| Summary: | 博士 === 國立成功大學 === 地球科學系碩博士班 === 96 === Creeping crustal faults often generate a number of microearthquakes, and less commonly, they may also produce large earthquakes that rupture the brittle crust. The Longitudinal Valley fault (LVF) in eastern Taiwan characterized by such behavior has been known to undergo 1-3 cm/yr surface creep, probably one of the most active creeping thrust faults known in the world. The understanding of fault behavior at depth as to how a creeping fault can generate large earthquakes has been limited due to sparse sampling of seismic and geodetic stations in this area. Repeating earthquakes study has been proposed to improve the understanding of deep fault deformation, however, it also has its limitation because of the poor station coverage. In this thesis, methodology of repeating earthquake identification is developed in the region where the station coverage is sparse and one-sided. The methodology is then applied to the earthquakes in eastern Taiwan. Two major clusters of repeating earthquake sequences (RESs) are found near the north and south ends of the LVF, with variable recurrence patterns and deep slip rates. The 25 M 2.3-4.6 RESs in the Hualien area (north of the LVF) are widely distributed along the strike of the LVF over a distance of 45 km at 10-22 km depth, whereas in the Chihshang area (north of the LVF) we found 30 M 2.2-3.4 RESs at 7-23 km depth with three times shorter along-strike extent. Slip estimates from the RESs in Chihshang and Hualien indicate the slip rate range of 2.5-7.8 cm/yr and 2.0-5.1 cm/yr respectively, which is consistent with those inferred from the surface measurements. With the fact that the aftershock of M 6 event were confined to the south half of the Chihshang fault, we infer that the 30-km stretch of the Chihshang fault is creeping at north half section and locked at south half section. For the Hualien area, M ≥ 3.8 RESs are observed to be spatially isolated and are characterized by quasi-periodic recurrences, where the surrounding earthquakes are small in both number and size. It suggests that these larger RESs are probably less influenced by nearby earthquakes and therefore can recur in a regular manner. The lack of M ≥ 3.5 RESs in the Chishang area likely results from the narrowness of the creeping area, as indicated by the limited spatial extent of the RESs. Thus it is demonstrated that repeating earthquakes observation can provide information of deep fault behavior, potentially leading to well-resolved fault model and better understanding of seismic hazard assessment. Recurrence properties of these repeating earthquakes in eastern Taiwan reveals a weak variation in recurrence interval (Tr ) with seismic moment (Mo). Compared to the scaling of Tr with Mo from repeating earthquake data near Parkfield in California, the repeating data from eastern Taiwan has recurrence intervals that are 2 times shorter. Also in northeastern Japan, Tr of repeating quakes are ~4 times shorter than those expected from the Parkfield scaling law. When adjusted to account for differences in the geodetically derived slip rates for the three fault zones, however, the Tr-Mo scaling is remarkably consistent among the three regions. It suggests that the tectonic loading rate is likely the most important factor that controls the repeat time. It also suggests that there seems to exist a universal rule on recurrence interval scaling of repeating earthquakes in diverse tectonic settings.
|
|---|
