
Full text loading...
Conventional seismic surface wave methods, usually used in engineering and environmental geophysics, might fail to image very shallow structures (few meters) in the presence of strong lateral variation. In such a case, Full Waveform Inversion (FWI) might be an alternative method. However, the predominance of very energetic surface waves, can induce specific issues that need adapted inversion strategies. Furthermore, in practice fitting accurately measured amplitudes is not always an easy task because of the source coupling effects, the differences between 2D/3D modeling, and the effect of attenuation which is often a poorly known parameter. For this purpose, phase-based observables are good candidates for robust (FWI) inversion. However, some limitations of this observable in the case of very shallow structure FWI imaging are shown in this study. To overcome these limitations, it is proposed to take advantage of the vectorial nature of multicomponent measurements by defining new observables related to the wave polarization. In this study, the shape of the cost function associated to different polarization observables is studied for a medium with a very shallow interface.