Broadband analysis, modelling and applications of seismic noise: from microseisms to Earth normal modes – MIMOSA

Seismic noise is recorded by broadband seismometers in the absence of earthquakes. It is generated by the atmosphere-ocean system with different mechanisms in the different frequency bands. Even though some mechanisms have been known for decades, an integrated understanding of the noise in the broadband period band 1-300 s is still missing. The causes of this poor knowledge involve a very limited quantitative knowledge on ocean wave properties at periods shorter than 5 s or longer than 25 s and difficulties in modeling the propagation of seismic waves in 3D earth models that include ocean-continent boundaries. Recent breakthroughs in numerical modelling of ocean waves and seismic propagation modelling will be used to solve these problems. In particular we will combine on-going developments in ocean wave modelling with extension towards both short and long periods, and advanced seismic modelling technique such as spectral element method to provide the first modelling of broadband seismic noise.

We will investigate different possible mechanisms for modelling the sources in the different period bands. The secondary microseisms (noise of 3-12s period) are generated by the interaction of ocean waves of similar frequencies and coming from opposite direction and we successfully modelled them. The modelling of noise at longer period is still an open question. The primary microseisms (noise of 10-20s period), is generated when ocean waves reach the coast. The mechanism is known for 50 years but a quantitative modelling is still missing and will be performed during this project. At even longer period, seismic noise is called hum and its source mechanism is still unclear. We will investigate several mechanisms that may explain it.

In order to verify key aspects of wave model parameterizations and noise source theory, we will carry out two types of in-situ experiments. First, we will perform detailed measurements of the wave directional spectrum from a fixed platform in shallow water, which will allow us to refine today's parameterizations of breaking waves, relevant for short period noise. A second experiment, with the year-long deployment of a vertical acoustic antenna in the Indian Ocean, will provide an absolute calibration of modelled noise sources with periods around 5s. These efforts will lead to more robust parameterization for the shape of the short wave spectrum and its impact on satellite measurements of wind, sea level and salinity.

Besides this work on sources, we will also improve on the seismic propagation modelling by using the spectral element method. This approach will be particularly used to investigate the effect of the ocean-continent boundary on seismic noise amplitude. This combination of calibrated sources and quantitative propagation should remove the need for empirical factors used in today's seismic noise models. We will further investigate whether propagation in 3D media that include bathymetry variability enables to generate Love waves. These waves are observed but the current models do not explain them. We will also model noise body waves to provide the first quantitative comparison between observed and modelled noise body waves.

We will analyse seismic data and synthetic seismograms in order to better understand the noise recorded by station on continents, on islands, at the ocean bottom and in the sofar channel. We will provide a well-documented catalogue of the strong noise sources. This catalogue can be useful for all noise related seismic studies. We will use seismic noise for monitoring velocity changes in 2 contexts: magma intrusion that occurred in the Canary island in 2012 and earthquakes related to ground water extraction. We will also use noise body waves to obtain a tomographic model of North America.

The MIMOSA project will thus provide results and models that will serve all geosciences, promoting seismic noise related research on solid Earth structure, ocean and atmosphere dynamics.