Presentation Information
[1Yin-A-47]Automated Earthquake Hypocenter Determination by Integrating Deep-Learning Phase Picking with Waveform Stacking
〇Mitsuyuki Ozawa1 (1. JGI,Inc.)
Keywords:
seismic,hypocenter determination,waveform stacking
We propose an analysis method for automated earthquake hypocenter determination that integrates deep-learning-based automatic P/S phase annotation with waveform stacking. Arrival times and detection probabilities estimated from continuous waveforms are used as weights to align and stack multi-station waveforms. By searching for the solution that maximizes the stack amplitude, the method simultaneously identifies events and estimates the hypocenter location and origin time. Application to real data confirms that, even under low S/N conditions and biased station geometries, the proposed method improves the detection rate compared with conventional onset detectors such as STA/LTA and reduces the scatter of the estimated hypocenters. These results indicate that the proposed approach offers high effectiveness and extensibility as an automated framework for earthquake detection and hypocenter determination with practical operational use in mind.