Modeling of fault reactivation and induced seismicity during hydraulic fracturing of shale-gas reservoirs

• We studied potential fault activation and seismicity in shale-gas hydraulic fracturing.
• Fault orientation, strength, permeability and stress field are critical parameters.
• Fault activation creating flow paths to shallow aquifers is a remote possibility.
• Shale-gas fracturing along faults may only give rise to (small) micro-seismic events.

We have conducted numerical simulation studies to assess the potential for injection-induced fault reactivation and notable seismic events associated with shale-gas hydraulic fracturing operations. The modeling is generally tuned toward conditions usually encountered in the Marcellus shale play in the Northeastern US at an approximate depth of 1500 m (~4500 ft). Our modeling simulations indicate that when faults are present, micro-seismic events are possible, the magnitude of which is somewhat larger than the one associated with micro-seismic events originating from regular hydraulic fracturing because of the larger surface area that is available for rupture. The results of our simulations indicated fault rupture lengths of about 10–20 m, which, in rare cases, can extend to over 100 m, depending on the fault permeability, the in situ stress field, and the fault strength properties. In addition to a single event rupture length of 10–20 m, repeated events and aseismic slip amounted to a total rupture length of 50 m, along with a shear offset displacement of less than 0.01 m. This indicates that the possibility of hydraulically induced fractures at great depth (thousands of meters) causing activation of faults and creation of a new flow path that can reach shallow groundwater resources (or even the surface) is remote. The expected low permeability of faults in producible shale is clearly a limiting factor for the possible rupture length and seismic magnitude. In fact, for a fault that is initially nearly-impermeable, the only possibility of a larger fault slip event would be opening by hydraulic fracturing; this would allow pressure to penetrate the matrix along the fault and to reduce the frictional strength over a sufficiently large fault surface patch. However, our simulation results show that if the fault is initially impermeable, hydraulic fracturing along the fault results in numerous small micro-seismic events along with the propagation, effectively preventing larger events from occurring. Nevertheless, care should be taken with continuous monitoring of induced seismicity during the entire injection process to detect any runaway fracturing along faults.