MACHINE LEARNING METHODS TO SOLVE THE PROBLEM OF DETECTING EARTHQUAKE PRECURSORS BASED ON SEISMIC NOISE MONITORING DATA FOR THE BAIKAL RIFT SYSTEM
MACHINE LEARNING METHODS TO SOLVE THE PROBLEM OF DETECTING EARTHQUAKE PRECURSORS BASED ON SEISMIC NOISE MONITORING DATA FOR THE BAIKAL RIFT SYSTEM
A. P. Grigoryuk,L. P. Braginskaya,V. Kovalevsky,A. A. Dobrynina
TLDR
It has been found that the change in spectral composition of MSN is statistically significant compared to background values, and the frequency range of 0.5–2.5 Hz exhibits an increase in power spectral density, whereas higher frequencies show its decrease.
Abstract
The paper deals with the monitoring-based determination of spectral composition of microseismic noise (MSN) several hours before weak and moderate seismic events in the Baikal rift system. Consideration is being given to 40 earthquakes with an energy class of K=9.5–14.5 at 10 to 120 km distances from the epicenters to the monitoring sites. It has been found that the change in spectral composition of MSN is statistically significant compared to background values. The frequency range of 0.5–2.5 Hz exhibits an increase in power spectral density, whereas higher frequencies, approximately 4 to 16 Hz, show its decrease. Machine learning methods were used to construct a binary classification model regarding the records of MSN whose spectral composition allows detecting the occurrence of short-term earthquake precursors. The studies were based on the digital platform (https://izk.sscc.ru). Seismic data is received at the digital platform from integrated monitoring of hazardous geological processes at the sites of the Institute of the Earth’s Crust SB RAS, Irkutsk.