Radon-222: A Potential Short-Term Earthquake Precursor

Petraki E1, Nikolopoulos D2*, Panagiotaras D3, Cantzos D4, Yannakopoulos P2, Nomicos C5 and Stonham J1 1Brunel University, Department of Engineering and Design, Kingston Lane, Uxbridge, Middlesex UB8 3PH, London, UK 2TEI of Piraeus, Department of Electronic Computer Systems Engineering, Petrou Ralli and Thivon 250, GR-12244 Aigaleo, Athens, Greece 3Department of Mechanical Engineering, Technological Educational Institute (TEI) of Western Greece, Alexandrou 1, 263 34 Patras, Greece 4TEI of Piraeus, Department of Automation Engineering, Petrou Ralli and Thivon 250, GR-12244 Aigaleo, Greece 5TEI of Athens, Department of Electronic Engineering, Agiou Spyridonos, GR-12243, Aigaleo, Athens, Greece


Introduction
Radon is a natural radioactive noble gas. It is generated by the decay of radium. There are thirty nine known isotopes of radon from 193 Rn to 231 Rn [1]. The most stable isotope is 222 Rn (hereafter radon) with a half-life of 3.823 days. Four isotopes, 222 Rn, 220 Rn, 219 Rn and 218 Rn occur in trace quantities in nature as decay products of, respectively, 226 Ra, 224 Ra, 223 Ra and 218 At [1,2]. 222 Rn and 218 Rn are intermediate steps in the decay chain of 238 U, 219 Rn is an intermediate step in the decay chain of 235 U and 220 Rn occurs in the decay chain of 232 Th [1][2][3]. 220 Rn is also known as thoron [1]. The half-life of thoron is 54.5 seconds [1]. Due to the short half-life, thoron disintegrates very quickly. For this reason, it is usually traced in smaller quantities compared to radon. 219 Rn is also called actinon [1]. It has lesser half-life time than 222 Rn and 220 Rn (3.92 seconds). It is traced in earth and atmosphere in smaller quantities in respect to radon and thoron [2,3]. Most of the radioactivity in the atmosphere at sea level is due to radon [3]. Radon is released primarily from the soil [1,3,4]. Approximately 10% of the radon in soil is diluted to the atmosphere [3]. Apart from soil, radon is present in fragmented rock, building materials, underground and surface waters [3,4]. While in fluids all generated radon atoms are diluted, in porous media and fragmented rock only a percentage of radon emanates, enters the volume of the pores and dissolves into the pore's fluid [1,4]. Once there, a macroscopic transport is possible, either by molecular diffusion advection or convection [1]. This transport is achieved through interconnected pores and water aquifers [4][5][6]. When the pores are saturated with water, radon is dissolved into water and is transported by it [1]. The transportation is achieved by means of fluid flow present in soil and fragmented rock [1,4,5]. Through these processes radon can travel to short, medium or long distances reaching water aquifers and air [7]. Various factors affect the whole process. The most important factors are the permeability of the soil, the temperature gradients and the pressure differences [3,7,8]. Radon is very important from radiological point of view, since it accounts for more than half of the natural exposure of the general public [2,6]. It is well known that among natural radioactivity (not man-made), the most dominant component is radon and, therefore, it is the major contributor to the effective dose equivalent.

Radon Signals and Earthquake Prediction Overview
Radon has been used as a trace gas in several studies of Earth, hydrogeology and atmosphere, because of its ability to travel to comparatively long distances from host rocks as well as the efficiency of detecting it at very low levels [9]. Significant variations of radon and progeny have been observed in geothermal fields [10], thermal spas [11], active faults [12][13][14][15][16], soil experiments [17], volcanic processes [18,19] and seismotectonic environments [5,7,17,[20][21][22][23][24][25][26][27][28][29]. Due to its importance, radon monitoring has become a continuously growing study area in the search of premonitory signals prior to earthquakes [5]. This falls more or less, in the general area of seismology where one most elusive goals is the short-term earthquake prediction [20]. By the mid-1970s the seismological community was confident that the short-term earthquake prediction would be achieved within a short period of time [20]. One area that may hold promise in advancing the science of short-term earthquake prediction is the study of earthquake precursors [20]. In fact, the short-term predictions are typically based on observations of these types of phenomena [20]. The term earthquake precursor is used to describe a wide variety of physical phenomena that reportedly precede at least some earthquakes [20]. Under this perspective, the real time radon monitoring can be viewed as an interesting possibility for credible earthquake precursors. However, the problem of earthquake prediction still remains unsolved. All the same, positive precursors recorded prior to earthquakes indicate there is evidence that they can be used for forecasting. For example, the strain changes occurring within the earth's surface during an earthquake enhance the radon concentration in soil gas [5,[25][26][27] and this renders impressive development in the study of the earth's crust which permits the estimate on the probabilities of earthquake risks [5]. In general, during earthquake rupture, certain precursory activity can be expected if the observation is made in the near vicinity of causative fracture [20]. The problem of the earthquake prediction however consists of the consecutive, step-by-step, narrowing of the time interval, space and magnitude ranges, where a strong earthquake should be expected [30,31]. In this sense, several investigators have attempted connections between earthquake-relating parameters (e.g. magnitude, precursory time, epicentral distance) and time-series characteristics (e.g. range, duration, number of radon anomalies) [5,20,[32][33][34][35][36].
The prediction of earthquakes is usually distinguished in five stages. The background stage provides maps with the territorial distribution of the maximum possible magnitude and recurrence time of destructive earthquake of different magnitudes. The four subsequent stages, fuzzily divided, include the time prediction; they are as follows [31]: longterm (10 1 years); intermediate-term (1 year); short-term (10 -1 to 10 -2 years), and immediate-term (10 -3 years or less). Such division into stages is dictated by the character of the process that leads to a strong earthquake and by the needs of earthquake preparedness; the latter comprises an arsenal of safety measures for each stage of prediction [30]. According to the classification of Hayakawa and Hobara [32] the prediction of earthquakes is grouped into three categories: long-term (timescale of 10 to 100 years); intermediate-term (time-scale of 1 to 10 years); short-term. Note, that even in short-term prediction there is no one-to-one correspondence between anomalies in the observations and the earthquake events [25][26][27]33]. Although much more difficult than the long-term and intermediate-term predictions, the short-term prediction of earthquakes on a time-scale of hours, days or weeks, is believed to be of the highest priority for social demands in seismoactive countries.
In the following, specific scientific evidence is presented regarding the possibilities of forecasting of earthquakes in terms of monitoring of radon gas emissions. The analysis is focused is on the short-term precursors of general failure since these are considered of higher prognostic value in terms of societal demands.

Radon Gas Emission and Pre-Earthquake Activity
In the late 1960s and early 1970s reports primarily from Russia and China indicated that concentrations of radon gas in the earth apparently changed prior to the occurrences of nearby earthquakes [34]. This stimulated a number of experiments in other parts of the world to monitor underground radon with time and to look for radon changes associated with earthquakes [20]. Table 1 presents a collection of relevant important data including: (1) the earthquake details; (2) the % (δα) disturbance or detected disturbances in radon concentration; (3) the duration of the detected anomaly or the recorded anomalies; (4) the precursory time; (5) the epicentral distance (4) the related references from 1980 and after.
In general, the anomalous radon variations observed prior to earthquakes have been reported in groundwater, soil gas, atmosphere and thermal spas [5,20,21,28,[57][58][59][60][61][62]. The seismological data of Table  1 are related to radon concentration data of wide fluctuations, peaks and downturns [25][26][27]. The earthquake-related connections of Table  1, namely the connections between the magnitude, the precursory time and the epicentral distance with the time-series characteristics, viz., the range, the duration and the number of radon anomalies vary significantly [5,20,33,34]. For example, the reported precursory times range from 3 months to some days prior to the earthquake event, whilst the epicentral distances range between 10 and 100 km. Similar ranges have been published also in [20] and [5] (please see also references therein). It is very important to note here that many precursory signals have been derived only with passive techniques [25][26][27] which integrate radon concentrations over long time intervals (at least >1-4 weeks), i.e., they provide coarse time-series estimations. This is a significant disadvantage for the reported estimations. On the other hand, the reported precursory signals with active techniques are limited. Note that the active techniques enable high radon recording rates (between 1/min and 1/hour) and in this manner they provide fine radon signals [5,9,16,20,[25][26][27]. Important is that there are also other parameters that affect and alter the radon-earthquake estimations. For example, radon concentration levels are influenced by geological and geophysical conditions, the seasonal variations and atmospheric changes such as the rainfall and the barometric pressure alterations (please see e.g. [1,2,20,[25][26][27]33,34]. For this reason the related timeseries data are usually presented in parallel to the radon precursory signals [5,20,[25][26][27]. As can be observed from Table 1, the majority of the associations between radon and earthquakes are based on events of small or intermediate magnitudes. This restricts the estimations more, since, up-to-date there seems not to exist, not only for the mild, but even for the intense earthquakes, a universal model to serve as a signature of a specific forthcoming seismic event [38][39][40][41][42][43][44][45][46][47][48][49][50][51][52][53][54][55][56][57]. Most of the disturbances of Table 1 were determined in terms of visual or simple statistical analysis. The most usual statistical criterion employed is the ±2σ one. Through this criterion, a radon disturbance is identified as such if it contains parts above the ±2σ zone. Although simple, this approach was used extensively in many papers of Table 1. Only few signals of Table 1 were analyzed through advanced techniques [25][26][27]57,58]. These reports are recent. They were published in the last five years, namely 2012 and after. Worth to notice is that the analysis was implemented in fine active signals recorded after significant earthquakes of near epicenters [25][26][27] fact which enhances the estimates of these reports further. A fact that reinforces the estimates of these reports is that the corresponding radon disturbances lasted long, i.e., between five and fifteen days. One of the advanced techniques of these reports [25][26][27]57,58] is the temporal Fractal Analysis based on a windowed version of the short-time wavelet transform of the density of the power spectrum in each window [25][26][27]. Note that the method was applied in both mono-fractals [25][26][27]57] and multifractals [58]. Another Taiwan advanced approach is the detrended fluctuation analysis (DFA) [27,57].
According to the reports [27,57] and several other papers [38,51] the detrended fluctuation analysis is the most advantageous technique to trace the long-memory of a system driven to rupture. Significant other techniques are the time-evolution of the fractal dimension [26] and the Hurst exponent [25][26][27]57] and the temporal changes of various metrics of entropy [26]. Note that the techniques can trace patterns of longmemory that are hidden in the pre-earthquake time-series. They can also identify features related to the self-organization of the earthquake generating system. It is also important to note that the vast majority of papers of Table 1 refer to measurement of radon in soil. Only some papers refer to radon in underground or thermal water and only one to radon detected in atmosphere prior to earthquakes [62]. Note that in this paper advanced Fourier based approach was implemented for a significant long-lasting signal retrieved prior to the Kobe earthquake, Japan.
Various physical mechanisms have been reported to relate the sub-surface physical changes with the variation in radon emanations [25]. Regarding modelling, the available models propose explanations in terms of strain changes within the earth's crust during preparation of earthquakes [5,33,34,40]. It is the displacement of rock mass under tectonic stress that opens up various pathways and exposes new surfaces when cracks open. The stress-strain developed within the earth's crust before earthquakes leads to changes in gas transportation from the deep earth to surface [41,42]. As a result, unusual quantities of radon emerge out of the pores and fractures of the rocks on the surface. Due to the seismic activity, changes in underground fluid flow may also render anomalous changes in concentration of radon and its progeny [43]. Under the so called compression model, according to [63] and [64] a small change in velocity of gas into or out of the ground causes a significant change in radon concentration at shallow soil depth as changes in gas flow disturb the strong radon concentration gradient that exists between the soil and the atmosphere. A slight compression of pore volume causes gas to flow out of the soil resulting to an increase in radon level. Similarly, when pore volume increases, gas flows into the soil from the atmosphere. Thus, an increased radon concentration occurs in the region of compression and radon concentration decreases in the region of dilation. As small changes in gas flow velocity causes significant change in radon concentration, soil radon monitoring is thus an important way to detect the changes in compression or dilation associated with an earthquake event.
Among the various theoretical models, the dilatancy diffusion model proposed by Martinelli [5,65] is a noteworthy approach. According to this model [5,[25][26][27] the earthquake generating medium is considered to consist of porous cracked saturated rocks. When a tectonic stress develops, the cracks extend and appear near the pores with the opening of favourably oriented cracks [5,[25][26][27]. As a result, the pore pressure decreases in the total preparation zone and water from surrounding medium diffuses into the zone. At the end of the diffusion period the main rupture occurs due to the appearance of pore pressure and increase in cracks [5,[25][26][27].
A well-accepted model is the the Crack-Avalanche model [5,[25][26][27]66]. According to the Crack-Avalanche model as tectonic stress increases during the earthquake preparation, a zone of cracked rocks is formed in the region of a future earthquake focal zone under the influence of the tectonic stresses. In the study of the surrounding medium this region may be considered as a solid inclusion with altered moduli. The inclusion appearance causes a redistribution of the stresses accompanied by corresponding deformations. As the tectonic stresses change with time, the shape and size of the zone change as well. According to the theory of stress corrosion, the anomalous behavior of radon concentration may be associated with this slow crack growth, which is controlled by the stress corrosion in the rock matrix saturated by groundwater. A very recent model has been proposed by [25][26][27] based on the aspects expressed by [38,39]. This model is called asperity model. According to the asperity model, the focal area consists of a backbone of strong and large asperities that sustain the earthquake-generating system. A strongly heterogeneous medium surrounds the family of strong asperities. The fracture of the heterogeneous system in the focal area obstructs the backbone of asperities. At this stage, critical antipersistent MHz electromagnetic anomalies and radon anomalies occur [25][26][27]38,39].
Comparing the aforementioned models, it can be claimed that as an earthquake approaches a region of several cracks is formed [8]. The  Table 1: Earthquake precursory data based on radon gas. Earthquake data, % of disturbance in reference to the baseline values (δα) or technique of analysis, together with duration of measurements, reported precursory time and effective distance (ED) from the epicenter of the earthquake and related references. earthquake is associated with deformations and as a result short or long term precursory phenomena like anomalies in radon concentration may occur. As mentioned already, radon can be considered as a shortterm earthquake precursor. Nevertheless, no universal model exists to serve as pre-earthquake signature [25][26][27]38,39]. Moreover, there is no definite rule to link any kind of pre-earthquake anomaly to a specific forthcoming seismic event, either if this is intense or mild [25][26][27]38,39]. For these reasons, despite the fairly abundant circumstantial evidence, the scientific community still debates the precursory value of premonitory anomalies detected prior to earthquakes [38]. On the other hand, well established criteria exist to identify pre-earthquake patterns hidden in time-series, which are based on the concepts of fractality, self-organisation, non-extensivity and entropy [25][26][27]. Especially according to [38], certain questions still remain: (i) How can a certain observation be recognised as pre-seismic? (ii) How can an individual precursor be linked to a distinctive stage of an earthquake preparation process? (iii) How can certain precursory symptoms in anomalous observations be identified so as to indicate that the occurrence of an earthquake is unavoidable? The above issues clearly indicate that radon monitoring in soil is a very important field of research from geological point of view.