Sensitivity enhancement of the ambient noise tomography and analysis of seasonal variations of noise sources to refine velocities in the lower crust in central Europe

The capability of the ambient noise tomography (ANT) to image subtle regional-scale velocity variations in the lower crust is limited by strong directionality of ambient noise sources in central Europe, which affects the quality of dispersion curves. Significant decrease of sensitivity kernels and sparse coverage of long interstation ray-pathes result in lower resolution at longer periods and thus increase uncertainty of the inversion solution in depth. If these well-known ANT limitations are properly addressed, the ANT is able to retrieve reliable high-resolution 3‑D shear velocities of the lower crust.

In this study we focus on seasonal variations of ambient noise sources in selected sites in different tectonic settings. We analyse ambient noise sources on continusly recorded wavefields from permanent observatories and temporary stations of AlpArray passive experiment with its complementary experiment and PACASE. These seismic networks with densely-spaced stations are well-suited for detailed analysis of period-dependent directionality of ambient noise sources and their effects on FTAN appearance and consequently on the quality of dispersion curves. In the second part of this study, we advocate a concept of layer-stripping during the stochastic inversion (enhanced ANT). It proved to be an efficient technique to explore the model space, particularly in the lower part of the crust. We discuss the sensitivity of the enhanced ANT to the imaged small-scale velocity features in the lower part of the crust, as well as the sensitivity to the sharp or gradational Moho in the models.