Ruptures in a Highly Fractured Upper Crust

The 1999 Chiayi M 6.4 earthquake sequence that occurred about one month after the 1999 Chi-Chi M 7.6 earthquake is remarkably similar to the Ruey-Li M 6.2 earthquake sequence which occurred one year before the big Chi-Chi event. Both sequences were energetic and short in duration; with events confined in small source volumes, and the fault-plane orientation of individual events in these sequences varied abruptly during the development of the dynamic rupture. Both occurred on the southern extension of the large Chelungpu rupture causal to the Chi-Chi event, and in an area of highly fractured upper crust. We consider that this faulting behavior is due to the rupture-induced dynamic stress that interacts with the pre-existing highly fractured upper crust and continually perturbs the local stress field. This phenomenon appears to be typical of particularly active plate boundary regions with a highly fractured upper crust under intense tectonic loading.     

A preliminary study on the relationship between precipitation and large earthquakes in Southern California

By comparing seasonal rainfall data from the past 90 years with the occurrence of large (M6) earthquakes along an arid stretch of the San Andreas fault system in southern California, certain correlations have been observed. Most large earthquakes are preceded by a pattern consisting of a few years of below normal precipitation (drought) terminated by one or more consecutive seasons of heavy (above normal) rainfall. While this drought-above normal rainfall cycle can be seen at times other than prior to major earthquakes, it precedes, to varying degrees, all of the twelve M6 events. This new precursor evidence, when combined with other premonitory signals, may offer a helpful diagnostic measure that could be useful in earthquake prediction in arid regions.     

A study of the human-fatality rate in near-fault regions using the Victim Attribute Database

After the 1999 Chi-Chi earthquake, the Taiwanese government immediately issued new guidelines prohibiting the construction of structures for human occupancy within the Chelungpu fault zone. However, these guidelines were not based upon an in-depth hazard analysis of the near-fault regions. The positions of more than 80% of the 2,492 victims of the Chi-Chi earthquake were found by our research team. A Victim Attribute Database has been compiled that includes the GPS coordinates of the positioned victims as well as other attribute data associated with the victims. The human-fatality rates in the near-fault regions have been analyzed with regard to distances from the Chelungpu fault, the hanging-wall and footwall areas, as well as building type. The severity at the human-fatality rates in the near-fault regions is inversely proportional to distances from the causative fault, i.e., the closer the distance, the higher the human-fatality rate observed. The human-fatality rate for victims who lived in closer proximity to the hanging-wall areas is also significantly higher than those who lived in closer proximity to the footwall areas, especially in areas on either side of the fault and within 1,000 m of the fault surface trace. In terms of different building types, factors that include the capacity of the buildings to resistant strong shaking and the level of strong ground-motion greatly affected the human-fatality rates in the hanging-wall and footwall areas. Therefore simply prohibiting the construction of buildings within the active fault zone would be an insufficient method of reducing the number of potential victims; a nationwide effort should be undertaken to upgrade the capacity of buildings to resist strong shaking.     

Receiver Functions for Three-layer Media

We have extended the H-k stacking method of receiver functions applicable to a three-layer model, which is useful in studying detailed crustal structures. We have demonstrated its application with two sample sites in Taiwan, making use of travel times of converted phases from the direct P waves of teleseismic events as the P reflects and refracts at different discontinuities in the crust. This three-layer extension allows a closer examination on the crust, as well as the relationship between crustal layers and the associated V_p/V_s ratios. Data were processed using the multiple-taper correlation technique to obtain the radial receiver functions (RRFs). The relative time delays of the converted phases measured from the RRFs were used to estimate the depths of the crustal discontinuities. Results not only yield the depths to principal crustal layers, including the Moho, but also give the corresponding ranges of V_p/V_s ratio which are comparable with findings from other tomographic studies.     

Significant crustal structural variation across the Chaochou Fault, southern Taiwan: New tectonic implications for convergent plate boundary

The Chaochou Fault, a major geological boundary in southern Taiwan is considered to be a part of the convergent plate boundary between the Eurasia Plate and the Philippine Sea Plate. We applied the Common Conversion Point stacking technique to teleseismic radial receiver functions and obtained Moho variation and crustal structure across the Chaochou Fault. In the Eurasia Plate to its west, the Moho depth is about 37 km and the crust is subducting to the east beneath the Philippine Sea Plate with a dip angle of about 30° between the Backbone Belt and the Tananao Schist. In the Philippine Sea Plate, the Moho depth is about 17 km. The Longitudinal Valley marks the collision boundary between the Eurasia Plate and the Philippine Sea Plate. The results suggest that the depth extent of the Chaochou Fault is about 30–35 km and the fault becomes a “shallow-angle” thrust fault at depth. The Common Conversion Point image also shows several bending interfaces of velocity contrast in the crust. We proposed a simple model to explain the Philippine Sea Plate and Eurasia Plate collision process and the observed crustal deformations.