Episodic venting of hydrothermal fluids From the Juan de Fuca Ridge

E.T. Baker, J.W. Lavelle, R.A. Feely, G.J. Massoth, and S.L. Walker

Pacific Marine Environmental Laboratory, National Oceanic and Atmospheric Administration, Seattle, Washington

J.E. Lupton

Marine Science Institute and Department of Geological Sciences, University of California, Santa Barbara

Journal of Geophysical Research, 94, 9237-9250 (1989)
Copyright ©1989 by the American Geophysical Union. Further electronic distribution is not allowed.

Discussion

The process (or processes) that creates megaplumes is as yet unknown. Any proposed model must account for an extraordinarily high flux of both mass and heat in the form of a fluid compositionally similar to that from normal hydrothermal vents. A precise measure of the megaplume fluxes depends on the duration of venting. Using the bounds of 2-20 days cited above, the mass and heat fluxes necessary to form megaplume 1 would have ranged from 30-3 × 10 g/s (at 350°C) and 70-7 × 10 W, respectively. The midpoints of these ranges are about a factor of 10 greater than the flux of all the hot springs at the 21°N vent field as estimated from vent orifice flow measurements [Converse et al., 1984], about a factor of 10 greater than the heat flux from the large vent field on the Endeavour Segment of the JDFR as calculated from temperature anomalies and fluid velocities [Baker and Massoth, 1987] and from radiochemical inventories [Rosenberg et al., 1988] in the neutrally buoyant plume, and about a factor of 10 greater than the hydrothermal volume flux from the entire Guaymas Basin as calculated by a silica box model [Campbell and Gieskes, 1984].

Two general types of processes suggest themselves: the creation of superheated fluid by the extrusion of molten basalt onto the seafloor or the sudden emptying of a reservoir of hydrothermal fluids in the axial crust [Baker et al., 1987; Cann and Strens, 1987]. Given a volume specific heat of 3.8 J cm °K and a latent heat of fusion of 10 J/cm, a seafloor extrusion of ~2.5 × 10 m of 1100°C basalt would yield the heat contained in megaplume 1. While a lava flow that was thin enough, of the order of 10 m or less, could cool within a few days, it is unlikely that it could produce either the observed plume height or chemistry. If one assumes that the total heat flux from the lava flow was distributed to many fissures over the lava flow, the heat flux per unit length of source would be much less than that calculated above for the single fissure model of the megaplume. Assuming the resulting individual line plumes do not influence each other until much of their buoyancy is depleted, the rise height of these many plumes from the lava flow would be much less than the rise height observed. By the same token, the interaction of seawater with an extensive high-temperature lava flow would likely produce a very open system, unlike laboratory batch experiments, where leaching of the basalt would be minimized because water would be driven away from the rock almost immediately upon coming in contact with it (K. Von Damm and M. Mottl, personal communication, 1989). This notion is consistent with the usual field characterization of recently extruded basalt as fresh and unaltered.

Although we think it likely that an event of this magnitude might be accompanied by a lava extrusion, the high discharge rate and well-developed hydrothermal chemical signature suggest that the observed megaplumes arose from the cataclysmic emptying of a reservoir of hydrothermal fluids in the axial crust. The volume of a high-temperature reaction zone in the Troodos ophiolite complex can reach ~1 km [Richardson et al., 1987], which could contain enough high-temperature fluid to form a megaplume if the porosity were high enough (~10%). In order to increase the mass flux from such a reaction zone by a factor of 10 to 10 the local crust must suffer a drastic increase in permeability. A permeability increase could occur passively, by tectonic fracturing unrelated to changes in the crustal fluid, or actively, by fracturing caused by an increase in the fluid volume or buoyancy pressure.

Tectonic fracturing, presumably caused by episodic rifting events, could increase the mass flux by coalescing fracture zones and creating new flow channels along irregularities on fault surfaces. Rapid heating of the country rock by a simultaneous dike intrusion might also accelerate local hydrothermal circulation. Sibson et al. [1975], for example, note that hydrothermal mineralization on terrestrial fault surfaces is usually episodic, with fluid pulses contemporaneous with fault motion, and that postseismic groundwater effusions of up to 10 m have been observed after large earthquakes. No earthquakes greater than magnitude 4, the approximate sensitivity limit of the land-based recording network, were associated with megaplume 1.

An alternate possibility is that the hydraulic force of the hydrothermal fluid itself occasionally exceeds the lithospheric strength and creates a temporary fracture permeability. Sibson [1981] has presented a model of this type that addresses periodic fault slippage. J.R. Cann and M.R. Strens (Modeling periodic megaplume emissions by black smoker systems, submitted to Journal of Geophysical Research, 1988) have adapted this concept specifically to a black smoker system. In their model, steadily elevating fluid temperatures increase the buoyancy pressure in the upflow zone until the strength of a low-permeability caprock is exceeded and the system catastrophically empties. The subsequent infusion of cold seawater into the reaction zone enhances sulfide precipitation, which reduces permeability and reestablishes normal black smoker flow. Hydraulic fracturing might also occur by phase separation of the hydrothermal fluid caused by a sudden decompression [Kelly and Delaney, 1987; Goldfarb and Delaney, 1988; Hulen and Nelson, 1988].

Fossil evidence of high fluid flow caused by rock fracturing and subsequent pressure release comes from the hydrothermal brecciation found in terrestrial hydrothermal fields [Hulen and Nelson, 1988], in ophiolite stockworks [Lydon and Galley, 1986], and in rock samples from the Mid-Atlantic Ridge [Delaney et al., 1987] and the Blanco Fracture Zone in the northeast Pacific [Hart et al., 1986]. Textural and chemical characteristics of hydrothermal breccias are compatible with megaplumelike events: flow channels wide enough to allow the movement of clasts up to tens of centimeters in diameter, a fluid flow rate high enough to transport large clasts, and episodic occurrences within the same hydrothermal system. The prevalence of hydrothermal breccias in the geologic record implies that cataclysmic fluid releases are not uncommon along the mid-ocean ridge system.

The occurrence of megaplumes on a ridge segment may be associated with other anomalies in the discharge rate of hydrothermal heat. Detailed plume observations on the Cleft Segment during the last few years are consistent with this idea. Over at least the last 2 years (assuming megaplume 2 originated on the Cleft Segment) episodic venting has released heat at an annual rate of ~2-4 × 10 W. Continuous venting is focused at two sites centered near 44°40N and 44°54N. From measurements of plume size, temperature anomaly, and advective transport, Baker and Massoth [1987] estimated the present-day heat flux from the vent field at 44°40N to be 6 ± 3 × 10 W. The plume at the northern end of the Cleft Segment [Baker and Massoth, 1986b] is even larger and more intense, and although its flux has not yet been determined, its size and temperature anomaly appear at least comparable to those of the plume evolved from the Endeavour Segment of the JDFR. The heat flux from that vent field is 1-5 × 10 W [Baker and Massoth, 1987; Rosenberg et al., 1988]. Thus a conservative estimate of the rate of heat loss by means of continuous hydrothermal discharge on the Cleft Segment is 2-6 × 10 W, plus any diffuse convective heat not accounted for by the plume measurements. The observed hydrothermal heat flux from episodic and continuous venting within the axial valley of the Cleft Segment is therefore at least 4-10 × 10 W.

This measured flux is considerably greater than various theoretical estimates of the available axial heat from the 55-km-long Cleft Segment. For example, at steady state, the latent heat of fusion from a column of dikes and gabbro 5 km thick (roughly the thickness of the brittle layer for a 3-cm/yr half-spreading rate ridge) plus the specific heat of cooling from 1 km of intruded dikes gives a flux of 1 × 10 W along the Cleft Segment. Axial-specific estimates are lower still. The hydrothermal flux within 3 km of the axis, calculated from the heat flow anomaly expression of Wolery and Sleep [1976], is 4 × 10 W. The thermal model of Morton and Sleep [1985], based on the conductive and convective cooling necessary to account for the depth of the axial magma chamber inferred from seismic reflection surveys, yields a flux of 5 × 10 W for hydrothermal heat sinks within 3 km of the ridge axis.

We suggest that the present imbalance between the observed and modeled heat flux on the Cleft Segment is a consequence of the fact that the hydrothermal release of heat and mass is in general an unsteady process. Is there an interpretable connection between hydrothermal episodicity and changes in the local magmatic budget? Lichtman and Eissen [1983] suggest that volcanism and hydrothermal activity along a ridge crest occur in evolutionary cycles controlled by random cycles of magma replenishment. They characterize the Cleft Segment as currently in the initial stage of a new high-effusion phase caused by a renewed magmatic pulse. Similarly, a morphological analysis of the JDFR by Kappel and Ryan [1986] led them to propose that ridge crest topography is not steady state but evolutionary, and that the Cleft Segment is currently in a phase of renewed volcanism, extending itself northward by a series of linear, rift-parallel belts of fissure eruptions. The high rate of hydrothermal discharge, and in particular the episodic megaplume events, might therefore reflect a recent surge in the local cycle of magmatic episodicity.

Acknowledgments. This research was supported by the National Oceanic and Atmospheric Administration (NOAA) Vents Program. We thank Terri Geiselman for conducting the scanning electron micrograph analyses of the particles, Geoff Lebon for XRF support, Maria J. Restrepo for performing the silica analyses, Kevin Roe for dissolved Mn determinations, and Karen Thornberry for He analyses. J.R. Cann provided helpful comments on the manuscript. J.E.L. was supported by the Ocean Sciences Division of NSF. Contribution 1050 from NOAA/Pacific Marine Environmental Laboratory.

Go back to previous section or forward to next section

Return to PMEL Outstanding Paper page

Return to Publications and Abstracts page

Return to PMEL Home Page