A quantitative model of fault reactivation due to depletion of hydrocarbon reservoirs has been developed based on accurate mechanical modeling. The reservoir has been modeled as a thin poroelastic inclusion represented by a discontinuity in a poroelastic matrix while the fault gouge obeys the Mohr-Coulomb failure criterion. To pose the boundary value problem for the stress perturbations caused by reservoir depletion, the boundary conditions on the discontinuity sides have been derived using asymptotic analysis. A general boundary collocation technique for the solution of this class of boundary value problems has been suggested. Using this technique, redistributed stresses around a depleted reservoir with non-uniform pressure distribution have been calculated and the fault instability caused by reservoir depletion has been analyzed. The magnitudes of the slip displacements developed in the fault zone have been estimated by modeling the slipped portion of the fault as a shear (mode II) fracture. Based on the fault slip magnitude, the seismic moment and the magnitude of the generated earthquake have been computed. The results show that even if the magnitude of induced earthquake is insignificant in terms of seismic activity, the slip still can be more than sufficient to shear the existing boreholes.
Geological discontinuities such as faults are inherent in most petroleum formations [e.g., Glennie, 1998]. There is a number of human activities, such as hydrocarbon production, that can sufficiently alter the in-situ stresses within a period of few years or even few months resulting in reactivation and slip of the nearby faults. Fault reactivation due to reservoir depletion may have various consequences ranging from shearing boreholes drilled through the fault zone [e.g, Bruno, 1992; Dusseault et al., 1998] to inducing seismicity [e.g., Segall, 1989; Baranova et al., 1999] to drastically affecting the formation permeability [e.g., Addis et al., 1998; Rudnic?', 1999]. Stress state perturbation resulting from subsurface fluid extraction clearly and robustly demonstrates the importance of the poroelastic effect (in this context first suggested probably by Geertsma [1966]). Indeed, the fault stability is not affected at all if the poroelasticity of the reservoir material is not taken into account.
Probably the most studied example of extractioninduced seismicity is that of the Lacq (France) gas field where seismic events have been continuously monitored for more than 25 years [e.g., Grasso and Wittlinger, 1990; Lahaie et al., 1998]. The largest reported deep seismic events, triggered by gas extraction, are three major earthquakes (with M> 7) near the Gazli gas field, Uzbekistan, in 1976-1984 [e.g., Simpson and Leith, 1985; Arnordse and Grasso, 1996].
Production induced casing damages have been reported in numerous publications [e.g., Bruno, 1992; Dusseault et al., 1998]. The most well known example of extraction-induced well damage is probably near Wilmington oil field near Los Angeles, California [e.g., Kovoch, 1974; Yerkes and Castle, 1976; Dusseault et al., 1998]. Reservoir compaction and stress perturbations caused by production damaged hundreds of wells [Dusseault et al., 1998] and resulted in 8 damaging earthquakes with M = 5.1 in 1949 and apparently reverse slip mechanism [Nicholson and Wesson, 1992]. A large number of well failures had been associated with the fault movements [e.g., Bruno, 1992].
