Seismic hum observations and modeling

Soutenance de these directrice de these: Eléonore Stutzmann Summary: The hum is the permanent free oscillations of the Earth at frequencies between 2 and 20 mHz. This thesis aims to increase the understanding of the hum observations and its generation mechanism. The long period hum signal can be used to image the Earth up to profound depths (up to 500 km). Including recordings from Ocean Bottom Seismometers (OBS) could improve tomographic models by increasing station coverage. At the vertical OBS trace, the small hum signal (ngal) is overshadowed by signals from ocean infragravity wave forcing (compliance) and from currents (tilt). In some cases, we also record transient electronic signals (glitch). We remove these signals from the vertical seismic trace by subtracting average glitch signals; performing a linear regression and using frequency-dependent response functions between pressure, horizontal and vertical seismic channels. After this noise reduction we observe the Earth’s permanent free oscillations for the first time at two OBS in the Indian Ocean. The hum originates from atmosphere-ocean-solid earth interactions. However, a clear consensus on the generation mechanism is not yet reached. We use the ocean bottom vertical pressure force as the source of the hum from a model that assumes the interaction of free infragravity waves with a coastal sloping topography. On average, we find strongest hum sources in the Southern Hemisphere, along the western coasts. We calculate the Rayleigh wave propagation from the pressure force to 17 seismic stations around the Indian Ocean, and the two OBS stations; and manage to fit the seismic data with the model. Local sources contribute most to the hum signal. Using the model, we identify storms and cyclones in the hum recordings. We conclude that infragravity wave-topography interactions is the main generation mechanism of the hum.