The potential of seismic methods for detecting cavities and buried objects: experimentation at a test site

One of the recurring problems in civil engineering and landscape management is the detection of natural and man-made cavities in order to mitigate the problems of collapse and subsurface subsidence. In general, the position of the cavities is not known, either because they are not recorded in a database or because location maps are not available. In such cases, geophysical methods can provide an effective alternative for cavity detection, particularly ground-penetrating radar (GPR) and seismic methods, for which pertinent results have been recently obtained. Many studies carried out under real conditions have revealed that the signatures derived from interaction between seismic signals and voids are affected by complex geology, thus making them difficult to interpret. We decided to analyze this interaction under physical conditions as simple as possible, i.e., at a test site built specifically for that purpose. The test site was constructed of a homogeneous material and a void-equivalent body so that the ratio between wavelength and heterogeneity size was compatible with that encountered in reality. Numerical modeling was initially used to understand wave interaction with the body, prior to the design of various data-processing protocols. P-wave imagery and surface-wave sections were then acquired and processed. The work involved in this experiment and the associated results are presented, followed by a discussion concerning the reliability of such a study, and its consequences for future seismic projects.