No abstract

We have carried out a comprehensive relocation of a total of 267,210 earthquakes in Taiwan that occurred during the past 15 yr. We based our relocation process on the earthquake catalog of the Taiwan Central Weather Bureau Seismic Network (CWBSN) and made improvements in three aspects. First, we incorporated a large dataset of the S-P times from 680 Taiwan Strong-Motion Instrumentation Program (TSMIP) stations distributed throughout the island of Taiwan to improve the coverage of earthquakes on the island. Secondly, we added 18 Japan Meteorological Agency (JMA) stations in the southern Ryukyu Island chain to enhance the station coverage for eastern offshore events, especially around the subduction zone northeast of Taiwan. Thirdly, we adopted 3D V P and V P =V S models in predicting the travel times of P and S waves. The effectiveness of these improvements in earthquake relocation can be seen in three aspects: (1) the reduction in the residuals of P-wave arrival times and S-P times, (2) a better understanding of the attenuation relationship between the peak-ground acceleration and epicentral distance, and (3) the geologically meaningful patterns of station corrections to P-wave arrival times and S-P times.

[1] The 2003 Chengkung earthquake (Mw 6.8) provided diagnostic evidence for a source model showing the deformation process of the seismogenic Chihshang fault in eastern Taiwan. The aftershocks show a fault-bend at a depth of 18 km. Coseismic ground displacements recorded by strong-motion records allow us to deduce instant rupturing of this event. Our resulting model shows a fault length of $33 km and dip-slip dominant rupture on fault-plane deeper than 18 km. Estimated coseismic displacements constrain two fault planes: one at 5 -18 km depth dipping 60°E and 18-36 km depth dipping 45°E. The uppermost fault-plane of the Chihshang Fault (0 -5 km) did not break immediately after the main shock; however, it may have a major role in after-slip and even interseismic ground deformation. The Taiyuan basin developed in the hanging wall is a geomorphic feature consistent with and adequately explained by coseismic ground displacements.

We analyze shear-wave splitting (SWS) in a high-quality waveform data set recorded at surface and downhole (0.2 km) seismometers in a region around the 20 September 1999 M w 7.6 Chi-Chi, Taiwan, earthquake sequence. The data set was generated by events in a 5-year period before, during, and after the mainshock. The purpose is to investigate the depth extent of the crustal anisotropy and its possible temporal evolution in relation to the occurrence of large earthquakes. Results from downhole records show a stable polarization direction of the fast shear wave that matches well the local Global Positioning System (GPS) velocity field. A slightly different polarization direction of the fast shear wave is obtained from surface data. This suggests a possible anisotropy change between the top 0.2 km structure and the deeper section of the crust. Measured time delays below the downhole station have an average value of 0.16 sec without systematic changes for sources from about 8 km to 20 km in depth. Estimates of time delays in the top 0.2 km of the crust based on shear waves reflected from the free surface give a constant 0.04 sec. A likely depth distribution inferred from these two types of measurements and an S-velocity model indicates that the crustal anisotropy in the region is dominated by the top 2 to 3 km. The measured polarization directions and time delays give essentially constant values over the 2.7-year premainshock and 2.3-year postmainshock periods in the region adjacent to the Chi-Chi rupture and within 10 km from the epicentral region of its two large M Ն 6.0 aftershocks. Analysis of SWS in waveforms produced by earthquake multiplets confirms further the lack of precursory temporal variations of crustal anisotropy in the immediate neighborhood of the Chi-Chi earthquake sequence. The results raise doubts on the general usefulness of SWS measurements for earthquake forecasting. An apparent coseismic increase in anisotropy time delay of approximately 10% is observed for depths Ͼ0.2 km; however, this value is clearly affected by spatial changes associated with different event locations before and after the Chi-Chi mainshock.

Progress has been made toward the goal of earthquake early warning in Taiwan. By applying the concept of a virtual subnetwork (VSN) to the Taiwan Central Weather Bureau seismic network, the earthquake rapid-reporting time has been reduced to about 30 sec or less by a new system called VSN. This represents a significant step toward realistic earthquake early-warning capability. The VSN system described here was put into operation from December 2000 to June 2001. A total of 54 earthquakes (100% correct detection) were detected and processed successfully during this period. Comprehensive earthquake reports are issued mostly in less than 30 sec, with an average of about 22 sec after the origin time. The 22-sec reporting time will offer more than 20 sec of early-warning time to cities at distances greater than 145 km from the source, for which the shear-wave strong-shaking arrival time is about 44 sec. This 20 sec is an adequate amount of time to carry out numerous preprogrammed emergency-response measures prior to the arrival of strong shaking.