HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text Full Text (PDF) Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (3) Citing Articles via Google Scholar Google Scholar Articles by Chen, H. Articles by Chiu, S.-C. Search for Related Content GeoRef GeoRef Citation

A Simple Algorithm for Local Earthquake Location Using 3D V_P and V_S Models: Test Examples in the Central United States and in Central Eastern Taiwan

¹ Computer Science Division
University of Memphis
Memphis, Tennessee 38152
 (H.C.)

² Center for Earthquake Research and Information
University of Memphis
Memphis, Tennessee 38152
 (J.-M.C., K.K., S.-C.C.)

³ Department of Earth Sciences
University of Memphis
Memphis, Tennessee 38152
 (J.P.)

⁴ Institute of Earth Sciences
Academia Sinica
Nankang
Taipei, Taiwan
 (K.K., K.-C.C., B.-S.H., Y.-H.Y.)

View larger version (27K): [in a new window]   Figure 1. Locations of PANDA stations (solid triangles) deployed in the central NMSZ from 1989 to 1991 and regional seismic-network stations (open triangles) in the NMSZ since 1991.

View larger version (32K): [in a new window]   Figure 2. Preliminary locations of earthquakes in the NMSZ (1989–2000) using the homogeneous-layer PANDA model (Chiu et al., 1992) and the HYPOELLIPSE program (Lahr, 1999).

View larger version (30K): [in a new window]   Figure 3. Simultaneously relocated epicenters in the NMSZ (1989–2000) determined from a 3D velocity inversion (H. Chen et al., unpublished manuscript, 2005).

View larger version (30K): [in a new window]   Figure 4. Relocated epicenters in the NMSZ (1989–2000) using the single-event- location program developed in this study with the recently available 3D VP and VS models (H. Chen et al., unpublished manuscript, 2005).

View larger version (10K): [in a new window]   Figure 5. An along-strike cross-sectional view of hypocenters in the northeast segment of the NMSZ (AA' in Figs. 2–4) showing that the earthquake locations determined using the PANDA model (crosses) are in general shifted toward larger depths when they are relocated using the newly developed single-event-location program with the available 3D VP and VS models (open circles). The thickness of the sediments in this region is much shallower than that in the PANDA model. The vertical and horizontal scales are equal.

View larger version (15K): [in a new window]   Figure 6. An along-strike cross-sectional view of hypocenters in the central segment of the NMSZ (BB' in Figs. 2–4) showing that the preliminary earthquake locations using the PANDA model (crosses) are very similar to those relocated using the newly developed single-event-location program with the available 3D VP and VS models (open circles). The thickness of the sediments in this region is very close to that in the PANDA model. Seismicity in the central NMSZ is the highest in the region, and it is impossible to visualize the differences in hypocenters before and after the relocation if all earthquakes are shown in the same cross-section. Therefore, only randomly selected earthquakes are displayed in this cross-section. The vertical and horizontal scales are equal.

View larger version (14K): [in a new window]   Figure 7. An along-strike cross-sectional view of hypocenters in the southwest segment of the NMSZ (CC' in Figs. 2–4) showing that the preliminary earthquake locations using the PANDA model (crosses) are mostly shifted toward shallower depths (with only a few exceptions) when they are relocated using the newly developed single-event-location program using the 3D VP and VS models (open circles). The thickness of the sediments in this region is slightly thicker than that in the PANDA model. The vertical and horizontal scales are equal.

View larger version (31K): [in a new window]   Figure 8. Uncertainties of earthquake relocation using 3D models (ERX, ERY, and ERZ) versus earthquake number in a chronological order in the database. The averaged ERX, ERY, and ERZ is 0.24 km, 0.30 km, and 0.48 km, respectively. The completion of the upgrade, expansion, and densification of the regional seismic network in the NMSZ since the time around earthquake number 500 is probably contributing to the significant reduction of the location uncertainties to mostly below the averaged value.

View larger version (105K): [in a new window]   Figure 9. Index map showing the locations of five additional transverse cross- sections (AA'–EE') of hypocenters in the NMSZ. (b) Cross-sectional views of hypocenters for the northeast (AA'), northwest (BB'), and southwest (EE') segments of the NMSZ showing vertical faults. In each section, earthquake locations from the pre- PANDA seismic-network catalog (Central Mississippi Valley Earthquake Bulletin, 1974–1991) are displayed on the top. Hypocenters obtained using PANDA-array data (Chiu et al., 1992) and regional seismic-network data and the PANDA model (1989– 2000) are displayed in the middle. Relocated hypocenters from this study (1989–2000) are displayed in the bottom. Earthquakes within –5 km from the cross-section lines are projected into the section. Lines representing the trend of seismicity are drawn from the relocated hypocenter display (bottom) and are shown in the top and middle displays for reference. (c) Cross-sectional views of hypocenters for the north-central (CC') and south-central (DD') segments of the NMSZ showing gently southwest-dipping faults. The format is the same as Figure 8b. Earthquakes within –1 km from the cross-section lines are projected into the section.

View larger version (35K): [in a new window]   Figure 10. Map showing island-wide seismic- network stations (solid triangles) in Taiwan. Thirty densely distributed PANDA II array stations (open triangles) were deployed from 1993 to 1995 in central eastern Taiwan near the Hualien area, marked by the rectangle. Light gray circles show background seismicity.

View larger version (66K): [in a new window]   Figure 11. Two cross-sectional views of seismicity along the AA' and BB' profiles (Fig. 9) in the Hualien area of central eastern Taiwan. Background is a projection in gray scale of the 3D VP model of Kim (2003). Data shown are from 1993 to 1995 for (a) the original CWB catalog located using a 1D model (Y. Chen, 1995) and CWB stations, (b) the relocated CWB catalog data using the 3D VP and VS models of Kim (2003) and CWB stations, (c) the original PANDA II data using the 1D model of Yeh and Tsai (1981) and PANDA II stations, and (d) the relocated PANDA II data using the 3D VP and VS models of Kim (2003) and PANDA II stations. The local seismic array recorded two to three times more earthquakes than the island-wide CWB network. Using only CWB stations, hypocenters along the AA' and BB' sections do not show significant differences between (a) and (b). Hypocenters shown in (b) may be relocated slightly deeper than in (a). There are significant difference in seismicity between (a, b) and (c, d) when local seismic-array data and array stations are used. Well-defined steeply west-dipping active faults can be identified easily from (c) and (d) but are not apparent from (a) or (b). Hypocenters are more clustered and extend deeper in (d) than in (c).

View larger version (15K): [in a new window]   Figure 12. Computation time required for one travel-time calculation using the 3D raytracing technique of Podvin and Lecomte (1991) (open circles) and for 1000 travel-time calculations by searching previously prepared disk files as discussed in this study (solid circles). The study region of 230 km x 150 km x 17 km is modeled in terms of 3D models with cubic blocks of various sizes to test the dependence of computation time on block size.