Seismic waves estimation and wavefield decomposition: application to ambient vibrations

SUMMARY Passive seismic surveying methods represent a valuable tool in local seismic hazard assessment, oil and gas prospection, and in geotechnical investigations. Array processing techniques are used in order to estimate wavefield properties such as dispersion curves of surface waves and ellipticity of Rayleigh waves. However, techniques presently in use often fail to properly merge information from three‐components sensors and do not account for the presence of multiple waves. In this paper, a technique for maximum likelihood estimation of wavefield parameters including direction of propagation, velocity of Love waves and Rayleigh waves, and ellipticity of Rayleigh waves is described. This technique models jointly all the measurements and all the wavefield parameters. Furthermore it is possible to model the simultaneous presence of multiple waves. The performance of this technique is evaluated on a high‐fidelity synthetic data set and on real data. It is shown that the joint modelling of all the sensor components, decreases the variance of wavenumber estimates and allows the retrieval of the ellipticity value together with an estimate of the prograde/retrograde motion.