Multimode Rayleigh wave profiling by hybrid surface and borehole methods

To improve the accuracy of shallow seismic shear wave velocity profiling, we propose a minimally invasive hybrid surface-and-borehole method that enhances the detection of higher modes of Rayleigh wave dispersion data. The new method combines techniques from the multichannel analysis of surface waves and multichannel simulation with one receiver (MSOR) methods to record components of Rayleigh wave motion at the surface as well as at shallow depths within the soil mass. The performance of the proposed method is demonstrated through computational and experimental studies. We show that individual modes of Rayleigh waves can exhibit different dominant depths at which their motion is most significant. This is demonstrated through a numerical study of eigenvectors of layered soil profiles via the stiffness matrix method, and confirmed by a finite element simulation of the apparent dispersion trends recorded at shallow depths using MSOR. Upon superimposing dispersion data recorded via the receivers at various depths, the resulting multimode dispersion data is used in a multi-objective inverse analysis, for which the difference between experimental and theoretical dispersive phase-velocity spectra are minimized for multiple modes simultaneously. In the numerical study, we demonstrate that the resulting inverted profiles and theoretical dispersion data have improved accuracy relative to single-mode inversion. Preliminary field tests are performed using the new hybrid method, and the results are shown to support the conclusions of the numerical study and confirm the feasibility of the proposed technique. Although the use of multiple modes in surface wave testing is not new, the proposed hybrid method can provide more accurate and complete multimodal dispersion data than achieved with surface-only Rayleigh wave methods. As a result, errors because of misidentification or partial measurement of higher modes may be minimized, thus reducing statistical uncertainty in the inverted profiles