Detecting Seismic Events Using Ensembles of Features
Abstract
Currently, many traditional methods are used to detect arrivals in three-component seismic waveform data collected at regional distances. Accurately establishing the identity and arrival of these waves in adverse signal-to-noise environments is helpful in detecting and locating the seismic events. STA/LTA, polarization analysis and Akaike Information Criterion (AIC)
are just a few of the various methods that may be used, each with their own performance benefits and drawbacks. In this work, we move to ensembles composed of outputs from these fundamental methods. It is assumed that combining key output quantities from these methods can significantly improve performance by leveraging the unique information contained in each of them. We compare the performance of the Hidden Markov Model method of ensembeling quantities with other methods, including neural network methods. We describe in detail results of these methods tuned on data from expert defined arrival picks. The dataset used is from the Dynamic Network Experiment 2018 (DNE18), and described elsewhere. Sandia National Laboratories is a multimission laboratory managed and operated by National Technology and Engineering Solutions of Sandia, LLC., a wholly owned subsidiary of Honeywell International, Inc., for the U.S. Department of Energy's National Nuclear Security Administration under contract DE-NA0003525.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFM.S53E0464F
- Keywords:
-
- 3355 Regional modeling;
- ATMOSPHERIC PROCESSESDE: 1040 Radiogenic isotope geochemistry;
- GEOCHEMISTRYDE: 6620 Science policy;
- PUBLIC ISSUESDE: 7219 Seismic monitoring and test-ban treaty verification;
- SEISMOLOGY