Retrieving 2-D laterally varying structures from multistation surface wave dispersion curves using multiscale window analysis

The analysis of multistation surface wave records is of increasing popularity in imaging the structure of the Earth due to its robustness on dispersion measurement. Since the representation of multistation surface wave dispersion curves (DCs) is uncertain in laterally varying media, average information beneath the receiver array is assumed to be obtained by inverting the dispersion curves with a horizontally layered model. To retrieve a more realistic 2-D laterally varying structure, we present a multiscale window analysis of surface waves (MWASW) method for analysing 2-D active-source surface wave data. The MWASW method is based on the use of a forward algorithm for calculating the theoretical DCs over 2-D models and multisize spatial windows for estimating the dispersion data. The forward algorithm calculates the theoretical dispersion considering the lateral variation beneath the receiver array; hence, the estimated DC is not treated as representative of the average properties but as data containing the lateral variation information. By inverting the dispersion data extracted from different spatial windows, the subsurface information at different depth ranges and lateral extensions are integrated to produce a shear wave velocity model. The dispersion curves analysed from smaller spatial windows retrieve the shallow structure with a higher lateral resolution, whereas the phase velocity data from larger spatial windows provide average information with a greater depth. We test the effectiveness of the MWASW method using three synthetic examples and two field data sets. Both results show the improved lateral resolution of the S-wave velocity structure retrieved with the MWASW method compared to the traditional multistation method in which the local horizontally layered model is adopted.