On geologic time scales, fluid movement within the Earth's crust
is driven by a number of mechanisms including sediment
compaction, gradients in (water-table) topography, lateral variations
in fluid density, seismogenic pumping, and the production of
diagenetic, metamorphic, and magmatic fluids (Fig. 1). Fluid
circulation within the shallow crust (<6 km) is characterized by
moderate to high permeability conditions (10
to 10
m
; Neuzil, [1994] and fluid pressure gradients that range
from hydrostatic (10 MPa/km) to near lithostatic levels (26
MPa/km; Person and Garven, [1994]. Here, topography-,
compaction-, and density-driven flow systems dominate. These
fluid flow driving mechanisms are reviewed in detail by
Garven [1995]. Relatively low permeability conditions
(<10
m
) and near lithostatic pore pressure gradients
characterize he middle and lower crust ( >10km) where overburden
stresses exceed the ductile limit of rocks. Flow systems are driven by
devolatilization reactions, mantle and magma degassing here [
Hanson, 1992, Furlong et al., 1991]. Several mechanisms
have been proposed to pump fluids into the middle and lower crust
during metamorphism, including permeability changes along active
deep fault zones and seismogenic pumping [ Sibson, 1994]. A
combination of the above mechanisms dominates within a transition
zone between a shallow groundwater regime (<6 km) and the
middle to lower crust where rock permeability is significantly
reduced. The depth to the transition zone is not fixed; it depends
on local thermal, hydrologic, rheological, and tectonic factors. All
of these processes are coupled within this transition zone and
deeper. Additional complications result from the complex feedback
between fluid flow, heat, and reactive mass transport processes [
Tsang, 1991; Lowell, et al., 1993; Ortoleva, 1993].
Compaction-driven flow within sedimentary basins and accretionary prisms results from mechanical loading of the sedimentary pile during subsidence and sedimentation and/or emplacement of thrust sheets [ Ge and Garven, 1994; Person and Garven, 1994]. The combination of low-permeability sedimentary fill, and pore-space collapse causes pore fluids to become anomalously pressured in compacting basins. Fluid pressures due to compaction-driven flow can approach lithostatic levels if sediment permeability is sufficiently low. Compaction driven flow systems can transport fluids over large lateral distances (>100 km) if a regional aquifer is present. In deeper crustal environments, porosity generated by fluid producing reactions and tectonic fracturing is inherently unstable, causing pore space collapse and fluid expulsion [ Rumble, 1994]. Tectonic compaction [ Ge and Garven, 1992], associated with continental collision, can occur at all depths (shallow, transitional, deep) producing pore space collapse. This fluid flow driving mechanism is referred to as the ``squeegee model'' by Oliver [1992]. Compaction driven flow is probably not an important agent for heat and chemical mass transfer within the crust due to the limited quantity of fluids associated with this mechanism.
Under subaerial conditions within sedimentary basins, tectonic and thermal processes operating within the lithosphere can create uplift that will develop a regional topography-driven groundwater flow system with flow from areas of higher to lower elevation, assuming the water-table configuration is a subdued replica of land surface (Fig. 1). Depending on the water-table configuration and subsurface permeability conditions, topography-driven flow systems can be continental in scale [ Garven et al., 1993]. Topography-driven groundwater flow may be the most important fluid flow inducing mechanism on a regional scale within sedimentary basins [ Sverjensky and Garven, 1992]. Ancient continental-scale, topography-driven groundwater flow systems are thought to have driven basinal brines [ Garven, 1995] and petroleum [ Bethke et al., 1991] hundreds of kilometers across the North American craton during ancient mountain building events, forming world class energy and mineral deposits. These topography-driven flow events remain active until erosional processes incise the landscape causing local flow cells to form.
Density-driven fluid flow arises from lateral fluid-density gradients [ Furlong et al., 1991; Hanson, 1992; Raffensperger and Garven, in press]. Density gradients can be produced by temperature or compositional variations in the fluids or due to boiling. Temperature perturbations within the crust can be caused by radiogenic decay, magmatic intrusions (Fig. 1), crustal thinning, and thermal conductivity contrasts. Density gradients associated with solute concentra-tion changes occur as a consequence of rock-water interactions such as evaporite dissolution [ Evans et al., 1991] or due to evaporative pumping within closed basin/playa lake environments [ Duffy and Al-Hassan, 1988]. However, vertical salinity gradients within basins have been shown to actually inhibit flow [ Deming and Nunn, 1991]. Recent studies of hydrothermal systems have focused on forced convection [ Ingebritsen et al., 1991, Deming et al., 1992; Hanson, 1992], the effects of aquifer heterogeneities on free convection [ Braster and Vadasz, 1993], and the hydrodynamics of boiling [ Ingebritsen and Rojstaczer, 1993]. Recent experimental and theoretical studies have shown that phase separation can occur up to very high temperatures and pressures (greater than 600 C and 200 MPa; see review by Labotka, [1991].
Prior to an earthquake, pore space dilation is thought to occur in
rocks around fault zones due to buildup of shear stress, frictional
heating, and/or metamorphic fluid production. Development of
fracture porosity normal to the least principal stress occurs at this
time producing decreased fluid pressure and causing pore water
infilling. During the subsequent earthquake, pore space collapse
drives fluids upwards towards the land surface along the most
permeable conduits through the fault zone. Based on volume
change estimates,10
liters of fluid can be produced around a 10
km spherical zone of dilation during an individual earthquake cycle
[ Sibson et al., 1975]. If seismogenic pumping occurs
repeatedly through geologic time, then this mechanism may provide
an efficient agent for focused fluid flow in the Earth's crust. The
occurrence of ore mineralization and chemical alteration of
sediments [ Wood and Boles, 1991; Glazner and Bartley,
1991] around fault zones as well as increased spring activity and
surface water flows [ Muir Wood, 1994], which typically peak
a few days after seismic events and continue for up to a year
afterwards, have frequently been cited as important field evidence
supporting this mechanism. However, increased surface flow after
a seismic event may also be due to increased fracture permeability
associated with earthquakes [ Rojstaczer and Wolf, 1992].
More recent mechanisms proposed to account for flow along fault
zones include the poro-elastic and fault-valve models [ Sibson
1994].
Progressive devolatilization of hydrous and carbonaceous mineral
phases upon heating is an important source of fluids in deeper
crustal environments, where it is often the only fluid source.
Kerrick and Caldeira [1993] calculated that more than 10
moles per million years of CO
were produced during the
Himalayan orogeny. A similar magnitude of H
O is probably
expelled, since shales comprise an important component of the
sedimentary pile [ Peacock, 1990]. Carbon dioxide and water
produced in such a manner will escape along shear zones [
Dipple and Ferry, 1992a], and by hydrofracturing [ Walther,
1990] towards the shallow groundwater system [ Kerrick and
Caldeira, 1993]. The considerable amount of methane generated
during the thermal decomposition of organic material [
Peacock, 1990] is probably oxidized in shallow ground water
environments to CO
and H
O [ Kerrick and Caldeira,
1993].
The different driving forces described above compete with each other for control of the hydrologic flow systems within different tectonic environments and during different stages of the plate tectonic cycle. The relative importance of these different driving forces on fluid flow varies depending on the tectonic and lithologic conditions (e.g. permeability, porosity, and mineralogy). Some studies have attempted to isolate the role of specific driving mechanisms on fluid flow (see for example, Ge and Garven, [1992]. Other studies have emphasized how several different driving mechanisms interact [ Furlong et al., 1991; Person and Garven, 1994]. The next section attempts to summarize much of the recent work that has been done to characterize fluid flow mechanisms within different tectonic environments.