Study Of Long-Memory And Fractal Trends In Variations Of Groundwater Radon Relating To Earthquakes

Prediction of the most destructive natural phenomenon has been one of the changeling tasks for the scientific communities of all times.Either on a local scale or global,natural disasters are significantly influencing the national economy and social development.Predictability is based on recognizable precursor signals,and our ability to provide warnings of both the timing and magnitude of many natural events has served to reduce damage to life and property.All-natural hazards other than earthquake are up to some extent predictable.The earthquake precursory research started about half a century ago and remains one of the most challenging and important problems.Evidence of radon emissions in soil gas and groundwater before earthquakes provide a potential data source for seismic predictions and research proposes that specific pre-seismic anomalous variations can be directly related to changes in stresses due to seismotectonic activities.Changes in the geochemistry of groundwater preceding an earthquake are not completely unexpected.In the time leading up to an earthquake,the crust of the earth dilates and deforms to expose a fresh surface of the rock to groundwater.This may permit chemical species to be released into solution,and dilation may release gases such as radon.In general,anomalous radon variations prior to earthquakes have been abundantly documented all over the world.Although,to precise links between these radon emissions and subsequent earthquakes is still an open issue.The research reported in this thesis is the aim of enhancing earthquake prediction methodology by correlating the anomalous groundwater radon variations with the occurrence of earthquakes.Earthquake and radon data of the period was retrieved by China Earthquake Data Centre and the research project is supported by National Science Foundation of China(NSFC).The first section comprised of groundwater radon anomalies associated with various earthquakes occurred near and far from a single radon monitoring station.This section reported the studies relating to chaotic and long-memory trends hidden in radon variation in groundwater of Panzhihua,Sichuan Province,China,between 2012 and 2017.The second section highlighted the influence of the Wenchuan Earthquake(May 12,2008;Mw 7.9)and the Gansu Earthquake(July 22,2013;Ms 6.6)on groundwater radon concentration recorded at various stations installed within earthquake preparation zones.The analysis was performed using sliding-window(a)detrended fluctuation analysis(DFA),(b)fractal dimension analysis(FDA)(c)residual radon concentration(RRC)(d)Rescale-Range Hurst analysis(R/S)and multifractal detrended fluctuation analysis(MFDFA).Several fractal Brownian motion(fBm) persistent time series segments of high predictability were found,with DFA slope above 1.5 and fractal dimensions,below 1.5.Comparison of MFDFA results for original with shuffled time series distinguished the multifractality due to long-range correlation from multifractality due to broad probability density function.The q dependence of generalized Hurst h(q),classical scaling ?(q) exponents and the broadness of the spectra indicate that the radon time series has multifractal behavior.Furthermore,the width of the multifractal spectrum for seismically active and quiescent phases clarified the influence of seismicity on the radon time series.Numerous seven-day segments exhibit RRC out of the ±2? limits and are of noteworthy precursory value.A novel approach of computationally combining Residual Radon Concentration data with the different chaos analysis methods provided strong evidence regarding the preseismic nature of the investigated groundwater radon variations.All methods were applied through a novel two-stage computational approach,where the persistent fractional Brownian motion(fBm)preseismic footprints hidden in time series segments were identified and separated from remaining segments of low-predictability and earthquake-related significance.It is concluded that the methodologies adopted in current research identified changes in radon time series data before the occurrence of an earthquake.Therefore,a combinatory approach will be helpful for better understanding of the behavior of subsurface changes.Groundwater radon can be a good indicator of earthquake induced changes and can be utilized to foresee the subsurface stresses behavior.As earthquake forecasting is not an easy task,therefore,it is needed to perform multiparametric research by applying various methodologies on every single parameter to obtain more acceptable results.