Seismic Noise Recorded by the Formosa Array

Summary

Seismic ambient noise records continuous ground motion from ocean waves, weather, and human activity, and is increasingly exploited for imaging and monitoring. We use the dense broadband Formosa Array in northern Taiwan to characterize the sources, directionality, and slowness distribution of ambient noise. We combine power spectral density–probability density functions (PSD–PDFs) with time-domain delay-and-sum beamforming of 10-minute vertical-component windows in the 0.1–0.4 Hz band. Array- and station-level PSD–PDFs show that secondary microseisms dominate 0.2–1 Hz. At longer periods (~0.02–0.06 Hz), only weak low-frequency peaks are visible, and no well-developed primary-microseism peak is observed. Above ~1 Hz, noise levels rise and become highly variable, with enhanced 2–6 Hz energy at urban and agriculturally intensive sites and lower levels in mountainous and volcanic regions, consistent with cultural and land-use sources. Beamforming distinguishes coherent arrivals from local earthquakes, teleseismic earthquakes, typhoon-related microseisms, and background noise in slowness–azimuth space. During Typhoon Mitag, beam azimuth and slowness show broad temporal changes consistent with storm-related coastal microseism excitation as the typhoon approaches and recedes. Ten-minute mean beam power in the 0.1–0.4 Hz band correlates strongly with significant wave height recorded at nearby buoys, with peak Pearson coefficients of ~0.8 when buoy wave height lags beam power by 6–7 hours, suggesting that effective microseism sources develop offshore of northern Taiwan along the incident wave direction. Beam-energy partitioning shows that recorded energy is broadly distributed across slowness annuli and azimuth sectors, rather than being dominated by a single persistent beam direction. This result reflects broad array-scale illumination at the Formosa Array after source excitation, propagation, scattering, and time averaging. These results demonstrate that dense onshore arrays adjacent to storm-exposed coastlines can resolve cultural, oceanic, and seismic contributions to ambient noise. They can also provide a land-based complement to oceanographic observations by identifying storm-related changes in coherent microseism energy associated with wave conditions, while supporting Green’s-function retrieval and other ambient-noise imaging applications.