Location of Body Wave Microseism Sources Using Three-Component Data From a Large Aperture Seismic Array in China

Abstract:

From September 2013 through October 2014 up to 100 Guralp CMG-3 broadband seismometers were deployed in the WT-Array (WTA) in northwest China. The aperture of WTA is about 700 km, with an average station spacing of approximately 50 km. Here, we process continuous, three-component WTA data to detect and locate body wave microseism sources in four distinct period bands: 1.0–2.5 s, 2.5–5 s, 5–10 s, and 10–20 s. We back-project vertical component data through a 1D reference Earth model (AK135) to a global grid of hypothetical source locations, assuming P-wave (30^o–90^o), PP-wave (60^o–180^o), and S-wave (30^o–75^o) propagation. At the same time, we rotate the horizontals and back-project the radial and transverse components of the wavefield. For each frequency band, grid point, and assumed origin time, the array power is calculated from the amplitude of a windowed, filtered, and tapered time domain beam constructed with fourth-root stacking. We find strong P-wave and S-wave noise sources in the North Pacific and North Atlantic Oceans. Shorter period sources (2.5–5 s) are mainly observed in the North Pacific Ocean, while both short and long period (2.5–20 s) sources are observed in the North Atlantic Ocean. Median power plots for each month during September 2013 through October 2014 show distinct seasonal variations. The energy peaks in the North Atlantic are visible from November to March and strong energy is also observed in the North Pacific from October to April. We also observe PP-waves in the Southern Ocean, especially for May-August 2014. Using classical f-k analysis and plane-wave propagation, we are able to confirm the back-projection results. To improve our understanding of body wave microseism generation, we compare the observed P, S, and PP wave microseism locations with the predictions of significant wave height and wave-wave interactions derived from the WAVEWATCH III ocean model.From September 2013 through October 2014 up to 100 Guralp CMG-3 broadband seismometers were deployed in the WT-Array (WTA) in northwest China. The aperture of WTA is about 700 km, with an average station spacing of approximately 50 km. Here, we process continuous, three-component WTA data to detect and locate body wave microseism sources in four distinct period bands: 1.0–2.5 s, 2.5–5 s, 5–10 s, and 10–20 s. We back-project vertical component data through a 1D reference Earth model (AK135) to a global grid of hypothetical source locations, assuming P-wave (30^o–90^o), PP-wave (60^o–180^o), and S-wave (30^o–75^o) propagation. At the same time, we rotate the horizontals and back-project the radial and transverse components of the wavefield. For each frequency band, grid point, and assumed origin time, the array power is calculated from the amplitude of a windowed, filtered, and tapered time domain beam constructed with fourth-root stacking. We find strong P-wave and S-wave noise sources in the North Pacific and North Atlantic Oceans. Shorter period sources (2.5–5 s) are mainly observed in the North Pacific Ocean, while both short and long period (2.5–20 s) sources are observed in the North Atlantic Ocean. Median power plots for each month during September 2013 through October 2014 show distinct seasonal variations. The energy peaks in the North Atlantic are visible from November to March and strong energy is also observed in the North Pacific from October to April. We also observe PP-waves in the Southern Ocean, especially for May-August 2014. Using classical f-k analysis and plane-wave propagation, we are able to confirm the back-projection results. To improve our understanding of body wave microseism generation, we compare the observed P, S, and PP wave microseism locations with the predictions of significant wave height and wave-wave interactions derived from the WAVEWATCH III ocean model.