Spatiotemporal Variations and Postseismic Relaxation Process Around Mt. Fuji, Japan, During and After the 2011 Tohoku‐Oki Earthquake.

To monitor the volcanoes at a high spatiotemporal resolution, we introduce the singular value decomposition‐based Wiener filter and the three‐component waveforms in ambient noise velocity monitoring. The continuous ambient noise data from 63 stations around Mt. Fuji and Mt. Hakone, Japan, during the January‐September 2011 were analyzed to estimate the seismic velocity variations at a 1‐day temporal resolution, allowing us to distinguish the velocity drops caused by the 2011 Mw 9.0 Tohoku‐oki and the Mw 6.0 East Shizuoka earthquake. The velocity drop during the Tohoku‐oki earthquake was large in volcanic areas and was larger around Mt. Hakone than Mt. Fuji. This difference is possibly due to the existence of fluid‐ and gas‐rich zones at shallower depths and a higher crack density around Mt. Hakone. In addition, the velocity drop at Mt. Fuji during the Tohoku‐oki and the East Shizuoka earthquake was the same level, despite larger static stress changes beneath Mt. Fuji during the East Shizuoka earthquake. We interpret this inconsistency between the velocity drops and static stress changes to arise from incomplete recovery of the generated cracks during the Tohoku‐oki earthquake when the East Shizuoka earthquake occurred. This study also investigates the spatial variations in recovery speed and recovery amount, finding slow recovery speeds in the volcanic areas and fault areas, possibly due to larger crack densities in the crust. Furthermore, we observe the lowest velocity recovery amount in the volcanic areas, which is likely attributed to the maintained increase in pore pressure due to the volcanic gas bubbles. Plain Language Summary: Ambient noise monitoring is a useful tool to estimate the dynamic behaviors of subsurface conditions, such as crustal stress and pore pressure. In recent years, this monitoring technique has been applied to volcanic eruption prediction, however, the temporal resolution is about 5 or 10 days, which is still insufficient for volcanic eruption prediction. Therefore, we introduce denoising techniques with combination of singular value decomposition and Wiener filter to stably estimate the velocity change in high temporal resolution. As a result, we are able to estimate velocity changes with 1‐day temporal resolution and distinguish the velocity changes caused by the 2011 Mw 9.0 Tohoku earthquake and the Mw 6.0 East Shizuoka earthquake that occurred 4 days after the Tohoku earthquake. The high temporal resolution of this monitoring approach could contribute to volcano monitoring systems, particularly through providing additional information on the state of the volcanic system. Moreover, we model the velocity changes and estimate the spatial variation of recovery speed and recovery amount, which is important to predict the postseimic recovery of velocity changes. We found slow recovery speed in volcanic areas and fault areas, and small recovery amount in volcanic areas. Key Points: Ambient noise monitoring with high temporal resolution (1 day) was applied to volcanic areas (Mt. Fuji and Hakone area)Slow velocity recovery speeds in the volcanic areas and fault areas, possibly due to larger crack densities in the crustThe lowest velocity recovery amount in the volcanic areas could attribute to the maintained high pore pressure due to the volcanic gas [ABSTRACT FROM AUTHOR]