The Micro-seismicity Study Of The Seismic GAP Between Wenchuan And Lushan Earthquakes In Southern LMS

The mechanism of why the rupture terminated sharply on both sides of the seismic gap between the Wenchuan and Lushan earthquakes is not yet known along the southwestern Longmen Shan fault zone.The accommodation modes of compressional strain and stress are still not known well in the gap subregion.Using the Match and Locate(M&L)method,we obtained a detailed catalog of background micro-earthquakes around the seismic gap from May to December 2015.The characteristics of spatiotemporal distribution of the events is highlighted with a complete magnitude of 0.2.These small events tend to concentrate at different depth ranges on two sides of the seismic gap,indicating that the gap serves as a transition zone between the Wenchuan and Lushan source regions.Strain release through micro-earthquakes is an unlikely candidate for accommodating modes in the gap,because of the sustained deficit of microseismicity.A strong anticorrelation between background seismicity and monthly precipitation variation rate is observed for different magnitude scales.Enhanced seismicity may be activated by the terrestrial water storage reduction next to a precipitation peak with a time lag of 1–2 months during the summer and autumn,indicating a seasonal modulation of seismicity by the water storage cycle.The background microseismicity exhibits prominent responses to precipitation variation due to a short nucleation duration of less than one quarter.We firstly try to investigate slow earthquakes in the seismic gap,using dense arrays deployed by Chengdu University of Technology.Applying a frequency scanning method,we find some tremor-like noise recorded only by single station in foreland basin,yet not tremor signals on multiple stations.Nevertheless,the possibility of tremor slip could not be ruled out based our preliminary study.We note that: high-quality permanent station network or borehole seismometers are indispensable for tremor detection in the region.And there is still a blank of slow slip events research through geodetic observation.Deep learning and machine learning have made great progress in several areas of artificial intelligence and shown promise in application of geoscience.We developed a CNN based small events detection method in the time-frequency domain,termed as TFcnnQuake.Our tests of the method show high identification rate and meanwhile very low mistake rate of recognizing noise as events,indicating the extensive application potential in earthquake warning.