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

Seafloor eruptions and evolution of hydrothermal fluid chemistry

D. A. Butterfield,1 I. R. Jonasson,2 G. J. Massoth,3 R. A.Feely,3 K. K Roe,1 R. E. Embley,4 J. F. Holden,5 R. E. McDuff,5 M. D. Lilley,5 and J. R. Delaney

1Joint Institute for the Study of Atmosphere and Ocean, University of Washington, Seattle, WA 98195
2Geological Survey of Canada, Ottawa, Ontario, Canada
3Pacific Marine Environmental Laboratory, National Oceanic and Atmospheric Administration, Seattle, WA 98115
4Pacific Marine Environmental Laboratory, National Oceanic and Atmospheric Administration, Newport, OR 97365
5School of Oceanography, University of Washington, Seattle, WA 98195

Philosophical Transactions of the Royal Society of London A 355, 369-386 (1997).
Copyright ©1997 by the Royal Society. Further electronic distribution is not allowed.

3. Results

(a) General comments

Table 1 contains the chemical data as analyzed in individual samples. Selected elements for Flow and Floc sites are shown in figure 2. Note that the measured Mg concentration in diffuse fluids is affected by subseafloor dilution of a Mg-depleted source fluid and by entrainment of ambient seawater at the vent orifice during sampling (i.e. sample quality).

No high-temperature fluids were found in the Flow or Floc areas. Observed post-eruption venting was strictly diffuse and low-temperature (maximum 51°C on 1 August 1993). Evidence for high-temperature (greater than 200°C) reactions is present in the diffuse fluids in the depleted Mg and elevated Li, Fe, Mn, Ca and Si concentrations. Flow, Floc, and Source sites each have distinctive chemical characteristics reflecting differences in hydrothermal conditions in the three locations.

fig02-sm.gif

Figure 2: Plots of (a) Cl, (b) HS, (c) Fe, (d) Mn, (e) Li, (f) SO, (g) NH, and (h) Si versus Mg for Flow and Floc diffuse fluids. Symbols: triangle, Flow site 1993; rectangle, Floc site 1993; circles, Floc site 1994; hourglass, Floc site 1995; open square, ambient seawater. Lines drawn to show range of element-magnesium trends: thin line, Flow site; dashed line, Floc 1993; dot-dashed line, Floc 1994; dotted line, Floc 1995. Arrow in (f) represents conservative mixing of seawater with a zero-sulphate, zero-magnesium endmember.

(b) Flow site

The Flow site vents appeared as shimmering water exiting cracks and interstices of a fresh lobate/pillow lava mound at ca. 2390 m depth. Extensive surface alteration and precipitation of hydrothermal sediment was already apparent within one week of the eruption, and globules of orange material could be seen coming directly out of cracks in some places. The initial samples recovered by ROPOS were depleted in Mg by only 0.5% relative to ambient bottom seawater (indicative of substantial seawater entrainment during sampling), did not differ significantly from seawater in chlorinity, but were significantly enriched in Li, HSiO, Ca, Fe, and Mn. HS was undetectable (below 0.5 µmol kg) in all samples collected by ROPOS and Alvin at the Flow site.

The Flow site was revisited approximately nine weeks after the first samples were taken, and some subtle changes in chemical composition had occurred. Recovered samples had Mg depletions ranging from 0.7-2.3% relative to ambient seawater, and temperatures ranged from 4-36°C. Though the October samples were more depleted in Mg, they had lower iron concentrations than the initial ROPOS samples, and the chlorinities were now clearly higher than seawater. Measured ammonia concentrations ranged from 1-11 µmol kg, compared to background seawater at less than 0.3 µmol kg (greater than 30 µmol kg nitrate in bottom seawater is more than sufficient as a nitrogen source for the ammonia). Slightly different element-magnesium trends are apparent in the different vents sampled (figure 2); vent 6a (46°31.609N, 129°34.713W) has higher Fe, Cl and temperature than vent 7a, located about 330 m to the southwest (46°31.451N, 129°34.811W). Two samples from lower temperature vents in the graben just south of the lava flow (46°30.581N, 129°35.33W) were closer to seawater in composition.

The Flow site was visited again by Alvin in July 1994. Venting had nearly ceased one year after the event. One vent (46°31.444N, 129°34.804W, marker 0) with very little flow and temperature of 9°C was sampled, and the samples did not differ significantly from ambient seawater in Mg or Cl. The fluids did have very slightly elevated silica, iron, manganese, and alkalinity, indicating that water-rock reaction and, possibly, microbial activity were still taking place at greatly reduced levels relative to the immediate post-eruption period. Flow site vents were not sampled in 1995, when water column surveys indicated no detectable thermal anomalies over Flow (E. T. Baker, personal communication 1995). (c) Floc site The seafloor vents giving rise to the water column 'Floc snowstorm' observed with ROPOS on 2 August were found with Alvin in October 1993, when fluids were collected from vents near marker 3A (16-18°C) and marker 11 (22°C), separated by a distance of ca. 580 m. The vents differed markedly in appearance from the Flow site, as they were associated with fissures in older basalt between 2220 and 2290 m depth, and were emitting white flocculent material thought to be bacterial in origin (Juniper et al. 1995). Also in contrast to the Flow site, Floc vent fluids had elevated levels of HS and were lower than seawater in chlorinity (figure 2), indicative of a vapour component. In contrast to the rapid exhaustion of heat and diminution of fluid flux seen at Flow site after one year, significant fluid fluxes were maintained at the Floc site up to two years after the volcanic event. Observations of the seafloor in the Floc area support a waning of venting at particular sites and a general decrease in the area of active venting over time, but our submersible surveys do not allow quantitative assessment of changes in fluid or heat flux. Fluid chemistry exhibits detectable temporal and spatial variation at Floc (figures 2 and 3). In October 1993, chlorinities were lower than seawater; in July 1994, more samples were recovered showing a range of chlorinities both higher and lower than the initial three samples indicated. There were slight changes in composition at reoccupied marker 11 and 3a sites. In July 1995, chlorinities of all the recovered samples from Floc were very close to seawater.

fig03-sm.gif

Figure 3: Temporal and spatial variation in diffuse fluids. Stacked bar graph of average element to heat ratios in nmol J (HS/heat values have been divided by three to match scales) and Fe/Mn ratio. Vent marker names (see figure 1 and  2) are given below the bars, which are ordered from N (left) to S (right) for each area. Absence of HS at Flow site indicates that the degree of subseafloor mixing caused the redox potential to be dominated by seawater. Iron oxidation is significant at the Flow site, as plentiful amorphous iron oxides and ferric phosphate coatings on bacterial sheaths (Juniper et al. 1995) attest. Note HS pulse in 1994 at HDV. Spatial variation in HS/heat (0.8-31 nmol J) and HS/Fe (10-340) in Floc vent fluids is enormous, with highest near HDV and marker 26. Lower values at the ends of observed venting may be due to a combination of precipitation and oxidation of HS toward the periphery of the Floc upflow zone. HS may be the biolimiting energy source for micro-organisms in some Floc vents (e.g. markers 11, 18 and 15 in 1995).

In October 1993, Floc samples had significant levels of HS and Fe (up to 55 and 27 µmol kg, respectively). In 1994, HS increased to a maximum of 380 µmol kg, while iron was below 3 µmol kg in all but one sample. In 1995, maximum HS was 230 µmol kg and iron was below 6 µmol kg. While iron was virtually unchaged at HDV vent from 1994-1995, hydrogen sulphide concentration apparently reached a peak in 1994, which coincides with maximum light attenuation in the plume over Floc (E. T. Baker, personal communication 1996).

The continued presence of HS, Li, Fe and Mn in the fluids indicates that high-temperature reactions continue beneath the Floc site. If the reaction zone produces a high-temperature fluid that has zero sulphate as well as zero magnesium, then sulphate reduction in the near-seafloor upflow zone is not important enough to cause significant depletion of sulphate below a conservative mixing line in the sulphate versus magnesium plot. Nearly all of the samples are slightly above the conservative mixing line, suggesting that oxidation of sulphide or dissolution of sulphate minerals (e.g., anhydrite or caminite) is quantitatively more important than sulphate reduction. (d) Source site There are four known high-temperature vents (Beard, Church, Mongo, and Twin Spires) at the Source site spread out over ca. 100 m along a fissure running ca. 020° NNE near the crest of a pillow lava ridge at ca. 2055 m depth. The Source site is at the southern end of the CoAxial neovolcanic zone, collinear with Flow and Floc sites. The chemistry of the vent fluids from this site indicates that it was unaffected by the 1993 eruptive activity and probably existed prior to the eruption.

fig04-sm.gif

Figure 4: Plots of (a) Cl, (b) Ca, (c) HS, (d) Li, (e) Fe, and (f) Mn versus Mg for source vent fluids. Symbols: triangle, 1993, rectangle, 1994, circle, 1995, open box, ambient seawater. Cl, Ca, and Li data show that there is a single primary endmember for this site that is not changing significantly over time. Lines drawn in (e) to show temperature dependence of Fe concentration: solid line for Twin Spires at 223°C dashed line for several vents in range of 250-275°C and dot-dashed line for a 294°C orifice at Church vent.

All four of the Source vents have virtually the same major element composition, which did not change significantly over the two years following the eruption (figure 4). Source fluids are moderate temperature (223-294°C) brines (695 mmol kg Cl) with low HS (1.0-1.7 mmol kg), iron (45-120 µmol kg), Fe/Mn (0.4-0.5), and zinc (6-17 µmol kg) relative to other MOR vents with similar temperature and chlorinity. The combination of low Fe and a pH of 4.5-4.8 means that significant iron sulphide precipitation has not occurred in the upflow zone near the vents (otherwise the pH would be lower) and implies that the fluids last equilibrated near the temperature of venting (Seyfried & Mottl 1995). These fluids have very high Ca (67 µmol kg), high Li (900 µmol kg), and low Na/Cl ratio (0.76 compared to 0.86 in seawater), indicative of extensive water-rock reaction (Li-derived water-rock ratio is 1.0, similar to many high-temperature MOR vents). These characteristics are consistent with a brine that formed at high temperatures (greater than 400°C), then remained in the crust, cooled or mixed with cooler seawater-derived fluid, and equilibrated at temperatures closer to 300°C before venting.


Return to previous section or go to next section

PMEL Outstanding Papers

PMEL Publications Search

PMEL Homepage