Multiple-Array Detection, Association and Location of Infrasound and Seismo-Acoustic Event-Utilization of Ground Truth Information (Postprint)
Abstract
Detectors were assessed that distinguish infrasound signals from both correlated and uncorrelated noise. Approaches to this problem are implementation of the F-detector, which employs the F-statistic, and cross-correlation techniques separating correlated and uncorrelated signals. We tested a modified F-detector that applies an adaptive procedure to identify variations in correlated noise, thus reducing false alarms. Using the modified F-detector, the temporal variation in the adaptive nature of the detector for a number of sites was investigated. The sensitivity of the detection procedure to the adaptive window used to estimate the C parameter that characterizes the correlated noise and impacts the false alarm rate was tested. The C values are found to be stable over long time periods for inland arrays within but show variations under high wind velocity conditions for adaptive windows as short as 1 hour for arrays on islands or near the coast. In order to further understand variations in background noise, the 10th, 50th, and 90th percentile noise spectral densities were estimated. The noise estimates show a strong wind speed effect on acoustic noise; for example, the spread between the 10th and 90th percentiles is about 40 dB at 0.1 Hz.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 07, 2012
- Accession Number
- ADA563111
Entities
People
- Brian W. Stump
- Christopher T. Hayward
- Il-young Che
- Junghyun Park
- Stephen J. Arrowsmith
Organizations
- Southern Methodist University