Source Scaling Relationship for M4.5-M9.1 Earthquakes: Specifically for Earthquakes in the Collision Zone

The finite fault source models derived from the dense strong motion or teleseismic stations were analyzed to investigate the scaling relationship of the source parameters. The spatial slip distribution over a rectangle finite-fault was considered for each event. The spatial slip distributions of 19 events derived from the dense strong motion stations in Taiwan and the Global Seismographic Network were determined to give the finite fault solution for the moment ranged from 7.8 × 1015 to 3.8 × 1020 N-m. Except for the 1999 Chi-Chi Mw7.6 earthquake, there are 10 blind thrust events, 1 normal event and 7 strike-slip events. Among them, there are 3 subduction events. The M7.9 2008 Wenchuan earthquake and M9.1 2004 Sumatra earthquake were also included in the scaling to provide broader scales for large earthquakes. The asperities in a finite-fault were lack of quantitative definition, and the dimensions of the finite-fault geometry were often overestimated due to the slip heterogeneity. The subjective definition of asperity for earthquakes leads to a less accurate scaling relationship for the source models. A mathematic method in which the autocorrelation function defined from quantifying spatial slips was utilized to evaluate the effective fault dimensions to systemically redefine the characteristics of the source parameters, including effective fault length (Le), width (We) and average slip (De). Our analysis shows two scaling relationships for Mo ∝ Le2 and Mo ∝ Le3. Effective length to width scaling behavior has a slope of Le ∝ We for moderate size events. This relationship might reflect the character for collision zone earthquakes, which events commonly from blind thrusts with deeper focal depths. Thus, this source scaling might provide additional information for further studies on the simulation of ground motion for earthquakes from the tectonic collision zone.