Hydrothermal activity on the eastern SWIR (50°–70°E): Evidence from core-top geochemistry, 1887 and 1998



[1] Evidence for hydrothermal activity on the eastern SWIR has been reported previously in the form of optical-backscatter anomalies interpreted to indicate the presence of hydrothermal plumes. Here, I report on a brief reconnaissance analysis of the geochemical composition of core-top samples collected from sites both beneath and away from those previously-reported plume signals to determine whether evidence for fall-out of hydrothermal plume material is discernible. Samples used for this study were collected using the deep-diving submersible SHINKAI 6500 in 1998 and from the tallow-coatings applied to lead sounding lines, 111 years earlier, aboard HMS Egeria. The data indicate hydrothermal input to all but one of eight SHINKAI 6500 cores along the length of the eastern SWIR rift-valley, including the site of strongest previously reported plume anomalies. Comparison with a recent MAR study suggests that the cores analyzed here, however, may predominantly lie distant from any current or recently-active source of venting.

1. Introduction

[2] First evidence for the presence of hydrothermal activity on the SW Indian Ridge was reported following the InterRidge cruise FUJI in 1997 (PIs: C.Mevel and K.Tamaki). In that work, miniature autonomous plume recorder (MAPR) instruments [Baker and Milburn, 1997] were suspended above and below the TOBI deep-tow vehicle as the latter conducted surveys at ca 300–500 m above the seafloor along two adjacent ∼200 km-long sections of this ultra-slow spreading ridge crest [German et al., 1998a]. Optical back-scatter records, interpreted to represent particle-laden plumes derived from high-temperature hydrothermal venting, were reported from six different locations within the TOBI/MAPR survey areas. This was the first study to argue for the presence of hydrothermal activity on any ultra-slow spreading ridge although subsequent investigations [Bach et al., 2002; Edmonds et al., 2003; D. P. Connelly, C. R. German, A. Egorov, N. V. Pimenov, and H. Dohsik, Hydrothermal plumes overlying the ultraslow spreading Knipovich Ridge, 72–78°N, Norwegian-Greenland Sea, manuscript submitted to Geochimica Cosmochimica Acta, 2003] have helped confirm the argument presented in that work that hydrothermal activity may actually be quite abundant along ultra-slow spreading ridge crests [cf. Baker et al., 1996; Van Dover, 1998].

[3] One particular limitation to the work conducted during the FUJI cruise, however, was the absence of any water column sampling and analysis to confirm that the particle-laden signals observed were, indeed, hydrothermal in origin [German et al., 1998a]. In subsequent studies, by contrast, care has been taken to ensure that such geochemical samples (e.g., for TDMn analysis) are, indeed, collected and analyzed [German et al., 2000; Edmonds et al., 2003].

[4] The purpose of this study, therefore, has been to analyze core-top samples collected from close to sites where particle-laden plumes were previously observed to investigate whether hydrothermal geochemical signatures can be detected that are consistent with hydrothermal plume particle fall-out. The majority of samples collected were recovered during an InterRidge cruise, INDOYO (PI's: H.Fujimoto and K.Fujioka) that returned to the FUJI SW Indian Ridge study areas in 1998 with the submersible SHINKAI 6500. Although no active sites of venting were discovered during that expedition, one site of extinct venting was identified atop Mt.Jourdanne (FUJI “East Box”) from which chimney-shaped hydrothermal sulfides were recovered [Münch et al., 2001]. In addition, two final dives of the INDOYO cruise were conducted close to the Rodrigues Triple Junction. Water column and seafloor observations gathered during those dives [Sohrin et al., 1999] were instrumental in the subsequent discovery of the Kairei hydrothermal field approximately 18 months later; the first active hydrothermal field to be found anywhere in the Indian Ocean [Hashimoto et al., 2001]. Thus the work presented here draws upon cores collected from sites where plume signals had previously been observed and from cores taken close to both an extinct (Mt. Jourdanne) and an active (Kairei) seafloor vent-field.

2. Sampling and Methods

[5] Sediment samples for this study have been analyzed from 13 discrete locations in the SW Indian Ocean (Table 1). Of these, eight were collected from the submersible SHINKAI 6500 in 1998. Those samples, whose precise locations are described in greater detail below, were collected from three principal areas: FUJI W.Box, FUJI E.Box and the site of the subsequently discovered Kairei hydrothermal field (Figure 1). In addition, a further suite of five core-top samples have been analyzed (E1–E5, Figure 1). These samples were collected aboard HMS Egeria, during October–November 1887. Sample E1 was collected from within the Central Indian Ridge rift-valley near 20°S while samples E2-E5 were collected closer to the SW Indian Ridge axis but, in all cases, were recovered from outside the SWIR rift-valley.

Figure 1.

Map of the SW Indian Ocean (10–40°S, 40–80°E) showing locations of the Central Indian Ridge (CIR), South West Indian Ridge (SWIR) and South East Indian Ridge (SEIR). Eight cores for this study were collected using the SHINKAI 6500 submersible in 1998 from three separate dive-areas along the SWIR/Rodrigues Triple-Junction: “West Box”, “East Box” and “Kairei”. Five further surface-sediment samples analysed in this work, E1-E5, were recovered from tallow-coated lead sounding lines by HMS Egeria in 1887.

Table 1. List of Samples Analyzed
SampleDateVesselLat., SLong., EDepth, fmsDepth, m
445-124/09/1998SHINKAI 650027°58.9′63°45.6′4,003
451-105/10/1998SHINKAI 650031°03.8′59°01.7′4,355
452-106/10/1998SHINKAI 650031°04.5′59°03.0′4,637
452-206/10/1998SHINKAI 650031°04.8′59°02.0′4,393
452-306/10/1998SHINKAI 650031°04.9′59°01.8′4,396
454-109/10/1998SHINKAI 650027°51.0′63°56.1′2,947
456-113/10/1998SHINKAI 650025°18.6′70°01.5′2,576
457-114/10/1998SHINKAI 650025°19.3′70°02.2′2,578
E130/10/1887HMS Egeria20°00′67°33′1,6122,953
E218/11/1887HMS Egeria33°13′52°59′2,5624,694
E320/11/1887HMS Egeria36°04′51°28′1,9413,556
E423/11/1887HMS Egeria37°23′56°43′2,5104,598
E527/11/1887HMS Egeria38°06′67°09′2,4734,531

[6] The HMS Egeria samples were recovered from the tallow-coatings applied to lead sounding lines - a practice dating back at least as far as the Challenger Expedition [Murray and Renard, 1891] but more recently “reintroduced” to ridge-crest science with the development of routine “rock-chipper” sampling [e.g., Reynolds et al., 1992]. In the case of the HMS Egeria collection, all samples were returned to the Natural History Museum, London where an annotation dating from 1922 had classified these samples as “too small for chemical analysis”; the samples had subsequently been archived at room temperature in an underground basement for approximately 75 years. The samples reported in this study (typically ≤1 g each) were not recovered from the vaults of the Natural History Museum until October 1996 when they were transferred to the SOC for geochemical analysis.

[7] In 1998, sediment samples were collected from within the rift valley of the SW Indian Ridge and the Rodrigues Triple-Junction by rather more sophisticated means. Sampling was conducted using short (30 cm) push-cores from the submersible SHINKAI 6500 at sites chosen by the diving scientist within the submersible.

[8] In the “West Box” area, four cores were collected during two dives (451, 452). These dives coincided with the strongest in situ optical records obtained during the previous FUJI cruise investigations close to 59°00′E (Figure 2a), one of three locations from which plume signals had been reported in this study area [German et al., 1998a]. Sample 451#1 was collected from a (relatively) shallow faulted block located toward the centre of the rift valley at this location while samples 452 #1–3 were collected along the base of the southern rift-valley wall. In the “East Box” area, two cores were collected (Figure 2b). The first of these (445#1) was collected south of the current spreading axis as part of an investigation of the “mega-mullion” structure which extends off-axis at this location (R. C. Searle, M. Cannat, K. Fujioka, C. Mével, H. Fujimoto, and L. Parson, FUJI Dome: A large detachment fault near 64°E on the very slow spreading South West Indian Ridge, manuscript submitted to Geochemistry Geophysics Geosystems, 2003). This sample, taken away from the current SWIR ridge-axis but at a depth comparable to the FUJI “West Box” cores, was chosen to provide a potential record of local “background” sediment deposition, averaged over the geologically-recent past. The second sample taken from the FUJI “East Box” was collected at 2947 m depth atop the Mt. Jourdanne axial volcanic feature, immediately adjacent to an extinct high-temperature hydrothermal field [Münch et al., 2001]. A final pair of cores was collected from close to the Rodrigues Triple-Junction (Figure 2c). Subsequent discovery of the Kairei hydrothermal field in late 2000 [Hashimoto et al., 2001], two years after completion of the INDOYO dive program, revealed that one of these cores (456#1) was collected approximately 2 km NW of the Kairei vent-field while the other (457#1) was located just 500 m down-slope from the site of active venting.

Figure 2.

Detailed maps of SHINKAI 6500 sediment coring sites, 1998. (a) FUJI West Box. Yellow circles show locations of sediment cores 451#1 and 452#1–3 within the SWIR rift-valley close to 59°00′E. Red circles and diamonds show locations of previously reported particle-rich plume signals interpreted to be hydrothermal in origin [German et al., 1998a]. (b) FUJI East Box. Yellow circles show locations of sediment cores 445#1, from one of three off-axis dives conducted in this area, and 454#3 collected close to an extinct hydrothermal field situated close to the summit of axial seamount Mt. Jourdanne (27°50′S, 63°55′E). Red circles and diamonds show locations of previously reported particle-rich plume signals interpreted to be hydrothermal in origin [German et al., 1998a]. (c) The Kairei vent-site, Rodrigues Triple Junction. Locations shown include the Kairei vent-site itself (red star), situated on the west flank of Hakuho Knoll (grey shading denotes water-depths shallower than 2500 m). Core locations are shown as yellow circles: core 457#1, ca.500m downslope from the active vent-site and core 456#1, approximately 2 km NW of the hydrothermal field. Red line striking NW-SE from 25°18′S, 70°01′E to 25°20′S, 70°03′E represents the track of the CTD tow-yo trace crossing the Kairei hydrothermal field discussed later (see Figure 6).

[9] After recovery of the submersible following each dive, SHINKAI push-cores were transferred to a wet-laboratory aboard RV Yokosuka where they were extruded from their corers and split longitudinally. Samples were then refrigerated at 4°C until the end of the cruise when an archive section of each core was returned to JAMSTEC (Japan), and a complementary section for geochemical analysis was transferred to SOC (UK).

[10] In the laboratory, all samples were dried for 24 hours at 60°C then homogenized by grinding in a tungsten carbide mill. Weighed sub-samples of the sediments (ca. 0.5 g) were then digested using a mixture of HClO4, HNO3 and HF, evaporated to dryness, and finally dissolved in 100mL of 1% HNO3. Concentrations of Fe, Al, Mn and Cu were determined by simultaneous ICP-OES, calibrated using matrix-matched standards. Precision of the measurements was ±3–5% in all cases.

3. Results

[11] Results for geochemical analysis of the SHINKAI 6500 push-cores are presented in Table 2. Although cores up to 22 cm in length were recovered from the INDOYO cruise only data from the upper 0–10 cm of each core are reported here to ensure that all interpretations are based upon geochemical signals observed in the surface mixed layer (Global average SML = 9.8 cm [Boudreau, 1998]) and, hence, relevant to the geologically (very)-recent past (102–104 yr [e.g., Boudreau et al., 2001]). Equivalent data for the “core-top” material recovered by HMS Egeria are listed in Table 3. Note, however, that Cu data are not reported for these samples; contamination of Cu was evident in these samples, perhaps arising from the lead sounding-line weights from which the samples were collected. The degree of this contamination, however, (≤200 ppm Cu in all cases) is insufficient to suggest that those major-element data are similarly compromised.

Table 2. SHINKAI 6500 Push-Core Analyses
SampleDepth, cmAl, %Mn, %Fe, %Mg, %Cu, μg/g
Table 3. HMS Egeria “Core-top” Analyses
SampleAl, %Mn, %Fe, %Mg, %

4. Discussion

[12] When high-temperature hydrothermal fluids erupt from the seafloor they mix with ambient seawater to generate buoyant turbulent plumes. Within these plumes a range of metal-rich precipitates are formed which are carried upward and then dispersed laterally as the plume's level of neutral buoyancy is reached. (For a recent review, see German and Von Damm [2003].) Eventually, this suspended particulate material will settle toward the seafloor delivering a hydrothermal geochemical input to the sedimentary record. Boström et al. [1969] proposed the index: 100*Al/(Al + Fe + Mn) for pelagic sediments to enable differentiation between samples which had and had not received any such metalliferous input. This “metalliferous sediment index” has been used previously to demonstrate the presence of abundant hydrothermal input to ridge-flank sediments along, e.g., the southern East Pacific Rise [Boström et al., 1969]. (Note that those early calculations were completed more than a decade before first discovery of high-temperature hydrothermal vents themselves!) Recently, Gier and Langmuir [1999] have used this “metalliferous sediment index” approach to investigate hydrothermal input to cores recovered from wax-tipped rock corers along both the northern Mid-Atlantic Ridge and the southern East Pacific Rise. Here we use the same technique to detect for the presence (or absence) of geologically recent hydrothermal input to near-surface sediments from the SW Indian Ridge.

[13] Boström et al. [1969] concluded that near-field hydrothermal sediments should exhibit metalliferous sediment index (MSI) values - 100*Al/(Al + Fe + Mn) - of less than 10%, while background pelagic sediments with no hydrothermal input should exhibit MSI values >50%. Averaged metalliferous sediment index values for the upper 0–10 cm of each SHINKAI 6500 core are plotted in Figure 3 (error bars represent ± 1s.d. of each average MSI value). What is immediately apparent is that all samples show average MSI values less than 50% with the majority falling in the range 30 < 100*Al/(Al + Fe + Mn) < 40%. Two cores, 454#3 and 457#1 exhibit much lower values: 100*Al/(Al + Fe + Mn) < 10%. This is entirely consistent with what is already known about those samples: core 457#1 was collected within ≤500 m of the Kairei hydrothermal-field and core 454#3 was collected from immediately adjacent to a (recently?) extinct vent-site. These observations are of particular importance because they demonstrate that the application of Boström et al.'s [1969] lower threshold value for “near-field” hydrothermal sediments does, indeed, hold valid for this suite of samples.

Figure 3.

Values of the metalliferous sediment index (MSI) as defined by Boström et al. [1969] for eight push-cores collected from the SWIR rift-valley by SHINKAI 6500 in 1998. Brown shading (MSI < 10%) indicates values representative of near-vent samples while pale-green shading (MSI > 50%) indicates values representative of pelagic sediments with no discernible hydrothermal plume input. Also shown for comparison are the range in MSI values for SWIR samples collected by HMS Egeria in 1887 (this study; dark green shading) and for cores collected at known distances down-plume from the Rainbow hydrothermal field, Mid-Atlantic Ridge (pale grey shading [Cave et al., 2002]).

[14] At the other extreme, core 445#1 exhibits the highest metalliferous sediment index values calculated from the SHINKAI 6500 data-set: 100*Al/(Al + Fe + Mn) = 44 ± 9% (Figure 3). While individual samples within the top 10cm of this core exhibit MSI values <40% the full range for this core also overlaps the upper threshold value - 100*Al/(Al + Fe + Mn) >50% - which Boström et al. [1969] defined as being representative of pelagic sediments free of any hydrothermal input. Further evidence that this “off-axis” SHINKAI 6500 sample (core 445#1) contains at least some “nonhydrothermal” pelagic sediment is provided by the observation that the surface-sediment samples collected by HMS Egeria, along the axis of the SWIR but well beyond the plume-confining topography of the modern rift-valley, also exhibit high, “background” MSI values (E2-E4: 100*Al/(Al + Fe + Mn) = 56 ± 2%). The exception to this rule is sample E1 (data not plotted in Figure 3) which was collected near 20°S on the Central Indian Ridge axis and which exhibits a much lower MSI value of <27% (cf. core 456#1). Interestingly, that sample - collected in 1887 - was located just ∼25 km north of a new hydrothermal plume site identified recently from in situ optical back-scatter and coregistered water column TDMn analyses [German et al., 2001].

[15] Returning our attention to the South West Indian Ridge, all four cores from the FUJI “West Box” plus the more distal core from the Kairei hydrothermal field exhibit MSI values that are significantly less than 50%, apparently indicating a significant hydrothermal input. Indeed, signals in these samples coincide very closely with those for a separate suite of cores analyzed recently at SOC as part of a detailed investigation of the Rainbow hydrothermal system [Cave et al., 2002]. In that work, sediment cores were collected within the rift-valley of the MAR at increasing distances away from an active vent-site, beneath a topographically controlled dispersing hydrothermal plume. Cores for that study were collected at a range of distances, from 2–25 km “down-plume” and exhibited metalliferous sediment index values spanning the range 100*Al/(Al + Fe + Mn) = 21–37%. Within that suite of cores, the box-core raised at station 316 [Cave et al., 2002] showed the closest geographic similarity to the location of core 456#1 because it was raised from a point approximately 5 km distant from the active Rainbow vent-site. This similarity is reflected well in the respective MSI values for the upper 10cm of each core: 23–26% at Rainbow versus 26–32% at Kairei.

[16] One should use caution when interpreting variations of the Metalliferous Sediment Index purely in terms of varying hydrothermal input, however. Previously, Kuhn et al. [2000] have investigated the geochemistry of sediment cores raised from both within and beyond the confines of the rift valley of the southern Central Indian Ridge. In their work it was shown that sediments from within that shallower and more volcanically active rift valley contained a higher proportion of relatively mafic basalt-derived detritus than samples in the off-axis ridge-flank cores. In the current study, therefore, it is important to confirm that similar systematic variations in core mineralogy are not also giving rise to apparent relative Fe and/or Mn enrichments in the rift-valley (SHINKAI 6500) samples, relative to the ridge-flank (HMS Egeria) cores, which might be quite independent of any hydrothermal input. This issue is addressed in Figure 4, which plots the relative Fe-Al-Mg distributions of the sediment cores from this study, following the approach of Kuhn et al. [2000]. The plot shows that samples from throughout this work all exhibit lower Mg:Al ratios than the southern CIR study, except for our single HMS Egeria sample from 20°S on the CIR which, reassuringly, falls directly within the field defined by the Kuhn et al. [2000] data. By contrast, the remainder of our HMS Egeria data are significantly more Al-rich, indicating a lower relative abundance of mafic minerals in this sample-suite, consistent with the lower magma-supply rates and greater depth (hence, less likely explosive volcanism) expected along the ultra-slow spreading SWIR when compared to the medium-fast spreading CIR.

Figure 4.

Normalized Fe-Al-Mg ternary diagram for the SWIR hydrothermal sediments [after Kuhn et al., 2000]. Solid circles represent samples collected by HMS Egeria of relatively Al-rich, (Fe and Mg-poor) sediments from the SWIR and one sample from the southern CIR which coincides directly with recently analyzed CIR sediment samples [Kuhn et al., 2000] represented here by the area shaded green. Two SHINKAI 6500 samples collected from immediately adjacent to active/extinct vent-sites on the SWIR show extreme Fe-rich concentrations (Fe-oxyhydroxide “end-member” of Kuhn et al. [2000]; shown here by red shading). All other samples from SHINKAI 6500 exhibit compositions indicative of simple two-component mixing between Fe-rich hydrothermal “end-member” material and local “background” sediments as exemplified by the HMS Egeria samples.

[17] For the rift-valley (SHINKAI 6500) samples, cores 454#3 and 457#1 show anomalously Fe-rich compositions, consistent with their proximity to (recently) active vent-sites; they plot directly within the previously defined “end-member” Fe-oxyhydroxide field [Kuhn et al., 2000]. What is interesting to note is that the remainder of the SHINKAI 6500 cores exhibit near-identical Mg/Al ratios which coincide with those of the HMS Egeria “background” samples but, again, are significantly lower than those reported by Kuhn et al. [2000] from the southern CIR. The simplest interpretation of the data presented in Figure 4, therefore, is that all samples collected from within the SWIR rift valley by SHINKAI 6500 represent simple two-component mixing between background (constant Mg:Al) pelagic sediments in this area (as represented by the HMS Egeria cores) and an Fe-rich (Mg-free, Al-free) hydrothermal end-member (cf. cores 454#3 and 457#1). This observation is important because it provides confirmation that variations in the MSI values observed in Figure 3 should reflect genuine variations in hydrothermal input rather than any variation in sediment mineralogy between the SWIR rift-valley and ridge-flank sample-sets.

[18] Although the measured metalliferous sediment index can provide a useful tracer of the presence or absence of hydrothermal input to a pelagic sediment core, however, it does not necessarily provide any sensitive indication of the distance of that sediment from the source of the hydrothermal input (see, e.g., the broad overlap in MSI values reported from Rainbow cores at 2–25 km from a common vent-source [Cave et al., 2002]). To discern additional “directional” information, different chemical indicators must be exploited. One elemental ratio that is particularly powerful in this regard is Cu:Fe. In buoyant hydrothermal plumes a range of Fe-rich sulfide and oxide particles are formed. Importantly, however, the sulfidic material within these plumes is much more chalcophile-element rich than the co-existant Fe-oxyhydroxide particles [Feely et al., 1987; Mottl and McConachy, 1990]. In an early study of the TAG hydrothermal plume, Trefry et al. [1985] noted a marked decrease in particulate Cu:Fe ratios at increasing distance from the vent-site and attributed this fractionation to preferential settling of sulfidic material, relative to lower-density oxyhydroxide material within the dispersing neutrally buoyant plume. Indeed, prior to discovery of the TAG hydrothermal field, at a time when active venting was still believed to be located high on the east wall of the MAR rift-valley near 26°N, Shearme et al. [1983] presented an analysis of core-top geochemical samples which exhibited maximum Cu:Fe ratios clearly demarcating the site of the subsequently discovered TAG hydrothermal mound [cf. Rona et al., 1986]. More recently, Cave et al. [2002] have confirmed that core-top Cu:Fe ratios decrease systematically “down-plume” away from the Rainbow hydrothermal field, exhibiting elemental ratios that decrease systematically Cu:Fe = 27*10−3, 23*10−3, 9*10−3 and 6*10−3 at ∼2 km, 5 km, 12 km and 25 km away from the active vent-site.

[19] Figure 5 presents the elemental Cu:Fe ratios from the upper 0–10 cm of the eight SHINKAI 6500 push-cores analyzed for this work. Also shown, for comparison, is the range of Cu:Fe ratios measured in the bulk sediments analyzed by Cave et al. [2002]. As for the metalliferous sediment index plot (Figure 3) the most striking features in Figure 5 are the anomalous compositions of cores 454#3 and 457#1 both of which lie close to sites of active or (recently?) extinct venting. Consistent with the settings from which they were collected, Cu:Fe ratios for these cores exceed even the highest values recorded by Cave et al. [2002]. Indeed, the highest values reported here are more directly comparable to another core, analyzed previously from the Mid-Atlantic Ridge: Core 2182, collected by Alvin from the flanks of the TAG hydrothermal mound in 1990 (Cu:Fe = 38–262*10−3 [German et al., 1993]).

Figure 5.

Values of the elemental Cu:Fe ratio for eight push-cores collected from the SWIR rift-valley by SHINKAI 6500 in 1998. Also shown for comparison are the values for cores collected at known distances down-plume from the Rainbow hydrothermal field, Mid-Atlantic Ridge (pale grey shading [Cave et al., 2002]).

[20] Only one other core collected by the SHINKAI 6500 in 1998 exhibits Cu:Fe ratios similar to those reported by Cave et al. [2002]. Core 456#1 exhibits an average core-top Cu:Fe ratio of ∼17*10−3. As was the case for the metalliferous sediment index values, the calculated Cu:Fe ratio value for this sample is also most directly comparable to that measured in core 316 from the Rainbow hydrothermal plume [Cave et al., 2002] which recorded an elemental ratio of 22*10−3 at a location ∼5 km downstream from the Rainbow vent-site. In the case of the Kairei hydrothermal field, core 456#1 was also located within a few km of an active vent-site and, again, lay directly beneath the path of the dispersing, particle-laden plume (Figure 6).

Figure 6.

NW-SE cross-section of the Kairei hydrothermal plume, as determined from in situ optical back-scatter; also shown here are the relative locations of cores 457#1 and 456#1. Increased back-scatter in hydrothermal plumes is caused by high concentrations of freshly precipitated Fe-Mn oxyhydroxide particles, introduced into the water column at vent-sites, which are subsequently sedimented to the underlying seafloor. This is the process believed to give rise to low MSI values along many ridge-crests, just as presented in Figure 3. The SE-NW shallowing of the nonbuoyant hydrothermal plume at Kairei (light and dark green) results from shoaling isopycnal surfaces (CTD data: RRS Charles Darwin cruise CD128/2002; Co-PIs, P.A.Tyler and C.R.German, SOC). The path of this tow-yo passes closest to the Kairei vent-site ca. mid-way along-slice where strongest optical back-scatter anomalies coincide with interception of the buoyant Kairei hydrothermal plume (for a more detailed discussion of the Kairei hydrothermal plume [see Rudnicki and German, 2002]).

[21] The remainder of the SHINKAI 6500 cores collected from the SW Indian Ridge in 1998 all exhibit Cu:Fe elemental ratios lower than the lowest values reported from beneath the Rainbow hydrothermal plume at ca. 25 km distance “down-stream” from the vent-site (Figure 5).

[22] How can we reconcile these apparently competing sets of evidence? “West Box” sediments underlie the strongest optical-backscatter/hydrothermal plume signals observed in the water column during the FUJI survey in 1997 [German et al., 1998a]. Further, they exhibit metalliferous sediment index values significantly less than 50% (Figure 3), consistent with sinking of hydrothermal precipitates into these rift-valley floor sediments. However, Cu:Fe ratios for these same samples are indistinguishable from background and suggest that any settling hydrothermal material must have traveled significant distances (≥25 km?) from any active vent-site. This, perhaps, is not too difficult to explain.

[23] On the Mid-Atlantic Ridge, suspended hydrothermal plume particles have already been traced up to 50 km down-plume from their Rainbow vent-field source [German et al., 1998b]. Further, Edmonds et al. [2003] have recently detected abundant particle-laden, TDMn-enriched plumes along the Gakkel Ridge (also an ultra-slow spreading mid-ocean ridge) which extend up to 75 km laterally away from 9–12 identified “plume-centres.” On the Gakkel Ridge such long length-scale dispersion has been attributed to some combination of the relatively un-segmented nature of the Gakkel rift-valley [Edmonds et al., 2003] and the relatively well-mixed (poorly stratified) nature of the local (fresh-formed, Arctic) deep-waters (E. T. Baker and C. R. German, On the Global Distribution of Hydrothermal Vent Fields, manuscript submitted to AGU Geophysical Monograph Series, 2003), up to a depth of ca. 3500 m. The latter observation is potentially of direct relevance to this study. A series of CTD deployments within the basket of the SHINKAI 6500 submersible in 1998 also indicated the presence of a well-mixed (poorly stratified) layer extending up to a depth of ca. 4500 m within the water column of both the “West Box” and “East Box” of the FUJI survey, indicative of inflow of fresh-formed Antarctic Bottom Water within the SWIR rift-valley [Sohrin et al., 1999]. Further, in the FUJI “West Box”, the uppermost limit of the well-mixed portion of the deep-water column from the 1998 studies coincided closely with the depth at which maximum plume particle anomalies were detected 12 months earlier, at the same location.

[24] By analogy with the conclusions reached for the Gakkel Ridge, Arctic Ocean, therefore, and recognizing that the degree of speculation becomes heightened in the absence of complete data sets, it can be hypothesized that a similar, weak stratification (albeit restricted to a deeper portion of the water column) may contribute to a widespread dispersion of nonbuoyant hydrothermal plume particles over significant distances within the SW Indian Ridge rift-valley.

5. Summary

[25] 1. Fe-Al-Mn data are consistent with hydrothermal input to all but one of eight push-cores collected with the submersible SHINKAI 6500 from the SW Indian Ridge rift-valley and the Rodrigues Triple Junction in 1998. Fe-Al-Mg distributions confirm that these enrichments are genuinely hydrothermal in origin and are not a function of varying sediment mineralogy (e.g., due to varying MORB input within the rift-valley, relative to ridge-flank settings).

[26] 2. At the Rodrigues Triple Junction both Fe-Al-Mn distributions and Cu:Fe ratios are consistent with the positions from which these cores were taken relative to the subsequently discovered active Kairei vent-site. Core 457#1 exhibits extremely high Cu:Fe ratios and very low MSI values, consistent with a near-field hydrothermal sediment core. Core 456#1, by contrast shows lower Cu:Fe and higher MSI values respectively, consistent with its position directly underlying the dispersing hydrothermal plume. Interestingly, data from core 456#1 also show great similarities with another hydrothermally influenced pelagic sediment core collected ∼5 km “down-plume” from the Rainbow hydrothermal field, 36°N Mid-Atlantic Ridge [Cave et al., 2002].

[27] 3. In the FUJI “East Box”, core 454#3, collected from the summit of an axial seamount (Mt. Jourdanne) exhibits both metalliferous sediment index and Cu:Fe values directly comparable to near-field Kairei core 457#1. This is consistent with the location of this core, collected from within ≤500m of an inactive hydrothermal field discovered during the 1998 SHINKAI dive programme [Münch et al., 2001]. Also in the FUJI “East Box”, core 445#1 - collected furthest from the active ridge-axis - exhibits near-background values for both Cu:Fe and the Boström et al. [1969] metalliferous sediment index. This core also shows an overlap with MSI values for “background” cores collected along the flanks of the eastern SWIR aboard HMS Egeria in 1887.

[28] 4. In the FUJI “West Box”, a series of four cores collected from beneath the strongest hydrothermal plume signals reported in 1997 [German et al., 1998a] all show evidence for settling input from hydrothermal plume-particle fall-out. However, low values of the metalliferous sediment index (37–40%) coupled with low (background) Cu:Fe ratios in these samples, indicate that this hydrothermal plume fall-out material may have dispersed some significant distance away from any site of venting, prior to deposition.


[29] I thank A. Fleet (Natural History Museum, London) and H. Fujimoto and K. Fujioka (Co-Chief scientists of the R/V Yokosuka “INDOYO” campaign) for the opportunity to gain access to the HMS Egeria and SHINKAI 6500 samples, respectively. I thank D.R.H. Green for support with geochemical analyses and M. D. Rudnicki, A. L. Evans, and N-C. Chu for assistance with the figures. This manuscript benefited from two helpful anonymous reviews.