Sources of mid-frequency ambient seismic noise

Project background

Ambient seismic noise is generated by many sources at the Earth’s surface. The ocean is the most prominent source up to a frequency of approximately 0.4 Hz. Human activities are an important source beyond 1 Hz. In between, it is often assumed that a location-specific combination of coastal waves, local meteorological sources, anthropogenic sources, and other sources like rivers, dominate the noise record.

A recent observation of ambient noise power spectral density between 0.4 and 0.5 Hz indicates a source that may be coherent over some distance, and may not be related to oceanic activity. Motivated by this enigmatic observation, the goal of this project is to investigate ambient seismic noise in the frequency range 0.4 – 1 Hz in Europe using existing power spectral density data from the recently established SeedPSD webservice of the European Integrated Data Archives (EIDA, e.g., https://seisdata.epos-france.fr/eidaws/psd/1/). This service gives us access to a large repository of ambient seismic noise data on the European scale.

Research questions

• Is it possible to identify short-lived, spatially coherent events directly from power spectral density data ?

• Can we approximately map distinct source regions based on power spectral densities ?

• Are there multiple occurrences of non-oceanic, coherent sources at 0.4-1 Hz over the space of one to two years ?

• What is the most plausible explanation for the previously observed event (in 2024) and other potential events (if further events are detected during the thesis work) ?

Representative tasks for the project, to give you an idea of the daily work :

• building a Python workflow to obtain and visualize power spectral density data ;

• building a workflow to spatially map potential source regions based on observed power spectral densities ;

• visual and potentially automated identification of noise “events” ;

• corroboration by additional processing (for example, polarization analysis or cross-correlation analysis) and interpretation.

Profile and skills, which can also be built during the internship :

• understanding of techniques to describe stochastic signals such as ambient seismic noise (e.g. Fourier spectra, power spectral density, cross-correlations) ;

• understanding of seismic wave propagation (wave types, understanding of concepts like attenuation, polarization, interference) ;

• Python programming and independent literature research.

For any questions, please contact us at the e-mail addresses above. To apply directly, please send an e-mail with your CV and a few informal sentences describing your motivation to the same addresses.