Simulation of Seismic Tunnel Detection Experiments in Heterogeneous Geological Media

Detecting covert tunnels and other underground openings is an important yet challenging problem for geophysicists, especially where geological heterogeneity is pronounced. A number of geophysical methods have been employed to solve this problem, each with varying degrees of success. We focus on the near-surface seismic techniques of surface wave backscattering, surface wave attenuation tomography, body wave diffraction imaging, and resonant imaging. We use the elastodynamic wave propagation code E3D to simulate tunnel detection experiments completed at this site for a range of synthetic fractal velocity models. The Black Diamond mine, located near Pittsburg California, is used for the field test of our analysis. Our results show that for the relatively low-frequency surface wave attenuation and backscattering methods, the maximum detectable tunnel depth in a homogenous medium is approximately equal to the wavelength of the probing Rayleigh wave. The higher-frequency body wave diffraction and resonant imaging techniques are able to locate tunnels at greater depths, but require more sophisticated analysis and are prone to greater attenuation losses. As is expected, for large values of heterogeneity amplitude, ɛ, the percent standard deviation from the mean velocity model, the average observed surface wave attenuation signal decreases and the maximum detectable tunnel depth decreases. However, for moderate values of heterogeneity amplitude (ɛ < 3%), the average surface wave attenuation signal increases and the maximum detectable tunnel depth increases. For the body wave diffraction and resonant imaging experiments, as ɛ increases the complexity of the observed signal increases, resulting in more difficult processing and interpretation. The additional scattering attenuation tends to degrade the signals significantly due to their reliance on lower amplitude and higher frequency waves.