Hydrothermal (Phreatic) Craters and Deposits

Steam eruptions that involve little or no juvenile tephra are termed hydrothermal eruptions (Muffler et al ., 1971), phreatic eruptions , or mud volcanoes (White, 1955); they are characteristic of the periodic behavior of many fumarolic areas. These eruptions form small craters, usually less than several hundred meters in diameter, which are

Fig. 3.4
(a) End-on view of 20- to 25-m-high travertine fissure ridge at Monon Hill, Creede caldera, Colorado.
The Creede travertine deposits are located along faults that intersect caldera margins and along
fractures located at the base of a resurgent dome. (b) Algal mounds in a travertine fissure ridge,
Creede caldera, Colorado. (c) Diagram of a travertine fissure ridge, showing the relationship
between springs that rose along faults and fractures and the layered travertine deposited there.
Where the travertine fissure ridge dammed a drainage, carbonate beds extend out as thin
wedges from the fissure ridge and are interbedded with clastic sediments.

surrounded by breccia, surge, and fallout deposits composed of lithic debris.

Hydrothermal eruptions occur in areas where the vapor pressure of geothermal fluids exceeds the hydrostatic boiling pressure for a given temperature. These eruptions take place at a point where the convective rise of geothermal fluids is impeded by a relatively impermeable layer termed caprock (Facca and Tonani, 1967). Caprocks are commonly formed when rock permeability is sealed by the precipitation of solids from geothermal fluids. Where a hydrothermal reservoir has a significant vapor-dominated region, additional over-pressure might be transmitted from the deep, cold-water hydrostatic head (Nelson and Giles, 1985). The vapor pressure of a geothermal fluid receives significant partial contributions from CO₂ and H₂ S, as well as H₂ O. Because the former two gases have lower sublimation temperatures than water does, they can increase the vapor pressure by up to several tens of bars more than that of pure water for any given temperature (Kieffer, 1982; Nelson and Giles, 1985). When expansion of the overpressured fluid is initiated by a trigger such as the chemical breakdown and failure of the caprock or seismic or hydraulic fracturing, the eruption proceeds as a vaporization wave propagates down through the fluid and accelerates the vapor and fluid out through the conduit in a manner similar to that proposed for geyser eruptions (Kieffer, 1982; 1984a; 1989).

Two examples of hydrothermal eruption models, developed by Hedenquist and Henley (1985), are illustrated in Fig. 3.5. The first model is for a shallow hydrothermal reservoir with a temperature of 195°C and a depth of 200 m. The second is for a reservoir with a temperature of 230°C and a depth of 400 m. In both models, a sealed caprock has developed at 100-m depth through deposition of silica—and possibly carbonate where lower temperatures have allowed exsolution of CO₂ . Because the overlying rock is sealed, fluid flow may be diverted at some greater depth and may follow another pathway to the surface. As vapor continues to accumulate below the sealed rock, the liquid water surface falls, and vapor pressure from the greater reservoir depths is transmitted to the seal (especially if significant noncondensible vapors are present).

In the case of eruption from the shallow hydrothermal reservoir, the transmitted vapor pressure is just above the lithostatic pressure. When the eruption occurs, hydrodynamic flow through the fracture conduit becomes hydrodynamically choked [u = u_sonic , as in Eq. (2-5)], and a pressure-balanced eruption occurs. Kieffer (1977) found that water-steam systems have greatly reduced sound speeds, ranging from 1 m to several hundred meters per second. In such an eruption, ejecta will follow ballistic trajectories to form fallout deposits.

For eruption from the deep hydrothermal reservoir, vapor pressure transmitted to the caprock can greatly exceed that required to lift the overburden. Overpressure builds because of the strength of the caprock. When the caprock fails, choked flow in the conduit is at a pressure above the lithostatic pressure and vent erosion occurs. The eruption is overpressured and supersonic at the surface; this circumstance produces blast conditions that form a crater and pyroclastic surges dominate ejecta dispersal.

In both shallow and deep reservoirs—but especially in shallow ones—a triggering mechanism is required to initiate caprock failure. The gradual breakdown of caprock strength through rate-limited chemical dissolution, sudden jarring by a seismic event, rapid heating by magma intrusion, the sudden influx of noncondensible gas, or unloading of overlying material through an avalanche or draining of a lake might trigger failure. An additional mechanism, hydraulic fracturing (discussed in Chapter 2), contributes to the failure of caprocks (Norton, 1984).

Hydrothermal eruption phenomena are strong indicators of active hydrothermal reservoirs. Phreatic craters and their deposits range from pits 1 m across to lake-filled depressions up to 1 km in diameter, as is seen in the example of the Eastern Kawerau Geothermal Field in New Zealand (Fig. 3.6). Ejecta deposits from these eruptions can extend >1 km from the crater center. Studies of lithic clasts within phreatic explosion breccias indicate that the foci for hydrothermal eruptions occur at many depths and have been observed as deep as 350 m (Bixley and Browne, 1988).

Most phreatic breccia deposits are massive, but they may also be bedded and may include graded bedding and pyroclastic surge dunes—features that are indicative of multiple steam blasts (Table 3.1). The massive deposits consist of poorly sorted angular tephra from sub-millimeter size to blocks several meters in diameter in a muddy matrix (Fig. 3.7). Nearly all the lapilli and blocks are hydrothermally altered and/or silicified; the glass has been altered to clay or hydrothermal quartz, and lithic and crystal components are replaced by clays, pyrite, chlorite, and other hydrothermal minerals. Many of the lithic clasts retain their relict textures and may contain several generations of fractures, filled with hydrothermal minerals, that formed in-situ during earlier hydrothermal

Fig. 3.5
Models of hydrothermal eruptions, showing (a) pressure vs depth and (b) a schematic cross section.
Pressure variations with depth of rock overburden (lithostatic), cold, vapor-unsaturated water
(hydrostatic), and hot, vaporsaturated water (saturated liquid) are indicated. The saturated liquid
curve depicts the effect of vapor pressure transmitted from a hot, underlying reservoir. For the
first model, involving a shallow reservoir at a temperature of 195°C, the pressurization history is
shown by a dashed line (A—A'). A gradual buildup of vapor pressure under the silica-cemented
cap continues until its failure, at which time steam erupts through a conduit with choked flow
and emerges at the surface at atmospheric pressure (dotted line). For the second model (B—B'),
which involves a deep reservoir, the transmitted vapor pressure greatly exceeds the lithostatic
overburden at the silicified cap. Ensuing eruptions are greatly overpressured as they emerge
from the vent conduit (dotted line), so that vent erosion promotes
the entrainment of lithic ejecta in expanding fluids.

explosions (Nairn and Solia, 1980). Bedding-plane sags are common; they resulted when blocks ejected during steam blasts impacted the muddy, fine-grained beds deposited earlier.

An analysis of the volume and variety of lithic clasts within a phreatic breccia provides a rough indication of both the stratigraphy under a geothermal area and the reservoir depth. One example of phreatic breccia, described by Espanola (1974), is from a minor hydrothermal eruption in early July 1966 within the Tikitere and Taheke hydrothermal fields of New Zealand:

"The eruption debris reached a maximum thickness of 0.3 m and consisted mainly of finely comminuted mudstone and sandstone with occasional large blocks up to 0.3 m diameter. The remaining debris was mostly pumice breccia and some rhyolite fragments from the Rotoiti Breccia Formation. The debris was found to have flattened the nearby surrounding scrub and damaged the tourist tracks. The old path, which partly encircled the amphitheater containing spring 6, was rendered impassable due to gravity slumping of the amphitheater banks towards the spring. The absence of Mamaku Ignimbrite in the ejecta indicates that the source of the eruption lay within the Rotoiti Breccia, i.e. within 90 m from the surface."

Phreatic craters and their deposits are undisputed indicators of the presence of a high-temperature hydrothermal system and therefore are excellent prospecting tools. If there is a sequence of phreatic breccia deposits, dating each deposit may provide information on thermal pulses that occurred during the history of the geothermal field. Because phreatic eruptions can be initiated accidentally by drilling or failure of a casing in a geothermal well (Bixley and Browne, 1988), these events must be considered potential hazards during the drilling and production processes.

Fig. 3.6
Simplified geological map of the eastern Kawerau
Geothermal Field, New Zealand, showing the
location of phreatic explosion craters, the
Rotoiti phreatic explosion breccia, and normal
faults that traverse the area
(bar and ball on downthrown side).
Thicknesses are shown in meters.
(Adapted from Nairn and Solia, 1980.)