ADVANCING SEISMIC CATALOGS IN NORTHERN CHILE:INSIGHTS FROM MODERN DEEP LEARNING

Northern Chile, a highly seismogenic area, has been under continuous monitoring since ۲۰۰۷. Thisstudy introduces a novel approach to build a comprehensive seismicity catalog, employing state-ofthe-art deep-learning algorithms for phase detection and association. We assessed EQTransformer andGaMMA, determining optimal parameters through precision-recall curves. By applying this method todata from the year ۲۰۲۰, we significantly augmented event detection, revealing ۸۳,۱۹۴ events, a nearlytenfold increase compared to previous catalogs. We refine the catalog by removing duplicate eventsand relocating associated events using advanced algorithms. Further, we extend this methodology toanalyze several years of IPOC data, aiming to cover the entire ۱۵-year operational period, facilitatingcomprehensive regional seismic analysis. Our findings highlight the potential of modern deep-learningtechniques in seismic catalog expansion, enabling deeper insights into regional seismic processes

کلیدواژه ها:

Phase Detection ، Association ، EQTransformer ، Northern Chile ، IPOC Data ، SeismicityCatalog