基于深度学习的湖南地区地震事件检测与定位OA

Based on the automated workflow of LOC-FLOW deep learning for earthquake detection and location,this study systematically processed continuous waveform data recorded by the Hunan Seismic Network from 2011 to 2024.Through automatic phase picking,phase association,absolute positioning and manual verification,a total of 2831 seismic events were identified,which was approximately 4.3 times the number of manually cataloged events.The minimum complete magnitude was reduced to ML1.5,significantly enhancing microseismic detection capability.1644 events were precisely relocated using the double-difference relocation method.The results indicate that some events are concentrated along known faults,while others may be associated with reservoir impoundment activities.This study provides fundamental data support for further understanding the characteristics of seismic activity,seismogenic mechanisms,and seismic hazard assessment in the Hunan region.