Reducing Systematic Errors for Seismic Event Locations in Anisotropic Regional Structures

In CTBT applications many events of interest are only detected at regional distances. Providing more accurate prediction of P-wave propagation at regional distances is therefore of particular importance in seismic event location. At such distances (2 -14 ) the phase Pn is the seismic phase that is most frequently reported and which thus controls the location accuracy. We are working on reducing systematic errors in Pn travel-times and thus seismic event location at regional distances. In recent work the P.I. has mapped lateral and anisotropic variations in Pn velocities beneath continents across the globe (Smith and Ekstrom, 1999). This work provides the most comprehensive and possibly most accurate mapping of anisotropic Pn velocities available to date. While the lateral variations in Pn velocities that were mapped were strong, and are likely to contribute to improved location capabilities, strong (up to 10%) anisotropic signatures were also observed. The horizontally travelling Pn phase should therefore accumulate large travel time residuals due to both heterogeneity and anisotropy, which would result in large systematic location errors. The question remains whether this new mapping can provide, in a practical sense, significant reductions in systematic event mislocation at the regional scale. Preliminary results indicate that, even in areas of good station coverage, a distinct difference in location is obtained using anisotropic models. OBJECTIVE In CTBT applications many events of interest are only detected at regional distances. Our objective is identification and reduction of systematic errors in the location of events determined using regional seismic data. At such distances (2 -14 ) the phase Pn is the seismic phase that is most commonly reported and which thus controls the location accuracy. In order to accurately locate seismic events, whether natural or artificial, by traditional travel-time methods one must first be able to accurately predict arrival times. Historically travel-times have been calculated using one-dimensional seismic velocity models (e.g. Jeffreys and Bullen, 1940; Herrin et al., 1968; Herrin and Taggart, 1968; Herrin, 1968; Dziewonski and Anderson, 1981; Kennett and Engdahl, 1991). However, the Earth is composed of rocks which vary laterally at varying length scales (e.g. Crosson, 1976; Engdahl et al., 1977, 1982; Engdahl and Billington, 1986; Dziewonski, 1984; Su and Dziewonski, 1993) and can be anisotropic (e.g. Christensen, 1966; Kumazawa and Anderson, 1969; Hess, 1964; Raitt et al., 1969; Forsyth, 1975; Tanimoto and Anderson, 1984), resulting in travel-times which do not match those predicted by these one-dimensional velocity profiles. In addition, at regional length scales global Earth models, which are largely based on long-period surface waves and vertically arriving body waves, provide poor first arrival travel-time predictions. Providing more accurate prediction of P-wave propagation at regional distances is therefore of particular importance in event location. When attempting to satisfy the location requirements of the CTBT it is essential to obtain the most accurate location possible, with the minimum necessary computing time The question remains as to whether the current generation of regional models can usefully contribute to relocation problems. While it has already been well established that variations in regional phases such as Pn can lead to large mislocations of the epicenter (Herrin and Taggart, 1962), progress has been slow in routinely applying regional models to locations for global catalogs. This is probably because most of the Pn velocity models produced are of a highly local nature (e.g. Hess, 1964; Raitt et al., 1969; Bamford, 1977; Fuchs, 1977; Hirn, 1977; Vetter and Minster, 1981), and no systematic global mapping of Pn velocities has been attempted. In addition although azimuthal anisotropy is a known feature of Pn Report Documentation Page Form Approved OMB No. 0704-0188 Public reporting burden for the collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to a penalty for failing to comply with a collection of information if it does not display a currently valid OMB control number. 1. REPORT DATE SEP 2000 2. REPORT TYPE 3. DATES COVERED 00-00-2000 to 00-00-2000 4. TITLE AND SUBTITLE Reducing Systematic Errors For Seismic Event Locations In Anisotropic Regional Structures 5a. CONTRACT NUMBER 5b. GRANT NUMBER 5c. PROGRAM ELEMENT NUMBER 6. AUTHOR(S) 5d. PROJECT NUMBER 5e. TASK NUMBER 5f. WORK UNIT NUMBER 7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) Washington University,Dept Earth & Planetary Sciences,St. Louis,MO,63130 8. PERFORMING ORGANIZATION REPORT NUMBER 9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSOR/MONITOR’S ACRONYM(S) 11. SPONSOR/MONITOR’S REPORT NUMBER(S) 12. DISTRIBUTION/AVAILABILITY STATEMENT Approved for public release; distribution unlimited 13. SUPPLEMENTARY NOTES Proceedings of the 22nd Annual DoD/DOE Seismic Research Symposium: Planning for Verification of and Compliance with the Comprehensive Nuclear-Test-Ban Treaty (CTBT) held in New Orleans, Louisiana on September 13-15, 2000, U.S. Government or Federal Rights. 14. ABSTRACT See Report 15. SUBJECT TERMS 16. SECURITY CLASSIFICATION OF: 17. LIMITATION OF ABSTRACT Same as Report (SAR) 18. NUMBER OF PAGES 7 19a. NAME OF RESPONSIBLE PERSON a. REPORT unclassified b. ABSTRACT unclassified c. THIS PAGE unclassified Standard Form 298 (Rev. 8-98) Prescribed by ANSI Std Z39-18 propagation (e.g. Beghoul and Barazangi, 1990; Hearn, 1996), most previous studies of Pn anisotropy have not mapped lateral variations in azimuthal anisotropy, but instead produced, if anything, a single estimate for an entire region. In recent work the P.I. has mapped lateral and anisotropic variations in Pn velocities beneath continents across the globe (Smith and Ekstrom, 1999). This work represents the most comprehensive and possibly the most accurate mapping of anisotropic Pn velocities available to date. This provides the first opportunity to truly test the possibility of applying an anisotropic Pn velocity model to calculation of travel-times to improve regional locations for events distributed in different parts of the world. The question remains whether this new mapping can provide, in a practical application, significant reductions in systematic event location at the regional scale. Our work is aimed at applying this new mapping of Pn anisotropic structure to investigate the possible systematic errors produced by lateral heterogeneity and azimuthal anisotropy RESEARCH ACCOMPLISHED As a preliminary investigation we have already implemented a grid search relocation algorithm and applied this to the Pn mapping of Smith and Ekstrom (1999) to test for systematic errors in location. In this study we have taken several events in western Europe as we have good mapping of both Pn velocities and anisotropy in this area. Figure 1 shows the lateral variations in Pn velocities of Western Europe extracted from our dataset. Figure 2 shows the fast azimuths of the Pn anisotropy from our model. In this preliminary relocation experiment we use travel-time data from the ISC database. The ISC location is used as a first estimate. The fit of travel times is then calculated for this location and for a set of points on a rectangular grid at 10-km spacing. The minimum in the rms of the travel times is then selected as the new location estimate and the travel-time misfits recalculated using a smaller grid spacing. This is repeated until the travel-time misfit appears to converge. This procedure has been performed for a selection of earthquake events for isotropic, laterally heterogeneous, and anisotropic structures. In this preliminary test for systematic effects great-circle raypaths were used. The results of this can be seen in Figure 3. Figure 3(a) shows the relocation vector of the events using laterally varying Pn velocities. The base of the arrow lies at the location found with an isotropic model. The arrow points in the direction of the location found by use of the laterally varying Pn velocities. The arrow length is proportional to the distance between the two locations, and a 10 km displacement is shown for scale. Similarly, Figure 3(b) compares the relocation vector of the events when using azimuthally anisotropic Pn velocities. Again the base of each arrow lies at the location found with an isotropic model. With these events “ground truth” is not known and so comparison to a “true” location is not possible. However, what is demonstrated is a noticeable systematic difference between the locations found with the different models. In both cases the revised locations have moved away from stations where the model predicts a fast traveltime anomaly, and towards stations where the traveltime anomaly would be slow. If we assume that the anisotropic locations are the most accurate, being based on the most detailed model, this would suggest that failure to account for anisotropic structure introduces significant systematic errors. Figure 1: Lateral variations in the isotropic Pn velocity across Western Europe. Upward pointing triangles are slow, downward pointing triangles are fast. The size of the triangle is proportional to the anomaly. Values are in the range 7.8 to 8.4 km/s. Figure 2: Pn anisotropy results for Western Europe. The center of each symmetric arrow pair is plotted at the center of the cap. Arrows point in the direction of fast Pn propagation and are proportional to the strength of anisotropy. Black arrows are higher quality estimates. Black and white center points indicate 1.5° and 3° radius caps respectively. Triangles show the location of null results. The absolute plate motion vector is also shown (Minster and Jordan, 1978). -15 -15