Integration of Microseism and Coda Correlations as a New Dataset for Deep Earth Seismic Imaging – TerraCorr

Two methodological approaches were explored on a global scale: microseism-based and coda-based correlations. On the one hand, the correlation of microseism signals with the aid of excitation models will be used to constrain the structure of the velocity of the entire lower mantle. On the other hand, the correlation of the reverberated coda of large earthquakes will be used to question the existence of a stratified layer at the top of the outer core.

The project consists of 3 workpackages: (1) correlation of microseimic events for imaging the lower mantle ; (2) correlation of reverberated Coda for imaging the last kilometers of the outermost outer core ; (3) methodological developments

We have obtained several important results concerning workpackage (1):

- A catalog of significant microseism events was built from an ocean model and seismological data. The catalog gathers thousands of sources, classified by effective strength, spatial extent, and location, over a period from 1994 to 2020.

- A processing workflow has been built and validated by observations to allow the measurement of particular interferences targeting the lower mantle and the core. Thus, for a single source in the North Atlantic Ocean spanning only over a few hours, it is possible to observe at great distance interferences of type PP-P, PKPPKP-PKP or PKPPcP-PKP.

For workpackage (2), the three elements required for the planned analysis are operational:

- The code for downloading and preprocessing of seismological data and the adaptations needed to calculate correlations for this particular case of earthquake coda interferometry is validated.

- A numerical tool for calculating synthetic seismograms on a global scale has been selected (Axisem) and adapted so that it can be combined with interferometric calculations. Comparisons with other approaches to constructing synthetic seismograms, particularly by modal summation, are currently underway.

- A ray-tracing tool, adapted to non-causal phases resulting from reverberation interferences, and allowing the interpretation of the observed interferences, is under validation.

Finally, several important advances have been made on workpackage (3), in relation with (1) and (2).

- A small-scale (crustal) opening work has allowed us to better understand how to evaluate the sensitivity of interferometric measurements in the context of a direct interference between two body waves (e.g. P and PP). Numerical tools to compute such sensitivity kernels at a large scale are under development.

- The code used for data pre-processing, calculation, and management of correlations has largely evolved. In particular, it contains an appropriate selection of station pairs according to the parameters of the source and the target phase.

- We have developed WMSAN (for “Wave Model Sources of Ambient Noise”), a Python package for modeling seismograms and their correlations from WAVEWATCHIII hindcasts, i.e., global sea state models. This is an important tool for continuing our work on ocean-generated seismic noise and helps bridge the gap between physical oceanography and seismology.

The main result of this project is the observation-based validation of the possibility of observing seismic phase interferences at long distances from a single oceanic event. For example, a storm lasting several hours in the North Atlantic allowed PcP waves (reflection on the outer core) to be observed between Australia and Antarctica by correlating recordings from pairs of carefully selected station, as well as P waves sensitive to the lower mantle were observed between different combinations of seismological stations. The latest work currently being finalized, as well as short- and medium-term prospects, generalize these observations to other regions/sources and suggest a possible and useful application for these new observables to complement conventional data for deep Earth imaging. Several avenues are being explored for the use of interferometric methods applied to the coda of large earthquakes. Results independent of the project suggest that this type of measurement is not entirely compatible with traditional arrival time measurements, at least for the outer core. This has helped to redirect our objectives towards a more fundamental understanding of the sensitivity of our interferometric measurements, for example with regard to effects related to the Earth's rotation. More generally, we believe that this type of data is largely underutilized for imaging our planet, which is driving us in the short and medium term to better understand the mechanisms underlying the reconstruction of the seismic phases we observe in this type of correlograms.

- Zhang, R., Boué, P., Campillo, M., Ma, J.. Quantifying P-wave secondary microseism events: a comparison of observed and modeled back projection. Geophysical Journal International, submitted

- Boué, P., Tomasetto, L.. Opportune Detections of Global P- Wave Propagation from Microseism Interferometry. Comptes Rendus Géosceinces, submitted

- P Boue, L Tomasetto, Detection of Teleseismic P-waves in the Correlation Wavefield from Significant Secondary Microseism
Events. AGU Fall Meeting Abstracts 2021, S35A-10

The main objective of the Project is to explore the possibility to use seismic wave?elds emerging from the correlation of long recordings (a.k.a. seismic interferometry) to probe the Earth’s deep interior. In practice, we want to build and mine a new robust dataset of seismic signals that are complementary to earthquake data, for improved deep-mantle and core imaging. The Project will focus on the core-mantle boundary (CMB) region. Two methodological approaches will be explored on a global scale: microseism-based and coda-based correlations. On the one hand, the correlation of microseism signals with the aid of excitation models will be used to constrain the structure of the velocity of the entire lower mantle. On the other hand, the correlation of the reverberated coda of large earthquakes will be used to question the existence of a stratified layer at the top of the outer core. The geodynamics of these two targets are still an on-going debate, mainly due to incomplete direct observables. TerraCorr will pave the way towards a much broader use of continuous seismic signals to improve our understanding of the Earth’s history and its present dynamics.