Item

Abstract

North Anatolian Fault Zone (NAFZ) is one of the major transform fault zones in Turkey. The NAFZ splits into two main branches in the east of the Marmara Sea. The northern strand also named as the Main Marmara Fault (MMF) accommodates most of the relative motion between Anatolia and Eurasia. Currently, the MMF is regarded as a seismic gap that can generate a large earthquake threatening Istanbul metropolitan area. In order to improve understanding of the possible future earthquake potential near Istanbul, it becomes crucial to understand fault segmentation and seismicity along the MMF.The seismicity in the Marmara Sea is continuously monitored by three observatories, namely; KOERI-RETMC, AFAD and a local network run by scientists from Helmholtz Centre Potsdam, GFZ German Research Centre for Geosciences (GFZ) and Boğaziçi University Kandilli Observatory and Earthquake Research Institute (KOERI) called Prince Islands Real Time Earthquake Monitoring System (PIRES). In this presentation, seismic data from all of these networks are used.In this study, we utilize EQTransformer, a deep learning algorithm which both detects earthquakes and phase arrivals. We also use classical STA/LTA to compare the performance of the machine learning algorithm. Our goal is to create a reliable earthquake catalog for the eastern Marmara, that will be created with the most suitable configuration and parameters, which then will be compared with the observatory center’s catalogs. In this way, in addition to the improved seismicity pattern of the MMF, we will test the performance of a machine learning algorithm in a high seismic risk area.