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

Title: LLNL's 3-D A Priori Model Constraints and Uncertainties for Improving Seismic Location

Abstract

Accurate seismic event location is key to monitoring the Comprehensive Nuclear-Test-Ban Treaty (CTBT) and is largely dependent on our understanding of the crust and mantle velocity structure. This is particularly challenging in aseismic regions, devoid of calibration data, which leads us to rely on a priori constraints on the velocities. We investigate our ability to improve seismic event location in the Middle East, North Africa, and the Former Soviet Union (ME/NA/FSU) by using a priori three-dimensional (3-D) velocity models in lieu of more commonly used one dimensional (1-D) models. Event locations based on 1-D models are often biased, as they do not account for significant travel-time variations that result from heterogeneous crust and mantle; it follows that 3-D velocity models have the potential to reduce this bias. Here, we develop a composite 3-D model for the ME/NA/FSU regions. This fully 3-D model is an amalgamation of studies ranging from seismic reflection to geophysical analogy. Our a priori model specifies geographic boundaries and velocity structures based on geology, tectonics, and seismicity and information taken from published literature, namely a global sediment thickness map of 1{sup o} resolution (Laske and Masters, 1997), a regionalized crustal model based on geology and tectonics (Sweeneymore » and Walter, 1998; Bhattacharyya et al., 2000; Walter et al., 2000), and regionalized upper mantle (RUM) models developed from teleseismic travel times (Gudmundsson and Sambridge, 1998). The components of this model were chosen for the complementary structures they provide. The 1{sup o} sediment map and regionalized crustal model provide detailed structures and boundaries not available in the more coarse 5{sup o} models used for global-scale studies. The RUM models offer improved resolution over global tomography, most notably above depths of 300 km where heterogeneity is greatest; however, we plan to test other published upper mantle models of both P- and S-wave velocity. We compute travel times through this integrated model for comparison with other standard 1-D models, as our goal is to evaluate whether the 3-D model can better predict the observed travel times. The arrival times are computed through the model using a 3-D finite-difference technique and are then compared with a declustered set of ISC P arrival times (Engdahl et al., 1998). Our ME/NA/FSU model predicts the P and Pn travel times very well, as measured by variance reduction, for three stations we tested: ARU, KVT, and GAR; these predicted times also resemble some patterns seen in Pn tomography models of this region. Such tests will allow us to identify parts of the model that may need modification. We also compute model-based correction surfaces for each station in the ME/NA/FSU regions that can be used as additional constraints in our event location algorithm to determine the improvement provided by using 3-D models. We use this method to relocate a set of ground truth events: the 1991 Racha aftershock sequence which was investigated by Myers and Schultz (2000) using empirical kriged correction surfaces and a 1-D velocity model. They find an epicenter mislocation bias of 42 km when no corrections are applied and that this mislocation is reduced to 13 km when their empirically derived correction surfaces are included. We relocate this same set of events using our model-based correction surfaces and produce a mislocation bias of only 26 km, a significant improvement. We are currently implementing methods to quantify uncertainties on the model-based corrections which will be required to compute representative error ellipses for the new locations. We also plan to combine both the model-based and empirical correction techniques to achieve the best improvement in location. This test case demonstrates the power of using 3-D velocity models to improve location capability for small, regionally recorded events. This example also shows how the model-based approach holds great potential for improving locations in aseismic regions where it may not be possible to compute empirical correction surfaces.« less

Authors:
; ; ; ;
Publication Date:
Research Org.:
Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
Sponsoring Org.:
USDOE
OSTI Identifier:
15013435
Report Number(s):
UCRL-JC-138984
TRN: US200601%%398
DOE Contract Number:  
W-7405-ENG-48
Resource Type:
Conference
Resource Relation:
Conference: 22nd Annual Department of Defense/Department of Energy Seismic Research Symposium, New Orleans, LA, Sep 12 - Sep 15, 2000
Country of Publication:
United States
Language:
English
Subject:
58 GEOSCIENCES; 98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL PROTECTION; AFTERSHOCKS; ALGORITHMS; CALIBRATION; CTBT; EPICENTERS; GEOLOGY; GROUND TRUTH MEASUREMENTS; MONITORING; REFLECTION; RESOLUTION; SEDIMENTS; SEISMIC EVENTS; SEISMICITY; TECTONICS; THICKNESS; TOMOGRAPHY; VELOCITY

Citation Formats

Flanagan, M P, Myers, S C, Schultz, C A, Pasyanos, M E, and Bhattacharyya, J. LLNL's 3-D A Priori Model Constraints and Uncertainties for Improving Seismic Location. United States: N. p., 2000. Web.
Flanagan, M P, Myers, S C, Schultz, C A, Pasyanos, M E, & Bhattacharyya, J. LLNL's 3-D A Priori Model Constraints and Uncertainties for Improving Seismic Location. United States.
Flanagan, M P, Myers, S C, Schultz, C A, Pasyanos, M E, and Bhattacharyya, J. 2000. "LLNL's 3-D A Priori Model Constraints and Uncertainties for Improving Seismic Location". United States. https://www.osti.gov/servlets/purl/15013435.
@article{osti_15013435,
title = {LLNL's 3-D A Priori Model Constraints and Uncertainties for Improving Seismic Location},
author = {Flanagan, M P and Myers, S C and Schultz, C A and Pasyanos, M E and Bhattacharyya, J},
abstractNote = {Accurate seismic event location is key to monitoring the Comprehensive Nuclear-Test-Ban Treaty (CTBT) and is largely dependent on our understanding of the crust and mantle velocity structure. This is particularly challenging in aseismic regions, devoid of calibration data, which leads us to rely on a priori constraints on the velocities. We investigate our ability to improve seismic event location in the Middle East, North Africa, and the Former Soviet Union (ME/NA/FSU) by using a priori three-dimensional (3-D) velocity models in lieu of more commonly used one dimensional (1-D) models. Event locations based on 1-D models are often biased, as they do not account for significant travel-time variations that result from heterogeneous crust and mantle; it follows that 3-D velocity models have the potential to reduce this bias. Here, we develop a composite 3-D model for the ME/NA/FSU regions. This fully 3-D model is an amalgamation of studies ranging from seismic reflection to geophysical analogy. Our a priori model specifies geographic boundaries and velocity structures based on geology, tectonics, and seismicity and information taken from published literature, namely a global sediment thickness map of 1{sup o} resolution (Laske and Masters, 1997), a regionalized crustal model based on geology and tectonics (Sweeney and Walter, 1998; Bhattacharyya et al., 2000; Walter et al., 2000), and regionalized upper mantle (RUM) models developed from teleseismic travel times (Gudmundsson and Sambridge, 1998). The components of this model were chosen for the complementary structures they provide. The 1{sup o} sediment map and regionalized crustal model provide detailed structures and boundaries not available in the more coarse 5{sup o} models used for global-scale studies. The RUM models offer improved resolution over global tomography, most notably above depths of 300 km where heterogeneity is greatest; however, we plan to test other published upper mantle models of both P- and S-wave velocity. We compute travel times through this integrated model for comparison with other standard 1-D models, as our goal is to evaluate whether the 3-D model can better predict the observed travel times. The arrival times are computed through the model using a 3-D finite-difference technique and are then compared with a declustered set of ISC P arrival times (Engdahl et al., 1998). Our ME/NA/FSU model predicts the P and Pn travel times very well, as measured by variance reduction, for three stations we tested: ARU, KVT, and GAR; these predicted times also resemble some patterns seen in Pn tomography models of this region. Such tests will allow us to identify parts of the model that may need modification. We also compute model-based correction surfaces for each station in the ME/NA/FSU regions that can be used as additional constraints in our event location algorithm to determine the improvement provided by using 3-D models. We use this method to relocate a set of ground truth events: the 1991 Racha aftershock sequence which was investigated by Myers and Schultz (2000) using empirical kriged correction surfaces and a 1-D velocity model. They find an epicenter mislocation bias of 42 km when no corrections are applied and that this mislocation is reduced to 13 km when their empirically derived correction surfaces are included. We relocate this same set of events using our model-based correction surfaces and produce a mislocation bias of only 26 km, a significant improvement. We are currently implementing methods to quantify uncertainties on the model-based corrections which will be required to compute representative error ellipses for the new locations. We also plan to combine both the model-based and empirical correction techniques to achieve the best improvement in location. This test case demonstrates the power of using 3-D velocity models to improve location capability for small, regionally recorded events. This example also shows how the model-based approach holds great potential for improving locations in aseismic regions where it may not be possible to compute empirical correction surfaces.},
doi = {},
url = {https://www.osti.gov/biblio/15013435}, journal = {},
number = ,
volume = ,
place = {United States},
year = {Fri Jul 14 00:00:00 EDT 2000},
month = {Fri Jul 14 00:00:00 EDT 2000}
}

Conference:
Other availability
Please see Document Availability for additional information on obtaining the full-text document. Library patrons may search WorldCat to identify libraries that hold this conference proceeding.

Save / Share: