Non-normal distribution of radon and residual radon and short-term abnormal precursors of residual radon before major earthquakes

The basis of short-term earthquake prediction is the identification of reliable precursors. Anomaly of radon concentration escaped from groundwater (C-Rn) is often used as a traditional earthquake precursor. However, the thresholds of anomaly are still controversial. Using the data series of long-term C-Rn, the data series of residual radon concentration (C-Rr) was obtained to exclude periodic environmental interference, and the distribution of long-term C-Rn and C-Rr was tested. It is verified that the distribution of C-Rn and C-Rr does not conform with the traditional default normal distribution. Therefore, the box plot of non-normally distributed is used to redefine the anomaly threshold. The temporal and spatial analysis results of C-Rn and C-Rr show that before the Wenchuan, Lushan and Jiuzhaigou earthquakes, C-Rn and C-Rr show similar short-term anomaly patterns as follows: (1) With the approach of main shocks, the C-Rn and C-Rr near the epicenter/fault zone increase sharply and are accompanied by obvious positive anomalies. (2) The duration of anomalies is closely related to the location of the epicenter/fault zone and the observation station. (3) The anomaly of C-Rr which filtered out periodic environmental disturbance can reflect the abnormal precursor caused by tectonic geological activity before earthquake more accurately than the C-Rn anomaly. (4) The non-normal anomaly analysis method is more consistent with the mathematical definition than the normal distribution anomaly analysis method. Therefore, analyzing the abnormal threshold of residual radon which is non-normally distributed may be a reliable method for the analysis of radon anomalies before large earthquakes.