Single Channel Seismic Event Detection on Alaskan Short-Period Data
Abstract
Three advanced detectors and a standard detector were tested on 15 of the weakest 40 teleseismic signals associated to events and received at an array in Alaska in the time interval 1-6 November 1980. The signals were originally detected on array beams and the detectors were given the task of detecting the signals on data from a single instrument. The false alarm rate was set at 3 to 4 per hour by comapring single instrument detections to beam detections. The optimum Z-log detector detected all 15; the Walsh detector, 13; and the MARS detector, 11. This relative ranking is in agreement with an earlier study using signals buried in phase-scrambled noise. The standard detector detected 14 out of the 15, whereas it was worse than all the other three in the early study. This may be because the standard detector takes the absolute value of the filtered data instead of squaring it and thus is more resistant to the effects of spikes and weak precursor signals. More likely the standard detector was at a disadvantage in the earlier study because the signal time window was held at 2.5 seconds while the signal windows of the other detectors were much larger as appropriate to the larger signals which were buried in noise.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 11, 1981
- Accession Number
- ADA116413
Entities
People
- Robert R. Blandford