Stress change and fault interaction of adjacent dip-slip, creeping faults in Taiwan

Understanding the aseismic slip and its interplay with seismic slip is central to seismogenesis as it ultimately controls the space-time patterns of seismicity. The aseismic slip has mostly been documented on subduction zones, where the aseismic slip occurs close to the plate rate during the interseismic period and accelerates after/preceding a nearby mainshocks. In tectonically-active continental regions, the intensive efforts of mapping and characterizing aseismic slip have been made to strike-slip faults. Due to lack of recognized creep on dip-slip faults, the nature of fault creep and its role in large earthquake generation in dip-slip creeping fault remains unclear. Whether the spatial and temporal distribution of earthquakes?

The two segmented, ~150-km-long, creeping fault systems in Taiwan are characterized by fast deep slip rate (4-5 cm/yr), large damaging earthquakes, repeating earthquakes, and swarm activities. They provide a rare opportunity for studying the nature of fault creep in the dip-slip faults. As a boundary between the Philippine Sea plate and the Eurasian plate, the Longitudinal Valley on the eastern Taiwan is composed of two parallel structures with opposite dipping direction: the east-dipping Longitudinal Valley fault (LVF) to the east and the west-dipping Central Range fault (CRF) to the west, with the surface separation of shorter than 10 km. Since 1990, thirteen M6 earthquakes have occurred along the two faults. To understand the characteristics and mechanisms of earthquake interaction between the two adjacent active faults, three major works are conducted: (1) Identifying earthquakes that are responsible for the LVF and CRF activities based on relocated seismicity (2) Identifying earthquake clusters using a statistics-based algorithm (3) Quantifying the interaction between seismicity on two separate faults using the spatiotemporal distribution of earthquake clusters (3) computing the static stress change to verify the stress triggering relationship between the two adjacent faults.

Each of the two adjacent faults can be both divided into three segments. We found that only the southern segments exhibit strong interaction in earthquake clusters. On December 10, 2003, a M6.4 earthquake in the southern LVF likely triggered a M5.3 earthquake in southern CRF that occurred 8 days later, as promoted by 0.8 bar stress change. On April 1, 2006, a M6.2 earthquake in southern CRF on the other hand, is capable of triggering a M6.0 earthquake in the south segment of LVF two weeks later, imported by 0.6 bar stress change. Given that the southern segment of the LVF is characterized by the creep rate of 2-3 cm/yr on the surface and ~4 cm/yr at greater depth below 10 km, while the other segments reveal stronger fault coupling, we argue that the nature of fault creep may control the triggering potential in the adjacent fault.