Integration of geoelectrical and seismic refraction data by means of fuzzy clustering for landslide investigation

Summary Seismic and geoelectrical methods are powerful tools to investigate the landslides. The effectiveness of the investigation will significantly increase if we can exploit the strength of each method and complement its information into the combination model. The question is how we can put the models together. In this work, we utilize the advancement of fuzzy clustering technique to integrate seismic refraction and geoelectrical datasets in a co-operative inversion process. The fundamental idea to use fuzzy clustering is to build a model that resembles geology, particular rock units. We apply our method to a dataset acquired at Doi Ong Tuong, Hoa Binh province, Vietnam. The dataset includes refraction seismic and direct current data for landslide investigation. Seismic refraction is good at defining a structure to assist geoelectrical inversion. In turn, the geoelectrical method is sensitive to low resistivity media that usually relates to weakened zones, but is not good at structure definition. Our results are consistent with the borehole information. Applying fuzzy clustering again to the models of velocity and resistivity, we can create a clustering map that is more interpretable than using directly the inverted models.