Surface wave attenuation from ambient noise correlation: methods and applications to 1D and 2D arrays

Abstract:

It can be shown that the field-field correlation function of an imperfectly diffuse wave field is equal to the (time derivative of) Green’s function times the specific intensity of the noise (Weaver, 2013). The theoretical understanding permits the interpretation of correlation amplitudes and promises to facilitate the retrieval of attenuation, site amplification factors, and scattering strengths from the noise correlation. In order to develop methods for extracting attenuation from ambient noise and apply to real data (particularly in Tibetan Plateau), we propose approaches with detailed formulations for a linear array and a more general 2D station network. A particular problem in retrieving amplitudes from noise is that seismic ambient noise source is not uniform and changes with time. We tested numerical simulations with azimuthally and temporally varying noise source, and have started to add internal scattering in the simulations. Our simulations validate that amplitudes and attenuations can indeed be extracted from noise correlations for a linear array or for a more general 2D array. We propose a temporal flattening procedure, which is effective in speeding up convergence while preserving relative amplitudes. For real data, we propose an “asynchronous” temporal flattening procedure that does not require all stations to have the data at the same time. Tests on real data suggest attenuations extracted are comparable with those from earthquakes.