ИСТИНА

In studying of the physical processes occurring in volcanoes and identifying upcoming eruptions, the main source of information about active processes in deep parts of volcanic systems is provided by geophysical observations. Nowadays the leading geophysical method is the seismological monitoring. Seismic manifestations of deep volcanic activity, so called volcanic earthquakes, are very numerous and diverse. Observed seismic signals correspond to different types of volcanic seismicity at different depths. Timely detection of these signals and their comparison with various activities is the basis of seismological monitoring of volcanoes. Volume of seismological data (seismograms) available for analysis is constantly increasing, which makes it possible to improve the quality of seismological monitoring. At the same time, a full analysis of all the available data requires the development and application of new automated methods for detecting and classifying recorded volcanic earthquakes and for measuring their main characteristics to assess the dynamic state of volcanoes and to predict the evolution of their activity. In this work we focus on Deep Long Period (DLP) earthquakes occurring in the vicinity of the crust mantle boundary below the Klyuchevskoy volcanic group in Kamchatka. This type of seismicity is believed to reflect the state of the deep magmatic reservoir and, therefore, this has been suggested that it can be used as one of the earliest precursors of eruptions. At the same time, the physical mechanism of these DLP earthquakes remains poorly understood. Therefore, this is necessary to analyze available seismic records in order to try to reconstruct focal mechanism and the spectral characteristics of the DLP sources. Additional important information about the physical processes generating these earthquakes can be obtained from studying their frequency size distribution and the scaling between their magnitudes and durations. In this work we present an automatic approach to the analysis of continuous seismic records by multiple stations in order to detect the DLP and to measure their relative magnitudes. This analysis strongly relies on the multiplet behavior of the DLP signals.