[S12-2] An interdisciplinary approach towards earthquake prediction studies II

[S12-2-03] Probability tomography and wavelet analysis of self-potential data and possible application in landslide monitoring

^○Qinghua Huang¹, Kaiyan Hu¹, Katsumi Hattori² (1.Peking University, Beijing, China, 2.Chiba University, Chiba, Japan)

invited

Self-potential method is a kind of near-surface geophysical technique, which has been adopted in exploration of metal ore, monitoring of contaminants and natural hazards. This study focuses on the self-potential data processing. The source element occurrence probability tomography can give the probability of the source location and the charge property. Due to the limited resolution, the probability tomography method might yield ambiguous or even misguiding results in case of the multiple sources with a short distance. In order to overcome the above disadvantage and enhance the tomography effect, we combine the charge occurrence probability tomography with the complex wavelet transform method in self-potential data processing. We adopt the commercial finite element method (FEM) software COMSOL Multiphysics in the forward modeling. We apply the complex wavelet analysis the synthetic self-potential data obtained from the forward modeling of some given models. The results show that the complex wavelet analysis not only can locate the electric sources, but also can identify the electric features (e.g., homogeneity, inclined angle of dipole or polarized body, etc.). Furthermore, we apply the combined probability tomography and the continuous complex wavelet analysis to the synthetic self-potential data. The model test shows that the combined method could reduce the computation of 3D current source probability and enhance the imaging resolution. We also apply the combined method to the data from sandbox experiments and test the possible time-lapse tomography. This study may provide an effective approach of monitoring ground water flow, with potential application in landslide monitoring.