Full wavefield decomposition of high frequency secondary microseisms reveals distinct arrival azimuths for Rayleigh and Love waves

In the secondary microseism band (0.1-1.0Hz) the theoretical excitation of Rayleigh waves (R-g/LR), through oceanic wave-wave interaction, is well understood. For Love waves (LQ), the excitation mechanism in the secondary microseism band is less clear. We explore high-frequency secondary microseism excitation between 0.35 and 1Hz by analyzing a full year (2013) of records from a three-component seismic array in Pilbara (PSAR), Australia. Our recently developed three-component waveform decomposition algorithm (CLEAN-3C) fully decomposes the beam power in slowness space into multiple point sources. This method allows for a directionally dependent power estimation for all separable wave phases. In this contribution, we compare quantitatively microseismic energy recorded on vertical and transverse components. We find the mean power representation of Rayleigh and Love waves to have differing azimuthal distributions, which are likely a result of their respective generation mechanisms. Rayleigh waves show correlation with convex coastlines, while Love waves correlate with seafloor sedimentary basins. The observations are compared to the WAVEWATCH III ocean model, implemented at the Institut Francais de Recherche pour l'Exploitation de la Mer (IFREMER), which describes the spatial and temporal characteristics of microseismic source excitation. We find Love wave energy to originate from raypaths coinciding with seafloor sedimentary basins where strong Rayleigh wave excitation is predicted by the ocean model. The total power of R-g waves is found to dominate at 0.35-0.6Hz, and the Rayleigh/Love wave power ratio strongly varies with direction and frequency. Plain Language Summary We focus on the continuous seismic background energy that has its origin in the oceans. Contrary to earthquakes, this seismic energy is generated by ocean wave processes as continuous earth vibrations. The origin and composition of the wavefield is well understood for acoustic and vertically polarized surface waves but less clear for transversely polarized surface waves. We make use of a spiral seismic array northwest of Australia and decompose the continuous seismic wavefield into its fundamental polarization and energy contributions. This allows us to study which geographical regions generate a certain type of seismic wave, and measure the energy and temporal variation of the signal. We find that contrary to previous beliefs, vertically and transversely polarized seismic waves are not observed from the same directions with the seismic array but show distinct arrival directions. We are able to correlate the transversely polarized seismic waves to regions with sedimentary basins.