Oscillation frequencies of long-period seismic events at Kusatsu–Shirane volcano, Japan, related to the volume of water vapour in a hydrothermal crack

<jats:title>SUMMARY</jats:title> <jats:p>Long-period (LP) seismic events at active volcanoes are characterized by damped harmonic oscillations, which are thought to be generated by resonances in fluid-filled cracks. LP source properties (crack geometry and fluid properties) are generally estimated by analytically calculating the ratios of spectral peaks in crack resonance frequencies and empirically calculating a quality (Q) factor. However, because this method is applicable only to LP events with more than four spectral peaks, we cannot use it to analyse LP events with fewer spectral peaks. To bridge this gap, we developed a new method to estimate source properties using the frequency (f), Q factor and seismic moment (M0) of the lowest spectral peak of an LP event. We assumed misty gas (water vapour containing small water droplets) as the fluid in the crack and analytically derived the geometrical relationships of the crack. M0 was estimated from observed amplitudes of the lowest spectral peaks at different stations using an assumed crack mechanism. We applied this method to LP events observed from 1989 to 1993 at Kusatsu–Shirane volcano, Japan. We found that the crack length was quite variable, but the crack width was almost constant at around 100 m during our study period. We also found a strong correlation between the inverse of f and the total mass of water in the crack, which can be explained theoretically using the acoustic properties of misty gas and crack geometrical relationships. The total mass of water is proportional to the volume of water vapour in the crack, and the water vapour has been interpreted to be derived from magmatic degassing at depth. The oscillation frequency of the LP event is thus a useful metric for monitoring magma degassing into the shallow hydrothermal system, which is an important constraint on the occurrence of phreatic eruptions. Because our simple approach is widely applicable to LP events, it can enable improved monitoring of hydrothermal activity and evaluation of the risk of phreatic eruptions.</jats:p>