Complex principal component analysis of volcanic earthquakes at Azuma volcano, Japan, recorded by a distributed acoustic sensing system (DAS) for the hypocenter determination

Distributed Acoustic Sensing System (DAS) can record ground motions with a measurement interval of a few to ten meters for a long-distance range of a few tens of kilometers. The sensor, which is a fiber-optic cable embedded underground, is protected from damage caused by lightning or pyroclastic materials. These features are useful for the seismic observation of active volcanoes. We analyzed DAS data recorded at the Azuma volcano, Japan, to determine the location of volcanic earthquakes and evaluate their reliability. Using the benefits of highly dense measurement points along the fiber optic cable, we applied complex principal component analysis (CPCA) to measure the arrival time differences of seismic waves from volcanic earthquakes. We used a relative hypocenter determination technique to accurately determine the source locations of the volcanic earthquakes recorded during our observation period in July 2019. We observed 148 small volcanic earthquakes and succeeded in determining 31 events with sufficient arrival time difference data. The obtained source locations were distributed around Jododaira, extending in the north-south direction with a length of approximately 2 km. These regions are consistent with the hypocenters determined from standard hypocenter determination using the arrival times of the P- and S-waves of volcanic earthquakes from 2017 to 2022. We also determine the source locations by using the arrival time difference measured by the cross-spectral method, which is often used in seismic wave analyses, and the number of located events is small and the obtained source locations are scattered. These results suggest that CPCA is useful for extracting signals from noisy raw seismic data and accurately measuring the arrival time difference for the source locations.