skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Integrated Seismic Event Detection and Location by Advanced Array Processing

Abstract

The principal objective of this two-year study is to develop and test a new advanced, automatic approach to seismic detection/location using array processing. We address a strategy to obtain significantly improved precision in the location of low-magnitude events compared with current fully-automatic approaches, combined with a low false alarm rate. We have developed and evaluated a prototype automatic system which uses as a basis regional array processing with fixed, carefully calibrated, site-specific parameters in conjuction with improved automatic phase onset time estimation. We have in parallel developed tools for Matched Field Processing for optimized detection and source-region identification of seismic signals. This narrow-band procedure aims to mitigate some of the causes of difficulty encountered using the standard array processing system, specifically complicated source-time histories of seismic events and shortcomings in the plane-wave approximation for seismic phase arrivals at regional arrays.

Authors:
; ; ;
Publication Date:
Research Org.:
Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
Sponsoring Org.:
USDOE
OSTI Identifier:
902233
Report Number(s):
UCRL-SR-228092
TRN: US200717%%498
DOE Contract Number:
W-7405-ENG-48
Resource Type:
Technical Report
Country of Publication:
United States
Language:
English
Subject:
58 GEOSCIENCES; ACCURACY; APPROXIMATIONS; DETECTION; PROCESSING; SEISMIC EVENTS

Citation Formats

Kvaerna, T, Gibbons, S J, Ringdal, F, and Harris, D B. Integrated Seismic Event Detection and Location by Advanced Array Processing. United States: N. p., 2007. Web. doi:10.2172/902233.
Kvaerna, T, Gibbons, S J, Ringdal, F, & Harris, D B. Integrated Seismic Event Detection and Location by Advanced Array Processing. United States. doi:10.2172/902233.
Kvaerna, T, Gibbons, S J, Ringdal, F, and Harris, D B. Fri . "Integrated Seismic Event Detection and Location by Advanced Array Processing". United States. doi:10.2172/902233. https://www.osti.gov/servlets/purl/902233.
@article{osti_902233,
title = {Integrated Seismic Event Detection and Location by Advanced Array Processing},
author = {Kvaerna, T and Gibbons, S J and Ringdal, F and Harris, D B},
abstractNote = {The principal objective of this two-year study is to develop and test a new advanced, automatic approach to seismic detection/location using array processing. We address a strategy to obtain significantly improved precision in the location of low-magnitude events compared with current fully-automatic approaches, combined with a low false alarm rate. We have developed and evaluated a prototype automatic system which uses as a basis regional array processing with fixed, carefully calibrated, site-specific parameters in conjuction with improved automatic phase onset time estimation. We have in parallel developed tools for Matched Field Processing for optimized detection and source-region identification of seismic signals. This narrow-band procedure aims to mitigate some of the causes of difficulty encountered using the standard array processing system, specifically complicated source-time histories of seismic events and shortcomings in the plane-wave approximation for seismic phase arrivals at regional arrays.},
doi = {10.2172/902233},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Fri Feb 09 00:00:00 EST 2007},
month = {Fri Feb 09 00:00:00 EST 2007}
}

Technical Report:

Save / Share:
  • In the field of nuclear explosion monitoring, it has become a priority to detect, locate, and identify seismic events down to increasingly small magnitudes. The consideration of smaller seismic events has implications for a reliable monitoring regime. Firstly, the number of events to be considered increases greatly; an exponential increase in naturally occurring seismicity is compounded by large numbers of seismic signals generated by human activity. Secondly, the signals from smaller events become more difficult to detect above the background noise and estimates of parameters required for locating the events may be subject to greater errors. Thirdly, events are likelymore » to be observed by a far smaller number of seismic stations, and the reliability of event detection and location using a very limited set of observations needs to be quantified. For many key seismic stations, detection lists may be dominated by signals from routine industrial explosions which should be ascribed, automatically and with a high level of confidence, to known sources. This means that expensive analyst time is not spent locating routine events from repeating seismic sources and that events from unknown sources, which could be of concern in an explosion monitoring context, are more easily identified and can be examined with due care. We have obtained extensive lists of confirmed seismic events from mining and other artificial sources which have provided an excellent opportunity to assess the quality of existing fully-automatic event bulletins and to guide the development of new techniques for online seismic processing. Comparing the times and locations of confirmed events from sources in Fennoscandia and NW Russia with the corresponding time and location estimates reported in existing automatic bulletins has revealed substantial mislocation errors which preclude a confident association of detected signals with known industrial sources. The causes of the errors are well understood and are primarily the result of spurious identification and incorrect association of phases, and of excessive variability in estimates for the velocity and direction of incoming seismic phases. The mitigation of these causes has led to the development of two complimentary techniques for classifying seismic sources by testing detected signals under mutually exclusive event hypotheses. Both of these techniques require appropriate calibration data from the region to be monitored, and are therefore ideally suited to mining areas or other sites with recurring seismicity. The first such technique is a classification and location algorithm where a template is designed for each site being monitored which defines which phases should be observed, and at which times, for all available regional array stations. For each phase, the variability of measurements (primarily the azimuth and apparent velocity) from previous events is examined and it is determined which processing parameters (array configuration, data window length, frequency band) provide the most stable results. This allows us to define optimal diagnostic tests for subsequent occurrences of the phase in question. The calibration of templates for this project revealed significant results with major implications for seismic processing in both automatic and analyst reviewed contexts: • one or more fixed frequency bands should be chosen for each phase tested for. • the frequency band providing the most stable parameter estimates varies from site to site and a frequency band which provides optimal measurements for one site may give substantially worse measurements for a nearby site. • slowness corrections applied depend strongly on the frequency band chosen. • the frequency band providing the most stable estimates is often neither the band providing the greatest SNR nor the band providing the best array gain. For this reason, the automatic template location estimates provided here are frequently far better than those obtained by analysts. The second technique is that of matched field processing whereby spatial covariance matrices calculated from large numbers of confirmed events from a single site can be used to generate calibrated narrow-band steering vectors which can replace the theoretical plane-wave steering vectors of traditional f-k analysis. This provides a kind of fingerprint which is specific to a given source region and is effective to higher frequencies than traditional beamforming since deviations from the theoretical planewave model are compensated for in the calibrations. The narrow-band nature of the technique makes the source identification most sensitive to the spatial nature of the recorded wavefield and less sensitive to the temporal nature. This may make the method far more suitable for events with very complicated seismic sources than full waveform-correlation methods.« less
  • When seismic events occur in spatially compact clusters, the volume and geometric characteristics of these clusters often provides information about the relative effectiveness of different location methods, or about physical processes occurring within the hypocentral region. This report defines and explains how to determine the convex polyhedron of minimum volume (CPMV) surrounding a set of points. We evaluate both single-event and joint hypocenter determination (JHD) relocations for three rather different clusters of seismic events; (1) nuclear explosions from Mururoa relocated using P and PKP phases reported by the ISC; (2) intermediate depth earthquakes near Bucaramanga, Colombia, relocated using P andmore » PKP phases reported by the ISC; and (3) shallow earthquakes near Vanuatu (formerly, the New Hebrides), relocated using P and S phases from a local station network. This analysis demonstrates that different location methods markedly affect the volume of the CPMV, however, volumes for JHD relations are not always smaller than volumes for single-event relocations.... Seismic event location, Seismic strain, Seismic discrimination.« less
  • Several location methods developed at SDAC are evaluated by comparing location errors of these methods against errors computed with the standard method. The new methods include: Using P travel times in laterally heterogeneous media: A station travel time residual correlation matrix in the normal equations; Location with simultaneous determination of Pn, Pg, and Lg velocities; Combination of the correlation matrix method and the simultaneous determination method. All the methods evaluated in this report appear to give smaller absolute epicenter and depth errors than the standard method of locating events. Furthermore, these methods do not use station corrections; therefore the locationsmore » obtained are not station- or region-dependent. They can be used to locate seismic events in any source region, with any set of stations. When average location errors are compared against errors with the standard method, the correlation matrix is about 3 km better, e.g., about 17 kilometers as compared to 20 kilometers; the regional models method is 3 to 6 km better, e.g., about 5 kilometers as compared to 9 kilometers, depending on whether near-regional stations are used. The average location errors from the true epicenters using regional data are approximately 6 km. The addition of Lg arrivals to locate events did not result in better locations. Typically, we added 6 Lg arrival times to a total of 15 Pn and Pg arrival times. The lack of improvement may be attributed to a greater variance in Lg travel times.« less
  • Three-component data from regional seismic events recorded by the former NRDC-Soviet Academy of Sciences regional seismic network in Kazakhstan, USSR, have been analyzed with the primary goal of improving regional seismic event location capability. Data from these events were used in the investigation of the following problems related to regional event location: (1) determination of wave arrival azimuth; (2) observability and value of secondary phase arrivals; (3) evaluation and improvement of regional event location algorithms; (4) independent determination of master event locations. Starting with very little prior information, the authors have demonstrated the potential for a sparse seismic network ofmore » three 3-component stations to locate events over a wide region with reasonable accuracy and precision, both for epicenter and depth. Their basic findings are: that arrival azimuth can be determined with reasonable precision, but the data provide little in the way of location constraint in most cases; secondary phase arrivals are routinely observable over a wide distance range, and they provide important location constraints; existing location algorithms perform well and provide appropriate estimates of location uncertainty, but must be modified for far-regional applications. Numerous master events have been identified and used to improve their location capability.« less