Non-parametric detection of atmospheric radon concentration anomalies related to earthquakes

Detection of atmospheric radon concentration anomalies and their potential for earthquake prediction using Random Forest analysis

Various anomalies occurring before earthquakes are currently being studied to predict seismic events, with one of them being the radioactive element radon (222Rn). Radon concentrations in the soil, water, and atmosphere fluctuate in response to crustal movement. Recent research has statistically detected anomalies by analyzing the fluctuations in radon concentrations before earthquakes and conducting quantitative evaluations of radon. However, the method used to determine the parameters in the analysis was problematic. Therefore, in this study, we compared observed atmospheric radon concentration data with predicted values based on typical annual patterns using Random Forest analysis. We conducted a more objective analysis by employing this method and statistically determining anomalies using thresholds. This analysis was conducted using atmospheric radon concentration observation data obtained at Kobe Pharmaceutical University (KPU) before the 1995 Kobe Earthquake, and ionization currents emitted when radon decays were obtained at Fukushima Medical University (FMU) before the 2011 Tohoku-oki Earthquake. Consequently, before the major earthquakes occurred at both locations, the difference between the predicted and observed values exceeded the standard deviation by a factor of three. These results indicate the potential of Random Forest analysis to identify anomalies in atmospheric radon concentrations before earthquakes occur.

Earthquake precursors: A review of key factors influencing radon concentration

Many factors influence the accurate identification of radon anomalies triggered by earthquakes to varying degrees. Therefore, this paper primarily provides a comprehensive review of the various factors influencing radon concentrations over the past two decades. In addition to examining the individual effects of these factors on radon concentrations, it explores the interactions among multiple factors, particularly the correlations among radon anomalies and seismic events as well as the environmental context. This review mainly includes the classification of groundwater radon anomalies and their potential formation mechanisms, the environmental impact on radon concentrations, the effects of soil and rock structures on radon migration, and the application of machine learning in detecting radon anomalies induced by earthquakes.

Radon transport carried by geogas: prediction model

This paper provides an overview and information on radon migration in the crust. In the past several decades, numerous studies on radon migration have been published. However, there is no there is no comprehensive review of large-scale radon transport in the earth crust. A literature review was conducted to present the research on the mechanism of radon migration, geogas theory, investigation of multiphase flow, and modeling method of fractures. Molecular diffusion was long considered the primary mechanism for radon migration in the crust. However, a molecular diffusion mechanism cannot explain the understanding of anomalous radon concentrations. In contrast with early views, the process of radon migration and redistribution within the Earth may be determined by geogas (mainly CO2 and CH4). Microbubbles rising in fractured rocks may be a rapid and efficient way of radon migration, as reported by recent studies. All these hypotheses on the mechanisms of geogas migration are summarized into a theoretical framework, defined as "geogas theory." According to geogas theory, fractures are the principal channel of gas migration. The development of the discrete fracture network (DFN) method is expected to supply a new tool for fracture modeling. It is hoped that this paper will contribute to a deeper understanding of radon migration and fracture modeling.

Estimating atmospheric radon deviation using statistical coefficients: Sulaymaniyah city, Iraq, as a case of study

The authors studied the atmospheric radon concentration with associated meteorological parameters variation during the dust events from July to November 2017. We obtained the meteorological parameters data in weather station of Sulaymaniyah city, Iraq. In the environmental monitoring plan, the atmospheric radon fluctuated from 15 to 48 Bq m^-3 around the mean value of 31.5 ± 7 Bq m^-3 within the summer. In autumn, varied from 22 to 46 Bq m^-3 with a mean value of 34 ± 12 Bq m^-3. We employed this to determine the radon level anomalously. Using the modified statistical coefficients, such as the residual deviation (RD), residual fluctuation ratio (RFR), F-test, and p-value coefficients. Among the atmospheric radon fluctuation values, particularly one anomalous (42 Bq m^-3) on 25 July was determined because the excessive value of the RD was 1.9 σ, and the RFR value was 66 %. Corresponding to our coefficients criteria, the minimum level of atmospheric radon (22 Bq m^-3) does not consider anomalous because of increasing wind speed. Based on this, our method for determining the atmospheric radon anomalies that are influenced by the missed factors beyond the mentioned meteorological parameters is accurate.

Radon Dynamics and Effective Dose Estimation in a Touristic Volcanic Cave: La Cueva del Viento, Tenerife (Canary Islands, Spain)

La Cueva del Viento is a volcanic lava tube located in Tenerife Island (Canary Islands, Spain). Its touristic section, 180 m long, receives more than 28,200 visitants each year. According to the European and Spanish legislation, a radon monitoring program is required to minimize the radon exposition of workers, tourists, and cavers. In this work, we studied the radon concentration dynamics in the touristic section of the cave for ca. 1 year, using both passive and active radon detectors. Pluviometry and external air temperature played an important role in the seasonal and daily variations of indoor radon concentrations. Daily fluctuations during the dry season were analyzed using time series (Box-Jenkins methodology) and frequency analysis (Fourier and Wavelet transforms) methods. The experimental radon time-series was well-fitted using a seasonal autoregressive integrated moving average model: Seasonal Auto-Regressive Integrated Moving Average (2,0,1) (2,1,0)₂₄, and its value, in a short-time window (ca. 1 week) was conveniently forecasted. Finally, this work revealed that the annual effective doses received, during the observation period (1 year), by the touristic guides and visitors was ca. 2 mSv/yr and 4 μSv/hr, respectively. We concluded that the touristic exploitation of La Cueva del Viento is safe for both tourists and guides. However, based on our results, La Cueva del Viento had to be classified as a "Monitoring zone" and a regular monitoring program should be implemented.

Observations on the stress related variations of soil radon concentration in the Gulf of Corinth, Greece

Our observations indicate a characteristic pattern in the long-term variation of soil radon concentrations, which seems to be consistent with the expected variation of regional stress in relation to seismicity. However, it seems that the major changes in radon level begin before the rock rapture, i.e. before the earthquake occurs. These conclusions have emerged after long-term observations with continuous and thorough real-time gamma-radiation monitoring in the seismically active area of the Gulf of Corinth, Greece. The recordings acquired close to a hot spring were of very high quality, implying that the deep hydraulic flow can possibly play a key role in the pre-earthquake variation of radon level. We were able to observe outstanding examples of radon level variations before significant seismic events in the Gulf of Corinth that cannot be attributed to other external factors such as atmospheric phenomena.

Preseismic atmospheric radon anomaly associated with 2018 Northern Osaka earthquake

Despite the challenges in identifying earthquake precursors in intraplate (inland) earthquakes, various hydrological and geochemical measurements have been conducted to establish a possible link to seismic activities. Anomalous increases in radon (²²²Rn) concentration in soil, groundwater, and atmosphere have been reported prior to large earthquakes. Although the radon concentration in the atmosphere is lower than that in groundwater and soils, a recent statistical analysis has suggested that the average atmospheric concentration over a relatively wide area reflects crustal deformation. However, no study has sought to determine the underlying physico-chemical relationships between crustal deformation and anomalous atmospheric radon concentrations. Here, we show a significant decrease in the atmospheric radon concentration temporally linked to the seismic quiescence before the 2018 Northern Osaka earthquake occurring at a hidden fault with complex rupture dynamics. During seismic quiescence, deep-seated sedimentary layers in Osaka Basin, which might be the main sources of radon, become less damaged and fractured. The reduction in damage leads to a decrease in radon exhalation to the atmosphere near the fault, causing the preseismic radon decrease in the atmosphere. Herein, we highlight the necessity of continuous monitoring of the atmospheric radon concentration, combined with statistical anomaly detection method, to evaluate future seismic risks.

Radon degassing triggered by tidal loading before an earthquake

The presence of anomalous geochemical changes related to earthquakes has been controversial despite widespread, long time challenges for earthquake prediction. Establishing a quantitative relationship among geochemical changes and geodetical and seismological changes can clarify their hidden connection. Here we determined the response of atmospheric radon (²²²Rn) to diurnal tidal (K1 constituent) loading in the reported 11-year-long variation in the atmospheric radon concentration, including its anomalous evolution for 2 months before the devastating 1995 Kobe earthquake in Japan. The response to the tidal loading had been identified for 5 years before the occurrence of the earthquake. Comparison between these radon responses relative to crustal strain revealed that the response efficiency for the diurnal K1 tide was larger than that for the earthquake by a factor of 21–33, implying the involvement of crustal fluid movement. The radon responses occurred when compressional crustal stress decreased or changed to extension. These findings suggest that changes in radon exhaled from the ground were induced by ascent flow of soil gas acting as a radon carrier and degassed from mantle-derived crustal fluid upwelling due to modulation of the crustal stress regime.

ANALYSIS OF RADON TIME SERIES RECORDED IN SLOVAK AND CZECH CAVES FOR THE DETECTION OF ANOMALIES DUE TO SEISMIC PHENOMENA

Anomalies in the radon (222Rn) releases in underground environments are one of the phenomena that can be observed before earthquake occurrence. Continuous measurements of radon activity concentration, and of meteorological parameters that influence the gas emission, were performed in three Slovak and Czech caves during 1-y period (1 July 2016-30 June 2017). The radon activity concentration in caves shows seasonal variations, with maxima reached during summer months. The anomalies in the radon time series are identified using a combination of three mathematical methods: multiple linear regression, empirical mode decomposition and support vector regression. The radon anomaly periods were compared with earthquake occurrences in Europe. Coincidences between both phenomena were found, since all monitored caves reflect contemporaneous local tectonic changes. The results indicate that radon continuous monitoring could assist a better understanding of radon emissions, along active tectonic structures, during seismic events.