Electrode Response in Seismo-Electric Measurements

Abstract:

Seismo-electric measurements consist in recording the transient electric fields generated by seismic waves propagating in fluid-filled porous or fractured media. These electric fields are usually measured by voltage differences between two electrodes. Unfortunately, the electrode spacing and their locations on the ground surface have a direct influence on the signal-to-noise ratio of the measurements, on the recorded waveforms and on their arrival times. Using a filter theory approach and full waveform numerical simulations of the coupled seismic and electromagnetic (EM) wave propagation in porous media, we show that the co-seismic electric arrivals and the small-amplitude EM interface response can be severely distorted and/or attenuated by conventional surface electrode layouts. To this end, we have computed synthetic electrograms providing the electric potential, to allow us to determine voltage differences between two arbitrary locations of electrodes. Unlike the low-pass filter obtained by connecting two geophones in series, the filter associated with a voltage difference is shown to be a band-pass filter. As a result, not only horizontally and obliquely propagating waves but also vertically propagating waves undergo selective frequency attenuation in the 0-150 Hz frequency band used in field measurements. It also turns out that electrode spacing cannot be optimized to enhance the electric signature of typical seismic reflections and EM interface response, neither with horizontal dipoles nor with reasonably sized vertical dipoles. To circumvent this problem, we consider arrangements of 3 and 5 electrodes analogous to multilayer capacitors in electronics. We show that such arrangements are ideally described by low-pass filters preserving the quasi-plane waves corresponding to the EM interface response. However, in reality, these benefits are challenged by the imperfect coupling between the electrodes and the ground, represented by an electrode contact resistance which is possibly frequency-dependent and complex. Somewhat paradoxically, poorly grounded electrodes can in some cases improve the measurement of the electric field. This suggests that electrode contact resistance should be measured whenever possible to be accounted for in the dipole or multi-electrode electric response.