig. 7A). The total spring flow rate estimated for the period 1965-1977, including the lower brine lost towards the UCL, has been around 358.4 kg s-1 (section 4.1). This implies a minimum spring temperature of --- 403 ° C, provided that all the lost lower brine was at equilibrium with the bulk reservoir, i.e. at a temperature of ---60°C (Fig. 7A). The temperature of the expelled brine, however, was probably significantly >/60 o C, and close to T~, because the top interface was disrupted by the hot brine plume itself during periods of enhanced hydrothermal activity (section 2.5). As the closer is T~ to T~, the higher is Ts (section 3.3.), hence it is concluded that brine overspill is not the required process for elucidating the heat-mass budget of the Atlantis II Deep. (2) Part of the heat supplied to the lower brine, instead of having a convective origin, could have originated from a basaltic magma directly in contact with the lower brine. A 25 cm thick sheet of basalt covering all the Deep bottom could have yielded by conduction the observed excess of calories. This volume of basalt is far too large however, given that the hydrothermal-chemical

205

sediments on the sea floor contain only basaltic shards. (3) A conductive heat flow >/3.1 W m -2 at the Deep bottom could account for a spring temperature lower than 390 ° C; such a mean heat flux over a period of 12 years, however, is unrealistically high. Hence it is concluded that, in the time interval 1965-1977, x~ being 1 m, the Atlantis II Deep has received excess calories due to convection, which cannot be simply accounted for on extrapolating to above 3 0 0 ° C present data on heat capacities of brines. On using the modified model of Born [60,61] which predicts a near exponential increase of the c f s above 300°C, a temperature of 353°C is obtained for the spring between 1965 and 1977, assuming that no mass has been lost out of the lower brine pool (q~ = 280 kg s 1; Fig. 7B), and of 342°C, taking into account the mass of lower brine transferred towards the U C L during that period (i.e. q~ = 358.4 kg s - I ; the temperature of the expelled brine being assumed to have been around 6 0 ° C : Fig. 7B). These mean maximum discharge temperatures of 353 ° or 342 ° C are quite consistent with the mean trapping temperature of the fluid inclusions of 333°C at 220 bar [64] in epigenetic anhydrites containing a 22 + 1 wt.% eq. NaC1 solution (Fig. 8) [21]. The calculations summarized in Figs. 6 and 7 therefore show that the heat-mass balance of the Atlantis II Deep cannot be coherently interpreted, and in particular reconciled with other available thermometric information, unless the deep has been fed by fluids presenting very high heat capacities above 300 ° C. Experimental measurements are obviously required to confirm the cp values predicted using the equation of Born modified, and in particular their near exponential increase between 300 o and 3 5 0 ° C (Fig. 6). The supply of superheated fluids to the lower brine reservoir could also account for the abnormal input of heat into the Deep. Superheated water (i.e., a metastable liquid nucleating no vapor although at a temperature above its unmixing point) indeed presents excess heat capacities. These exponentially increase in the metastable region, i.e., above the unmixing temperature [64]. The supply of superheated fluids to the Atlantis II Deep seems realistic, as prolonged boiling at the sea floor of the southwest Basin, with vaporization of = 20 wt.% of the fluid, is

206

C. R A M B O Z A N D M. D A N I S

recorded by anhydrite fluid inclusions [21]. This proves that large amounts of calories are supplied to a limited volume of fluid in the deep reservoir. Fluid inclusions in barite and anhydrite crystals collected 1 m above the boiling zone described by Ramboz et al. [21], seem indeed to support the latter interpretation (Ramboz, Oudin and Thisse, unpublished data). The sensitivity of the model is further evaluated hereafter using the exponentially increasing Cp values predicted by the modified equation of Born for the hydrothermal brine. Xn" ~o being proportional to 1 / x n, the uncertainty in x n of about _+0.2 m [23] results in an uncertainty in ~ of about + 4 ° C . This uncertainty increases as the temperature decreases to = 300° C and remains at about _+25 o C below 300 ° C (Fig. 7B), because of the exponential form of the cp relation (Fig. 6B). Between 1965 and 1977, x n may have varied outside of the range 1_+0.2 m. If x n is taken to vary in the range 0.5-1.5 m, the maximum calculated spring temperature varies from 360 ° to 343°C (Fig. 7B). It is concluded that such an uncertainty in x n does not result in a large error in T~, as long as the spring temperature remains around 350 o C. - - M 0' Mr. The uncertainty in Ts related to a poor knowledge of the mass of the lower brine is minor. This is because the calculation involves the changes in the mass of the lower brine with time and this is better known than the mass of the system at a given time t. - - TLc L. The temperature of the LCL over the whole Deep at a given time t is difficult to average. The temperature of this reservoir varies both vertically and horizontally, i.e. from one basin to another. In particular, a cooler layer of lower brine, whose temperature has remained unchanged with time [24,39], exists at the bottom of the north and east basins. The related uncertainty in ~ is, however, minor because heat-mass balance equations only involve the increase in the temperature of these layers. The temperature of the LCL over the Deep was averaged at a given time using the temperature profiles measured in each basin, taking into account all the lower brine layers in the separate basins as a function of their specific temperature and volume [24]. The difference between the average value obtained and the temperature measured in the southwest basin _

_

was found to be ~ 0.8°C, i.e. less than the analytical uncertainty in temperature [23]. T~. Calculations show that the calculated spring temperature does not depend at all on the exact temperature between TO and Tf at which the brine was transferred towards the UCL. -

-

4.3. Changes in the hydrothermal activity with time In this paragraph, heat-mass balance equations are solved for short periods of time between 1965 and 1979, using either of the two cp models shown in Fig. 6. The exponentially increasing c? values predicted by the equation of Born modified, are used to provide qualitative evidence for periods of enhanced hydrothermal activity, involving the influx of fluids with unusually high heat transport properties into the Deep. It is beyond the scope of this paper to determine whether the high enthalpy fluids discharging into the Atlantis II Deep are stable liquids (T~< 390°C), as predicted by the equation of Born modified, or are superheated brines. Two contrasted phases in the hydrothermal activity of the Atlantis II Deep are characterized: (1) The period between 1965 and 1976 implies enhanced geyser spring activity. The spring temperatures calculated using either of the two ce models are higher than 350 ° C and separated by a gap of 1 2 0 ° C (474 ° and 354°C: Fig. 9). This proves that the input of heat was abnormally high all over that period, and compatible with the influx of "superheated-like" fluids into the Deep, as discussed earlier. The average discharge temperatures, calculated using the Born modified Cp model [60,61] and neglecting a possible loss of mass, vary between 350 ° and 355°C, i.e. they remain quite constant (Fig. 9). This is because, although both the amount of heat required to increase the temperature of the lower brine (AQ~) and the heat lost through the lower interface (AQ S) have both varied independently by - 15% during this period, the sum of these two heat flows has remained nearly constant. (2) From 1976 to 1979 in contrast, the spring temperatures dropped to 332°C or lower, because both AQ S and AQ~ decreased simultaneously. Note in particular that the data of Bubnov et al. [37] and H a r t m a n n [23], showing a decrease in the temperature of the lower brine of 0.38°C from

207

S U P E R H E A T I N G IN T H E R E D SEA?

Ts(OC) Flew rate

350

0!li::)°::]47;)::? °:°:::? : :°°:t )°" : ;~O ::° : °D :b:~7~°°

0.29

• 0.28

0.27

300

1965

•

1970

1975

abnormal heat supply (superheating?)

~

1981 decreased heat supply

Fig. 9. Spring temperatures (solid line) and flow rates (dotted line) calculated for the time intervals 1965-1976, 1976-1977, 1977-1979, assuming that no mass was lost out of the lower brine pool (xli =1 m and M e = 0), and using the data shown in Fig. 3.

June 1976 to November 1977, imply a spring temperature of 272 ° or 285 ° C, depending on the model chosen for the cp's. The fact that, in the time interval 1976-1979, the temperatures obtained with either of the two cp models do not differ by more than 55 ° C and remain lower than 3 5 0 ° C (Fig. 9), proves that the supply of fluids having abnormal heat transport properties was greatly reduced during that period. The present work therefore confirms Hartmann's [30] conclusions based on the changing chemistry of the brines, that the hydrothermal activity in the Atlantis II Deep has decreased between 1976 and 1979 as compared to that between 1965 and 1976. The flow rates calculated without brine overspill over short periods of time from 1965 to 1977 continuously increase from 275 to 281 kg s-1 (Fig. 9). If one admits that some lower brine can be transferred towards the UCL only when the top interface of the LCL is disrupted by the plume, i.e. during periods of enhanced hydrothermal activity, then the mean spring flow rate was probably the highest between 1965 and 1976 (---367 kg s 1. section 4.1), and it dropped to around 280 kg s-1 between 1976 and 1977, when the hydrothermal activity decreased.

Finally, the increasing hydrothermal activity from 1977 to 1979 suggested in Fig. 9 must be taken with caution, as the depth of the LCL top interface in 1979 was extrapolated from earlier data. 5. Conclusions

The major term in the heat budget of the Atlantis II Deep lower brine is the conductive output of heat through the lower interface on top of the system [22]. Most previous heat-mass balance calculations have neglected this term. They have therefore in general underestimated the temperature of the spring in the Atlantis II Deep at values below 215 ° C (Table 1). The maximum temperature that can be reached by the spring is that fixed by the unmixing conditions of a 6 M eq. NaC1 solution at 220 bar, i.e. 3 9 0 ° C [21]. The present work shows that the hydrological changes of the Atlantis II stratified brine pool documented by the Preussag group from 1965 to 1977, cannot be interpreted in terms of a realistic spring temperature (T~<390°C), unless the Deep has been fed by fluids presenting abnormally high heat transport properties. Two heat capacity models are discussed for the hydrothermal fluid: (1) The cp's of the brine are predicted to present a near exponential increase above 300°C, using the equation of Born [60] modified after Wood et al. [61] near the critical point of pure H 2 0 (unpublished calculations by Quint, Grolier and Coxham). (2) Superheated liquids (i.e. metastable liquids nucleating no vapor above their unmixing temperature) indeed present exponentially increasing c e values in the metastable region [64]. Brine superheating can be a likely process in the Atlantis II Deep given that, below the southwest basin, there is a deep hot brine reservoir where fluids are boiled out [21]. In case of superheating, however, the heat-mass budget of the Atlantis II Deep cannot be solved unequivocally in terms of a spring temperature, as it depends both on the amounts of superheated brines supplied to the Deep and on the temperature they reached above 390 o C. Heat-mass budget equations reveal two distinct

208

C. R A M B O Z A N D M. D A N I S

periods in the recent hydrothermal activity of the A t l a n t i s II D e e p . (1) F r o m 1965 to 1976, t h e D e e p h a s r e c e i v e d a n excess h e a t s u p p l y , c o m p a t i b l e w i t h t h e i n f l u x o f f l u i d s p r e s e n t i n g e n h a n c e d h e a t c a p a c i t i e s , e.g., superheated liquids. The spring flow rate, including the mass of lower brine expelled towards the U C L [30], w a s a r o u n d 367 k g s 1 (2) T h e h y d r o t h e r m a l a c t i v i t y t h e n d e c r e a s e d f r o m 1976 to 1979. T h e i n f l u x o f f l u i d s w i t h abnormal heat transport properties was then greatly reduced and, probably, the mean spring flow r a t e d r o p p e d t o ~ 280 k g s -1. I n p a r t i c u l a r , t h e m e a n s p r i n g t e m p e r a t u r e d e c r e a s e d to 278 + 6°C during the period 1976-1977, depending on t h e m o d e l s c h o s e n to c a l c u l a t e t h e of the brine.

Cp'S

Acknowledgements T h e a u t h o r s a r e g r a t e f u l to C. M o y n e a n d J.C. Batsale for their help in putting the heat-mass balance equations in a proper thermodynamic f o r m , a n d to F. A l b a r r d e a n d S . M . F . S h e p p a r d f o r d e t a i l e d review. H. B a c k e r , P. G u e n n o c a n d E. Oudin greatly improved the manuscript by their k n o w l e d g e o f t h e g e o l o g y o f t h e A t l a n t i s II D e e p a r e a . M a n y d i s c u s s i o n s w i t h P. W a t r e m e z c o n c e r n ing the physics of the Deep were greatly apprecia t e d . F i n a l l y , J.R. Q u i n t , J.-P.E. G r o l i e r a n d J.Y. C o x h a m p r o v i d e d a m o d e l f o r t h e cp's o f t h e hydrothermal brine which was decisive in solving heat-mass balance equations in terms of a realistic spring temperature. This work was financed by C.R.P.G. (Nancy, France).

References 1 P.E. Rona, Hydrothermal mineralization at seafloor spreading centers, Earth Sci. Rev. 20, 1-104, 1984. 2 E. Oudin, Minrralogie des gisements et indices lids 5. des zones d'accrrtion ocraniques actuelles (rive Est-Pacifique et Met Rouge) et fossile (Chypre), Chron. Rech. Min. 470, 43-56, 1983. 3 F. Albarrde, A. Michard, J.F. Minster and G. Michard, 87Sr/86Sr ratios in hydrothermal waters and deposits from the East Pacific Rise at 21°N, Earth Planet. Sci. Lett. 55, 229-236, 1981. 4 G. Michard, F. Albar~de, A. Michard, J.F. Minster, J.L. Charlou and N. Tan, Chemistry of solutions from the 13°N East Pacific Rise hydrothermal site, Earth Planet. Sci. Lett. 67, 297-307, 1984.

5 H. B~icker, Rezente hydrothermal-sediment~ire Lagerst~ittenbildung, Erzmetall 32, 544-555, 1973. 6 J.L. Bischoff and F.W. Dickson, Seawater-basalt interaction at 200°C and 500 bars; implication for origin of sea-floor heavy metal deposits and regulation of sea-water chemistry, Earth Planet. Sci. Lett. 25, 383-397, 1975. 7 M.J. Mottl and H.D. Holland, Chemical exchange during hydrothermal alteration of basalt by sea water, I. Experimental results for major and minor components of seawater, Geochim. Cosmochim. Acta 42, 1103-1115, 1978. 8 M. Arnold and S.M.F. Sheppard, East Pacific Rise at latitude 21°N: isotopic composition and origin of the hydrothermal sulfur, Earth Planet. Sci. Lett. 56, 148-156, 1981. 9 D.R. Converse, H.D. Holland and J.M. Edmond, Flow rates in the axial hot springs of the East Pacific Rise (21 ° N): implications for the heat budget and the formation of massive sulfide deposits, Earth Planet. Sci. Lett. 69, 159-173, 1984. 10 H. Craig, Geochemistry and origin of the Red Sea brines, in: Hot brines and recent heavy metal deposits, E.T. Degens and R.A. Ross, eds., pp. 208-242, Springer-Verlag, New York, N.Y., 1969. 11 F.T. Manheim, Red Sea geochemistry, in: R.B. Whitmarsh, O.E. Weser, D.A. Ross et al., Initial Reports of the Deep Sea Drilling Project 23, 975-998, U.S. Government Printing Office, Washington, D.C., 1974. 12 B. Dupr6, G. Blanc, J. Boulegue and C. Allrgre, Metal remobilization at a spreading centre using lead isotopes, Nature 333, 165-167, 1987. 13 M.H. Drlevaux and B.R. Doe, Preliminary report on uranium, thorium, lead contents, and lead-isotopic composition in sediment samples from the Red Sea, in: R.B. Whitmarsh, O.E. Weser, D.A. Ross et al., Initial Reports of the Deep Sea Drilling Project 23, 943-946, U.S. Government Printing Office, Washington, D.C., 1974. 14 W.C. Shanks III and J.L. Bischoff, Ore transport and deposition in the Red Sea geothermal system: a geochemical model, Geochim. Cosmochim. Acta 43, 15071519, 1977. 15 I.R. Kaplan, R.E. Sweeney and A. Nissembaum, Sulfur isotope studies on Red Sea Geothermal brines and sediments, in: Hot Brines and Recent Heavy Metal Deposits, E.T. Degens and R.A. Ross, eds., pp. 474-498, SpringerVerlag, New York, N.Y., 1969. 16 R.J. Pottorf and H.L. Barnes, Mineralogy, geochemistry and ore genesis of hydrothermal sediments from the Atlantis II Deep, Red Sea, Econ. Geol. Monogr. 5, 198-223, 1983. 17 R.J. Pottorf, Hydrothermal sediments of the Red Sea, Atlantis II Deep--a model for massive sulfide-type ore deposits, 192 pp., Ph.D. Thesis, Pennsylvania State University, 1980 (unpublished). 18 Y. Thisse, Srdiments mrtalfifrres de la fosse d'Atlantis II (Mer Rouge). Contribution ~ l'6tude de leur contexte morpho-structural et de leurs caractrristiques mindralogiques et grochimiques, 155 pp., Thesis, U n i v e r s i t 6 0 r l r a n s et B.R.G.M., 1982 (unpublished). 19 P.A. Zierenberg and W.C. Shanks, Mineralogy and geo-

SUPERHEATING IN THE RED SEA?

20

21

22

23

24

25

26

27 28

29

30

31

32 33

34

35

36

37

chemistry of epigenetic features in metalliferous sediments, Atlantis II Deep, Red Sea, Econ. Geol. 78, 57-72, 1983. E. Oudin, Y. Thisse and C. Ramboz, Fluid inclusion and mineralogical evidence for high-temperature saline hydrothermal circulation in the Red Sea metalliferous sediments: preliminary results, Mar. Min. 5, 3-31, 1984. C. Ramboz, E. Oudin and Y. Thisse, Geyser-type discharge in Atlantis II Deep (Red Sea): evidence from fluid inclusions in epigenetic anhydrite, Can. Mineral. 26, 765-786, 1988. J.S. Turner, A physical interpretation of the observations of hot brine layers in the Red Sea, in: Hot Brines and Recent Heavy Metal Deposits, E.T. Degens and R.A. Ross, eds., pp. 164-173, Springer-Verlag, New York, N.Y., 1969. M. Hartmann, Atlantis II Deep geothermal brine system. Hydrographic situation in 1977 and changes since 1965, Deep-Sea Res. 27A, 161-171, 1980. M. Schoell, Valdivia VAO1 Rotes Meer-Golf Von Aden--Hydrography II+ III, Daten, 1063 pp., Bundesanstalt fiir Bodenforschung, Hannover, 1974. H. B~icker, K. Lange and H. Richter, Morphology of the Red Sea Central graben between Subair Islands and Abul Kizaan, Geol. Jahrb. 13, 79-123, 1975. P.A. Zierenberg and W.C. Shanks, Isotopic constraints on the origin of the Atlantis II, Suakin and Valdivia brines, Red Sea, Geochim. Cosmochim. Acta 50, 2205-2214, 1986. J.S. Turner, Buoyancy Effects in Fluids, 368 pp., Cambridge University Press, 1979. T.J. McDougall, Fluid dynamic implications for massive sulphide deposits of hot saline fluid flowing into a submarine depression from below, Deep-Sea Res. 31,145-170, 1984. A.D. Voorhis and D.L. Dorson, Thermal convection in the Atlantis II hot brine pool, Deep-Sea Res. 22, 167-175, 1975. M. Hartmann, Atlantis II geothermal brine system. Chemical processes between hydrothermal brines and Red Sea deep water, Mar. Geol. 64, 157-177, 1985. M. Schoell and M. Hartmann, Detailed temperature structure of the hot brines in the Atlantis II deep area (Red Sea), Mar. Geol. 14, 1-14, 1973. D.T. Pugh, Origin of hot brines in the Red Sea, Nature 214, 1003-1004, 1967. D.A. Ross, Temperature structure of the Red Sea brines, in: Hot Brines and Recent Heavy Metal Deposits, E.T. Degens and R.A. Ross, eds., pp. 148-152, Springer-Verlag, New York, N.Y., 1969. R.G. Munns, R.J. Stanley and C.D. Densmore, Hydrographic observations of the Red Sea brines, Nature 214, 1215-1217, 1967. P.G. Brewer, C.D. Densmore, R. Munns and R.J. Stanley, Hydrography of the Red Sea brines, in: Hot Brines and Recent Heavy Metal Deposits, E.T. Degens and R.A. Ross, eds., pp. 138-147, Springer-Verlag, New York, N.Y., 1969. A.J. Erickson and G. Simmons, Thermal measurements in the Red Sea hot brine pools, in: Hot Brines and Recent Heavy Metal Deposits, E.T. Degens and R.A. Ross, eds., pp. 114-121, Springer-Verlag, New York, N.Y., 1969. V.A. Bubnov, V.S. Fedorova and A.D. Shcherbinin, New

209

38

39

40 41

42

43

44

45

46

47

48 49

50

51

52

53

54

data on the deep-water brines in the Red Sea, Okeanologia 17, 603-611, 1977. D.T. Pugh, Temperature measurements in the bottom layers of the Red Sea brines, in: Hot Brines and Recent Heavy Metal Deposits, E.T. Degens and R.A. Ross, eds., pp. 158-163, Springer-Verlag, New York, N.Y., 1969. M. Schoell, Heating and convection within the Atlantis II Deep Geothermal system of the Red Sea, Proc. 2nd U.N. Symp. on the Development and Use of Geothermal Resources, San Francisco, Calif., 20-29 May 1975, 2, 583-590, 1976. D.A. Ross and J.M. Hunt, Third brine pool in the Red Sea, Nature 213, 687-688, 1969. H. B~icker and H. Richter, Die Rezente HydrothermalSedimentare L~igerst~itte Atlantis II-Tief im Roten Meer, Geol. Rundsch. 3, 697-741, 1973. J.W. Murray, Iron oxides, in: Marine Minerals, R.G. Burns, ed., Mineral. Soc. Am. Short Course Notes, pp. 47-98, 1979. G. Krause and J. Ziebenbein, Die Struktur des heissen salzreichen Tiefenwassers im zentralen Roten Meer, "Meteor" Forschungsergebn. 1, 53-58, 1966. M. Schoell and M. Hartmann, Changing hydrothermal activity in the Atlantis II Deep geothermal system, Nature 274, 784-785, 1978. J.L. Bischoff, Red Sea geothermal brines deposits: their mineralogy, chemistry and genesis, in: Hot Brines and Recent Heavy Metal Deposits, E.T. Degens and R.A. Ross, eds., pp. 368-401, Springer-Verlag, New York, N.Y., 1969. J.L. Bischoff, Goethite-hematite stability relations with relevance to sea water and the Red Sea brine system, in: Hot Brines and Recent Heavy Metal Deposits, E.T. Degens and R.A. Ross, eds., pp. 402-406, Springer-Verlag, New York, N.Y., 1969. J.P. Hackett, Jr. and J.L. Bischoff, New data on the stratigraphy, extent and geologic history of the Red Sea geothermal deposits, Econ. Geol. 68, 553-564, 1973. D.A. Ross, Red Sea hot brine area: revisited, Science 175, 1455-1457, 1972. A.S. Monin and E.A. Plakhin, Stratification and space-time variability of Red Sea hot brines, Deep-Sea Res. 29, 1271-1291, 1982. P.G. Brewer, T.R.S. Wilson, J.W. Murray, R.G. Murms and C.D. Densmore, Hydrographic observations on the Red Sea brines indicate a marked increase in temperature, Nature 231, 37-38, 1971. N. Sval'nov, V.P. Strizhov, Y.A. Bogdanoff and A.B. lsayeva, Hydrothermal barite crusts on basalts from Atlantis II Deep (Red Sea), Oceanology 24, 716-720, 1984. D.E. White, Hydrology, activity and heat flow of the Steamboat Springs thermal system, Washoe County, Nevada, Geol. Surv., Prof. Pap. 458-C, 109 pp., 1968. G. Blanc, Grochimie de la fosse Atlantis II (Mer Rouge): ~volution spatio-temporelle et rrle de l'hydrothermalisme, 140 pp., Ph.D. Thesis, Universit6 P.M. Curie, Paris, 1987. M. Urvoy, Apport de l'estimation grostatistique de l'rpaisseur des unitrs m~tallif~res dans la comprrhension des m~canismes de mise en place des srdiments de la fosse

210

55 56

57

58

59

60 61

d'Atlantis II (Mer Rouge), Thesis, Universit6 Orlrans, 1987 (unpublished). L. Merlivat, C. Beauverger, C. Brunet and A. Poisson, Rapport de campagne MD3/MEROU, 82-04 (Fasc. 1). L.-G. Danielsson, D. Dyrssen and A. Graneli, Chemical investigation of Atlantis II and Discovery brines in the Red Sea, Geochim..Cosmochim. Acta 44, 2051-2065, 1980. K.S. Pitzer, J.C. Peiper and R.H. Busey, Thermodynamic properties of aqueous sodium chloride solutions, J. Phys. Chem. Ref. Data 13, 1-102, 1984. S.L. Phillips, A. Igbene, J.A. Fair, H. Ozbek and M. Tavana, A technical databook for geothermal energy utilization, 45 pp., Lawrence Berkeley Laboratory, University of California, Berkeley, Calif., 1981. H.C. Helgeson, Prediction of the thermodynamic properties of electrolytes at high pressures and temperatures, in: Chemistry and Geochemistry of Solutions at High Temperatures and Pressures, D. Rickard and F.E. Wickham, eds., pp. 133-175, Pergamon Press, New York, N.Y., 1982. M. Born, Volumen und hydratationsw~,irme der Ionen, Z. Phys. 1, 45-49, 1920. R.H. Wood, J.R. Quint and J.-P.E. Grolier, Thermody-

c. RAMBOZAND M. DANIS

62

63

64

65

66

narnics of a hard sphere in a compressible dielectric fluid. A modification to the Born equation to include the compressibility of the solvent, J. Phys. Chem. 85, 3944-3949, 1981. D.J. Bradley and K.S. Pitzer, Thermodynamics of electrolytes, 12. Dielectric properties of water and Debye-Hiickel parameters to 350°C and 1 kbar, J. Phys. Chem. 83, 1599-1603, 1979. Y.G. Zhang and J.D. Frantz, Determination of the homogenization temperatures and densities of supercritical fluids in the system NaC1-KCI-H20 using synthetic fluid inclusions, Chem. Geol. 64, 335-350. C.A. Angell, M. Oguni and W.J. Sichina, Heat capacity of water at extremes of supercooling and superheating, J. Phys. Chem. 86, 998-1002, 1982. M. Hartmann, Sound velocity data for the hot brines and corrected depth of interfaces in the Atlantis II Deep, Mar. Geol, 12, 1116-1120, 1972. A.R. Miller, C.D. Densmore, E.T. Degens, J.C. Hathaway, F.T. Manheim, P.F. McFarlin, R. Pocklington and A. Jokela, Hot brines and recent iron deposits in deeps of the Red Sea, Geochim. Cosmochim. Acta 30, 341-359, 1966.