The first serpentine mud volcano drilled in the Mariana forearc was Conical Seamount, the summit and flanks of which were cored during ODP Leg 125 in 1989. The unusual interstitial waters recovered were inferred to be ascending from the décollement zone of the subducting Pacific plate, some 29 km below the seafloor (Mottl, 1992). The pore waters recovered from Site 1200 on the summit of South Chamorro Seamount are similar in many ways to those from summit Site 780 on Conical Seamount and doubtless have a similar origin, by dehydration and decarbonation of the crust of the downgoing slab (Fryer et al., 1999). The décollement is nearly as deep beneath South Chamorro Seamount, about 27 km, as beneath Conical Seamount. The South Chamorro Seamount fluids differ from those at Conical Seamount, however, in some significant ways, chiefly within the uppermost 20 mbsf. These differences apparently result from microbial activity within the shallow regions of the South Chamorro Seamount summit. There is also a macrofaunal community at the summit springs on South Chamorro Seamount (Fryer and Mottl, 1997), the likes of which was not seen during 21 dives in the manned submersible Alvin at Conical Seamount summit, in spite of the discovery of springs and chimneys there (Fryer, 1992a; Fryer et al., 1990, 1995). We detail here the chemistry of interstitial water, sediment, and headspace gas from Site 1200, which tell a coherent story about processes within and beneath the summit of South Chamorro Seamount.
Interstitial waters were squeezed from serpentine mud from four of the six holes drilled at Site 1200 (Table T7). A fifth hole (1200C) yielded a single water sample from the open borehole at 30 mbsf and 2.69°C, which proved to be drill water (surface seawater), but that will be useful as a starting composition for long-term sampling by the osmotic sampler that was left in this cased, CORKed, and instrumented hole. The chlorinity of this sample is indistinguishable from that of two surface seawater samples taken on 18 and 23 March, which bracket the dates of sealing of this hole, and show no change in chlorinity over this time period. Only a single uncontaminated sample of interstitial water was recovered from Hole 1200A, but this sample from 71 mbsf proved to be the deepest we obtained. The other three holes produced multiple samples from which depth profiles could be constructed. Hole 1200E is located within 10 m of Hole 1200A and a spring identified by the presence of mussels, whelks, small tubeworms, galatheid crabs, and a dive marker left by M.J. Mottl on Dive 351 of the Shinkai 6500 submersible in November 1996 (Fryer and Mottl, 1997). This spring was located during Leg 195 using the VIT camera mounted near the end of the drill string. As at Conical Seamount, the three known springs on the South Chamorro Seamount summit are cool, within a degree of the bottom-water temperature of 1.67°C. Temperature gradients estimated from Holes 1200E and 1200F are 0.0092 and 0.0724°C/m, respectively (Table T18), such that none of the interstitial water exceeded 3.0°C at the sampling depth (Table T7). Holes 1200F and 1200D are ~20 and 80 m north of Hole 1200E, respectively, and thus constitute a transect northward from the spring.
The data in Table T7 include chloride measured by ion chromatography (IC) as well as chlorinity measured by electrochemical titration with AgNO3, which is normally a much more accurate and precise technique. A third column gives the preferred value for chloride; these values are those measured by IC except where titration chlorinity yielded smaller values. The reason for this somewhat unusual presentation is that many of the interstitial water samples from Site 1200 proved to have exceptionally high concentrations of dissolved sulfide, which would be present mainly as the bisulfide ion (HS-) at pHs between 7.05 and 14 at 25°C. The high concentrations overwhelmed the technique intended to measure dissolved sulfide, colorimetry as methylene blue (Parsons et al., 1984). It was not feasible to bubble off this sulfide because of the risk of evaporation from the minuscule aliquots devoted to these analyses. From late in the sampling of Hole 1200E through Hole 1200F, therefore, we precipitated abundant dissolved sulfide from different aliquots using HgCl2, AgNO3, and Zn acetate to be measured on shore along with sulfur isotopic composition. In the meantime, we have estimated the dissolved sulfide concentration as the difference between chlorinity by titration and chloride by IC, as the latter measures only chloride and the former titrates bisulfide as well as chloride, bromide, and iodide (the usual contributors to chlorinity), according to the following reaction:
2 AgNO3 + 2 HS- = Ag2S
+ H2S + 2 NO3-.
Except for the 2 deep samples from Hole 1200A, 11 of the 17 samples from Hole 1200D farthest from the spring and the single uppermost samples from Holes 1200E and 1200F, chlorinity measured by titration exceeded chloride measured by IC, indicating the presence of dissolved sulfide (along with the putrid smell, the blackness of some of the core, and the abundant precipitate on addition of the reagents noted above, including a salt-and-pepper precipitate that formed during the chlorinity titration instead of the usual white AgCl precipitate).
We attempted to titrate the supernatant solution from above the Zn acetate precipitate (white colloidal ZnS) for chlorinity but determined that many ions co-precipitated with ZnS, including chloride, sulfate, Na, and K. Accordingly, the relatively imprecise (1%-2%) chloride data measured by IC proved to be our best estimate of the true chloride concentration for the 39 sulfide-bearing samples (of 54 total). Calculations of charge balance (Na + K + 2 Mg + 2 Ca - Cl - 2 SO4 - alkalinity) using the preferred chloride values and Na measured by IC showed that anions exceeded cations in 49 of 54 samples by an average of 2%. The charge imbalance results mainly from an overestimate of chloride; comparison with titration chlorinity from samples lacking dissolved sulfide suggests that the IC chloride data are too high by ~1%. Comparison of Na measured by IC with the Na concentration estimated from charge balance indicates that the IC Na values are too low by ~2%, almost half of which is due to the contribution from the overestimate of IC chloride to the calculated charge balance. For this reason, the calculated Na values are considered to be more accurate. Neither of these probable discrepancies from the true value is sufficiently large or certain to justify a correction, except that chloride measured by IC has been corrected downward by 1% in the calculation of dissolved sulfide concentrations only; without this correction, the concentrations of dissolved sulfide reported as near zero would become negative by 5 mmol/kg. Accordingly, the estimated uncertainty in the calculated bisulfide concentrations is 5 to 10 mmol/kg.
Bisulfide concentrations do not contribute to charge balance as calculated above because they are titrated by HCl and have already been counted in the alkalinity. The same is true for boron, which would be present entirely as borate ions (B[OH]4-) at a pH much greater than 9.3. Given the composition and high pH of the samples, the measured alkalinity results mainly from bisulfide, carbonate, and hydroxyl ions and their complexes, in that order, for the high-bisulfide samples, with a small contribution from borate. Oxidation of dissolved sulfide would potentially cause measured sulfate to be too high. To test for this possibility, we analyzed the supernatant solution from 18 samples from Holes 1200E and 1200D to which AgNO3 had been added immediately upon recovery from the squeezer, prior to any exposure to the atmosphere. These supernatants yielded sulfate concentrations identical to those from the normal IC aliquots, indicating that sulfide oxidation had not affected the IC aliquots at the time of analysis and that the sulfate data reflect the actual concentration in the pore water at the time of sampling.
The concentration of chloride in the interstitial water is lower than that in the bottom seawater by as much as 7% (Fig. F50) but not nearly as low as the minimum of 234 mmol/kg observed at Conical Seamount. Chloride exhibits a steep decrease within the uppermost 4.5 mbsf. Mottl (1992) interpreted the freshening and the steep surficial gradient at Conical Seamount to result from upward flow from great depth of an aqueous fluid derived from dehydration of the downgoing slab. The steepness of the surficial gradient implies upwelling at 1 cm/yr or more. The chloride maximum in Hole 1200D at 12 mbsf represents a significant deviation from an ideal one-dimensional advection-diffusion profile and probably results from lateral intrusion of seawater within the summit knoll; it corresponds to a depression in an otherwise ideal profile for Na (Fig. F50) that appears in both the Na calculated from charge balance and measured by IC. The chloride variations in Hole 1200D likewise appear in both the IC data, in replicate analyses, and in the chlorinity data determined by titration. Also reproducible was the chloride maximum in Hole 1200E at 20-25 mbsf, but this increase is within the imprecision of the IC method and is not matched by a deviation in Na. It is difficult to explain this latter maximum except as analytical imprecision.
Depth profiles of Na, Na/Cl, K, and B are near-ideal vertical advection-diffusion profiles that strongly support the inference of upwelling (Fig. F50). The curvature of the profiles implies that upwelling is fastest in Holes 1200E and 1200F near the spring and slowest in Hole 1200D farther away. As at Conical Seamount, these three elements, along with H2O, are apparently strongly enriched in the upwelling deep slab-derived fluid. Except for H2O, the degrees of the enrichments relative to seawater also agree well with those at Conical Seamount (Table T8).
The pH of the samples (Fig. F51), measured at 25°C, is as high as 12.5; along with the Site 780 samples from the Conical Seamount summit, these are the highest pH values ever measured in deep-sea pore water. The high pH results from equilibration of the solution with brucite and/or serpentine at low temperature, which requires an extremely low hydrogen ion concentration (i.e., high pH) at the low Mg concentrations observed. The depth profiles for pH resemble those for Na, K, and B and likewise indicate upwelling at different velocities for the three holes. The only major discrepancy from a typical advection-diffusion profile is the small maximum in pH at 4-9 mbsf in Holes 1200E and 1200F (which shows up better in a plot of hydroxyl ion vs. depth, not shown). This maximum corresponds to a similar maximum in alkalinity of 130 meq/kg at 3-16 mbsf (centered at 10 mbsf) in Hole 1200E near the spring (Fig. F51). At greater depths, alkalinity drops precipitously to 60 meq/kg, less than half the maximum and the same concentration to which alkalinity rises in distant Hole 1200D, without an intervening maximum. This value is also similar to the deep alkalinity at Site 780 on Conical Seamount (Table T8). The intermediate Hole 1200F shows a maximum of 112 meq/kg at 12 mbsf but did not penetrate deeply enough to exhibit the expected decline with increasing depth.
A shallow alkalinity maximum is a common feature in deep-sea pore waters where microbial sulfate reduction is oxidizing organic matter. It is typically accompanied not only by a decrease in sulfate but also by an increase in ammonium ions, which would dissociate to ammonia at the high pH of the Site 1200 solutions. All three features are found at Site 1200 (Fig. F51). Ammonia profiles with depth are quite similar to those for alkalinity, although the ammonia maximum at 3-16 mbsf in Hole 1200E is more clearly bimodal, with peaks at 2 and 9-13 mbsf. Below the maximum, ammonia decreases with depth to a concentration nearly identical to that in the deep pore water at Conical Seamount (Table T8). Sulfate decreases in all three holes and is essentially absent in Hole 1200E at the spring, between 2 and 20 mbsf. Below that depth, it increases back to seawater concentration. In the deepest sample, from 71 mbsf in nearby Hole 1200A, sulfate actually exceeds seawater concentration by 7% when adjusted to the chlorinity of ocean bottom water. Summit Site 780 at Conical Seamount also yielded sulfate-rich pore waters, with up to 1.7 times seawater concentration, even when uncorrected for seawater chlorinity. Sulfate in intermediate Hole 1200F decreases to nearly zero at 1 mbsf but then abruptly rebounds. It shows a second strong minimum at 14 mbsf. Distant Hole 1200D shows a pronounced minimum at 3 mbsf and a second minimum at 14 mbsf that would be even larger if this sample had not been contaminated by seawater during core recovery. At greater depth, sulfate rebounds to near seawater concentrations in this hole as well.
Microbial sulfate reduction produces dissolved sulfide that typically precipitates as metastable iron monosulfide minerals that eventually convert to pyrite. Dissolved sulfide, estimated as the difference between chlorinity measured by titration and chloride (adjusted) measured by IC, as discussed above, defines clearly bimodal maxima at the same depths as the extrema for alkalinity, ammonia, and sulfate (Fig. F51). These maxima reach very high concentrations in spring Hole 1200E and smaller maxima in successively more distant Holes 1200F and 1200D. Whereas the calculated concentrations have a large uncertainty (±5-10 mmol/kg), the shape of the profiles including the bimodality of the peaks has been confirmed by semiquantitative estimates of the amount of ZnS and Ag2S produced in the treated samples, as well as from the distribution of metal sulfide precipitates formed during colorimetric analyses for phosphate. The deepest sample, from 71 mbsf in Hole 1200A, shows no evidence of dissolved sulfide. At Conical Seamount summit, a single sample from 2.7 mbsf yielded a dissolved sulfide concentration of 2 mmol/kg (by methylene blue colorimetry), whereas samples from 3 to 35 mbsf in the same Hole 780D had concentrations below the detection limit of 0.25 mmol/kg.
Depth profiles of Mg, Ca, Sr, and Li are similar to one another (Fig. F52). All decrease sharply within 1-4 mbsf from the seawater value to low concentrations that persist to greater depth. The profiles are consistent with upwelling of a fluid that is highly depleted in these elements, with the fastest upwelling in Holes 1200E and 1200F near the spring and the slowest in Hole 1200D farther away. Depletion of these elements was also observed in the upwelling fluid at Conical Seamount (Table T8). The low concentration of Mg is typical of fluids in equilibrium with serpentine-bearing assemblages, such as the altered interstitial seawater recovered from serpentinite within Torishima Forearc Seamount, an inactive mud volcano in the Izu-Bonin forearc (Table T8). The low concentrations of Ca and Sr result from precipitation of CaCO3 at the high observed alkalinity. Ca could not be detected by either titration or IC at the depth of the alkalinity maximum, but it rebounds to low values as the alkalinity decreases with depth. Ca and Sr in Hole 1200D, likewise, are higher where alkalinity is lower, except in the surficial carbonate precipitation zone discussed below. Li may be substituting for Mg and Fe in minerals of the serpentinite assemblage. It is the only one of the alkali elements to show depletion rather than enrichment relative to its concentration in seawater.
Mn is enriched near the seafloor as the sediment becomes reducing, but it decreases abruptly and drastically below 0.5-2 mbsf in the various holes (Fig. F52). As the sediment remains reducing below these depths, this drop is surprising. The decrease is probably not in response to carbonate precipitation, based on depth profiles from the inactive Torishima Forearc Seamount (Mottl, 1992); at the transition from pelagic sediment to serpentinite there, Mn plummets but Ca and Sr rise and alkalinity is little changed. Like Li, Mn may be substituting for Mg and Fe in minerals of the serpentinite assemblage.
Si and F (Fig. F53) both decrease sharply within the upper meter below the seafloor and then rebound just as sharply, especially in Holes 1200E and 1200F near the spring. Below that they fluctuate and decrease again to a minimum at ~15 mbsf in Hole 1200D and ~20 mbsf in Hole 1200E before converging toward intermediate concentrations in the deeper samples. The structure of these profiles, with multiple maxima and minima, indicates that these elements are highly reactive within the upper 20 mbsf and these reactions provide both sources and sinks at different depths. Al is generally low but appears to increase at depths below 20 to 30 mbsf.
The zone of complete sulfate reduction at 2-20 mbsf in Hole 1200E suggests the possibility of microbial methanogenesis within this zone. Samples of gas collected from the core liner immediately on recovery (Table T9) and from sediment plugs taken for headspace gas analysis (Table T10) show abundant methane as well as higher hydrocarbons, up to C3 in the sediment, and C6 in the core-liner gas, in each case the heaviest hydrocarbon detectable by the methods used. Whereas methane is generally in the range of 1-5 mM at this depth, there is no obvious peak in methane there (Fig. F54); instead, there are peaks above and below at the same depths as the extrema in dissolved sulfide, sulfate, and ammonia. There are small peaks in methane and ethane at 3 and 9-12 mbsf in Hole 1200E and much larger peaks at 1.5 and 14 mbsf in Hole 1200F; the latter depth produced the only detectable propane in the headspace gas samples. The C1/C2 ratio, moreover, increases with depth, suggesting selective consumption rather than production of methane at shallow depth.
The light hydrocarbon concentration range at South Chamorro Seamount is similar to that at Conical Seamount only when the hydrocarbon-rich flank Site 779 is included; there, the high concentrations are deeper than 200 mbsf. At the respective summit sites, South Chamorro Seamount has much higher methane and ethane at 1-20 mbsf (Fig. F54). The deeper fluids are more similar; methane is nearly identical, whereas ethane appears to be higher at Conical Seamount (Table T8). Another major difference is a larger and more rapid increase with depth in the C1/C2 ratio at South Chamorro Seamount, to 780 at 54 mbsf. At Conical Seamount summit Site 780, this ratio never exceeds 300, and at flank Site 779, it is higher in only two samples, from 227 and 266 mbsf.
The steep gradients in pore water composition near the seafloor, where the deep upwelling fluid interacts chemically with overlying seawater, is a zone of intense reaction. The most obvious of these reactions is reflected in the bulk composition of the surficial sediment (Table T11). The H2O that is bound into mineral structures is constant at 12-14 wt%, approximately the value in serpentine, over the depth interval sampled (Fig. F55). This percentage drops near the seafloor as other phases precipitate at this reaction front. Chief among them are aragonite, formed when the high-alkalinity, Ca-poor fluid from depth meets seawater, which has low alkalinity but plenty of Ca, and iron sulfides, formed from microbially produced dissolved sulfide. Bulk analyses of the serpentine mud show a large spike in CaCO3 in the uppermost 2 mbsf to concentrations as high as 36 wt%. The same effect was seen at all the Conical Seamount sites (Mottl, 1992). At other depths CaCO3 is <1%, except for two samples from 12 and 16 mbsf, which also have high sulfur. The presence of calcareous fossils at these two depths suggests that they represent a paleoseafloor surface that became enriched in CaCO3 and S in the same manner as the present seafloor today but then was buried by serpentine mud protruded from the throat of the mud volcano. Calcareous fossils at this depth show the darkening typical of incipient pyritization. The sulfur content of the serpentine mud near the present seafloor spikes at 2.4 wt%, compared with concentrations of <0.01 wt% at greater depth. Between the surficial spike and the low values at depth, all three holes exhibit a plateau defined by an intermediate sulfur content of 0.2-0.3 wt%. This plateau extends as deep as 25 mbsf in Hole 1200E, but to only 5 mbsf in Hole 1200D. At Conical Seamount summit Site 780, by contrast, the highest S concentration measured in the mud was only 0.14 wt% and S generally was not detectable.
Organic carbon (Table T11;
Fig. F55) was detectable in
only 13 of the 49 samples of serpentine mud analyzed. Of these 13, nine had 
0.1
wt%. The remaining four samples are all from the uppermost 2.7 mbsf of Hole
1200D, and the highest two values of 0.4-0.5 wt% are from the uppermost 0.75
mbsf. These organic carbon concentrations are lower than those from the summit
Site 780 on Conical Seamount, which were typically 0.2 wt%, except for a similar
enrichment to 0.56 wt% within the uppermost 2 mbsf.
The two major processes exhibited by the chemical data from Hole 1200A are upwelling of a fluid from a deep source and microbial sulfate reduction near the seafloor. Both processes are more active in Hole 1200E near the spring and become less so with increasing distance toward Hole 1200D. As was inferred by Mottl (1992) for Conical Seamount, the upwelling fluid can best be explained as originating at the décollement 26.5 km below South Chamorro Seamount summit by dehydration and decarbonation of the subducting Pacific plate. The upwelling fluid is quite similar to that at Conical Seamount for most components (Table T8), including its high pH; its enrichment in alkalinity, Na/Cl, K, B, ammonia, and hydrocarbons through C6, all components that are virtually absent in depleted harzburgite and therefore require a different source; its depletion in Mg, Ca, Sr, and Li; and its low concentrations of Si, Mn, Fe, Ba, and phosphate. The South Chamorro Seamount fluid differs from that at Conical Seamount chiefly in its much higher chloride content, which is still 6% lower than in the ambient ocean bottom water, and its lower sulfate, which is still slightly higher than in seawater. As a result of its higher chloride, the South Chamorro Seamount fluid has more Na than seawater, whereas the Conical Seamount fluid has less, although both are highly enriched in Na relative to Cl and by approximately the same amount.
The deep fluids from Conical and South Chamorro Seamounts contrast markedly with the interstitial water recovered from Torishima Forearc Seamount, an inactive mud volcano in the Izu-Bonin forearc (Table T8) (Mottl, 1992). The Torishima Forearc Seamount pore water represents the end product of reaction between seawater and partially serpentinized harzburgite. Compared with the Conical and South Chamorro Seamount fluids, it is enriched rather than depleted in chloride, Na, Ca, and Sr and depleted rather than enriched in alkalinity, sulfate, K, and B. Its pH, at 9.6, is not nearly as high and its Na/Cl ratio is unchanged from seawater values. Its only resemblance to the slab-derived fluids is its low concentrations of Mg, Si, Fe, and Mn, and even then, Mn is considerably higher than at the active mud volcanoes. The Torishima Forearc Seamount pore water has lost some Li but not nearly as much as Conical and South Chamorro Seamounts. It also has substantially less ammonia, less methane by three orders of magnitude, and no higher hydrocarbons. The deep fluids at South Chamorro and Conical Seamounts certainly cannot have originated by simple reaction between ultramafic rock and seawater.
The deep fluids from South Chamorro and Conical Seamounts also differ from interstitial waters sampled from six other serpentine seamounts in the Mariana forearc that were gravity and piston cored in 1997 (Fryer et al., 1999). All these other seamounts are closer to the Mariana Trench, such that the distance to the underlying décollement is correspondingly shallower, ranging from 16 to 25 km. Only the farthest from the trench, Big Blue Seamount, yielded pore waters depleted in Ca; the others showed large increases in Ca and decreases in alkalinity. Fryer et al. (1999) attributed this difference to depth-dependent decarbonation of the downgoing plate. If this interpretation is correct, it implies that the deep slab-derived fluids are continuously saturated with CaCO3 during their ascent. If they originate deep enough that decarbonation reactions are active, they will be enriched at their source in carbonate alkalinity, which will drive down the Ca concentration, as in the South Chamorro and Conical Seamount fluids, by precipitation of CaCO3. If not, they will be enriched in Ca, leached from noncarbonate rocks in the source region and during ascent, and become correspondingly poor in carbonate alkalinity. In either case, however, little CaCO3 will actually form in the conduits, as one or the other of the components necessary to form CaCO3 will always be in short supply. This supply shortage is only alleviated at the seafloor by the abundant Ca in seawater, which still cannot precipitate much carbonate from the shallower-sourced, alkalinity-poor fluids. Fryer et al. (1999) noted that only the most deeply sourced fluids, such as those at South Chamorro and Conical Seamounts, have produced CaCO3 chimneys at the seafloor and the corresponding zone of aragonite precipitation in the surficial muds, as observed at Site 1200 (Fig. F55).
The much smaller decrease in chloride relative to seawater in the South Chamorro vs. Conical Seamount fluids could be explained if the conduits through which the deep fluids ascend were more mature (i.e., more pervasively altered) at Conical Seamount than at South Chamorro Seamount. Consider that these fluids ascend tens of kilometers through a matrix that consists mainly of partially serpentinized, depleted harzburgite. This rock is still undergoing serpentinization, which will consume H2O and leave the remaining solution saltier, yet the fluid reaching the seafloor at both Conical and South Chamorro Seamount summits remains fresher than seawater, as measured by chloride content (but not by total equivalents of cations or anions, which are 4% higher in the South Chamorro Seamount deep fluid than in seawater vs. 47% lower in the Conical Seamount fluid). How fresh must surely be a function, in part, of how much serpentinization the fluid has effected during its ascent, which depends in turn on how much fresh harzburgite the fluid has encountered. A more thoroughly altered conduit, as hypothesized for Conical Seamount, would deliver a fresher fluid to the seafloor. In this scenario, the higher sulfate in the Conical Seamount fluid could also result from less reaction; perhaps the greater serpentinization accomplished by the South Chamorro Seamount fluid also involved greater loss of sulfate, presumably by reduction to sulfide. It is well known that serpentinization generates extreme reducing conditions by oxidation of ferrous iron in the primary ultramafic minerals to ferric iron in magnetite. As the deep fluids ascending beneath Conical and South Chamorro Seamounts appear to have little or no dissolved sulfide, this reaction product would have to have precipitated as pyrite or some other sulfide mineral within the conduit. Of course, this hypothesis does not explain how the fluids acquired their high sulfate content in the first place.
The ultimate freshness of the fluid must also be a function of its original salinity. We do not know the chlorinity of the fluid at its source, but if it originates by dehydration of various minerals, it is not inconceivable that it is essentially fresh. The probable source of chloride is mixing with seawater-derived pore solutions, most likely by dewatering of sediment and basement pore solutions (as opposed to dehydration of minerals) at the décollement but possibly also by admixture of downwelled seawater during ascent. The biggest problem is in explaining the extreme enrichment in some other major components, especially Na, sulfate, and alkalinity. In both the South Chamorro and Conical Seamount fluids, Na is so enriched as to be nearly the only cation (along with a much smaller amount of K) and its concentration exceeds that of chloride by 20%-40% (Table T8), much more than the sum of cations exceeds chloride in seawater (11%). Sulfate is barely enriched at South Chamorro Seamount, but the sulfate-rich fluid at Conical Seamount requires a major source. Substantial (chiefly carbonate) alkalinity has been added to both fluids. The sources of these components are unknown in detail, but given the relative chemical simplicity of the depleted harzburgite in the overriding plate through which the solutions ascend and react, the only real candidate is the sediment and altered basalt at the top of the subducting plate.
Although the composition of the deep upwelling fluids is generally similar, at 0-20 mbsf the compositions of the interstitial water, headspace gas, and surficial sediment at South Chamorro Seamount summit all differ considerably from those at Conical Seamount. All of the differences can be attributed to microbial activity within the uppermost 20 mbsf at Site 1200. This activity is remarkable, considering the prevailing pH of 12.5; this is clearly a community of extremophiles. The zone of zero sulfate at 2-20 mbsf in Hole 1200E requires that sulfate is being actively reduced at two depths, one just above the zone of sulfate exhaustion and the other just below it. The presence of two such zones is confirmed by the data from Holes 1200D and 1200F, which clearly show two sulfate minima, the first at 1-3 mbsf and the second at 14 mbsf. The sulfate maximum that lies between these minima in both holes will eventually be erased by diffusion unless it is externally supplied. In Hole 1200D this maximum lies at 12 mbsf, the same depth as the chloride maximum attributed earlier to a lateral incursion of seawater. In Hole 1200E, however, there is no evidence of such an incursion, and it is likely that this maximum is unsupported; the depth profile is thus non-steady state and implies that sulfate reduction was initiated relatively recently in this hole.
In most marine sediments,
once dissolved sulfate goes to zero with depth, as in Hole 1200E, it stays
there, as there is usually no deeper source of additional sulfate. The only
previously identified exceptions are young ridge flanks, such as the eastern
flank of the Juan de Fuca Ridge drilled during ODP Leg 168, where unaltered
seawater downwells through basement outcrops and flows laterally, carrying a new
supply of sulfate to the base of the sediment column, and settings such as
Middle Valley, drilled during ODP Leg 139, in which previously deposited gypsum
or anhydrite is redissolving in the sediment, thereby supplying sulfate to the
interstitial water. Site 1200 represents a third case, in which sulfate is being
supplied by upwelling of sulfate-rich water from a deep slab source. The two
depths of sulfate reduction in the holes at Site 1200, therefore, each have a
different source of sulfate; the upper level is supplied by downward diffusion
of seawater sulfate, against advective upwelling, whereas the lower level is
supplied by advection of sulfate from a deep source below. These different
supplies should be reflected in the sulfur isotopic composition of the sulfide
produced. Mottl and Alt (1992) found that the 
34S
of dissolved sulfate in the deep fluid from Conical Seamount Site 780 was 14
± 1, considerably
lighter than the value of 20.5
in seawater and suggestive of a contribution from a basaltic source. (By
contrast, pore water from serpentinite within the inactive Torishima Forearc
Seamount yielded 34S
values as high as 31.2,
suggestive of a residue from microbial sulfate reduction.)
The products of sulfate reduction by oxidation of organic matter include carbonate alkalinity, dissolved sulfide, and ammonia. These products tend to exhibit maxima corresponding to those of the sulfate minima (Fig. F51), confirming that there are upper and lower zones of microbial sulfate reduction. The simplified reaction is usually written as follows:
2 CH2O + SO42- = H2S + 2 HCO3-.This reaction produces two equivalents of (carbonate) alkalinity. At the high pH of the present solutions, the reaction is more appropriately written as follows:
2 CH2O + SO42- + 3 OH- = 2 CO32- + HS- + 3 H2O.This reaction also produces a net of two equivalents of alkalinity, the difference between four equivalents of carbonate plus one of bisulfide alkalinity produced vs. three of hydroxyl alkalinity consumed. Because this reaction consumes hydroxyl, effectively producing hydrogen ions, it cannot cause the pH maximum observed in Holes 1200E and 1200F. The most likely hydroxyl-generating reactions involve dissolution of Fe-bearing minerals in the serpentinite, giving rise to the observed maximum in dissolved Fe within the uppermost 10 mbsf (Fig. F53). The likely candidates are magnetite and the ferrous silicate components in olivine and orthopyroxene. At high pH, these reactions are as follows:
4 Fe3O4 + HS - + 11 H2O = 12 Fe2+ + SO42- + 23 OH -,The dissolution of magnetite is probably the predominant reaction because secondary magnetite is abundant in the altered rock, and its dissolution produces by far the most hydroxyl ions per mole of starting phase. As this reaction involves reduction of Fe and oxidation of dissolved sulfide, it is almost certainly microbially mediated.
So far, we have assumed that the microbiota are consuming normal organic matter in the process of reducing dissolved sulfate. Yet, this is not normal deep-sea sediment; the organic carbon content of partially serpentinized, depleted harzburgite squeezed up from below is essentially zero. As noted earlier, it could not be detected in 36 of 49 samples analyzed. Values exceeding 0.1 wt% were found only in the uppermost 3 mbsf. What, then, are the microorganisms eating? The answer must be that the ultimate source is the C1 through C6 hydrocarbons supplied with the deep upwelling fluid. Whereas methane can be biogenic, these higher hydrocarbons can only be thermogenic and, thus, must have a deep source. The reaction at high pH can be written as follows:
CH4 + SO42- + OH - = CO32- + HS- + 2 H2O.These simple hydrocarbons contain no nitrogen, however. This, too, must ultimately be supplied from depth, in the form of ammonia in the deep fluids. Once a microbial fauna has been established that utilizes these resources, it can intercept the organic carbon and nitrogen streaming in with the ascending fluid and effectively trap them within the ecosystem, where they can be recycled and continually enriched. This process may explain the enrichment in organic carbon in the uppermost sediment (Fig. F55). The molar ratio of organic carbon (methane) to nitrogen (ammonia) supplied by the deep fluid is ~8 at both Conical and South Chamorro Seamounts (Table T8). This compares with the Redfield ratio of 106/16 = 6.6, implying that the deep fluids represent a good food source for the microbiota. The CH4/NH3 molar ratio within the alkalinity maximum at 2-10 mbsf in Hole 1200E ranges from 4 to 16 and averages 11 (average = 10 for all of Hole 1200E). These values are similar to both the Redfield ratio and the ratio in the supply and are consistent with a deep source for C and N in this ecosystem.
Most of the dissolved sulfide generated by these reactions surely precipitates as the iron sulfides that blackened the cores from Holes 1200E and 1200F. The plateau of sulfur enrichment in the mud extends to 25 mbsf in Hole 1200E at the spring to at least 15 mbsf in Hole 1200F, but to only 5 mbsf in the distant Hole 1200D (Fig. F55). The highest sulfur enrichments, however, are restricted to the uppermost 2 mbsf, where the gradients in both dissolved sulfide (Fig. F51) and Fe (Fig. F53) are steepest, presumably because upwelling of the fluid compresses the gradients at the seafloor. (Mn gradients are also steepest in this chemically reduced interval, implying strong redox gradients.) This combination of dissolved species would precipitate abundant iron sulfides according to the following reaction:
Fe2+ + HS- = FeS
+ H+.
Why are there such active communities of both micro- and macrobiota at the South Chamorro Seamount summit and apparently none at Conical Seamount? The macrofauna almost certainly depend on the microbiota for their existence. The overall fluid flow rates are not obviously dissimilar at the two sites, and the organic carbon and ammonia content of the fluids are virtually identical (Table T8), implying a similar food supply. Although the drill holes at Site 780 were not located at any known spring, upwelling velocities of 1 cm/yr were documented there from the curvature in the chlorinity-depth profiles (Mottl, 1992). On the other hand, even Hole 1200D, the most distant from the South Chamorro Seamount spring, has three times the concentration of methane within the uppermost 10 mbsf that Site 780 has on Conical Seamount (Fig. F54). (This higher methane, however, is probably more an effect of the microbial activity than a cause, if, in fact, these materials are recycled within the community.) At present, this question is unanswered. Perhaps the lower salinity of the Conical Seamount fluids (410 mmol/kg total equivalents of cations or anions compared with 604 mmol/kg in the bottom seawater and 630 mmol/kg in the South Chamorro Seamount deep fluid) inhibits microbial colonization and growth. Perhaps there are similar communities on Conical Seamount that have just not been discovered yet. It is a large seamount, and 21 Alvin dives could cover only a small fraction of even its summit area.
