Application of Dense Array Analysis to Strong Motion Data Recorded at The SMART-1 Array

This paper is part of a project to design an optimal strong motion dense array in New Zealand. The overall project looks at developing a dense network of strong motion seismometers in order to measure directly the rupture process of major seismogenic sources such as the Alpine Fault and strands of the Marlborough Fault System defining the South Island sector of the Australia-Pacific plate boundary zone. This work shows the application of dense array analysis to a set of seismic data recorded at the SMART-1 array in Taiwan (data kindly provided by the Institute of Earth Sciences, Academia Sinica Data Management Center for Strong Motion Seismology - Taiwan). The data have been processed and analysed applying modified MUSIC algorithm with higher computing capabilities giving higher resolution results. The SMART-1 array is an ideal dense array of 37 strong motion instruments set up in the following configuration: 3 concentric circles of radii 200m, 1 km and 2km, and one central station. The studied event called Event 5 was recorded on January 29th 1981 and had a magnitude 6. Event 5 is an ideal case study as its epicentral distance (about 30 km) is comparable to epicentral distances for expected events on the Alpine Fault or on the Hope Fault in New Zealand. Event 5 has been previously widely analysed using strong motion array studies and aftershocks studies but with disagreeing results; this new study hopes to bring new insights in the debate. Using simple fault and velocity models, this latest analysis of Event 5 has given the following rupture properties. It has confirmed one of the hypotheses that the fault ruptured from southeast to northwest. The higher resolution of the computation has improved the location of the hypocentre depth and the location of the propagating rupture front. This allowed resolving changes of velocities in the rupture process and locating asperities in the fault plane. Contrary to the previous array studies, the inferred size of the fault plane is in agreement with a magnitude 6 earthquake. Finally, this event 5 case study has shown significant improvement in imaging the fault rupture using strong motion dense array analysis. But are we truly imaging the real process of a fault rupturing? This question brings on the need to compare known synthetic rupture processes with their computed projection using synthetic strong motion accelerograms.