HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text Full Text (PDF) Citation Map Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Tibuleac, I. M. Articles by Britton, J. M. Search for Related Content GeoRef GeoRef Citation

An Automated Short-Period Surface-Wave Detection Algorithm

¹ Weston Geophysical Corporation
181 Bedford Street, Suite 1
Lexington, Massachusetts 02420
ileana{at}westongeophysical.com
brittonj{at}westongeophysical.com

View larger version (29K): [in a new window]   Figure 1. The Meyer wavelet (upper plot) at scale 14 (center period, 1 sec) and its frequency-domain representation (middle plot) normalized to the maximum value. The frequency response of a zero-phase (three poles, two passes) Butterworth filter between 0.4 and 1.3 Hz (dotted line) and between 0.55 and 2 Hz (continuous line) is presented in the bottom plot. The sample rate is 20 samples/sec.

View larger version (66K): [in a new window]   Figure 2. Location of station HYB (white triangle), the event clusters (gray squares), and the GT15 events (white squares) used in this study. The figure inset shows a satellite image of the open pits of the Ramagundan mine (inside the circles) and the estimated locations for cluster 1 (18.803° N, 79.570° E), cluster 2 (17.57° N, 80.33° E), cluster 3 (18.623° N, 79.635° E), and cluster 4 (18.675° N, 79.600° E), represented as gray squares. Median values for latitude and longitude are calculated for each cluster. The HYB single-station locations of the recorded events are represented by white dots. Natural seismicity listed by the U.S. Geological Survey in the region between 1980 and 2000 is represented as grey dots. Known mines in the area are shown as stars. Single-station locations of 114 events with Rg phases, detected at HYB over a 13-day period, are added as black dots.

View larger version (27K): [in a new window]   Figure 3. Backazimuth residuals estimated by using Fourier prefiltering (dotted line with circles) and wavelet prefiltering (continuous line with triangles) with a Meyer wavelet centered on an Rg period of 1 sec. The residuals are relative to each cluster backazimuth. The cluster backazimuth is determined by using median values of event-location parameters for each cluster. Events 1 and 2 are from cluster 1, events 3 to 5 are from cluster 2, events 6 to 9 are from cluster 3, and events 10 to 23 are from cluster 4.

View larger version (34K): [in a new window]   Figure 4. (top) Wavelet (top plots) and Fourier (bottom plots) prefiltered data from the 19 December 1999 10:12:34 event from cluster 4, designated 21. For each prefiltering method, the upper plot represents the filtered waveforms (E, N, Z from top to bottom). The abscissa represents the right side of the STA and LTA windows. The detection-function value (larger than 10% of the detection-function maximum) is represented below as a function of time and backazimuth. Also, Rg is detected after wavelet prefiltering (top plots), but Rg is not the choice of the detector after Fourier prefiltering (bottom plots). The empirical detection threshold value was 670 at HYB for Fourier prefiltering and 2500 for wavelet prefiltering. Note larger waveform amplitudes after wavelet prefiltering, which lead to larger values of the empirical detection threshold when using wavelet prefiltering compared with Fourier prefiltering.

View larger version (16K): [in a new window]   Figure 5. (Left) Comparison of P and Rg estimated backazimuth for data recorded at HYB over a 13-day period. (Right) Rg and P arrival-time difference as a function of distance. Slopes of the L1 norm regression lines were estimated by using 2.9 km/sec Rg group velocity, 6.2 km/sec Pg velocity, and 8 km/sec Pn velocity.

View larger version (28K): [in a new window]   Figure 6. Uneven distribution of the "Rg noise" shown as a histogram of number of "detections" in 10-min backazimuth intervals as a function of Rg-type wave backazimuth.