Vertical, horizontal, and temporal changes in temperature in the Atlantis II and Discovery hot brine pools, Red Sea

https://doi.org/10.1016/j.dsr.2012.02.006Get rights and content

Abstract

In October 2008, we measured temperature and salinity in hot, hypersaline brine filling the Atlantis II and Discovery Deeps on the Red Sea spreading center west of Jeddah, Saudi Arabia. In agreement with previous observations in the Atlantis II Deep, we found a stack of four convective layers with vertically uniform temperature profiles separated by thin interfaces with high vertical temperature gradients. Temperature in the thick lower convective layer in the Atlantis II Deep continued to slowly increase at 0.1 °C/year since the last observations in 1997. Previously published data show that the temperature of all four convective layers increased since the 1960s at the same rate, from which we infer that diffusive vertical heat flux between convective layers is rapid on time scales of 3–5 years and, thus, heat is lost from the brine pools to overlying Red Sea Deep Water. Heat budgets suggest that the heat flux from hydrothermal venting has decreased from 0.54 GW to 0.18 GW since 1966. A tow-yo survey found that temperature in the upper convective layers changes about 0.2 °C over 5–6 km but the temperature in the lower brine layer remains constant. Temperature in the lower convective layer in the Discovery Deep remains unchanged at 48 °C. To explain these results, we hypothesize that heat flux from a hydrothermal vent in the floor of the Discovery Deep has been stable for 40 years, whereas temperature of the brine in the Atlantis II Deep is adjusting to the change in hydrothermal heat flux from the vent in the Southwest Basin. We found no changes in the upper transition layer at 1900–1990 m depth that appeared between 1976 and 1992 and suggest that this layer originated from the seafloor elsewhere in the rift.

Highlights

► Maximum temperatures are 68.3 °C in Atlantis II Deep and 45.0 °C in Discovery Deep. ► Temperature of all convective layers increased since the 1960s at the same rate. ► Heat is lost from the brine pools to overlying Red Sea Deep Water. ► New upper brine layer at 1900–1990 m appeared sometime between 1976 and 1992. ► Hydrothermal heat flux decreased since 1966 from 0.54 GW to 0.18 GW.

Introduction

In October 2008, we surveyed water mass properties in and around the Atlantis II and Discovery brine pools (Fig. 1). This paper describes our results and our effort to use these observations to extend the understanding of processes affecting the brine pools determined by previous surveys.

The Atlantis II and Discovery Deeps in the Red Sea were the first hydrothermal sites discovered in the oceans (temperatures of about 56 °C and 45 °C, respectively, in 1965), and their brine pools have been monitored the longest. Anomalously high temperatures were serendipitously observed near the seafloor along the axis of the Red Sea in 1948 by the R/V Albatross and in 1958 by the R/V Atlantis. The results, however, were not published in the open literature (Bruneau et al., 1953, Neuman and Densmore, 1959, Miller, 1969, Swallow and Crease, 1965). The anomaly was confirmed in 1963 by sampling from the R/V Atlantis II and R/V Discovery as they transited through the Red Sea during the International Indian Ocean Expedition (Miller, 1964, Swallow and Crease, 1965). The hot, saline brines of the Atlantis II and Discovery and Chain Deeps were first systematically investigated and mapped by the R/V Chain in 1966 and the results presented in Degens and Ross, 1969, Swallow, 1969. Munns et al. (1967) reported an increase of 0.56 °C between February 1965 and October 1966. Brewer et al. (1971) showed that temperatures in two vertically uniform brine layers of the Atlantis II Deep had increased by 2.7 °C (lower convective layer, LCL) and 5.6 °C (upper convective layer, UCL1) from November 1966 to February 1969. Subsequent measurements showed that temperatures continued to increase reaching 67.1 °C in 1997, although the rate of increase slowed somewhat during the 1970s (Bubnov et al., 1977, Schoell and Hartmann, 1978, Hartmann, 1980, Monin and Plakhin, 1982, Blanc and Anschutz, 1995, Hartmann et al., 1998a, Hartmann et al., 1998b). In contrast, the temperature of the lower layer in the Discovery Deep remained constant at about 44.7 °C.

The vertical temperature structure of the brine in the two basins differs, as well. Munns et al. (1967) showed that brine in the Atlantis II Deep in 1965 comprises (1) a high-temperature (56.5 °C) bottom layer below about 2050 m depth with thickness greater than 140 m and (2) a 100 m thick transition layer characterized mid-way by a 30–40 m thick layer with uniform temperature (44 °C). From continuous temperature measurements obtained in 1992, Blanc and Anschutz (1995) recognized three uniform layers in the transition layer (UCL1–3) separated by thin interfaces with thicknesses of 1–2 m. Although early observations in the Discovery Deep by Ross and Hunt (1967) also found a single uniform layer at about 36 °C interrupting the vertical temperature gradient in the transition layer, subsequent investigations found a continuously varying transition layer (Bubnov et al., 1977, Danielsson et al., 1980, Winckler et al., 2001; Schmidt et al., 2003). Thus, the transition layer in the Atlantis II basin became more structured, whereas the transition layer in the Discovery Basin became less structured.

Another temporal change observed since the 1960s is the growth in thickness of the transition layer of both basins. Early measurements placed the base of Red Sea Deep Water above both basins at about 1945 m depth (Swallow and Crease, 1965, Brewer et al., 1965, Munns et al., 1967, Ross, 1972, Bubnov et al., 1977). In 1992 Blanc and Anschutz (1995) observed a second transition layer with a distinctly lower vertical temperature gradient that extended up to about 1900 m depth. Observations reported by Winckler et al., 2001 and Schmidt et al. (2003) indicate that this new transition layer was present in 1997. In calculating heat and salt fluxes for the deep, Anschutz and Blanc (1996) treated the new layer as an expansion of the transition zone due to addition of new heat from the seafloor.

Spatial as well as vertical variations in the temperature of the lower layer were observed in 1971. Schoell and Hartmann (1973) found a drop of about 1.7 °C in maximum temperature of the lower brine layer over >12 km from the Southwest Basin of the Atlantis II Deep to the North Basin. To explain these observations, they hypothesized a hydrothermal vent in the Southwest Basin and a clockwise “spreading” of recently injected hot water through the sub-basins of the Atlantis II brine pool. In 1980, Monin and Plakhin (1982) observed a similar spatial change in temperature of the LCL and reported a northward decrease in temperature of the UCL1, as well. The gradient in the lower brine layer, however, is not a permanent feature of the basin. Blanc and Anschutz (1995) re-sampled the basins in 1992 and found no significant lateral temperature differences.

Previous investigators concluded that the vertical structure of brine fluids is consistent with processes of turbulent convective mixing in the uniform layers and double diffusion across the interfaces between the convecting layers (e.g., Turner, 1969, Schoell and Hartmann, 1973, McDougall, 1984, Blanc and Anschutz, 1995). Double diffusion is characterized by faster molecular diffusion of heat than salt with opposing effects on density (Schmitt, 1994). Turner (1969) first applied this model to the Atlantis II Deep based on similarities of early field observations to the results of theoretical and laboratory models of double diffusion. He suggested that the addition of heat and salt from below occurs in a turbulent plume, presumably from a hydrothermal vent, although no such vent has yet been directly observed in the Atlantis II Deep. Turner further supposed that vent turbulence established the lower convective layer, whereas the intermediate uniform layers originated by mixing accompanying the breaking of internal waves along the sides of the basin. In the model, convection in the intermediate layers is maintained by more rapid diffusion of heat than salt across their lower boundaries. Later field observations confirmed several features of Turner's hypothesis. Internal waves on a brine surface were observed by submersible (Monin and Plakhin, 1982). The system of mixed layers and high gradient interfaces also resembles the laboratory experiments of Huppert and Linden (1979) that involved heating a stable salinity gradient from below. Application of a heat flux caused layering to propagate upward into the salt gradient, with lower layers growing by merging and diffusive fluxes across the interfaces driving thermal convection in the mixed layers. In the Atlantis II Deep, Voorhis and Dorson (1975) placed an instrument that recorded vertical current flow in the upper convecting layer, and recovered the current meter 3.5 day later after a lateral drift of 3 km to the east for a mean drift of about 1 cm/s. The results showed turbulent vertical motion on scales of 1 m or less and unsteady convection on the scale of the layer (30 m).

Erickson and Simmons (1969) measured sediment temperature in both the Atlantis II and Discovery Deeps. In some cores, they found non-linear increases in temperature, whereas in others, they found no change in temperature with depth below seafloor. The maximum temperature was 62.5 °C. Assuming that some cores over-penetrated, they linearized temperature gradients just below the seafloor and computed conductive heat flow of 630–840 mW/m2 for the Atlantis II Deep and 500–630 mW/m2 for the edge of the Discovery Deep.

There have been no measurements of the hydrographic properties of the Atlantis II or Discovery Deeps since a visit by the R/V Sonne in 1997 (Hartmann et al., 1998b, Winckler et al., 2001, Schmidt et al., 2003). In October 2008, the R/V Oceanus visited the Red Sea to make hydrographic and microbiological observations. During this cruise, a number of opportunistic hydrocasts were made into the brine layers to determine whether changes in temperature and salinity had occurred during the past decade.

Section snippets

Methods

Most of the new data collected from the Atlantis II and Discovery Deep brine pools were obtained using one of two internally-recording temperature sensors manufactured by Onset Computer Corporation of Onset, Massachusetts. The “Hobo” Stainless Temperature Data Loggers with 5-inch probes, Model U12-015-02, were set to sample at 1 Hz. PVC housings were manufactured to provide some protection for the external probes during deployment and recovery. The Hobo logger was attached either to the

Results – temperature and salinity structure

Fig. 3 shows the five Hobo temperature profiles from the center of the Southwest Basin of Atlantis II Deep (location in Fig. 1). The 2008 observations show that the four convective layers documented by Blanc and Anschutz (1995) and Hartmann et al., 1998a, Hartmann et al., 1998b still exist, including the Lower Convective Layer (LCL) and Upper Convective Layers UCL1, UCL2 and UCL3. A new convective layer UCL4 is developing above UCL3 (∼42.5 °C, 2002–2006 m; orange box in Fig. 3(a)). Above UCL4,

Convective layers

The rate of change in temperature for all four convective layers in the Atlantis II Deep is almost uniform since the early 1970s. This is particularly remarkable for time spans as short as 5 years between 1992 (Blanc and Anschutz, 1995) and 1997 (Hartmann et al., 1998b). This observation reveals an important aspect about the heat balance of the brine: upward diffusion of heat across the thin interfaces with high temperature gradients is rapid on the scale of years. If the exchange of heat

Conclusions

Brine in the Atlantis II Deep is vertically structured in paired layers comprising a vertically homogeneous convective layer (each 5–120 m thick) and a thinner interface with steep vertical temperature gradient (thickness 1–8 m) across which heat and salt are transported by diffusion. This system of convective and double diffusion layers fills both the Atlantis II and Discovery Basins up to the sill leading between the basins (∼1990 m). Historical records indicate that the 80–90 m thick Upper

Acknowledgments

We gratefully acknowledge the financial and logistical support for the R/V Oceanus cruise provided by King Abdullah University of Science and Technology (KAUST), as well as the support of an international group of scientists and technicians during the cruise. This research is based on work supported by Award Nos. USA 00002, KSA 00011 and KSA 00011/02 made by KAUST. Guidance and support in Saudi Arabia was provided by Y. Kattan, A. Al-Suwailem, H. Al-Jahdali, and J. Luyten. In the US, ship

References (47)

  • M. Schmidt et al.

    High-resolution methane profiles across anoxic brine-seawater boundaries in the Atlantis-II, Discovery, and Kebrit Deeps (Red Sea)

    Chem. Geol.

    (2003)
  • M. Schoell et al.

    Detailed temperature structure of the hot brines in the Atlantis II Deep area (Red Sea)

    Mar. Geol.

    (1973)
  • A.D. Voorhis et al.

    Thermal convection in the Atlantis II hot brine pool

    Deep-Sea Res.

    (1975)
  • G. Winckler et al.

    Constraints on origin and evolution of Red Sea brines from helium and argon isotopes

    Earth Planet. Sci. Lett.

    (2001)
  • R.A. Zierenberg et al.

    Isotopic constraints on the origin of the Atlantis II, Suakin and Valdivia brines, Red Sea

    Geochim. Cosmochim. Acta

    (1986)
  • R.N. Anderson et al.

    The relationship between heat flow, sediment thickness, and age in the Eastern Pacific

    J. Geophys. Res.

    (1976)
  • P. Anschutz et al.

    The development of layering, fluxes through double-diffusive interfaces, and the location of hydrothermal sources of brines in the Atlantis II Deep: Red Sea

    J. Geophys. Res.

    (1998)
  • H. Bäcker et al.

    New deeps with brines and metalliferous sediments in the Red Sea

    Nature

    (1972)
  • H. Bäcker et al.

    Morphology of the Red Sea central graben between Subair Islands and Adul Kizan

    Geol. Jahrb. Reihe D

    (1975)
  • G. Blanc et al.

    New stratification in the hydrothermal brine system of the Atlantis II Deep, Red Sea

    Geology

    (1995)
  • Bower, A., 2009. R/V Oceanus voyage 449-6 Red Sea Atlantis II Deep complex area. Woods Hole Oceanographic Institution...
  • P.G. Brewer et al.

    The chemical composition of the hot salty water from the bottom of the Red Sea

    Nature

    (1965)
  • P.G. Brewer et al.

    Hydrography of the Red Sea brines

  • Cited by (41)

    • Chemical characteristics of modern deep-sea metalliferous sediments in closed versus open basins, with emphasis on rare-earth elements and Nd isotopes

      2021, Earth-Science Reviews
      Citation Excerpt :

      The metalliferous sediments of the Atlantis II Deep constitute a major deposit of base and precious metals with an estimated in-situ value of US$ 11 billion (Brueckmann et al., 2017); mining schemes have been proposed for piping the sediments to a surface ship, with primary processing carried out at sea (Mustafa and Amann, 1980; Nawab, 1984; Thiel et al., 2015; Hamer, 2017). Detailed investigations have been made on the temperature and chemical composition of the brine layers that overlie the metalliferous sediments in the Atlantis II Deep (Ross, 1972; Schoell and Faber, 1978; Danielsson et al., 1980; Hartmann, 1985; Zierenberg and Shanks, 1986; Anschutz et al., 1995, 1998, 2000; Anschutz and Blanc, 1996; Winckler et al., 2000, 2001; Pierret et al., 2001; Swift et al., 2012; Anschutz, 2015; Schardt, 2016), and on microbial communities both in the hot brines (Eder et al., 2001; Qian et al., 2011; Wang et al., 2011, 2013) and in the underlying sediments (Siam et al., 2012). Several other brine pools along the spreading axis of the Red Sea (shown in Fig. 2) have also been studied, although in less detail than the Atlantis II Deep (Bignell, 1975; Blum and Puchelt, 1991; Scholten et al., 1991; Pierret et al., 2010).

    • Molecular and functional characterization of unique thermo-halophilic thioredoxin from the metagenome of an exotic environment

      2020, International Journal of Biological Macromolecules
      Citation Excerpt :

      The deep-sea brine pools like Atlantis II Deep, Discovery Deep, and Kebrit Deep are particularly attractive due to their distinctive physical and geochemical characteristics. The Atlantis II Deep is known for its harsh and extreme environment with a temperature around 67.1 °C and a salinity of 252 PSU [3]. The Atlantis Deep is the largest brine pool in the Red Sea, which covers a surface area of 60 km2.

    • Dynamics of diffusive layering and chemocline formation in Lake Kivu and brine pools

      2019, Journal of African Earth Sciences
      Citation Excerpt :

      In Lake Vanda (Antarctica) with its temperature and salinity profiles being similar to those of Lake Kivu, the frequency of the annual oscillations of solar heat input seems to be too high significantly to affect the DL profile (Hoare, 1966). Sometime after the year 1990, hydrothermal activity in the upper part of the Atlantis II brine pool might have generated a new developing chemocline (at 1995–2002 m depth) and a new developing mixed layer (at 2002–2006 m depth, Swift et al., 2012). This system has not yet been observed to develop towards a steady state.

    • Nitrous oxide in the northern Gulf of Aqaba and the central Red Sea

      2019, Deep-Sea Research Part II: Topical Studies in Oceanography
    • The red sea

      2018, World Seas: An Environmental Evaluation Volume II: The Indian Ocean to the Pacific
    View all citing articles on Scopus
    View full text