We evaluate the output from the time-domain adaptive beamformer (ABF) developed by Shen (1979) coupled to both incoherent (short-term average/long-term average [STA/LTA]) and coherent F -statistic (FSTAT) detectors. Detection test results on regional phases are compared to the baseline conventional beamformer output coupled to both STA/LTA and FSTAT detectors. Examples and test results are presented that demonstrate reduced detection thresholds, improved clutter rejection, and lower false-alarm rates with a time-domain ABF coupled to either STA/LTA or FSTAT detectors with significant reductions in the false-alarm rate observed for the ABF coupled to the FSTAT detector.

Evaluation of a Time-Domain Adaptive Beamformer for Regional Phase Detection

: This report presents time-domain maximum likelihood adaptive beamforming results for data from the Alaska Long-Period Array and the Korean Short-Period Array. Adaptive-beamforming detection gain relative to beamsteering is investigated for unmixed long-period seismic events in the presence of background noise and for simulated mixed events where two data samples, each containing a signal, are summed to create a composite sample containing an interfering event. In the adaptive beamforming studies for background noise, the performance of two different adaptive algorithms, the effect of using closely-spaced partial arrays instead of the full ALPA array, and the effect of different frequency filters applied to the input channels before array beamforming are examined. In the interfering-event analysis, the relative on-azimuth to off-azimuth event strength at which an on-azimuth event detection is possible determines the detection gain of adaptive multichannel filtering over time-shift-and-sum beamforming.

Thomas E Barnard

Papers

An evaluation of adaptive-beamforming techniques applied to recorded seismic data