Assesment of the response of the meteorological/hydrological parameters on the soil gas radon emission at Hsinchu, northern Taiwan: A prerequisite to identify earthquake precursors

Highlights

• The response of meteorological parameters on soil radon emission is assessed and quantified.

• Radon and Ground Water Head data from the Hsinchu monitoring station is used.

• Singular Spectrum Analysis and BAYTAP-G statiscial filters are used to remove noise.

• Data is simulated using: Tank Model, Exponential Decay and Double Exponential Decay.

Abstract

The present study is an attempt to assess and quantify the influence of the meteorological (atmospheric temperature and pressure) and hydrological (rainfall and ground water head-GWH) parameters on the soil gas radon emission at Hsinchu, northern Taiwan. The quasi-periodic variations corresponding to diurnal and semi diurnal periods were estimated and eliminated by decomposing the time series for the period of September 16, 2009 to March 5, 2010 to singular spectrum analysis. The reconstructed non-periodic variations, which reproduce the salient feature of recorded time series, were searched for meteorological/hydrological influences in radon emission. The combined response of barometric pressure and atmosphere temperature are found to be small when compared to the total variability in radon. The influence of rainfall on radon is found to be strongest. At the onset of rainfall, radon shows a step-jump that attains peak with a time lag of 12–15 h. This enhancement is attributed to entrapment of soil gas in the top soil cover as increased soil moisture prevents escape of radon into the atmosphere (capping effect). The decay of radon after the recession of rainfall is approximated by double exponential decay terms, one corresponding to the natural decay of radon with half life of 3.84 days and second representing slow weakening of capping effect. The third effect related to internal loading due to rise and fall of groundwater modulates the propagation of radon in overlying strata, accounting for the long term variations in radon. The rainfall inflicted changes in radon look strikingly similar to earthquake related precursory or co-seismic perturbations, inferred by long term synotopic observations. It is surmised that unless radon variations are corrected for meteorological/hydrological contamination, some precursory signals are masked on one hand while on the other hand some anomalies are falsely viewed as earthquake precursors.

Introduction

As the stress level increases during earthquake preparatory cycle, the rocks in the impending focal zone experience opening of micro-cracks (Scholz et al., 1973). With the opening of cracks, exposed surface area of rocks also increases that leads to enhanced emanation of radon (Thomas, 1988). If this increase in radon intensity can be measured at the surface, it can serve as a possible precursor to earthquakes. With the realisation of the physical mechanism, some of the early field examples, e.g. 1966 M5,5 Tashknet earthquake (Ulomov and Mavashev, 1967), reporting several fold increase in radon in association earthquake occurrences has given major impetuous for initiating the monitoring of radon across the word wide seismic active (Barbosa et al., 2015, Heinicke et al., 1992, Igarashi et al., 1995, Virk et al., 2001, Steinitz et al., 2003, Zmazek et al., 2005, Kumar et al., 2009, Kumar et al., 2015). So rapid is the growth, that number of reviews have appeared at regular time interval, attempting to characterize the nature of precursors, their physical validation or to establish a scaling relation of anomalous changes in radon with the magnitude and distance of impending earthquakes (Wakita et al., 1988, Thomas, 1988, Toutain and Baubron, 1999, Hartmann and Levy, 2005, King et al., 2006, Cicerone et al., 2009). Despite the rapid growth, yet no universal ones to isolate and document the diagnostic signature of a precursory signal in radon definition of earthquake precursory signal has emerged and hence radon precursors are seldom used in real time forecasting of earthquakes. Further, even if reported radon precursors are real, how they are physically related to the preparation process of an impending earthquake remain an open issue (Woith, 2015). More intriguing is the observation that there is apparent negative correlation between the numbers of reported anomalies and the length of the time series processed to isolate precursory signals (Woith, 2015). In other words, most of the so claimed success stories that have often emerged from short lengths of data and hence it is likely that influence of non-dynamic factors influencing radon variability may not be neutralised. However, as a result of continuing monitoring in different environments, it has been well recognised that the propagation and transportation of radon from the focal zone to surface is controlled by diffusion (gradient in concentration), advection (pressure- Clements and Wilkening, 1974), convection (temperature- Finkelstein et al., 2006) processes. In addition, modulation of soil moisture (capping effect- Schumann et al., 1988, Schumann et al., 1992), squeezing of water (rinsing effect- Hesselbom, 1985) following events of rainfall produce strong contamination in radon intensity (Schumann et al., 1992, Fujiyoshi et al., 2006). The real assessment and quantification of these influences is a major pre-requisite to isolate weak precursory and co-seismic signals. The present study is a fresh attempt to quantify the nature and extent of influence of meteorological and hydrological parameter on radon emission in the seismically active belt of Taiwan, where dense network of soil gas radon monitoring network has been in operation for a long time (Walia et al., 2009a).