资源环境科技发展态势分析平台

We tested the performance of an infrasonic array consisting of three microphones with a 20-m aperture at Stromboli volcano, Italy. There were four active vents separated by approximate to 10. We employed multiple signal classification (MUSIC) to estimate direction of arrival (DOA) of the detected signals. Using test signals of which the source vents were identified by visual observation, the resolution of DOA estimation of MUSIC is compared with those of Capon beamforming, grid search, and semblance. We confirmed that MUSIC and grid search gave better resolution of DOA than the other two methods. Also, MUSIC provided the best resolutions in time and frequency. It was shown that the DOA switched between different vents or fluctuated in short time scales and can vary with frequency, which indicate multiple active sources. Possible DOA estimation errors were evaluated. A small aperture infrasonic array combined with MUSIC will become a powerful tool for studying and monitoring active volcanoes.

Plain Language Summary Infrasound is an acoustic wave in subaudible frequency ranges (<20 Hz) and is used to detect various phenomena, such as volcanic eruptions, tsunami, snow avalanche, and nuclear test bans. Multiple sensors distributed in short distances, which is called an array, are useful for the detection and for the estimation of source directions. Here we examined the performance of a small-size infrasonic array combined with a high-resolution source location algorithm. The field experiment was conducted at Stromboli volcano, Italy, recording infrasound signals emanated from four active vents. Signals from different vents were successfully distinguished. The proposed method provided better resolution in time and frequency as well as the source direction than other conventional methods, which enabled us to observe that the sources of the main signals detected by the array change in short time duration and can be variable with frequency. These observations are important in clarifying the source processes of infrasound associated with volcanic activity. The proposed signal processing method also has good extensivity to resolve multiple sources and compatibility with the analyses used for seismic arrays. We believe a small-size infrasonic array combined with the proposed processing method will become a powerful tool in studying and monitoring active volcanoes.