DISCUSSION

The pore fluids and serpentinized materials from the South Chamorro and Conical seamounts are unique in several respects. Values of ¹¹B in the fluids (Wei et al., this volume; Savov et al., 2004) indicate slab sources derived from depth rather than a shallow-marine sediment source. In addition, the low D and high ¹⁸O of the Conical Seamount pore waters are clearly different from seawater (Benton, 1997) and point to mineral dehydration/metamorphic reactions as a water source. Moreover, the pore fluid concentrations of boron in the South Chamorro Seamount (Mottl et al., 2003) are also similar to those along the décollement of the Nankai Trough (You et al., 1993), indicating that perhaps the fluids encountered in the décollement of the Nankai Trough are also sourced from greater depths.

Whereas local productivity in the Mariana region could potentially produce an iodine-rich sediment component, the organic content of the sediments around the South Chamorro Seamount is <0.5% in the upper meter and generally <0.2% in deeper sediments (Salisbury, Shinohara, Richter, et al., 2002) and generally <0.3% throughout the core sections of the Conical Seamount (Fryer, Pearce, Stokking, et al., 1990). Depth profiles for pore fluid iodine (Fig. F2) are similar to those found in sediments along other margins with high productivity (Martin et al., 1993; Mahn and Gieskes, 2001; Egeberg and Dickens, 1999; Fehn and Snyder, 2003; Fehn et al., 2003). In these cases, however, the increase in iodine concentration with depth tends to be paralleled by an increase in bromine concentration. The Mariana serpentinite mud volcanoes, however, have bromine concentrations in the pore fluids that actually decrease slightly with depth (Wei et al., this volume). Given that seawater is depleted in iodine and local productivity is also an unlikely source, the pore fluid iodine most likely originates from deeply derived slab fluids.

Another unique aspect of this study is that both the serpentinite muds and clasts from the South Chamorro and Conical seamounts have the highest iodine concentrations ever recorded for nonsedimentary rocks (6–102 µmol/kg). Muramatsu and Wedepohl (1998) note that hydrothermal iodine enrichment occurs only modestly in altered granites (1.3 µmol/kg) and show that even metamorphic rocks do not have iodine concentrations >0.4 µmol/kg. The possibility that a significant and previously undocumented iodine reservoir exists in low-temperature altered ultramafic forearc materials has several important implications. First, it provides a mechanism for transport of iodine to subarc depths. This explains, in part, both the widespread presence of iodine in island arc volatiles and the presence of subducted ¹²⁹I signatures in fumaroles and crater lakes (Fehn and Snyder, 2003; Snyder and Fehn, 2002; Fehn et al., 2002). A similar mechanism of the involvement of heavy ¹¹B mantle wedge serpentinites in the arc magma sources was recently proposed by Straub and Layne (2002) to explain the very heavy ¹¹B of tephra glasses from the Izu island arc.

Second, the presence of a significant iodine reservoir in the forearc, and possibly in subarc conditions, can have significant implications in our understanding of the chronology of the Earth's accretion and differentiation. One measure of the time involved in planetary accretion is by comparing the ratio of ¹²⁹Xe_excess/¹²⁷I of the bulk silicate earth (BSE) to that of primitive meteorites (e.g., Allègre et al., 1995; Zhang, 1998, 2002). These time estimates depend on a reliable estimate of the fraction of Xe presently degassed from the whole mantle, as well as the amount of iodine presently residing in the BSE. In addition, there are cases where excess ¹²⁹Xe (the daughter product of the decay of ¹²⁹I) has been observed, presumably migrating from beneath continental crust. The transfer of iodine-bearing serpentinized peridotites to great depths may be a mechanism for initiating heterogeneities in the iodine reservoir of the upper mantle and lower crust, although it does not explain the concurrent abundance of the lighter xenon isotopes where excess ¹²⁹Xe is observed (Caffee et al., 1999).

Finally, the release of iodine in forearc regions may have a significant impact on the marine iodine cycle. A recent study of the iodine flux from a nonserpentinite mud volcano in the Black Sea (Aloisi et al., 2004), when extrapolated to a global scale, suggests that the input of iodine into the oceans from mud volcanoes is on the same order of magnitude as the total riverine input (Muramatsu and Wedepohl, 1998). Given the even higher iodine concentrations present in the South Chamorro Seamount (Fig. F2A), it is possible that the flux is even greater.

A comparison of iodine and bromine concentrations in muds, clasts, and pore fluids (Fig. F4) offers some insight into the degree to which these systems are coupled. In order to compare the solid and fluid components, both were normalized to the sediment volume. The end-members suggested in Tables T1 and T2 were also normalized to the volume. With only one exception, the mud samples fell along a mixing trend between depleted mantle (DM) values and marine sediment (MS) values. The concentration of boron in the clast samples exceeds that of the mud and is roughly a third that of the MS end-member. Despite the boron enrichment, the clasts are not more enriched in iodine than the muds. The clasts plot below the mixing line between DM and MS, whereas the muds plot above the mixing line. The clast data is consistent with the recycling of a serpentinized source in island arc systems because volcanic rocks also have a similar ratio of iodine to boron. Pore fluid boron concentrations in the Mariana serpentinite mud volcanoes have pore fluid boron concentrations that all fall along a restricted range of 1000–2000 mmol/m³, despite large variations in iodine concentration. This trend is distinct when compared with the average for the relatively shallow décollement fluids of the Nankai. From these data, we can conclude that significant amounts of iodine are incorporated into the serpentinized clasts, although less readily than boron, and that at greater depths, iodine is released into the pore fluids, which are released in the Mariana mud volcanoes. Nonserpentinite mud volcanoes may also show similar enrichments in iodine (Aloisi et al., 2004).

BSE values for halogens are generally calculated from crust and mantle values (Déruelle et al., 1992; Muramatsu and Wedepohl, 1998) and by comparison to the relative ratios of volatile to nonvolatile elements in meteorites; however, it has been noted that adsorption and entrainment may complicate matters (Allègre et al., 2001). Although the marine sedimentary pile has been considered the major reservoir, our data also suggest that the failure to account for adsorption to serpentinized ultramafic materials may also produce an underestimate of the crustal iodine reservoir. Likewise, large uncertainties exist in the estimation of BSE values for boron (McDonough and Sun, 1995; Allègre et al., 2001). Because BSE values generally plot between mantle and sediment values, a large altered ultramafic reservoir would shift the BSE for these two elements away from the DM and toward the MS/3 end-member.

Marine sediments make up a scant 4.9% of the oceanic crust by mass, yet they compose 68% of the total iodine inventory in crustal and lower crustal rocks (Muramatsu and Wedepohl, 1998). When combined with sedimentary rocks on the continental crust, >90% of the crustal inventory is presumably accounted for (Déruelle et al., 1992; Muramatsu and Wedepohl, 1998). The overall iodine budget may be subject to some scrutiny, however, because it does not consider the enrichment of ultramafics through hydrothermal alteration. Enrichment of boron in oceanic crust is well documented (e.g., Seyfried et al., 1984; Spivack and Edmond, 1987), although supporting data to confirm or refute iodine enrichment in the altered oceanic crust are conspicuously lacking. Because iodine concentrations in sediments and associated pore waters are a function of both marine productivity and deposition rates (Kennedy and Elderfield, 1987; Mahn and Gieskes, 2001), fluids associated with hydrothermal circulation in the open ocean may not have high concentrations of iodine available to alter the oceanic crust in areas of low productivity. Where iodine is available in formation waters, a modest increase in the I/Cl ratios of mineralizing fluids associated with the quartz veins of Cu porphyry deposits has been observed (Kendrick et al., 2001); perhaps this process may be analogous to the process of iodine enrichment in the oceanic crust as well as in ultramafic material. Thus, the phenomenon of iodine enrichment in ultramafic materials is possibly limited to forearc serpentinization along active convergent margins

The South Chamorro and Conical seamounts provide a unique window to determine the presence of iodine in the deep forearc. The similarity between boron and iodine profiles in both the fluids and serpentinized material suggests that these two elements behave somewhat similarly in forearc settings, although boron may be preferentially preserved during deep serpentinization. Despite this, both the ultramafic clasts and the serpentinized muds point to a previously undocumented reservoir that may contribute significantly to the overall crustal iodine budget. Although our data are limited in quantity and distribution, they point to an area that merits future research. Further investigations should combine halogen, stable isotope, and ¹²⁹I determinations to ascertain the source and residence time of the halogens within these unusual fluids and the serpentinized material.