ALBERT

Description: Inversion of surface wave dispersion properties is commonly used to derive shear wave velocity depth profiles. However, one of the critical and yet rarely considered issues in this ill-posed inversion process is mode contamination. Rayleigh dispersion modes are the theoretically possible solutions of motion. Experimentally, we define Rayleigh dispersion properties from spectra energy maxima in some domain (as f–k), thus possibly producing only apparent experimental dispersion curves, where energy spreads onto several modes. If this phenomenon is not recognized, the inversion of an apparent dispersion curve can produce results unrelated to the actual subsurface structure. In this work, we present the results of synthetic tests that highlight the most common subsoil conditions and acquisition pitfalls that can give rise to surface wave mode contamination. In particular, we consider three typical subsoil structures that can produce this phenomenon: 1) a simple two-layer system with a strong impedance contrast, 2) a layered system with velocity inversion, and 3) a two-layer system having a lateral discontinuity. It is known that all such systems produce, to some extent, situations in which Rayleigh wave energy propagates at higher modes rather than on the fundamental mode. We also consider the impact of acquisition parameters, which influence mode contaminations mainly as a consequence of inadequate spatial sampling. This exercise leads to identifying the most critical subsoil characteristics that control ambiguous mode inversion. We finally propose possible a priori detectors and warnings for the most critical subsoil conditions leading to mode contamination.