Singular spectrum analysis on soil radon time series (222Rn) in Kachchh, Gujarat, India: detection of periodic oscillations and earthquake precursors

Abstract

The influence of various meteorological parameters on the observed time series of soil radon (222Rn) in the form of periodic oscillations is profound and well established. An attempt is made in the present study to identify and remove the periodic oscillations using singular spectrum analysis (SSA) to identify earthquake precursors in the time series of soil radon (222Rn) recorded at Badargadh, Kachchh region of Gujarat, India, for the period from February 21 to June 10, 2011. The temperature has shown diurnal oscillations, whereas radon and pressure show the mixture of diurnal and semi-diurnal oscillations. A Butterworth digital filter has been applied to the observed data of all parameters (222Rn, pressure, temperature, and humidity) to remove the long periodic trends (more than 30 days). The periodic and aperiodic components in the time series of soil radon are identified and separated using SSA. While correlating the radon with other parameters, we found that the aperiodic radon has negative correlation with temperature and the radon has positive correlation with humidity and pressure. Two earthquakes of magnitudes M3.7 and M4.2 which occurred on March 26, 2011, and May 17, 2011, respectively, are being investigated to identify anomalies before these earthquakes. These earthquakes are located within a distance of 100 km from Badargadh. Based on the statistical analysis of mean+2*sigma, we observed the anomalous radon signal before the occurrence of these earthquakes with relative percentage of 94–96% at Badargadh.