Reducing parameter coupling effect for subsoil density estimation using 3D seismic full-waveform inversion with horizontal-force source

Abstract

The information of density (ρ) of the shallow subsoil layers or structures is crucial in hydrogeology, geotechnical engineering, agriculture and engineering seismology. 3D seismic full-waveform inversion (FWI) can be a promising tool for estimating the spatial distribution of density in soils. The observed seismic response is relatively less sensitive to density than to seismic velocities. In order to estimate accurately the density distribution, it is necessary to mitigate the effect of deterioration of the inverted model caused by the interaction of the different model parameters, which is commonly known as the parameter coupling effect. In this research, we have found that the use of a horizontal-force source instead of a vertical-force source mitigates the parameter coupling effects in 3D FWI. We perform synthetic studies on a spatially uncorrelated subsoil model using different force directions. We find that the gradients of the misfit function for horizontal source show much less parameter coupling than vertical source, which results in better reconstruction of the ρ model as well as the S-wave velocity (VS) model. This can be achieved using a standard gradient-based optimization without implementing complex algorithms or computationally expensive accurate Hessian inverse, in order to mitigate the coupling effects. Our results and discussions are crucial to estimate the spatial distribution of the soil physical parameters such as porosity or small- strain shear modulus from ρ and VS obtained from 3D FWI.