**1. Introduction**

178 Earthquake Research and Analysis – Statistical Studies, Observations and Planning

Yu, S. B. & Kuo, L. C. (2001). Present-day crustal motion along the Longitudinal Valley

Radon is one of many geophysical and geochemical phenomena that can be considered to be an earthquake precursor. Due to the non-linear dependence of earthquakes' initial conditions, the question about the predictability of earthquakes often arises (Geller, 1997). The successful prediction of earthquakes is yet to be accomplished, in terms of their magnitude, location and time, and much effort has been spent on this goal.

The term "earthquake precursor" is used to describe a wide variety of geophysical and geochemical phenomena that reportedly precede at least some earthquakes (Cicerone et al., 2009). The observation of these types of phenomena is one recent research activity which has aimed at reducing the effects of natural hazards. Among the different precursors, geochemistry has provided some high-quality signals, since fluid flows in the Earth's crust have a widely recognised role in faulting processes (Hickman et al., 1995). The potential of gas geochemistry in seismo-tectonics has been widely discussed by Toutain and Baubron (1999).

In the late 1960s and early 1970s, reports from seismically active countries such as the former USSR, China, Japan and the USA (Ulomov & Mavashev, 1967; Wakita et al., 1980) indicated that concentrations of radon gas in the earth apparently changed prior to the occurrence of nearby earthquakes (Lomnitz, 1994). The noble gas radon (222Rn) originates from the radioactive transformation of 226Ra in the 238U decay chain in the Earth's crust. Since radon is a radioactive gas, it is easy and relatively inexpensive to monitor instrumentally, and its short half-life (3.82 days) means that short-term changes in radon concentration in the earth can be monitored with a very good time resolution. Radon emanation from grains depends mainly on their 226Ra content and their mineral grain size, its transport in the earth being governed by geophysical and geochemical parameters (Etiope & Martinelli, 2002), while exhalation is controlled by hydrometeorological conditions. The stress-strain developed within the Earth's crust before an earthquake leads to changes in gas transport and a rise of volatiles from the deep earth up to the surface (Ghosh et al., 2009; Thomas, 1988), resulting in anomalous changes in radon concentration. The mechanism of observed radon anomalies is still poorly understood, although several theories have been proposed (Atkinson, 1980; King, 1978; Lay et al., 1998; Martinelli, 1991). Over the past three decades, the occurrence of anomalous temporal

Radon as an Earthquake Precursor – Methods for Detecting Anomalies 181

Radon concentration in soil, gas or water is not only controlled by geophysical parameters, but it also changes due to other external effects. Meteorological effects – such as soil humidity, rainfall, temperature, barometric pressure and wind – control radon concentrations in soil gas. These parameters change the physical characteristics of soil and rock, thus influencing the rate of radon transport and, consequently, perturbing eventual radon variations caused by geophysical processes originating in the deeper parts of the Earth's crust. Shallow soil levels are more affected by changing meteorological conditions than deeper ones. Radon concentrations with no larger variations present are usually observed at depths of 0.8 m or deeper. Besides the effects of meteorological parameters on radon in soil gas, considerable variations in the gas composition of thermal springs have been shown to be the result of fluctuations of local hydrologic regimes (Klusman &

The significant influence of barometric pressure has been discussed by several authors, who clearly pointed out an inverse relationship between barometric pressure and radon concentration in soil gas (Chen et al., 1995; Clements & Wilkening, 1974; Klusman & Webster, 1981). A decrease in barometric pressure, with the values of other environmental parameters remaining constant, generally causes an increase in radon exhalation from the ground, whereas during periods of rising pressure, air with low radon concentration is forced into the ground, thus diluting radon. Temperature-related fluctuations of soil gas radon concentration have also been proven to be very important. Klusman & Jaacks (1987) found an inverse relationship between soil temperature and radon concentration. They suggested that lower air temperatures as compared with soil temperatures during winter months promoted an upward movement of radon by convection, whereas during the summer, lower soil temperatures as compared with air temperatures and an inversion layer below the level of sampling reduces the upward flux and observed concentration. In general, the behaviour of soil gas migration in different types of soil is seasonally dependent (King & Minissale, 1994; Washington & Rose, 1990). In systems where gas movement is driven by diffusion or slow advection processes, radon activity in soil might be controlled by soil moisture and rainfall through the opening of cracks in the surface (Pinault & Baubron, 1996; Toutain & Baubron, 1999). On the other hand, barometric pressure has the major influence on radon concentrations in soils in advective systems, which display generally higher gas flows. However, micro-scale soil heterogeneities in permeability, porosity and lithology can cause significant heterogeneities in the response of radon concentration to changes of atmospheric parameters (King & Minissale, 1994; Neznal et al., 2004). Numerous and often divergent results in studies related to the effect of external factors on soil gas radon concentration suggest that no general predictive model for excluding meteorological effects can be proposed, and studies of radon in soil gas need a

Both mechanisms of radon transport – diffusion and advection – depend on both soil porosity and permeability, which at the same time vary as a function of the stress field (Holub & Brady, 1981). However, migration by diffusion is negligible, where a component of advective long-distance transport exists (Etiope & Martinelli, 2002). The high permeability

**2.1 External effects on radon in soil gas and water** 

simultaneous record of meteorological parameters.

**3. Anomalous radon concentration and seismicity** 

Webster, 1981).

changes of radon concentrations has been studied by several authors specialising in soil gas (King, 1984, 1985; Kuo et al., 2010; Mogro-Campero et al., 1980; Planinić et al., 2001; Ramola et al., 2008; Ramola et al., 1990; Reddy & Nagabhushanam, 2011; Walia et al., 2009a; Walia et al., 2009b; Yang et al., 2005; Zmazek et al., 2005; Zmazek et al., 2002b) and groundwater (Barragán et al., 2008; Favara et al., 2001; Gregorič et al., 2008; Heinicke et al., 2010; Kuo et al., 2006; Ramola, 2010; Singh et al., 1999; Zmazek et al., 2002a; Zmazek et al., 2006). However, radon anomalies are not only controlled by seismic activity but also by meteorological parameters like soil moisture, rainfall, temperature and barometric pressure (Ghosh et al., 2009; Stranden et al., 1984). This makes it complicated and, for small earthquakes, often impossible to differentiate between those anomalies caused by seismic events and those caused solely by atmospheric changes. Therefore, the application of theoretical and empirical algorithms for removing meteorological effects is necessary (Choubey et al., 2009; Ramola et al., 2008; Ramola et al., 1988; Torkar et al., 2010; Zmazek et al., 2003). In this chapter, the different approaches to distinguishing between those anomalies in radon time series caused by seismic activity and those caused solely by hydrometeorological parameters are presented and discussed.

#### **2. Radon migration in the Earth's crust**

Only a fraction of the radon atoms created by radioactive transformation from radium are able to emanate from mineral grains and enter into the void space, filled either by gas or water. Radon ascends towards the surface mainly through cracks or faults, on a short scale by diffusion and, for longer distances, by advection - dissolved either in water or in carrier gases. Gas movement should be ascribed to the combination of both processes. Diffusive movement is driven by a concentration gradient and is described by Fick's law. Considering gas diffusion in porous media, it is necessary to take into account that the volume through which gas diffuses is reduced and the average path length between two points is increased, thus altering the diffusion coefficient (Etiope & Martinelli, 2002). Nevertheless, the velocity of radon transport in the earth is quite low (≤ 10-3 cm/s) and the concentration of radon is reduced by radioactive decay to the background level before even 10 m are traversed (Etiope & Martinelli, 2002; Fleischer, 1981). Diffusion is only important in capillaries and small-pored rocks. On the other hand, the velocity and space scales of advective movements are much higher than those of diffusive ones. Advective transport is driven by pressure gradients, following Darcy's law. The amount of radon itself is, however, too small to form a macroscopic quantity of gas which can react to pressure gradients. Therefore, it must be carried by a macroscopic flow of carrier gases (Kristiansson & Malmqvist, 1982). A gas mixture formed by carrier gases (e.g., CO2, CH4, and N2) and rare gases (e.g., He, Rn) can be referred to as "geogas" (Etiope & Martinelli, 2002; Kristiansson & Malmqvist, 1982). In dry, porous or fractured media, gas flows through an interstitial or fissure space (gas-phase advection) whereas in saturated, porous media gas can dissolve and then be transported in three ways: by groundwater (waterphase advection), by displacing water (gas-phase advection) or by forming a bubble flow by means of buoyancy in aquifers and water-filled fractures. The bubble movement has been theoretically and experimentally recognised as a fast gas migration mechanism governing the distribution of carrier and trace gases over wide areas on the Earth's surface (Vàrhegyi et al., 1992).

#### **2.1 External effects on radon in soil gas and water**

180 Earthquake Research and Analysis – Statistical Studies, Observations and Planning

changes of radon concentrations has been studied by several authors specialising in soil gas (King, 1984, 1985; Kuo et al., 2010; Mogro-Campero et al., 1980; Planinić et al., 2001; Ramola et al., 2008; Ramola et al., 1990; Reddy & Nagabhushanam, 2011; Walia et al., 2009a; Walia et al., 2009b; Yang et al., 2005; Zmazek et al., 2005; Zmazek et al., 2002b) and groundwater (Barragán et al., 2008; Favara et al., 2001; Gregorič et al., 2008; Heinicke et al., 2010; Kuo et al., 2006; Ramola, 2010; Singh et al., 1999; Zmazek et al., 2002a; Zmazek et al., 2006). However, radon anomalies are not only controlled by seismic activity but also by meteorological parameters like soil moisture, rainfall, temperature and barometric pressure (Ghosh et al., 2009; Stranden et al., 1984). This makes it complicated and, for small earthquakes, often impossible to differentiate between those anomalies caused by seismic events and those caused solely by atmospheric changes. Therefore, the application of theoretical and empirical algorithms for removing meteorological effects is necessary (Choubey et al., 2009; Ramola et al., 2008; Ramola et al., 1988; Torkar et al., 2010; Zmazek et al., 2003). In this chapter, the different approaches to distinguishing between those anomalies in radon time series caused by seismic activity and those caused solely by

Only a fraction of the radon atoms created by radioactive transformation from radium are able to emanate from mineral grains and enter into the void space, filled either by gas or water. Radon ascends towards the surface mainly through cracks or faults, on a short scale by diffusion and, for longer distances, by advection - dissolved either in water or in carrier gases. Gas movement should be ascribed to the combination of both processes. Diffusive movement is driven by a concentration gradient and is described by Fick's law. Considering gas diffusion in porous media, it is necessary to take into account that the volume through which gas diffuses is reduced and the average path length between two points is increased, thus altering the diffusion coefficient (Etiope & Martinelli, 2002). Nevertheless, the velocity of radon transport in the earth is quite low (≤ 10-3 cm/s) and the concentration of radon is reduced by radioactive decay to the background level before even 10 m are traversed (Etiope & Martinelli, 2002; Fleischer, 1981). Diffusion is only important in capillaries and small-pored rocks. On the other hand, the velocity and space scales of advective movements are much higher than those of diffusive ones. Advective transport is driven by pressure gradients, following Darcy's law. The amount of radon itself is, however, too small to form a macroscopic quantity of gas which can react to pressure gradients. Therefore, it must be carried by a macroscopic flow of carrier gases (Kristiansson & Malmqvist, 1982). A gas mixture formed by carrier gases (e.g., CO2, CH4, and N2) and rare gases (e.g., He, Rn) can be referred to as "geogas" (Etiope & Martinelli, 2002; Kristiansson & Malmqvist, 1982). In dry, porous or fractured media, gas flows through an interstitial or fissure space (gas-phase advection) whereas in saturated, porous media gas can dissolve and then be transported in three ways: by groundwater (waterphase advection), by displacing water (gas-phase advection) or by forming a bubble flow by means of buoyancy in aquifers and water-filled fractures. The bubble movement has been theoretically and experimentally recognised as a fast gas migration mechanism governing the distribution of carrier and trace gases over wide areas on the Earth's

hydrometeorological parameters are presented and discussed.

**2. Radon migration in the Earth's crust** 

surface (Vàrhegyi et al., 1992).

Radon concentration in soil, gas or water is not only controlled by geophysical parameters, but it also changes due to other external effects. Meteorological effects – such as soil humidity, rainfall, temperature, barometric pressure and wind – control radon concentrations in soil gas. These parameters change the physical characteristics of soil and rock, thus influencing the rate of radon transport and, consequently, perturbing eventual radon variations caused by geophysical processes originating in the deeper parts of the Earth's crust. Shallow soil levels are more affected by changing meteorological conditions than deeper ones. Radon concentrations with no larger variations present are usually observed at depths of 0.8 m or deeper. Besides the effects of meteorological parameters on radon in soil gas, considerable variations in the gas composition of thermal springs have been shown to be the result of fluctuations of local hydrologic regimes (Klusman & Webster, 1981).

The significant influence of barometric pressure has been discussed by several authors, who clearly pointed out an inverse relationship between barometric pressure and radon concentration in soil gas (Chen et al., 1995; Clements & Wilkening, 1974; Klusman & Webster, 1981). A decrease in barometric pressure, with the values of other environmental parameters remaining constant, generally causes an increase in radon exhalation from the ground, whereas during periods of rising pressure, air with low radon concentration is forced into the ground, thus diluting radon. Temperature-related fluctuations of soil gas radon concentration have also been proven to be very important. Klusman & Jaacks (1987) found an inverse relationship between soil temperature and radon concentration. They suggested that lower air temperatures as compared with soil temperatures during winter months promoted an upward movement of radon by convection, whereas during the summer, lower soil temperatures as compared with air temperatures and an inversion layer below the level of sampling reduces the upward flux and observed concentration. In general, the behaviour of soil gas migration in different types of soil is seasonally dependent (King & Minissale, 1994; Washington & Rose, 1990). In systems where gas movement is driven by diffusion or slow advection processes, radon activity in soil might be controlled by soil moisture and rainfall through the opening of cracks in the surface (Pinault & Baubron, 1996; Toutain & Baubron, 1999). On the other hand, barometric pressure has the major influence on radon concentrations in soils in advective systems, which display generally higher gas flows. However, micro-scale soil heterogeneities in permeability, porosity and lithology can cause significant heterogeneities in the response of radon concentration to changes of atmospheric parameters (King & Minissale, 1994; Neznal et al., 2004). Numerous and often divergent results in studies related to the effect of external factors on soil gas radon concentration suggest that no general predictive model for excluding meteorological effects can be proposed, and studies of radon in soil gas need a simultaneous record of meteorological parameters.

#### **3. Anomalous radon concentration and seismicity**

Both mechanisms of radon transport – diffusion and advection – depend on both soil porosity and permeability, which at the same time vary as a function of the stress field (Holub & Brady, 1981). However, migration by diffusion is negligible, where a component of advective long-distance transport exists (Etiope & Martinelli, 2002). The high permeability

Radon as an Earthquake Precursor – Methods for Detecting Anomalies 183

effective sensitivity to the impending earthquake. The ideal circle with its theoretical radius can be transformed into an ellipse or characterised by shadow areas where no precursory phenomena are observable due to crustal anisotropy, discontinuities or loose contacts along some faults, which prevent further stress transfer (İnan & Seyis, 2010; Martinelli, 1991). Although radon anomalies can be studied in soil gas and thermal waters, thermal waters could be much more representative of the geologic environment and could be more reactive to stress/strain changes acting at depth than soil gases. The disadvantage of soil gases lie in weak gas concentrations, generally due to the thickness of the sedimentary cover and the

An anomaly in radon concentration is defined as a significant deviation from the mean value. Due to the high background noise of radon time series, it is often impossible to distinguish an anomaly caused solely by a seismic event from one resulting from meteorological or hydrological parameters. For this reason, the implementation of more advanced statistical methods in data evaluation is important (Belyaev, 2001; Cuomo et al., 2000; Negarestani et al., 2003; Sikder & Munakata, 2009; Steinitz et al., 2003). In our research, radon has been monitored in several thermal springs (Gregorič et al., 2008; Zmazek et al., 2002a; Zmazek et al., 2006) and in soil gas (Zmazek et al., 2002b) and different approaches to

A very common practice in determining radon anomalies is the use of standard deviation. The average radon concentration is calculated for different periods with regard to the nature of yearly cycles of radon concentration. In the case of radon in soil gas, the mean value of radon concentration is calculated separately for four seasons (spring, summer, autumn and

Fig. 1. Continuous radon concentration recorded in soil gas at Krško basin. Straight lines represent the mean value and two standard deviations of the radon concentration. Local seismicity is expressed in terms of local magnitude (*M*L) and the distance between the measuring location and the earthquake epicentre (*D*). Radon anomalies are *C*Rn values

high level of atmospheric perturbations (Toutain & Baubron, 1999).

**4. Methods for detecting anomalies in radon time series** 

distinguishing radon anomalies were applied.

winter) based on the air and soil temperature.

**4.1 Standard deviation** 

outside the ±2σ region.

of the bedrock and soil in areas of crustal discontinuities, such as fractures and fault zones, promotes intense degassing fluxes, which causes higher soil gas radon concentrations on the ground surface above active fault zones. Although several measurements, experiments and models have been performed, the understanding of the mechanism of radon anomalies and their connection to earthquakes is still inadequate (Chyi et al., 2010; King, 1978; Ramola et al., 1990). Before the earthquake stress in the Earth's crust builds up causing a change in the strain field; the formation of new cracks and pathways under the tectonic stress leads to changes in gas transport and a rise in volatiles from the deep layers to the surface. In fact, fluids play a widely recognised role in controlling the strength of crustal fault zones (Hickman et al., 1995). Anomalous changes of radon concentration are closely linked to changes in fluid flow and, therefore, also to highly permeable areas along fault zones. Large faults are not discrete surfaces but rather a braided array of slip surfaces encased in a highly fractured and often hydrothermally altered transition – or "damage" – zone. Episodic fracturing and brecciation are followed by cementation and crack healing, leading to cycles of permeability enhancement and reduction along faults (Hickman et al., 1995).

Several mechanisms have been proposed, which could explain the relationship between radon anomalies and earthquake. Two models of earthquake precursors are discussed by Mjachkin et al. (1975), with a common principle: at a certain preparation stage, a region of many cracks is formed. According to the dilatancy-diffusion model (Martinelli, 1991; Mjachkin et al., 1975), the increase in tectonic stress causes the extension and opening of favourably-oriented cracks in a porous, cracked, saturated rock. Water flows into the opened cracks, drying the rock near each pore and finally resulting in a decrease of pore pressure in the total earthquake preparation zone. Water from the surrounding medium diffuses into the zone. The increased water-rock surface area, due to cracking, leads to an increase in radon transfer from the rock matrix to the water. At the end of the diffusion period, the appearance of pore pressure and the increased number of cracks leads to the main rupture. According to the crack-avalanche model (Mjachkin et al., 1975), the increasing tectonic stress leads to the formation of a cracked focal rock zone, with slowly altering volume and shape. At a certain stage – when the whole focal zone becomes unstable – the cracks quickly concentrate near the fault surface, triggering the main rupture. An alternate mechanism for earthquake precursory study, based on stress-corrosion theory, has been proposed by Anderson and Grew (1977). According to them, the observed radon anomalies are due to slow crack growth controlled by stress corrosion in a rock matrix saturated by ground waters. King (1978) has proposed a compression mechanism for radon release, whereby anomalous high radon release may be due to an increase of crustal compression before an impending earthquake, that squeezes out soil gas into the atmosphere at an increasing rate.

Toutain and Baubron (1999) observed that gas transfer within the upper crust is affected by strains less than 10–7, much smaller than those causing earthquakes. According to Dobrovolsky (1979), the radius of the effective precursory manifestation zone depends on the earthquake magnitude and can be calculated using the empirical equation:

$$R\_{\rm D} = 10^{0.43 \times M\_{\rm L}} \tag{1}$$

Where *R*D is the strain radius in km and *M*L is the magnitude of the earthquake. Considering the Earth's crust to be an anisotropic medium, this law can be modified according to the effective sensitivity to the impending earthquake. The ideal circle with its theoretical radius can be transformed into an ellipse or characterised by shadow areas where no precursory phenomena are observable due to crustal anisotropy, discontinuities or loose contacts along some faults, which prevent further stress transfer (İnan & Seyis, 2010; Martinelli, 1991).

Although radon anomalies can be studied in soil gas and thermal waters, thermal waters could be much more representative of the geologic environment and could be more reactive to stress/strain changes acting at depth than soil gases. The disadvantage of soil gases lie in weak gas concentrations, generally due to the thickness of the sedimentary cover and the high level of atmospheric perturbations (Toutain & Baubron, 1999).
