Improving the Detection of Low‐Magnitude Seismicity Preceding the Mw 6.3 L’Aquila Earthquake: Development of a Scalable Code Based on the Cross Correlation of Template Earthquakes

Vuan, A.; Sugan, M.; Amati, G. (Cineca), Kato, A.

A scalable procedure to detect microseismicity based on the cross correlation of well‐located template events is developed using Python, Numpy, and ObsPy toolkits. With this technique, we investigate the spatiotemporal evolution of the seismicity before the 2009 L’Aquila earthquake. Expanding on Sugan et al. (2014) to improve detection ability, we apply additional quality controls to the input data, sample tolerance in the calculation of the stacked correlograms, and detection thresholds. The methodological improvements enhance the recognition of small‐magnitude earthquakes (~6%) and decrease false positives (~9%). Because most of the events we found are clustered near the AQU station, located above the main fault in the hanging wall, we rerun the matching algorithm only using this station and templates in a radius of 25 km to enhance the microseismicity recorded close to the hypocenter of the 6 April 2009 Mw6.3 earthquake. In this further analysis, we found 938 new events that highlight times and magnitudes of creeping at the fault bottom around the mainshock nucleation. Template events that repeated over time migrated from northwest to southeast surrounding the low b‐value stressed patch (b‐value = 0.8) described in Sugan et al. (2014). Within the main normal fault, the quasi‐repeating earthquakes are frequent when b‐values are equal to or greater than 1. This behavior indicates that fault creeping around the mainshock nucleation caused stress loading on the locked fault portion that had a low b‐value.