SANS: Daily Global Seismic Ambient Noise Source Maps

With the prospect of amplitude and full-waveform based methods applied to ambient noise tomography and monitoring, the importance of modelling underlying noise sources has only increased. Heterogeneous noise source distributions can influence full waveform or travel time measurements leading to possible misinterpretations, e.g., in terms of time-varying Earth structure. Especially secondary microseismic sources (0.1 to 0.2 Hz) have strong spatio-temporal variations on multiple time scales. We present a framework where Seismic Ambient Noise Source (SANS) maps are computed daily on a regional to global scale and made available to the public.

This is possible due to recent methodological developments including pre-computed wavefields and spatially variable grids (Igel et al., 2021) and a combination of two methods to locate noise sources: Matched Field Processing (MFP) and nonlinear finite-frequency inversion. Despite differences in the two methods, Bowden et al. (2021) show that these can be derived from one another. Using the computationally inexpensive MFP algorithm, we produce an initial model, which accelerates convergence of a non-linear finite-frequency inversion. In collaboration with the Swiss National Supercomputing Centre, the SANS framework runs two daily inversions for stations surrounding the North Atlantic and a global distribution of stations. Subsequently, they are made available on the website (sans.ethz.ch) where users can view and download the inversion results and possibly implement them in their own research.

Analysis of a full year of SANS inversions shows the seasonal spatio-temporal variations of the secondary microseismic sources. Modelling cross-correlations for various different noise source distributions demonstrates the significant effect changing noise sources have on the cross-correlation waveforms which should not be neglected. We strongly encourage the implementation of noise source distribution knowledge into ambient noise tomography and monitoring methods to avoid unwanted artefacts and misinterpretations.