资源环境科技发展态势分析平台

We present a method for automatic location of dominant sources of seismovolcanic tremor in 3-D, based on the spatial coherence of the continuously recorded wavefield at a seismic network. We analyze 4.5 years of records from the seismic network at the Klyuchevskoy volcanic group in Kamchatka, Russia, when four volcanoes experienced tremor episodes. After enhancing the tremor signal with spectral whitening, we compute the daily cross-correlation functions related to the dominant tremor sources from the first eigenvector of the spectral covariance matrix and infer their daily positions in 3-D. We apply our technique to the tremors beneath Shiveluch, Klyuchevskoy, Tolbachik, and Kizimen volcanoes and observe the yearlong preeruptive volcanic tremor beneath Klyuchevskoy from deep to shallow parts of the plumbing system. This observation of deep volcanic tremor sources demonstrates that the cross-correlation-based method is a very powerful tool for volcano monitoring.