ABOUT THE POTENTIAL OF WELLS TO REFLECT STRESS VARIATIONS WITHIN INHOMOGENEOUS CRUST

Numerous pre-, co-, and post-seismic well level changes, that are believed to reflect stress variations in the crust, have been observed in various parts of the world. So far, analysis of these data has not favoured a distinct model of the processes that could be involved. The benefit from the signals is, in fact, limited because understanding of the underlying physical mechanisms is poor. The paper discusses the impact, that inhomogeneities in the brittle crust can have on the observations. In particular, the following problems are considered: (1) Could the well level signals be induced by the invasion of pressurized pore fluid at some place? (2) Could the hydrological conditions close to the observation well modulate any remote pore pressure variation? (3) Which features of the crust are most likely to dominate the process of stress transfer in the crust? (4) How should a stress-sensitive well be designed? The results can be summarized as follows. (1) Stress diffusion due to the invasion of pressurized pore fluid can lead to significant well level changes, only when the source region is at closer distance than a few kilometres (probably less than a few hundred metres) from the well. (2) Local inhomogeneities in the vicinity of the observation well may strongly influence the well's response to remote pressure changes. Even sign reversal of the anomaly may be encountered. (3) There is increasing evidence that the wide range of stiffness values of natural rocks (from 100 MPa or less, till 100 GPa), manifested in a fractal distribution of fault-like weak zones within relatively stiff formations, dominates stress transfer in the brittle crust. Uncertainty about the effective rheology for time scales from 10(-4) to 10(3) years limits the inversion of many geophysical observations at present. (4) Of various configurations, a well tapping an extensive hydraulic fracture could act best as a sensitive stress sensor, provided that the fracture volume is isolated from the pore space of the host rock. However, experimental evidence of the practicability of such configuration is still due. The weakness of conventional wells as stress sensors results from the fact that the process of fluid flow often lacks a sufficient control on attenuation mechanisms for low-frequency well level variations, regardless of whether the wells end in stiff or weak formations. To a large extent, the presented findings are compiled from publications of investigators working in various fields.