Ancient subsurface flow systems within the Earth's crust can be recognized by the pervasive thermal and chemical alteration of rocks through which they once flowed. The many independent lines of evidence from field and laboratory based studies leave little doubt that much of the Earth's crust has interacted with large volumes of fluids. Indeed, Oliver [1992] correctly points out that this should come as no surprise since continental rocks above 10 km depth are continuously exposed to fluids through most of geologic time. Field evidence for crustal-scale flow can be deduced from chemical and mineralogical alteration of sediments, thermal anomalies recorded by rock thermochronometers, and the remagnitiza-tion of rocks. This evidence is discussed below and by Oliver [1992], Ferry [1994], and Garven [1995].
Thermochronometers record the time-temperature history of crustal rocks. If fluid circulation rates are high enough to produce convective heat transfer effects (>0.1 m/yr.; Garven et al., [1993], then rock thermochronometers can provide important information regarding the timing and migration patterns of ancient fluid flow systems. For example, Cook and Bowman [1994] used calcite-dolomite solvus thermometry to docuement convective heat transport in siliceous dolomites. A variety of thermochronometers have been used to trace ancient flow systems through sedimentary basins including apatite fission track analysis [ Arne, 1991; Steckler et al., 1993], vitrinite reflectance profiles [ Person and Garven, 1992], fluid inclusion freezing points [ Roedder, 1984], and petroleum generation in sediments at shallow (0.5-1.1 km) depths [ Peter et al., 1991; Hulen et al., 1994]. Depending on the technique used, convective heat transfer effects can be preserved for 200 million years or longer after flow has ceased [Person et al., in press].
Development of new techniques in radioisotope geochemistry has
also provided important new constraints on the timing of crustal
scale fluid migration [ Halliday et al., 1991]. The ability to
relate the timing of a mineralizing event to the formation of distant
orogenic belts or the emplacement of localized igneous intrusions
is of great value in establishing the driving mechanisms associated
with ancient groundwater flow systems. The most widely utilized
dating techniques rely on a number of different parent-daughter
decay systems including 
Rb-
Sr, 
Sm-
Nd,
Ar-
Ar, and 
U-
Pb. Radiogenic isotope dating
techniques have been used as a chronometer for the formation of
diagenetic minerals in clastic rocks, metal ore deposits, and the
formation of diagenetic clays Roden et al. [1993] and
Elliot and Aronson [1993]. Radiometric techniques also show
promise for dating the timing of oil migrations [ Parnell and
Swainbank, 1990]. Application of these dating techniques has not
been without complications; many of the age dates for the timing
of migration events are not entirely consistent. This is undoubtedly
due to complexities in the evolution of these (episodic) systems and
due to inherent problems with the dating methods.
Crustal-scale fluid flow systems play a critical role in the formation of diagenetic cements and destruction of porosity within sedimentary basins and crystalline rocks. Many recent field [ Budd and Vacher, 1991; Montanez and Read, 1992; McCullough and Land, 1992; Gregg et al., 1993; McManus and Hanor, 1993] and modeling [ Sanford and Konikow, 1989; Phillips, 1991; Kaufman, 1994] studies have attempted to quantify the volumes of fluid flow and circulation patterns associated with diagenesis. These studies have helped to establish the fluid circulation patterns, integrated fluid fluxes, and chemical mass transport processes associated with diagenesis.
Another line of evidence documenting paleo-groundwater flow can be deduced from the chemical remagnitization of crustal rocks. The timing of remagnitization events is determined by correlating paleopole directions of the remagnetized rocks with igneous rocks that cooled at the time of the flow event or by independently dating associated minerals. For example, research conducted during the past decade has yielded unequivocal evidence that sedimentary rocks within North America and Europe have been remagnetized during the Pennsylvanian and Permian periods [ McCabe and Elmore, 1989]. This remagnitization is thought to have been associated with the formation of diagenetic iron oxides and authigenic magnetite cements during during a period of regional fluid flow at relatively low temperatures (<200 C). The remagnitization event is also coincident with the formation of world class lead-zinc deposits within the midcontinent, USA. Fluids responsible for this chemical remagnitization event could be saline groundwater, hydrocarbons, or both [ Elmore and Leach, 1990].
The presence, composition, and amount of fluid in deep crustal
settings is inferred from fluid inclusions [ Roedder, 1984], and
changes in whole rock and mineral chemistry. Fluid inclusions in
minerals represent the only actual sample of fluids from ancient
crustal flow systems. Much effort has been directed towards
understanding and documenting fluid flow in contact aureoles
surrounding shallow intrusions [ Kerrick, 1991; Nabelek,
1991; Bickle, 1992; Nabelek and Labotka, 1993;
Hanson et al, 1993; Lasaga and Rye, 1993; Cook and
Bowman, 1994] and medium to high pressure metamorphic
terrains [ Dipple and Ferry, 1992; Baumgartner and Ferry,
1991; Rumble et al, 1991; Criss and Fleck, 1990;
Bickle, 1992; Ague, 1994; Ferry, 1994; Manning
and Bird, 1991; Leger and Ferry, 1993; Barnett and
Chamberlain 1991; Wickham et al., 1993]. In all
metamorphic environments, arguments for pervasive fluid flow, as
well as intense focusing along shear zones and veins have been
presented. Pervasive fluid flow parallel to original sedimentary
layering is dominant, but flow perpendicular to it has been
described. Time integrated fluid fluxes in regional and contact
metamorphic environments reported in the above studies are as
high as 10
cm
/cm
(see also Rumble, 1994; Ferry,
1994]! The nature of porosity, and hence fluid flow mechanisms,
as well as the applicability of Darcy's Law in rocks of the lower
crust have all been topics of much controversy [ Rumble,
1994]. Experimental sintering of mineral aggregates in the presence
of fluids [ Laporte and Watson, 1991; Holness and
Graham, 1991; Watson and Lupulescu, 1993; Holness,
1993] shows that re-crystallization proceeds rapidly under deep
crustal temperature and pressure conditions, equilibrium pore
geometry is obtained within a few days. Fluid-mineral wetting
angles are high, and for most systems fluids are confined to
isolated pores. Similarly, the healing of fluid filled cracks
progresses rapidly (10-100
m/h) at pressures in exess of one
hundred MPa and temperatures of several hundred degrees
centigrade. All these processes act to quickly eliminate porosity,
thereby increasing the likelihood of hydro-fracturing and buoyancy
driven fluid flow [e.g. Nakashima, 1993]. The observed
presence of zones of high electrical conductivity and certain
seismic reflectors in the middle and lower crust have led some
geophysicists [e.g. Bailey, 1994] to suggest the permanent
presence of a continous, lithostatic pressured fluid phase. While
lithostatic pressured fluids are certainly present during at least some
time of active burial and metamorphism [ Peacock, 1990;
Ferry 1994] it is likely that the fluids would escape towards
shallow crustal environements due to boyency. In fact, many
petrologists have argued for the presence of essentially dry lower
[e.g. Yardley and Valley, 1994].
The stable isotope composition of fluids, particularly those of oxygen, carbon, strontium, and hydrogen, are fingerprints of the origin of fluids [e.g. Valley et al., 1986]. The composition of most fluid reservoirs is significantly different from that of mid and deep crustal rocks. Recent interpretation of spatial stable isotope composition patterns have been based on one-dimensional equilibrium [ Baumgartner and Rumble, 1988; Bickle, 1992] and kinetic formulations [ Bowman and Willet., 1991; Bowman et al., 1994] for stable isotope fluid-rock exchange. Two-dimensional models are currently being developed [ Gerdes et al., 1993]. Applications of stable isotope techniques to contact metamorphic environments are reviewed by Nabelek [1991].
Recent advances in laser and ion probe micro-analytical techniques [ Elsenheimer and Valley, 1992; Kohn et al., 1993; Valley and Graham, 1991] permit the measurement of isotopic mineral zonation in individual grains. Pronounced prograde zoning in garnets was interpreted to be due to infiltration [ Chamberlain and Conrad, 1991, 1993; Jamtveit and Hervig, 1994], though some doubt remains [ Kohn and Valley, 1994]. Eiler et al. [1992] developed a theoretical framework for stable isotope diffusion between minerals to interpret closed system stable isotope retrogression, which was subsequently used by Eiler et al. [1993] to identify open system retrogression of a granulite facies rock. Trace element zonation is used in a similar fashion to document fluid mobility during subduction [ Hickmott et al., 1992].
The reaction progress of volatile producing reactions is sensitive to fluid infiltration and was used to estimate fluid fluxes and flow directions [ Baumgartner and Ferry, 1991; Ferry and Dipple, 1991; Symes and Ferry, 1991; Ferry, 1994] using one-dimensional mass conservation equations.