S31C-01
Improved Detection and Location of Ocean Microseism Signals using Array Techniques

Wednesday, 16 December 2015: 08:00
307 (Moscone South)
Anya Marie Reading1, Martin Gal1, Keith D Koper2 and Hrvoje Tkalcic3, (1)University of Tasmania, Hobart, TAS, Australia, (2)University of Utah, Salt Lake City, UT, United States, (3)Australian National University, Canberra, ACT, Australia
Abstract:
We present and evaluate a range of approaches that may be used to investigate ocean microseisms using seismic array data. At amplitudes below the dominant incoming signal, the ambient seismic energy (background noise) associated with microseisms arrives from multiple directions at any one time. Thus we address the challenge of detecting weaker signals from unpredictable directions in the presence of other strong signals. Our aim is to extract the most accurate information possible from such weaker signals in order to expand the capability of ocean storm studies, using seismology, including the ability to extract storm patterns from archive seismic array records.

Detection of weaker microseism signals may be improved using algorithms widely used in astronomy. One example is the CLEAN algorithm which has wide usage in radio astronomy. This algorithm operates by finding the position and strength of point sources and iteratively deconvolving their contribution to the image. It may be combined to optimum effect with the previously published (Incoherently Averaged Signal) IAS Capon implementation for an accurate detection of weaker sources. Having detected weaker sources, they may be backprojected using a suitable Earth model, taking into account a correction for the mislocation due to slowness-azimuth station corrections. The microseism generation locations inferred in this manner are strongly frequency dependent, even within relatively restricted frequency ranges (0.325-0.725 Hz) for some arrays.

Our advances in seismic array processing, with a focus on methods appropriate to weaker ambient noise signals, have led to insights, for example, regarding the generation of seismic noise. We find that secondary microseisms in the lower frequency band are generated mainly by ocean swell whereas higher frequency bands are generated by local wind conditions. These arrivals are investigated over a two-decade time frame for the Southern Ocean and west Pacific Ocean.