Clustering Analysis and Visualization of Synthetic Earthquake Events

Advisor: David A. Yuen

Studying seismic patterns with synthetic data is important for analyzing the structure of San Andreas Fault. Using clustering analysis and a visualization software, Amira, we have examined 4 different simulated models and revealed several spatial-temporal fault zone disorder. The synthetic models generated by Ben-Zion, a professor at the University of Southern California, are based on 3D dislocation theory of the central San Andreas Fault. His 4 different types of model represent various stress drop distribution. My co-worker created C program for MNN (mutual nearest neighbor) clustering method and I ran the program and used visualization software to analyze seismic patterns based on the proximity of each earthquake. We showed that the each synthetic model has distinct spatial-temporal cluster structure. The result represents the clear correlation between M > 6 earthquakes and clusters of small earthquakes in high-density earthquake region and in many cases, the clusters of small earthquakes precede the large earthquakes.