A review of recent developments concerning the structure, mechanics and fluid flow properties of fault zones

Deep drilling and induced seismicity experiments at several locations worldwide indicate that, in general, the brittle crust in intraplate regions is critically stressed, pore pressures are close to hydrostatic, and in situ bulk permeability is ∼10 −17 to 10 −16 m 2 . This high permeability, three or four orders of magnitude higher than that measured on core samples, appears to be maintained by critically stressed faults and greatly facilitates fluid movement through the brittle crust. We demonstrate that such high permeabilities can maintain approximately hydrostatic fluid pressures at depths comparable to the thickness of the seismogenic crust. This leads to the counterintuitive result that faulting keeps intraplate crust inherently strong by preventing pore pressures greater than hydrostatic from persisting at depth.

How faulting keeps the crust strong

The standard model of injection-induced seismicity considers changes in Coulomb strength due solely to changes in pore pressure. We consider two additional effects: full poroelastic coupling of stress and pore pressure, and time-dependent earthquake nucleation. We model stress and pore pressure due to specified injection rate in a homogeneous, poroelastic medium. Stress and pore pressure are used to compute seismicity rate through the Dieterich (1994) model. For constant injection rate, the time to reach a critical seismicity rate scales with t ∼ r2/(cfc), where r is distance from the injector, c is hydraulic diffusivity, and fc is a factor that depends on mechanical properties, and weakly on r. The seismicity rate decays following a peak, consistent with some observations. During injection poroelastic coupling may increase or decrease the seismicity rate, depending on the orientation of the faults relative to the injector. If injection-induced stresses inhibit slip, abrupt shut-in can lead to locally sharp increases in seismicity rate; tapering the flux mitigates this effect. The maximum magnitude event has been observed to occur postinjection. We suggest the seismicity rate at a given magnitude depends on the nucleation rate, the size distribution of fault segments, and if the background shear stress is low, the time-varying volume of perturbed crust. This leads to a rollover in frequency-magnitude distribution for larger events, with a “corner” that increases with time. Larger events are absent at short times, but approach the background frequency with time; larger events occurring post shut-in are thus not unexpected.

/pdf/injection-induced-seismicity-poroelastic-and-earthquake-54uxtxkl5e.pdf

Injection‐induced seismicity: Poroelastic and earthquake nucleation effects

We systematically describe an approach to estimate the large-scale permeability of reservoirs using seismic emission (microseismicity) induced by fluid injection. We call this approach seismicity-based reservoir characterization (SBRC). A simple variant of the approach is based on the hypothesis that the triggering front of hydraulically-induced microseismicity propagates like a diffusive process (pore pressure relaxation) in an effective homogeneous anisotropic poroelastic fluid-saturated medium. The permeability tensor of this effective medium is the permeability tensor upscaled to the characteristic size of the seismically active heterogeneous rock volume. We show that in a homogeneous medium the surface of the seismicity triggering front has the same form as the group-velocity surface of the low-frequency anisotropic, second-type Biots wave (i.e., slow wave). Further, we generalize SBRC for 3-D mapping of the permeability tensor of heterogeneous reservoirs and aquifers. For this we apply an approach similar to the geometric optics approximation. We derive an equation describing kinematic aspects of triggering-front propagation in a way similar to the eikonal equation for seismic wavefronts. In the case of isotropic heterogeneous media, the inversion for the hydraulic properties of rocks follows from a direct application of this equation. In the case of an anisotropic heterogeneous medium, only the magnitude of a global effective permeability tensor can be mapped in a 3-D spatial domain. We demonstrate the method on several field examples and also test the eikonal equation-based inversion.

Characterization of fluid transport properties of reservoirs using induced microseismicity

A review of mechanical models of dike propagation: Schools of thought, results and future directions

SUMMARY

During the hydraulic-fracturing experiment in the German Continental Deep Drilling Borehole (KTB) in December 1994, microseismic activity was induced. Here we develop a technique for estimating permeability using the spatio-temporal distribution of the fluid-injection-induced seismic emission. The values we have obtained for the KTB experiment (0.25times10-16 to 1.0times10-16 1.0times10-16 m2) are in a very good agreement with the previous hydraulic-type permeability estimates from KTB deep-observatory studies. In addition, our estimates of the hydraulic diffusivity support the previously calculated value for the upper crust, which is of the order of 1 m2 s-1. However, this estimate now relates to the depth range 7.5-9 km.

Günter Borm

Papers

Estimating the crust permeability from fluid-injection-induced seismic emission at the KTB site