LLNL Calibration Program: Data Collection, Ground Truth Validation, and Regional Coda Magnitude
Lawrence Livermore National Laboratory (LLNL) integrates and collects data for use in calibration of seismic detection, location, and identification. Calibration data is collected by (1) numerous seismic field efforts, many conducted under NNSA (ROA) and DTRA (PRDA) contracts, and (2) permanent seismic stations that are operated by national and international organizations. Local-network operators and international organizations (e.g. International Seismic Center) provide location and other source characterization (collectively referred to as source parameters) to LLNL, or LLNL determines these parameters from raw data. For each seismic event, LLNL rigorously characterizes the uncertainty of source parameters. This validation process is used to identify events whose source parameters are accurate enough for use in calibration. LLNL has developed criteria for determining the accuracy of seismic locations and methods to characterize the covariance of calibration datasets. Although the most desirable calibration events are chemical and nuclear explosions with highly accurate locations and origin times, catalogues of naturally occurring earthquakes offer needed geographic coverage that is not provided by man made sources. The issue in using seismically determined locations for calibration is validating the location accuracy. Sweeney (1998) presented a 50/90 teleseismic, network-coverage criterion (50 defining phases and 90{sup o} maximum azimuthal gap) that generally results in 15-km maximum epicenter error. We have also conducted tests of recently proposed local/regional criteria and found that 10-km accuracy can be achieved by applying a 20/90 criteria. We continue to conduct tests that may validate less stringent criteria (which will produce more calibration events) while maintaining desirable location accuracy. Lastly, we examine methods of characterizing the covariance structure of calibration datasets. Each dataset is likely to be effected by distinct error processes that result in a distinct covariance structure. We present covariance models for select data sets and demonstrate how these datasets can be integrated into one calibration-event catalog. LLNL has developed a robust magnitude calibration methodology for sparsely distributed regional stations using narrow band coda envelopes. This technique provides stable magnitudes for small events that makes detection and identification calibration possible at low magnitudes. This approach has most recently been applied to IMS stations located in Israel, Jordan and Egypt for events that span local and near regional distances. Our preliminary results show that a magnitude estimate from one station using the coda is equivalent to a network average of roughly 9 stations when using traditional magnitudes (e.g., m{sub b}(P), M{sub L}, M{sub d}). The stability of the coda comes from measuring a long length of coda using a calibrated synthetic envelope as an empirical metric. We relate the non-dimensional coda amplitudes to an absolute scale by tying them to independent moment estimates from larger waveform-modeled events. Unlike most narrow band magnitudes, this approach yields an azimuthally averaged, moment-rate spectrum that is completely corrected for path and site effects. The resultant magnitudes from the spectra (e.g., M{sub w} and m{sub b}) are fully transportable and do not suffer from regional bias.
- Research Organization:
- Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
- Sponsoring Organization:
- US Department of Energy (US)
- DOE Contract Number:
- W-7405-ENG-48
- OSTI ID:
- 15001980
- Report Number(s):
- UCRL-JC-144307; TRN: US200406%%392
- Resource Relation:
- Conference: 23rd Department of Defense/Energy Seismic Research Review, Jackson Hole, WY (US), 10/02/2001--10/05/2001; Other Information: PBD: 28 Aug 2001
- Country of Publication:
- United States
- Language:
- English
Similar Records
SEISMIC SOURCE SCALING AND DISCRIMINATION IN DIVERSE TECTONIC ENVIRONMENTS
LLNL Location and Detection Research