Episodic and rapid flow of hydrothermal fluid in a fault zone. An example from a Variscan fault in the Peak District, UK.


There is abundant evidence that pore fluids and fracture processes in the upper crust are physically and chemically coupled (Hubbert et al., 1959; Frank, 1965; Nur, 1973; Sibson, 1981). Increases in pore fluid pressure, for example due to fluid injection, can lead to rupture and an increase in seismic activity. This is readily explained by the Navier-Coulomb criteria for brittle failure and the decrease in effective stress as a result of the elevated pore fluid pressure (Price, 1966; Sibson, 1981). Conversely, the changes in groundwater levels and the surface effusions of warm water that sometimes occur along fault traces following earthquakes show that failure can also have a profound effect on fluid flow, heat and mass transport (Nur, 1974; Sibson et al., 1975; Sibson, 1981). The flow is interpreted to result from either seismic pumping as a result of dilatancy diffusion type processes or a seismic valve mechanism in which fault rupture leads to leakage of an overpressured aquifer or reservoir of fluid (Nur et al., 1972; Sibson, 1981).

In the geological record direct evidence for episodic and rapid flow of fluids associated with fracture and faulting is more equivocal. The presence of banded hydrothermal mineralisation of varying degrees of complexity in exhumed fault systems is often taken as evidence of pulsed fluid flow driven by seismic activity. Failure often results in brecciation of earlier generations of veins and the opening of large dilation voids that are then cemented by mineral precipitation from upwelling hydrothermal fluids (Wright et al., 2009). Observations of epithermal mineralisation associated with dilation jogs between en-echelon fault segments suggests that in this structural setting fluid flow is rapid, and accompanied by rapid pressure fluctuations which triggers high level boiling, or effervesence of hydrothermal fluids. This promotes precipitation of common gangue quartz and calcite as well as metalliferous minerals of economic importance (Sibson et al., 1975; Sibson, 1987; Henley et al., 2000). In other geologic settings, however, the exten to which mineral precipitation in the veins is contemporaneous with, and directly coupled to failure is still open to question.

A characteristic feature of systems with episodic pulsing of hot fluids is the development of a thermal anomaly along the high permeability paths in which flow is focused. An example is the perturbation of the temperature field observed in the Mississippi Valley Type (MVT) mineralization districts that lie around the margins of major Palaeozoic sediment basins in the continental USA (Sangster et al., 1994). Similar temperature anomalies have been reported for other sedimentary basins, for example to the south east of the Massif Central in France (Charef et al., 1988). The anomaly is seen as a difference between the temperature of precipitation of hydrothermal minerals and that of the host rock at the time and depth of burial that the mineralization took place. Using simple thermal modelling Cathles and co-workers show that the temperature difference is due to heat advection associated with episodic, rapid release of hot fluids from deeper regions of the basins (Cathles et al., 1983; Cathles III et al., 2005). In these example the fluids are thought to originate from overpressured formation water in the compacting sedimentary basins. Importantly, the fluid velocities required to produce the thermal anomaly are more than 1000 times greater than could be produced by the steady subsidence, compaction, and dewatering of the basins (Cathles et al., 1983). This suggests to us that the pulses result from a coupling between the pore fluid pressure and rock failure. We envisage that when pore fluid pressures approach lithostatic either mode I hydraulic fracturing or shear failure and development of dilation jogs along fault surfaces allows rapid dewatering of shales within the sediment pile. The fluid is channelled into high permeability aquifers located at the base of the basins where it flows outwards towards the basin margins.

Mineralization with MVT affinities occurs in the Peak District area of the southern Pennines in the UK. Here strata bound deposits (flats) of dominantly Pb-Zn and fluorite mineralization are closely associated with near vertical veins (scrins) that lie along strike-slip fault surfaces and fractures of Variscan age (Quirk, 1991). The minerlization is restricted to lower Carboniferous shelf carbonates that lie on the margins of half-graben sedimentary basins filled with lower and upper carboniferous silici-clastic sediments. As with the upper Mississippi valley sedimentary basins it is widely held that the mineralization results from basin scale migration of sedimentary formation waters (Ixer et al., 1993). However, the driving force for, flow paths and rates of fluid migration are poorly constrained. Opinion ranges from slow gravity driven flow as a result of tectonic uplift associated with the Variscan orogeny (Quirk, 1991) to a seismic valve type process with rapid dewatering of the over-pressured basin fill triggered by fault activity (Hollis et al., 2002; Frazer et al., 2014). Evidence for a thermal anomaly associated with rapid advection of fluids is not conclusive. Fluid inclusion homogenization temperatures for fluorite and calcite span a wide range from <70\(^{\circ}\) to >240\(^{\circ}\)C (Atkinson, 1983; Hollis et al., 2002; Kendrick et al., 2002). Mineralization is thought to have occurred at depths between 1 and 2km (Colman et al., 1989). Thus, assuming a geothermal gradient of 30\(^{\circ}\)\(^{-1}\), the fluid inclusion homogenization temperatures are at, or greater than the maximum expected host rock temperature at the time of mineralization. This suggests that fluid movement was rapid and therefore unlikely to be associated with slow, gravity driven flow or a gradual dewatering of the basin fill. There are questions, however, as to the reliability of some of the reported temperatures that are derived from fluid inclusion analysis with little agreement amongst researchers as to the temperature associated with different paragenetic phases.

To help us better understand the possible coupling between faulting and fluid flow in the Peak District we have used clumped isotope thermometry to determine the temperature at which a Variscan hydrothermal calcite vein precipitated. Clumped isotope thermometry is based on the fact that the rare, heavy isotopes of carbon (\(^{13}\)C) and oxygen (\(^{18}\)O) are ordered in the carbonate lattice. This is a result of the greater stability of the \(^{13}\)C-\(^{18}\)O bond compared to bonds involving either no, or a single isotopic substitution. The degree of ordering is an inverse function of temperature. As temperature increases the isotopes tend towards a more random or stochastic distribution (Eiler, 2007). Measurement of the degree of ordering allows us to estimate the temperature at which the distribution of isotopes in the calcite structure are locked in (Ghosh et al., 2006). This is analogous to the concept of a closure temperature for radiogenic isotopes or for cation ordering in minerals. A key advantage of the method is that the temperature estimate is based on the distribution of carbon and oxygen isotopes within a single phase and not on the partitioning of oxygen isotopes between calcite and it’s parent fluid as in the conventional oxygen isotope geothermometer. Thus determination of the mineral precipitation temperature is decoupled from knowledge of the parent fluid oxygen isotope composition. Combining the clumped isotope temperature (T(\(\Delta_{47}\))) with the bulk oxygen isotope composition of the carbonate represented by it’s \(\delta^{13}\)C value we can constrain the isotopic composition of the parent fluid. Recently two studies have demonstrated the use of clumped isotopes to constrain the temperature and isotopic composiiton of fluids associated with faulting and fractures in the upper crust (Swanson et al., 2012; Bergman et al., 2013).

We find that:

(i) the calcite precipitated at temperatures between 40° and 100°C.

(ii) the parent fluids range in isotopic composition from -4‰\(_{VSMOW}\) to +5‰\(_{VSMOW}\) and represent mixtures of a cool, meteoric water and a more evolved formation water.

(iii) the temperature at which the calcite precipitated is a conservative tracer for the fluid mixing. This implies that heat is rapidly advected as the hydrothermal fluid flow is focussed along the fault plane.

(iv) The calcites exhibit zoned development characterised by cyclic and pulsed evolution of precipitation temperatures and fluid compositions as a result of varying mixing ratios of the two fluid end-members.

(v) The high temperature formation water, may be connate, and has evolved under low fluid:rock ratios and is in apparent equilibrium with the silici-clastic basin fill.

Simple thermal considerations indicate that fluid flow was episodic and highly focussed along the fault plane. We conclude that rising pore fluid pressures as a result of rapid sedimentation, hydrocarbon generation and silicate diagenesis coupled with increasing tectonic stress as the basin inverted during the Variscan orogeny led to fault movement and release of the pore fluid pressure as the sediment pile dewaters. As with the example of the upper Mississippi Valley sedimentary basins the dewatering is episodic with extended periods during which the pore fluid pressure increases. These periods are punctuated by short duration episodes of fluid release. This resembles very closely a seismic valve type process in which rising pore fluid pressure reduces the effective stress on fault surfaces, ultimately leading to failure and the formation of a high permeability path along which the pore fluid is released and the pressure dissipated (Sibson, 1981).