The possibility of electrical earthquake precursors (EEPs) has long been appreciated, but to date there still exists neither a solid theory to describe their generation and expected characteristics, nor proven techniques to identify and discriminate true precursors from noise. Experimental studies have produced a prolific variety of signal shape, complexity and duration, but no explanation for the apparently indefinite diversity. Statistical analyses on the basis of such poorly constrained data were inconclusive, leading to scepticism and intense debate. The most objective means of EEP identification would be to construct generic models of their source(s) and compare the model predictions with field observations. We attempt to show the merits of this approach with two studies. The first study expands on the phenomena of spontaneous electric field generation during crack propagation (microfracturing), demonstrated by laboratory experiments. Large-scale microfracturing may appear at the terminal stage of earthquake preparation. We apply a generic, qualitative model, based on a kinetic theory of crack interaction and propagation. The model suggests that EEP signals from such a type of source may have a limited class of permissible waveforms, with characteristic bay- or bell-shaped curves of variable width and duration. We provide two examples consistent with this model: the VAN claims of precursors on 15/1/1983 and 18/1/1983. The magnetic field that may accompany an anomalous electric signal is the subject of the second study. This has been a grossly overlooked quantity, although valuable for identification and discrimination, because it is considerably less sensitive to distortion than the electric field, less sensitive to inhomogeneities along the propagation path, insensitive to the local geoelectric structure and sometimes, telltale of the source (for instance, external magnetic fields can only be generated by (sub)horizontal current configurations). We investigate the 18/4/1995 and 19/4/1995 electric and magnetic signals observed at Ioannina, Greece, used for the prediction of the 13/5/1995 M6.6 Kozani event by the VAN group. The electric and magnetic waveforms are inconsistent with the crack propagation model. By their observed characteristics, the magnetic signals preclude any (sub)vertical electrokinetic current. Using analytic formulations, we investigate whether they might have been generated by an electrokinetic source across a lateral interface, either at the focal area or locally, at Ioannina. We conclude that the magnetic field properties are also inconsistent with such a type of source. Conversely, we cannot rule out their local industrial origin. The examples presented herein indicate that the successful identification and discrimination of EEP and noise may be possible by working out plausible theories of the source.
Read full abstract