In the preceding sections, I have attempted to highlight some of the studies within the past four years that have significantly advanced our understanding of submarine hydrothermal systems, and to indicate a number of topics that need to be pursued to further the eventual development of a quantitative predictive model of hydrothermal circulation within the oceanic crust.
The recent studies on the East Pacific Rise and the Juan de Fuca Ridge have emphasized the rapid rates and magnitudes of changes observed in hydrothermal systems in response to volcanic and tectonic events. The initial processes accompanying crustal accretion events, their time scales, their variability, and their global importance are currently unknown, and will require time-series measurements beginning immediately following such an event and continuing over periods of years. A major breakthrough in our ability to conduct such studies occurred in 1993 when a real-time acoustic monitoring system for low-level seismicity along the Juan de Fuca Ridge using a network of permanent deep-ocean hydrophone arrays (SOSUS) owned by the U.S. Navy became operational. Within four days, a seismic event was detected north of Axial Seamount [ Fox et al., 1994]. A response effort was launched and documented a lateral diking event associated with a small volcanic eruption similar to those observed on Hawaii and in Iceland [ Embley et al., 1994]. This detection capability, combined with plans for in situ instrumentation over the long-term to investigate the coupling between volcanic, tectonic and hydrothermal events and the temporal variability of hydrothermal systems, will lead to a better understanding of crustal accretion processes.
In addition, we need to further our understanding of the spatial distribution of hydrothermal systems, particularly along slow-spreading ridges, and along mid-ocean ridges systems that are not well explored. Remote sensing technologies that detect anomalies in the physical and chemical characteristics of the water column are being used with some success to delineate ridge segments that appear to be hydrothermally active. However, pin-pointing the exact location of a vent field is problematic, and still somewhat serendipitous. This means that two approaches are necessary. The first is to conduct detailed multidisciplinary studies of known hydrothermal fields in different settings in order to derive simple models that predict the combinations of tectonic and volcanic features that result in hydrothermal activity. The second approach is to combine water column surveys, which will define hydrothermally active segments, with high-resolution mapping and acoustic imaging surveys, which will provide information on tectonic and volcanic structure, of previously unexplored sections of the mid-ocean ridges.
Although alluded to only briefly in this review, another important type of axial hydrothermal system occurs in sedimented rifts---the setting of some of the largest deposits in the geologic record. At Middle Valley on the northern Juan de Fuca Ridge, active venting occurs through hundreds of meters of turbidites, and sulfides are also precipitated within the sediment column [ Goodfellow and Franklin, 1993]. During ODP Leg 139 in 1991, a 94 m section of massive sulfides was drilled at one location, and downhole pressure and temperature sensors were installed in two holes to monitor the rebound of the system from the perturbations caused by drilling. Results from these studies, including models of the hydrothermal circulation system are expected soon; however, this drilling leg demonstrated not only the feasibility of drilling into sea floor sulfide deposits to investigate subsurface hydrothermal processes, but also the potential of using the perturbations caused by drilling to learn more about the physical characteristics of the hydrothermal system. This concept was expanded in the fall of 1994, when drilling at the TAG active hydrothermal mound on the Mid-Atlantic Ridge was combined with in situ instrumentation of the mound to document changes in the circulation of the hydrothermal system during and after drilling. This approach of sea floor instrumentation combined with drilling looks promising, and may represent a powerful way of investigating the physical properties of the oceanic crust.
One of the outstanding problems---and arguably one of the most difficult to solve---is quantifying the relative importance of thermal and chemical fluxes associated with focused high temperature fluid flow and diffuse lower temperature discharge. Black smoker fluids can be readily sampled, whereas methods for sampling and determining the areal extent of diffuse flow, which results from subsurface mixing of hydrothermal fluids with seawater, are just being developed. Consequently, flux calculations tend to overestimate the importance of high temperature fluid compositions and underestimate fluxes associated with diffuse flow [ Alt, 1994]. This is further complicated by our lack of knowledge of the distribution of hydrothermal circulation cells both on-axis and off-axis. Off-axis, discharging fluids may circulate through, and react with, overlying sediments resulting in significantly different heat and chemical fluxes. In order to understand the role of hydrothermal processes in regulating the chemistry of seawater, the relative proportions of these different styles of venting, and the pervasiveness of the associated seawater-rock reactions within the crust must be determined. Since these cannot be determined by direct observations, flux estimates must rely on the development of models that relate permeability structure and its evolution to focused vs. diffuse flow while incorporating data on chemical exchanges during seawater-rock interactions obtained from studies of seafloor and ophiolitic rocks, and from experimental studies.
The recognition that seawater-rock reactions in the lower sheeted dikes and upper gabbros may determine vent fluid chemistry has been an important step forward. A next important step is to investigate further the role played by phase separation of both magmatic and seawater-derived hydrothermal fluids in determining the chemistry of the fluids discharging at the sea floor. Vent fluid chemistry, analyses of fluid inclusions in altered rocks from the sea floor and from ophiolites, and experimental studies all suggest complex processes involving phase separation occur within the reaction zone [e.g. Berndt and Seyfried, 1990; Kelley et al., 1992, 1993; Von Damm et al., 1992; Butterfield and Massoth, 1994]. Additional studies and experiments are needed, in conjunction with the development of numerical models of two-phase flow of magmatic fluids and seawater, to critically examine this process and assess its importance.
The study of hydrothermal systems is still relatively young, and there are many fundamental questions that remain to be addressed. The upcoming years will provide a number of exciting opportunities to constrain the physical and chemical nature of hydrothermal circulation, and to refine our models of submarine hydrothermal processes.
Acknowledgments. I would like to express my gratitude to the many colleagues who provided their views on the most significant advances in their own fields of research, and offer my apologies to those whose work could not be included because of the brevity of this review. K. Becker and P. Saccocia reviewed and offered helpful comments on sections of this paper. Discussions with R. Detrick and K. Gillis were useful and are much appreciated. I gratefully acknowledge NSF grants OCE-9013150 and OCE-9314697 which in part supported this work. This is WHOI Contribution No. 8890.