U.S. Dept. of Commerce / NOAA / OAR / PMEL / Publications

Hydrothermal Plumes Over Spreading-Center Axes: Global Distributions and Geological Inferences

Edward T. Baker

NOAA Pacific Marine Environmental Laboratory, Seattle, Washington, USA

Christopher R. German

Institute of Oceanographic Sciences, Wormley, Surrey, UK

Henry Elderfield

Department of Earth Sciences, Cambridge University, Cambridge, UK

Seafloor Hydrothermal Systems: Physical, Chemical, Biological, and Geological Interactions, Geophysical Monograph 91, AGU, 47-71, 1995.
Copyright 1995 by the American Geophysical Union. Further electronic distribution is not allowed.


Precursors of Modern Plume Surveys

Even before the discovery of submarine hydrothermal vents in 1977 [Corliss et al., 1979] and 1979 [RISE Project Group, 1980], a few investigators had recognized large-scale physical and chemical anomalies that were rightly attributed to hydrothermal discharge. Over 30 years ago, Knauss [1962] suggested that abnormally warm water between 3000 and 3500 m over the East Pacific Rise (EPR) near 20S was probably related to a recently mapped zone of high seafloor heat flow in the same area [Von Herzen, 1959]. Ten years later, Warren [1973] noticed the same feature in data from the SCORPIO expedition at 28S and 43S, although he was reluctant to attribute the anomaly to geothermal effects since purely oceanographic effects might also lead to the same temperature distribution. Lonsdale [1976] made perhaps the earliest attempt to quantify the temperature anomaly of a hydrothermal plume when he plotted temperature-salinity curves of profiles west, east, and directly over the EPR at 8S. The axial stations revealed near-bottom temperature increases of several hundredths of a degree Celsius relative to off-axis water of the same salinity (Figure 2).

Fig. 2. Plot of potential temperature vs. salinity for discrete samples near the East Pacific Rise at 8S showing hydrothermal heating at the rise crest (reprinted from Lonsdale [1976]).

Measurements of the helium isotopes 3He and 4He in the Kermadec Trench [Clarke et al., 1969] and across the EPR [Craig et al., 1975] provided early evidence that hydrothermal venting could produce widely dispersed chemical anomalies. Shortly thereafter, sampling near the Galapagos Ridge by Klinkhammer et al. [1977] and Bolger et al. [1978] showed that hydrothermal plumes also contain significant anomalies of dissolved and particulate Mn. Since then, plume surveys have contributed to the discovery and characterization of hydrothermal discharge sites on ridges and hotspots throughout the oceans and marginal seas.

Eastern Pacific Spreading Centers

Hydrothermal plumes have been mapped in more detail over eastern Pacific ridge crests than anywhere else in the ocean. Virtually continuous surveys of plume hydrographic, optical, and/or chemical anomalies have been conducted along much of the Explorer-Juan de Fuca-Gorda Ridge complex in the northeast Pacific, in the Gulf of California, and along the EPR from about 9 to 13N and 13 to 22S. Full-rate spreading speeds at these sites range from a medium rate of 60 mm yr-1 in the northeast Pacific and the Gulf of California, to intermediate rates of 110-120 mm yr-1 near 10N, to superfast rates of >150 mm yr-1 near 20S [DeMets et al., 1990]. The comprehensive nature of plume maps in the eastern Pacific results in large part from the narrowness of the axial ridge crest at medium and greater spreading rates, which allows a single along-axis transect to detect and locate most discharge sources.

Fig. 3. Location map for northeast Pacific spreading centers.

The Explorer-Juan de Fuca-Gorda Ridge complex is an isolated set of spreading ridges that extend for almost 1000 km along the west coast of Canada and the United States (Figure 3). The Juan de Fuca Ridge (JDFR) has been surveyed extensively, while Explorer and Gorda Ridges have benefitted only from site-specific work. McConachy and Scott [1987] used real-time nephelometer signals to map the distribution of plumes over a 40 km2 area of Southern Explorer Ridge. Specific plume maxima were linked to two known vents sites near 4946'N, 13016'W, plus undiscovered vent fields farther north. On the Gorda Ridge, plumes from high-temperature venting appear restricted to the northern third of the axial valley [Baker et al., 1987a; Collier and Baker, 1990]. Discharge from vents within the thickly sedimented Escanaba Trough in the southern third of the Gorda Ridge produces plumes difficult to detect chemically, since conductive cooling and precipitation removes most of the dissolved metals during passage through the sediments [Campbell et al., 1994; German et al., in press].

Plumes emanating from the JDFR have been mapped more intensively and extensively than those over any other group of ridge-crest segments in the ocean. The first indication of JDFR hydrothermal plumes came in 1980 from observations of dissolved Mn [Jones et al., 1981] and 3He [Lupton, 1990] at five widely spaced stations stretching from 4431'N to 4747'N. In 1982, Crane et al. [1985] conducted the first large scale, continuous survey of hydrothermal plumes anywhere on the global midocean ridge (MOR). They attempted to measure hydrothermally induced temperature anomalies in plumes by hanging a 50-m-long thermistor chain from a side-scan sonar vehicle while towing it along the entire JDFR. Their measurements were successful in identifying a few major thermal anomalies but did not fully reveal the actual plume distribution because of calibration problems and the very slow (~2 min) response time of the thermistors. Nevertheless, the methods and strategy of Crane et al. [1985] were precursors for many studies that later used two-dimensional transects of temperature and chemical anomalies to approximately locate vent sources, estimate hydrothermal heat and mass fluxes, and relate spatial and temporal variability of venting to the magmatic budget of the ridge.

Plume mapping on the JDFR has focused on three areas: the Endeavour segment, Axial Volcano, and the Cleft segment (Figure 3). Along the Endeavour segment, Baker and Massoth [1987] and Thomson et al. [1992] used transects of temperature and light attenuation anomalies to characterize the plume and discover vent locations. They combined these anomalies with concomitantly collected records of plume advection from moored current meters to estimate mean heat fluxes of 1700 1000 MW [Baker and Massoth, 1987] and 800 490 MW [Thomson et al., 1992]. Thomson et al. [1992] also reported an "instantaneous" heat flux of 12,000 6000 MW based on temperature and plume velocity measurements directly over the central vent site. Rosenberg et al. [1988] combined temperature anomaly profiles with discrete sampling of 222Rn and 3He to make a geochemical estimate of heat flux based on the premise that the loss of 222Rn from the plume by decay is exactly balanced by the constant addition of hydrothermal 222Rn. Their best estimate of heat flux was 3000 MW. Although the 3He plume from the summit of Axial Volcano is traceable for tens of kilometers [Lupton, 1990], temperature and light attenuation anomalies are detectable only within the summit caldera. Baker et al. [1990] used temperature anomaly and plume advection measurements to estimate a heat flux 800 MW from the caldera.

Fig. 4. CTD tow-yo transects of hydrothermal temperature anomaly (C) along the Cleft segment axial valley, 1986-1991. Triangles show location of known venting sites. The lower half of event plume EP86 ("megaplume") is apparent in the 1986 transect. The anomaly at 1800 m in 1988 is a hydrographic rather than hydrothermal feature [Baker, 1994].

The plume distribution over the entire Cleft segment has been mapped one or more times every year since 1986 [Baker and Massoth, 1986; Baker and Massoth, 1987; Baker, 1994], making it a unique source of information about the temporal variability of hydrothermal venting. Discharge from the Cleft segment is concentrated in two areas centered at 4438'-4443'N and 4454'-45N (Figure 4). The first plume survey to estimate hydrothermal fluxes was conducted at the southern site in 1985 [Baker and Massoth, 1986; 1987], finding a heat flux of 580 350 MW. A six-year average of plume extent and temperature anomaly at the northern site, combined with long-term current meter measurements of advection at plume depths [Cannon et al., 1991], yielded a flux estimate of 660 280 MW [Baker, 1994]. Continuous monitoring of the plume at the northern site was begun in 1991 by means of moored temperature sensors and current meters [Baker and Cannon, 1993].

Fig. 5. (A) Cross-section of EP86 and underlying chronic plume (reprinted from Baker et al. [1989]). Contours are isolines of hydrothermal temperature anomaly (C). Potential density (kg m-3) isolines shown as dashed lines. Thin line in sawtooth pattern is track of the CTD tow-yo path. (B) Plan view showing locations of EP86 and EP87 in relation to known hydrothermal fields (stars) and volcanic eruption mounds emplaced between 1981 and 1987 (solid beads). Megaplume contours in temperature anomaly. Origin of EP87 is unknown.

Fig. 6. (A) Cross-sections of two event plumes mapped over the CoAxial segment in July, 1993 [Baker et al., 1995]. Both plumes were smaller and had lower temperature anomalies than EP86 or EP87. (B) Plan view of all three CoAxial event plumes showing their relative locations when first discovered. EP93A and EP93B were found directly over the lava eruption mound (solid bead) and warm-water fissure (heavy line) near 4630'N. EP93C was found two weeks after EP93A and EP93B; its discharge site is unknown.

The Cleft segment time series is proving especially useful in revealing the linkage between hydrothermal and magmatic activity, because it began simultaneously with the observation of a new class of plume phenomena: "event plumes." In 1986, Baker et al. [1987b] discovered a "megaplume" at the northern end of the Cleft segment (Figure 5). Analysis of this and a second megaplume found over the Vance segment in 1987 [Baker et al., 1989; Gendron et al., 1993] demonstrate that event plumes are near-instantaneous releases of enormous volumes of hydrothermal fluid; about108 m3 in the case of the 1986 megaplume. The subsequent discovery of volcanic extrusions emplaced at the north Cleft megaplume site sometime between 1983 and 1987 [Chadwick et al., 1991] left little doubt that event plumes are a signature of the creation or reinvigoration of a vent field by local magmatic activity. This hypothesis was confirmed in 1993 by a series of spectacular observations on the CoAxial segment of the JDFR, just north of Axial Volcano. Between June 26 and July 13, 1993, NOAA/VENTS monitoring of the U.S. Navy's SOSUS hydrophone network detected seismic signals from the JDFR between 4614'N, 12950'W and 4631'N, 12936'W [Fox, in press]. A series of cruises to this area over the next four months found new volcanic flows, apparently new vent fields, and several examples of event plumes with volumes up to 30% of the 1986 megaplume (Figure 6) [Baker et al., 1995; Embley et al., 1995].

Fig. 7. Summary of hydrothermal temperature anomalies observed along the crest of the Juan de Fuca Ridge, including Axial Volcano, from 1985 to 1989. For clarity, segments and the overlying plumes are displayed individually, without overlap. Triangles show location of known venting sites. Anomalies at the northern end of the Cobb segment, 47.4-47.6N, are believed to mark an intrusion of geothermally heated bottom water from Cascadia Basin to the east, rather than axial hydrothermal heating (reprinted from Baker and Hammond [1992]). Significant hydrothermal plumes ( 0.02C) cover ~25% of the Juan de Fuca Ridge. This figure does not reflect recent work at the N. Axial (CoAxial) segment.

In addition to intensive work along the Cleft segment and at other specific sites, the NOAA/VENTS Program also undertook to survey, at least once, as much of the entire JDFR as feasible. As of 1989, about 90% of the ridge from the Cleft to Endeavour segments had been surveyed at least once, and more than 50% at least twice, by a continuous tow-yo transect. The goal of this program was to test the hypothesis that the distribution and intensity of hydrothermal activity are directly related to the variable magmatic budget along the ridge. Baker and Hammond [1992] reported that hydrothermal discharge is strongest on those segments, or portions of segments, where the apparent magmatic budget is highest, as indicated by the degree of along-axis inflation or other morphological characteristics (Figure 7).

Fig. 8. Location map for East Pacific Rise spreading centers. Solid circles indicate locations of axial discontinuities.

Three areas on the EPR, one medium (Gulf of California, 50-60 mm yr-1), one fast (9-13N, 100-120 mm yr-1), and one superfast (14-23S,155 mm yr-1) spreading, have been surveyed extensively enough to produce maps of plume distributions at multisegment scales (Figure 8). Two of the sites (the fast and superfast regions) have hosted multichannel seismic investigations of the axial crest [Detrick et al., 1987; 1993], so the extent and depth of the axial magma chamber (AMC) furnishes some indication of the apparent magmatic budget along these ridge crest sections.

Fig. 9. A (3He) section through the Gulf of California showing evidence of hydrothermal venting in the Guaymas Basin (reprinted from Lupton [1979]). Triangle shows location of known venting sites in the Guaymas Basin.

The Gulf of California is a series of en echelon spreading centers opening at 50-60 mm yr-1 and creating a string of deep, semi-enclosed basins (Figure 9). Helium isotope profiles in each of the six major basins [Lupton, 1979] imply that only Guaymas Basin, with (3He) values of 65-70% ((3He) = ((R/Ratm) - 1) 100, where R = 3He/4He in the fluid or atmosphere, i.e., the 3He/4He ratio was 65-70% higher than in atmospheric helium), sustains active venting (Figure 9).

Plumes over the fast-spreading section of the EPR between 8 and 13N have been systematically studied since at least 1983. Crane et al. [1988] used the same techniques employed by Crane et al. [1985] on the JDFR to identify perturbations in the temperature field between 820' and 1310'N along the ridge crest, but the results were similarly compromised by the equipment used. This same area has since been resurveyed in detail, by Dynamic Hydrocast and CTD casts from 1208' to 1312'N in 1986 [Bougault et al., 1990; Charlou et al., 1991b], and by CTD/transmissometer tows and casts from 840' to 1150'N in 1991 [Lupton et al., 1993; Baker et al., 1994; Feely et al., 1994; Mottl et al., 1995]. These studies precisely mapped the distribution of hydrothermal plumes along the ridge crest, revealing a fine-scale agreement between hydrothermal activity and the apparent magmatic budget, as inferred from ridge morphology and subsurface structure [Bougault et al., 1990; Baker et al., 1994].

Fig. 10. Transects of methane and dissolved manganese along the East Pacific Rise determined from three Dynamic Hydrocast tows [after Bougault et al., 1990]. Connected triangles show area of known venting sites.

Fig. 11. Transects of light-attenuation anomaly (continuous tow-yo) [Baker et al., 1994], dissolved Mn, and CH4 (both from discrete samples) [Mottl et al., 1995] obtained in November 1991 along the East Pacific Rise south and north of the Clipperton Transform Zone. Contour intervals are 0.004 m-1 (with an additional dotted contour at 0.002 m-1) for light-attenuation anomaly, 5 nmol/L for Mn, 5 nmol/L for CH4 south of the Clipperton Transform Zone, and 1 nmol/L for CH4 north of the Clipperton Transform Zone. Triangles show location of known venting sites; double-headed arrow shows extent of April 1991 lava eruption [Haymon et al., 1991, 1993]. Note strong CH4 and relatively weak Mn anomalies directly above the eruption zone, in contrast to the reverse trend over presumably "mature" vent fields near 1118'N. Significant hydrothermal plumes (light-attenuation anomaly >0.004 m-1 or Mn > 10 nmol/L) cover ~38% of the EPR crest surveyed in Figures 10 and 11.

Between 1208' and 1312'N, Bougault et al. [1990] and Charlou et al. [1991b] found plumes with dissolved Mn >10 nmol kg-1and CH4 >2.5 nmol kg-1 from 1224' to 1253'N, plus a small zone of high values from 1211' to 1215'N; plume maxima were centered between 1240' and 1250'N (Figure 10). Between 840' and 1150'N, plumes were mapped from distributions of temperature and light-attenuation anomalies [Baker et al., 1994], dissolved Mn and CH4 [Mottl et al., 1995], and particulate chemistry [Feely et al., 1994]. Based on temperature and light-attenuation anomalies from continuous tow-yo data, intense and spatially continuous plumes were found from 848' to 858'N, 929' to 1001'N and 1105' to 1127'N. Chemical data are available only at discrete locations, but the plume maps based on chemical species such as CH4 and dissolved Mn agree with the hydrographic and optical data (Figure 11).

Plume sampling on the southern EPR has been concentrated along the superfast-spreading (150-156 mm yr-1) portion of the EPR between the Garret Transform Fault at 1330'S and the Easter microplate at ~23S. The first unequivocal indicator of active hydrothermal venting there was a westward trending 3He plume mapped along a cross-axis transect at 15S [Lupton and Craig, 1981]. A series of five along-axis hydrocasts between 1924' and 2215'S found concentrations of total dissolvable Mn as high as 49 nmol kg-1, with the highest values occurring over an axial high at 2122'S [Klinkhammer and Hudson, 1986]. More recent and detailed work in this same area (2120' to 2240'S) revealed intense plume anomalies in light attenuation near 2140'S [Krasnov et al., 1992].

Fig. 12. Tow-yo transects of light attenuation along the East Pacific Rise just south of the Garrett Transform Zone [Baker and Urabe, 1994]. Nephelometer data are combined from five tow-yos, comprising >1000 vertical profiles. Contour interval is 0.01 volts, 0.02 m-1 as measured by a simultaneously recording beam transmissometer. Plume signals are present almost everywhere, but increase sharply south of 1730'S. Significant hydrothermal plumes (nephelometer volts >0.022) cover ~60% of this survey area.

The first large-scale plume survey in this area was completed in 1993 by the Japanese-NOAA Ridgeflux Expedition using detailed chemical sampling and nearly continuous CTD/SUAVE/optical tow-yos between 1350' and 1840'S [Urabe et al., 1994; Baker and Urabe, 1994; Massoth et al., 1994; Ishibashi et al., 1994; Feely et al., 1994; Lupton et al., 1994]. Preliminary analyses show that hydrothermal plumes covered ~60% of the entire study area (Figure 12). Plumes were most common south of ~1630'S, becoming virtually continuous between 1720' and 1840'S. The southern portion of the study area was also characterized by CH4/Mn ratios as high as 3.9, particulate S/Fe ratios >1, and Fe/Mn ratios >8. Similarly high CH4/Mn and S/Fe ratios were found in 1991 in plumes at 950'N on the EPR [Lupton et al., 1993], 6 months after a seafloor lava eruption at that site [Haymon et al., 1993]. An increased release of heat and volatiles is typical of hydrothermal systems perturbed by a magmatic intrusion [Baker, 1995].

Western Pacific Marginal Basins

The western boundary of the Pacific Ocean is a broad expanse of marginal basins and troughs fragmented by trench-arc systems that mark subduction zones of several tectonic plates [Taylor and Karner, 1983] (Figure 13). Many of these basins contain active spreading centers that likely host hydrothermal activity. Systematic though limited plume sampling has been conducted in the Harve Trough, Lau Basin, North Fiji Basin, Woodlark Basin (Solomon Sea), Manus Basin (Bismark Sea), Mariana Trough, and Okinawa Trough.

Fig. 13. Location map for western Pacific marginal basins. Thin lines are fracture zones, thick lines are spreading centers, and triangled lines are convergent plate edges. 1, Harve Trough; 2, Lau Basin/Valu Fa Ridge; 3, North Fiji Basin; 4, Woodlark Basin; 5, Manus Basin; 6, Mariana Trough; 7, Okinawa Trough.

The 1986 Papatua Expedition [Craig and Poreda, 1987] occupied four hydrocast stations in the Harve Trough, five in the Lau Basin, and three in the Woodlark Basin. Only weak to negligible hydrothermal indications were found in each basin. Subsequent seafloor explorations in the Lau [Fouquet et al., 1991] and Woodlark [Binns et al., 1993] basins, however, have discovered active hydrothermal sites, indicating that sampling during the Papatua Expedition was too sparse to adequately survey these large basins.

Fig. 14. Transects of methane and dissolved Mn along the central spreading axis in the North Fiji Basin. Data compiled from published and unpublished sources (see text) spanning several years. Note that the data base is not identical for each tracer. Contour intervals are 1 nmol kg-1 (with additional dotted contour at 0.5 nmol kg-1) for CH4, and10 nmol kg-1 (with addition dotted contour at 5 nmol kg-1) for Mn. Triangle shows location of known venting sites. Insets show geographic location of profiles. The distributions suggest other venting sites north and south of the triple junction at 17S.

Plumes in the North Fiji Basin have been sampled more intensively than in any other back-arc basin. Geophysical surveys [Auzende et al., 1994] indicate two north-south spreading axes: a medium-rate (50-60 mm yr-1) ridge running along ~17330'E and a slow-spreading ridge along ~176E. Only the westerly ridge has been sampled. Between 21S and a triple junction at 1650'S the morphology is typical of a fast-spreading ridge, shallowing from an axial depth of ~2800 m south of 18S to <2500 m at the ridge-ridge-transform triple junction. North of the triple junction the ridge trends NNW and assumes a slow-spreading ridge morphology of 3500-4000-m-deep grabens. South of the triple junction, CH4 and dissolved Mn were first sampled on SEAPSO 3 in 1985 [Auzende et al., 1988], again in 1986 (CH4 only) by the Papatua Expedition [Craig and Poreda, 1987; Craig et al., 1987], and in 1987 on Kaiyo 87 [Nojiri et al., 1989] (J. Ishibashi, unpublished data, 1994). Sedwick et al. [1990] sampled total dissolvable Mn north of the triple junction in 1987. An along-axis distribution of the reported values (Figure 14) shows several isolated maxima between 19 and 17S. The particularly intense CH4 plume at 188'S also carried a (3He) maximum of 71%, confirming its magmatic origin [Craig et al., 1987]. Nojiri et al. [1989] interpreted as an event plume a strong plume 600 m above bottom at 1848'S with a high CH4/Mn ratio that was observed in 1987 but not in 1988. Sedwick et al. [1990] proposed that intense Mn and weak CH4 signals high above the ridge north of the triple junction were advective features from venting near the triple junction [Grimaud et al., 1991], attributing the absence of CH4 to in situ consumption or a deficit of volatiles in the original discharge (e.g., hydrothermal brine from a phase-separated system).

Fig. 15. Transects of methane and manganese collected in 1990 along the spreading center in the eastern Manus Basin (reprinted from Gamo et al. [1993] with kind permission from Elsevier Science Ltd., The Boulevard, Langford Lane, Kidlington OX5 16B, UK). The Papatua Expedition [Craig and Poreda, 1987] collected samples near stations 37 and 29. Triangle shows location of known venting sites.

The Manus Basin is a fast-spreading (>100 mm yr-1) basin with two east-northeast trending spreading axes connected by a northwest trending transform fault (Figure 13) [Taylor, 1979]. The Papatua Expedition occupied three stations over the northern ridge, finding CH4 anomalies up to 3.5 nmol kg-1 and a (3He) maximum of 58%. The Papatua Expedition also found CH4 anomalies over the southern ridge, which was sampled more extensively by Gamo et al. [1993] in 1990. Gamo et al. [1993] mapped multiply layered CH4 and Mn plumes between 1000 and 2000 m depth that indicated at least two distinct hydrothermal sites (Figure 15). CH4 concentrations in the deep plume remained steady at 1-2 nmol kg-1 between the 1986 Papatua Expedition and 1990, but no indication of shallower plumes was seen in 1986. This appearance of a shallow plume and its high CH4/Mn ratio led Gamo et al. [1993] to suggest that it was an event plume. The extremely small thickness/width ratio (<0.004) of the shallow plume relative to known event plumes (~0.03 [Baker, 1994]), however, argues against this hypothesis.

Farther to the north, plumes have been mapped in the Okinawa and Mariana Troughs. In the Okinawa Trough, a relatively shallow basin behind the deep Ryukyu Trench, strong CH4, Mn, and 3He (up to a (3He) of 65%) anomalies at depths between 1300 and 1600 m were found in 1987 [Ishibashi et al., 1988] and 1988 (J. Ishibashi, personal communication, 1994). Horibe et al. [1986] occupied a series of hydrocasts between 1811'N and 1815'N in 1982 in the Mariana Trough, mapping CH4 plumes near-bottom and 700-800 m above the seafloor. A resurvey of this same area in 1986 [Horibe and Craig, 1987] again found CH4 plumes, though somewhat differently distributed. The CH4 plumes in both years had no 3He signature.

Mid-Atlantic Ridge

Unlike fast-spreading ridge axes such as the EPR, the neovolcanic zone of the slow-spreading MAR (Figure 16) sits not at a narrow axial high, but within a broad (5-10 km wide) axial rift valley. The typical valley is bounded across axis by steep and faulted valley walls that rise 1000-2000 m above the valley floor and along axis by frequent transform-fault offsets. The height of rise of buoyant hydrothermal plumes is nearly always less than the bounding heights of the MAR rift valley. Consequently, hydrothermal plumes are trapped and transported within the axial valley. These bathymetric characteristics allow easy detection of hydrothermal tracers within a particular segment, but make precise location and enumeration of the discharge sources difficult [e.g., Klinkhammer et al., 1985; Nelsen et al., 1986/87; Charlou and Donval, 1993; Murton et al., 1994]. Plume mapping on the MAR has typically covered a greater linear distance at significantly less detail than similar studies in the eastern Pacific.

Fig. 16. Location map for the northern Mid-Atlantic Ridge and Reykjanes Ridge. Known and suspected hydrothermal sites identified in Figure 19.

Fig. 17. Along-axis maxima in total reactive (~dissolved) Mn [Klinkhammer et al., 1985] and CH4 [Charlou and Donval, 1993] from vertical casts on the Mid-Atlantic Ridge. High values in parentheses at 26N and 29N are representative samples over the TAG [Klinkhammer et al., 1986] and Broken Spur [Elderfield et al., 1993] hydrothermal fields, respectively. Thin dashed line on each plot marks the regional background value of Mn and CH4.

Until 1984 it was widely predicted that hydrothermal activity might be restricted to fast-spreading ridges and that at slow-spreading ridges, such as the MAR, heat flow would be insufficient to support active high-temperature black smoker vent fields. It is now clear that this is not the case. In 1984, Klinkhammer et al. [1985] carried out systematic discrete sampling and shipboard Mn analysis of seawater samples from stations occupied throughout the MAR rift valley between 11N (the Vema Fracture Zone) and 26N (Figure 17). Nine discrete segments were occupied during this survey and dissolved Mn anomalies--interpreted as the product of active hydrothermal venting--were found at every station. As a conservative estimate, Klinkhammer et al. [1985] concluded that a minimum of at least five discrete hydrothermal fields must characterize this section of the MAR.

Fig. 18. Geochemical plume transect for total reactive Mn at the TAG hydrothermal site in 1985 (reprinted from Klinkhammer et al. [1986] with kind permission from Elsevier Science Ltd., The Boulevard, Langford Lane, Kiddington 0X5 16B, UK). Solid circles indicate sample positions.

These predictions were confirmed in 1985 when a series of CTD-nephelometer hydrocasts (Figure 18) and deep-tow video camera deployments confirmed the first active black-smoker hydrothermal field at the northernmost station occupied by Klinkhammer et al. [1985], the TAG hydrothermal field at 26N [Rona et al., 1986]. In 1986, black-smoker hydrothermal venting was also found at the Snakepit hydrothermal field at 23N [Ocean Drilling Program Leg 106 Scientific Party, 1986], close to where Klinkhammer et al. [1985] had reported water column dissolved Mn anomalies.

Between 1985 and 1988 another series of cruises mapped the distribution of CH4 between 12 and 26N [Charlou and Donval, 1993] (Figure 17). High CH4 concentrations were found over the TAG and Snakepit sites, as well as the 1520' fracture zone. Because of very low Mn in the plumes near the 1520' fracture zone, and the occurrence of local outcrops of serpentinized ultrabasic diapirs, Charlou and Donval [1993] hypothesized that CH4 anomalies there arise from fluid circulation in ultrabasic rocks rather than basalt-seawater interactions typical of faster-spreading ridges. The changes of mechanical properties and density caused by the serpentinization of deep crustal rocks, as inferred from these CH4 anomalies, may play an influential role in the construction of slow-spreading ridges [Charlou et al., 1991a; Charlou and Donval, 1993; Bougault et al., 1993].

Fig. 19. Summary of evidence for hydrothermal venting along the Mid-Atlantic Ridge compiled from water-column investigations between 11 and 40N [German et al., 1995]. Only four segments or segment portions have confirmed active venting.

In 1992, as part of a joint FARA (French American Ridge Atlantic) cruise aboard the R/V Atlantis II, Project FAZAR, systematic water-column sampling was carried out along a more northerly section of the Mid-Atlantic Ridge between 32N and the Kurchatov Fracture Zone close to 40N [e.g., Wilson et al., in press; Klinkhammer et al., in press]. Stations were occupied approximately every 15 miles along axis using a CTD/nephelometer/transmissometer array augmented with the ZAPS Mn probe. Eleven of 19 segments were investigated and evidence for high-temperature black-smoker-type hydrothermal activity was found in seven, including the AMAR segment at 36N and the confirmed Lucky Strike hydrothermal field at 3717'N [Langmuir et al., 1993; Wilson et al., in press; Klinkhammer et al., in press]. Further studies south of the Atlantis fracture zone [Murton et al., 1994; Chin et al., 1994] have used a combination of a transmissometer and the ZAPS sensor mounted on IOS Deacon Laboratory's Towed Ocean Bottom Instrument (TOBI) to identify discrete hydrothermal signals in three further segments, between 27N and 30N, including the Broken Spur hydrothermal field at 29N. Thus, systematic water-column surveys have been carried out along most of the Mid-Atlantic Ridge rift valley between 11N and 40N and evidence for at least 15 sites of high-temperature hydrothermal venting has been found. This yields an average incidence of hydrothermal activity of at least one site every ~175 km for the ~2,500 km of ridge-crest surveyed (Figure 19) [German et al., 1995].

Reykjanes Ridge

German et al. [1994] carried out a systematic survey of the Reykjanes Ridge between 5745'N and 6309'N. Approximately 120 stations were occupied at 10-20 km intervals (i.e., at closer spacings than the 11-26N and 32-40N surveys), using a CTD/nephelometer/transmissometer array combined with shipboard analyses for total dissolvable Mn and dissolved Si, CH4, and H2. The only evidence of hydrothermal activity throughout this 750 km of ridge crest was found at the Steinahll vent field at 6306'N. The site is situated in just 250-350 m of water and is notable for the formation of bubble-rich plumes that have been imaged using a high-frequency (38 kHz) echo sounder (Figure 20). It is important to understand why such a lengthy section of ridge, patently affected by the local thermal flux of the Icelandic hot-spot, should appear to be so devoid of high-temperature hydrothermal activity.

Fig. 20. 38 kHz echo-sounder trace from a transect along the axis of the Reykjanes Ridge between 6306.04'N and 6306.10'N, indicating the presence of gas-rich hydrothermal plumes rising from the seabed close to 6306.06'N (reprinted from German et al. [1994] with kind permission from Elsevier Science Ltd., The Boulevard, Langford Lane, Kidlington OX5 1GB, UK).

Indian Ocean

Very few studies of hydrothermal activity have been carried out in the Indian Ocean, even though sediments exhibiting hydrothermal metal enrichments along the Indian Ocean ridge system have been known for almost thirty years [Bostrm et al., 1969]. To date, however, discovery of evidence for hydrothermal activity along the ridges of the Indian Ocean has been limited to three specific studies. The Central Indian Ridge between 21S and 24S, close to the Rodriguez Triple Junction, was investigated between 1983 and 1988 by cruises of the R/V Sonne. A series of CTD hydrocast stations occupied between 2115'S and 2130'S in 1986 found distinct plumes of dissolved Mn and CH4 anomalies at several stations, with concentrations up to 27.5 nmol kg-1 and 2 nmol l-1 respectively [Herzig and Plger, 1988]. Evidence for an additional site of hydrothermal activity was revealed in 1988 by the presence of ~20 nmol kg-1 dissolved Mn anomalies and up to 8.9 nmol l-1 CH4 anomalies at ~24S [Plger et al., 1990].

In 1993, a group of Japanese investigators conducted water-column observations around the Rodriguez Triple Junction [Gamo et al., 1994]. Hydrothermal optical and chemical (CH4 and dissolved Mn and Fe) anomalies were found centered around the 2250 m horizon near 2520'N over the eastern rift wall of the Central Indian Rift, the northern arm of the Rodriguez Triple Junction. Because of the shallow depth of these plumes and the lack of strong anomalies directly over the 4000-m deep rift valley axis, the source is likely located on the eastern rift valley wall [Gamo et al., 1994]. Two casts at the 24N site of the plume anomalies discovered by Plger et al. [1990] found CH4 and dissolved Mn anomalies similar to the 1988 profiles. Only negligible anomalies were found on casts in the Southeastern and Southwestern Indian Ridge valleys, the other two arms of the Rodriguez Triple Junction. Based on the distribution of deep 3He anomalies in the Indian Ocean, the Rodriguez Triple Junction area appears to be the largest source of hydrothermal effluent from Indian Ocean ridges [Jamous et al., 1992].

Return to previous section or go to next section

PMEL Outstanding Papers

PMEL Publications Search

PMEL Homepage