Systematic Analysis of Shear-Wave Splitting in the Aftershock Zone of the 1999 Chi-Chi, Taiwan, Earthquake: Shallow Crustal Anisotropy and Lack of Precursory Variations

Abstract

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.

Seismically invisible fault zones: Laboratory insights into imaging faults in anisotropic rocks

“…Strong anisotropy may also occur in fault zones where phyllosilicate material is not a major component. Stress-induced anisotropy is often observed in fault zones [e.g., Liu et al, 2004;Cochran and Kroll, 2015]. If the extent of stress-induced anisotropy is significant, then this may also affect seismic imaging of faults in a similar way.…”

The Erua earthquake cluster and seismic anisotropy in the Ruapehu region, New Zealand

“…However, Liu et al [2004] demonstrated that spatial variations in seismic sources can be misinterpreted as temporal changes in anisotropy as different ray paths sample different regions of the anisotropic medium. Even though SWS analysis can be used as an indicator of stress and of fluid saturation in the crust, surprisingly few studies have been conducted on shear wave splitting around volcanoes.…”

Ruptures in a Highly Fractured Upper Crust

“…Secondly, the phenomenon of principal stress permutations of (r 1 , r 2 ) and (r 2 , r 3 ) is commonly observed in nature, and it may be caused by strong inhomogeniety and/or anisotropy in the upper crust (HU and ANGELIER, 2004), especially for the highly fractured area. Research indicated that strong anisotropy exists in the upper crust beneath the Chia-Nan area (LIU et al, 2004;CHEN and YEN, 1998). Nevertheless, the results strongly suggest that mixed modes of reverse/strike-slip focal mechanism can exist in a compressional region, which is similar to the study of HU and ANGELIER (2004), and also can be observed in the Costal Range of eastern Taiwan (BARRIER and ANGELIER, 1986).…”