Cyclone Signatures in the South-West Indian Ocean from Two Decades of Microseismic Noise

Tropical Cyclones (TC) represent the most destructive natural disaster affecting the islands in the South-West Indian Ocean (SWIO) each year. Monitoring ocean activity is therefore of primary importance to secure lands, infrastructures and peoples, but the little number of oceanographic instruments makes it challenging, particularly in real time. Long-term seismological records provide a way to decipher and quantify the past cyclonic activity by analyzing microseisms, seismic waves generated by the ocean activity and propagating through the solid Earth. In the present study, we analyze this microseismic noise generated by cyclones that develop in the SWIO basin between 1999 and 2020, using broadband seismic stations in La Reunion. The power spectral density (PSD), together with the root mean square (RMS) analyses of continuous seismic data recorded by the permanent Geoscope RER seismic station, indicate the intensification of the microseismic noise amplitude in proportion to the cyclone intensity. Thus, we establish a relationship between the cyclone intensity and the PSD of the Secondary Microseisms (SM) in frequency band similar to 0.14 to 0.25 Hz (4 to 7 s period). The Pearson coefficient between the observed and estimated TC intensity are >0.8 in the presence of a cyclone with mean wind speeds >75 km/h and with a seismic station distance-to-storm center D < 3000 km. A polarization analysis in the time and frequency domains allows the retrieval of the backazimuth of the SM sources during isolated cyclone events and well-polarized signal, i.e., CpH > 0.6. We also analyzed the RMS of the Primary Microseisms (PM frequency between similar to 0.05 and 0.1 Hz, i.e., for 10 to 20 s period) for cyclones passing nearby La Reunion (D < 500 km), using the available temporary and permanent broadband seismic stations. We also found high correlation coefficients (>0.8) between the PM amplitude and the local wave height issued from the global hindcast model demonstrating that the PM amplitude can be used as a robust proxy to perform a real-time wave-height monitoring in the neighboring ocean. Transfer functions are calculated for several cyclones to infer wave height from the seismic noise amplitude recorded on land. From the analysis of two decades of data, our results suggest that it is possible to quantify the past ocean activity for as long as continuous seismic archives are available, emphasizing microseismic noise as a key observable for quantifying and understanding the climate change.