We use a waveform matched filter technique to detect microseismicity recorded by a 5200-station array in the Long Beach area around the Signal Hill oil reservoir. Because the computation cost increases linearly with number of stations, it is challenging to apply the technique to this large data set. Here we conduct a series of synthetic tests to explore how to best perform waveform detection with a large array. First we show that a large array can successfully detect very small signals, even they are totally buried in the background noise with very low signal to noise ratios. However, when only part of the array records the event, direct stacking of correlation-coefficients (CCs) for all stations results in lower mean CCs. Finally, given the temporal sparsity of events we can decrease the data volume of cross-correlation by a large amount without sacrificing detection reliability. The proposed strategy can efficiently cut down the computational cost and can be used to detect seismic events in similar types of dataset.
|Number of pages||5|
|Journal||SEG Technical Program Expanded Abstracts|
|State||Published - 2015|
|Event||SEG New Orleans Annual Meeting, SEG 2015 - New Orleans, United States|
Duration: 18 Oct 2011 → 23 Oct 2011