Testing of Rn application for recognizing tectonic events observed on water-tube tiltmeters in underground Geodynamic Laboratory of Space Research Centre at Książ (the Sudetes, SW Poland).

Research on relationships between variation in Rn activity concentration and tectonic events recorded using the instruments of the Geodynamic Laboratory of SRC PAS at Książ (the Sudetes, SW Poland) had been conducted since 2014. The performed analyses of variation have demonstrated the spatial character of changes in Rn activity concentration. Their time-course is comparable in all parts of the underground laboratory. This means that gas exchange between the lithosphere and the atmosphere occurs not only through fault zones but also through all surfaces of the underground workings: the floors, the sidewalls and the roofs. Further, some relationships between Rn activity concentration and tectonic activity of the orogen have been demonstrated with the use of Pearson's linear correlation coefficient. The comparison between temporal distribution (times series) of radon activity concentration and water-tube tiltmeters (WTs) demonstrated that radon data have regular oscillations which can be approximated using the sine function with a 12 month cycle (seasonal changes) and amplitude in the range of 1000-1500 Bq/m. To compare the collected radon signal data and tectonic activity, we used linear function as the simplest method of trend assessment. Pearson's correlation coefficient r cannot be accepted as appropriate for assessing the interdependencies between variables because they do not have a normal distribution, and the relationship between them is not linear. It was noted that each series of data, namely radon activity concentration and tectonic activity determine the series of deviations above and below the trend function. Because of the non-fulfillment of the above assumptions, we used nonparametric equivalents such as Spearman's rank correlation coefficient r and Kendall's tau. The obtained results showed that the value of the r coefficient ranges from 0.38 to even 0.43. The best relationship at the level of r = 0.43 was determined between the radon activity concentration recorded by detector no. 3 and the tectonic activity of the rock mass registered on the WT-2 channel. Similar at the r level of 0.37-0.38 between detector no. 5 and 4 and the WT-2 channel. A bit higher than r = 0.39 between detector no. 3 and the WT-2 channel. In each case, these were positive correlations. The obtained Spearman's r coefficients indicate the correlation between Rn activity concentration and tectonic activity of the rock mass. The t-statistic, which analyzes the significance of Spearman's coefficient r is a descriptive measure of the accuracy of regression matching to empirical data. It takes values in the range of percentage and provides informations about which part of the total variability of the radon activity concentration (Y) observed in the sample has been explained (determined) by regression in relation to tectonic activity of the rock mass (X). In our case, approximately f 40% to more than 50% of the radon activity concentration (Y) was explained by regression in relation to the tectonic activity of the rock mass. We obtained similar results with the use of Kendall's tau coefficient. Precise description of the character of this relationship requires further, more detailed analyses, such as comparing characteristics of the distributions based on trend variation like Monte Carlo simulation, Multivariate Adaptive Regression Splines or neural networks.