Search

EGU General Assembly 2019, in Viena, Austria, 7–12 April 2019, Many ambient noise tomography applications are based on measuring surface wave group velocities from timedomain cross-correlations. This approach has several downsides, among them being: (A) reliable group velocity measurements require at least an inter-station distance of 3 wavelengths, (B) usually, only the fundamental mode of Rayleigh wave propagation is considered, disregarding the possible contribution to the observed group/phase velocity of higher modes, and (C) surface wave dispersion measurements only offer weak constraints for the very shallow and the very deep shear-wave velocity structure at the periods used in most regional studies. To overcome these limitations, we propose an inversion scheme based on Aki’s spectral formulation. First, we measured Rayleigh wave phase velocities by fitting Bessel functions to the observed cross-correlation spectra, and used these measurements in a tomographic inversion to produce a set of phase velocity maps. Then, we defined a grid of regularly spaced nodes on the phase velocity maps, and used the dispersion information on a non-linear inversion to obtain a 1D shear-wave velocity profile at each node. Our inversion scheme considers the influence of higher modes in the observed phase velocities. The same noise recordings used for measuring surface wave dispersion can be used to determine the H/V spectral ratios at the receivers. The fundamental frequency and the shape of the peaks in the H/V ratios can be related to the thickness of the sediment layers and the impedance contrasts between the sediments and the basement, thus providing additional constraints for the deep shear-wave velocity structure in the vicinity of the receivers. Therefore, for the nodes closer to the receivers, we performed a joint inversion of the Rayleigh wave phase velocities and the H/V spectral ratios. We applied this methodology to the data provided by the SISCAN-MISTERIOS network, a regularlyspaced grid of 40 broad-band stations with inter-station distances ranging from 27 to 440 km, deployed in the Basque-Cantabrian basin (N Spain) during the period 2014-2018. The Basque-Cantabrian basin (N Spain) is one of the major Mesozoic extensional basins developed during the opening of the Bay of Biscay. In Cenozoic times, the Alpine collision between the European plate and Iberian sub-plate caused the inversion of the normal faults, the reactivation of older (Variscan) thrusts and the creation of new structures. As a consequence, both the depth to the Paleozoic basement and the velocity structure vary greatly across the area, providing a challenging scenario to test the proposed methodology. In spite of some difficulties (i.e. weak impedance contrasts between the Mesozoic sediments and the Paleozoic basement in parts of the basin, problems to retrieve accurate spectral ratios for some stations at the considered frequencies, etc), the application of the methodology described above provided good results in areas where the structure was known in detail, and allowed us to achieve a better constrained image of the 3D shear-wave velocity structure across the whole basin.

AGU Fall Meeting 2019, in San Francisco, 9-13 december 2019, Surface waves in seismic noise correlation functions are routinely exploited to image structures at a broad range of scales and resolutions. In most applications, the correlation functions are considered (1) to be estimates of the Green¿s function between pairs of receivers and (2) to be composed mainly of fundamental-mode surface waves. This approach requires at least an inter-station distance of 3 wavelengths to produce reliable phase/group velocity measurements, which can be a limiting factor when working at local or regional scales. Additionally, as shown by many previous studies, the presence of superficial low-velocity layers may cause a dominance of higher modes over certain frequency ranges, which means that only considering the fundamental mode may not be enough to accurately model the observed dispersion curves. In this work, we overcome the limitations by measuring Rayleigh wave phase velocities directly from the cross-correlation spectra. This approach does not depend on the far-field approximation and is, therefore, not subject to the three-wavelength rule. We use these measurements to produce a set of phase velocity maps using a standard tomographic algorithm. Finally, for a set of regularly spaced nodes, dispersion curves are compiled from the phase velocity maps and used in a non-linear inversion to obtain the 1D shear-wave velocity structure. Previous geological and geophysical studies are used to place constraints on the non-linear inversion. We apply this methodology to seismic data acquired in the Basque-Cantabrian region (N Spain) between 2014 and 2018 (by the SISCAN-MISTERIOS network, with inter-station distances ranging from 7 to 400 km) and obtain a 3D model of this structurally complex area (a thick Mesozoic basin inverted and incorporated into the Pyrenean-Cantabrian mountain belt in the Cenozoic) which agrees with the existing geological knowledge but significantly extends the area for which high-resolution information is available.