[S-SS08] Fault Rheology and Earthquake Physics

[SSS08-24] Hotspots of Hypocenters Illuminated by a Newly-Developed Cross-Correlation-Based Hypocenter and Centroid Relocation Method

*Ta-Wei Chang¹, Satoshi Ide¹ (1.The University of Tokyo)

Keywords:Earthquake relocation, Repeating earthquakes

The location of earthquake occurrences has been extensively discussed for its importance. With the advances in methods of earthquake relocation, we now possess methods to determine earthquake centroids with lower uncertainties, using cross-correlation-based methods with data in regional distances. Hypocenters of earthquakes, on the other hand, are traditionally determine using onset picking of event waveforms, which tend to results in poorer resolution as compared to centroid locations. Here, by including an extra step to search for suitable window to perform cross-correlation on the onsets of waveforms, we developed a new cross-correlation-based relative hypocenter relocation method based on the Network Correlation Coefficient (NCC) method. After obtaining relative locations among hypocenters and centroids, we then perform a joint inversion to directly compare their relative locations under a unified coordinate. We have applied this methods to 3 regions that include repeating earthquakes, showing not only the repetitive rupturing of centroids for repeating events, but also the newly-illuminated hotspots of hypocenters, from which earthquakes growing into selective final sizes initiate. Joint hypocenter and centroid locations enables quick estimations of rupture developments without the need of slip inversion analyses. Uncertainties of the locations are estimated using bootstrap, where the 95% confidence intervals yielding sub-100 m error in both centroid and hypocenter locations for horizontal components under best station coverage and event availability. The results of these case studies point to some but limited correspondence between hypocenter and centroid locations or final event sizes, as well as complexity in hypocenter distributions even within repeating earthquakes where rupturing regions mostly overlap.