Radon concentrations in soil air are variable depending on factors that are considered external (planetary) and internal (geodynamic) relative to the Earth. In active fault zones, variations of gas emanations are most intense. A permanent monitoring station was established near Tyrgan settlement in Western Pribaikalie to study temporal variations of soil radon concentration, Q, in the faults of the Baikal rift, East Siberia. This station is located in the zone of the Primorsky normal fault that is the largest in the region. The station is equipped with radon radiometer PPA01M03 that records Q values every 85 minutes and also monitors a number of meteorological parameters, including atmospheric pressure, humidity, and air temperature.
We analysed records of two measurement sessions (148 and 66 days) covering a part of the year during which field measurement of Q are possible in the cold climate conditions of the area under study. According to the available monitoring data, variations of radon concentrations in the Primorsky fault zone may vary by more than one order of magnitude through a springsummerautumn period, and such variations are oscillatory. Significant changes of permeability in time occur due to intensive changes in the state of stresses of the rock massives under the impacts of the planetary and geodynamic factors. The influence of the first group of factors, i.e. planetary ones, is manifested by synchronous oscillations of radon concentrations and atmospheric pressure, which phases of occurrence are opposed. Domination of daily and fourday periods gives evidence that the state of stresses of the rock massives is impacted by the lunar tides and cyclonic phenomena associated with the interaction between the Earth and the Sun. The influence of the second group of factors, i.e. geodynamic ones, is suggested by an evident relation between radon emanations and seismic events, including the catastrophic earthquake in Japan (March 11, 2011, M=9.0).
Tectonophysics
The external and internal factors are acting together, but their roles are different with regard to variations of radon concentrations in different periods of time. In the monitoring periods, radon emanation variations were mainly controlled by the planetary factors. Radon exhalation increases and decreases according to periodic variations in atmospheric pressure, which, in additional to ‘pumping’ effects, may lead to opening/closure of pores and cracks in the rocks. While external pressures are reduced, internal stresses are released by relatively weak earthquakes. The guiding influence of atmospheric pressure on the yield of radon is disturbed when internal stresses are in excess of a certain level due to intensive movements along faults in the Baikal rift or displacements of plates in neighbouring active zones (for example, due to the strongest earthquake in Japan). In such relatively short periods of time, when seismic activity is increased, the influence of tectonic stresses on permeability of rocks and radon emanations becomes dominant.
Based on our analysis of the measurements of soil radon concentrations obtained on the local site in the Primorsky fault zone through the monitoring period, it became possible, for the first time for Pribaikalie, to reveal and theoretically model the principal specific features of variation of soil radon concentrations, Q, in time and the dependence of such variations on the external and internal factors. Prospects of these studies are related to installation of a network of monitoring stations in the territory of the Baikal rift and assurance of longterm monitoring sessions.

This paper describes the statistical thermo-dynamical evolution of an ensemble of defects in the geomedium in the field of externally applied stresses. The authors introduce ‘tensor structural’ variables associated with two specific types of defects, fractures and localized shear faults (Fig. 1). Based on the procedure for averaging of the structural variables by statistical ensembles of defects, a self-consistency equation is developed; it determines the dependence of the macroscopic tensor of defects-induced strain on values of external stresses, the original pattern and interaction of defects. In the dimensionless case, the equation contains only the parameter of structural scaling, i.e. the ratio of specific structural scales, including the size of defects and an average distance between the defects.

The self-consistency equation yields three typical responds of the geomedium containing defects to the increasing external stress (Fig. 2). The responses are determined from values of the structural scaling parameter. The concept of non-equilibrium free energy for a medium containing defects, given similar to the Ginzburg-Landau decomposition, allowed to construct evolutionary equations for the introduced parameters of order (deformation due to defects, and the structural scaling parameter) and to explore their solutions (Fig. 3).

It is shown that the first response corresponds to stable quasi-plastic deformation of the geomedium, which occurs in regularly located areas characterized by the absence of collective orientation effects. Reducing the structural scaling parameter leads to the second response characterized by the occurrence of an area of meta-stability in the behavior of the medium containing defects, when, at a certain critical stress, the orientation transition takes place in the ensemble of interacting defects, which is accompanied by an abrupt increase of deformation (Fig. 2). Under the given observation/averaging scale, this transition is manifested by localized cataclastic deformation (i.e. a set of weak earthquakes), which migrates in space at a velocity several orders of magnitude lower than the speed of sound, as a ‘slow’ deformation wave (Fig. 3). Further reduction of the structural scaling parameter leads to degeneracy of the orientation meta-stability and formation of localized dissipative defect structures in the medium. Once the critical stress is reached, such structures develop in the blow-up regime, i.e. the mode of avalanche-unstable growth of defects in the localized area that is shrinking eventually. At the scale of observation, this process is manifested as brittle fracturing that causes formation of a deformation zone, which size is proportional to the scale of observation, and corresponds to occurrence of a strong earthquake.

On the basis of the proposed model showing the behavior of the geomedium containing defects in the field of external stresses, it is possible to describe main ways of stress relaxation in the rock massives – brittle large-scale destruction and cataclastic deformation as consequences of the collective behavior of defects, which is determined by the structural scaling parameter.

Results of this study may prove useful for estimation of critical stresses and assessment of the geomedium status in seismically active regions and be viewed as model representations of the physical hypothesis about the uniform nature of deve­lopment of discontinuities/defects in a wide range of spatial scales.

This paper presents results of the study of attenuation of seismic waves in the lithosphere and upper mantle of the northern part of the Basin and Range Province (BRP) (Fig. 1). In this study, the coda-wave method [Aki, Chouet, 1975] is applied to process data collected in the seismic experiment conducted in 1988–1989, PASSCAL Basin and Range Passive Seismic Experiment [Owens, Randall, 1989], including records of 66 earthquakes and explosions (Mb=1.1–5.0) which occurred in BRP (Fig. 2).

The effective seismic quality factor by the coda is calculated using the single-backscattering model [Aki, Chouet, 1975]. The QC values are calculated for 18 values of the lapse time window W from 10 to 95 sec with the step of 5 sec at six (6) central frequencies (0.3, 0.75, 1.5, 3.0, 6.0, and 12.0 Hz). In total, 7776 individual measurements of QC were done. It is observed that the quality factor QC is strongly dependent on the frequency and the lapse time window W: QC increases from 12±6 to 359±17 for the central frequencies of 0.3 and 12.0 Hz when the lapse time window is W=10 sec and from 87±6 to 1177±87 for the same frequencies when W=95 sec (Fig. 6). On the basis of the QС values obtained for all the lapse time windows W empirical relationships of quality factors and frequencies are calculated according to [Mitchell, 1981], and values of quality factor Q0 at reference frequency f0 (f0=1 Hz) and frequency parameter n (which is close to 1 and varies depending on the heterogeneity of the medium [Aki, 1981]) are obtained. In this study, Q0 varies from 60±8 to 222±17, the frequency parameter ranges from 0.57±0.04 to 0.84±0.05, and the attenuation coefficient δ varies from 0.015 to 0.004 km–1, depending on W (Fig. 8); similar values of attenuation parameters are typical of regions with high tectonic activity [Mak et al., 2004].

In the single-backscattering model, the dependence of the attenuation parameters from the lapse time window can be explained in terms of the depth of formation of the coda [Pulli, 1984]: a larger value of W corresponds to a greater depth through which the coda-waves go. As shown by the analysis of variations of attenuation coefficient δ and frequency parameter n for the Basin and Range Province, both parameters decrease irregularly with depth – the slope of the curve showing variations of δ is considerably changed at the depth of 150 km. At the top of the graph (to the depth of 150 km), an abrupt change of δ with depth is observed; it is clearly seen in the graph of gradient δ (Fig. 9 and Fig. 10); such behaviour is also characteristic of n. At the depth of 140 km, parameter n is increased. In the middle section (at depths of 150–200 km), the slope of the δ curve increases, and gradients of δ and the frequency parameter are significantly reduced. At the bottom of the profile (> 200 km), the value of δ is almost constant, and an abrupt increase of n is observed (Fig. 9 and Fig. 10). Figure 10 shows the high-speed profile of the area under study, which is published in [Wagner et al., 2012]. The profile shows the low velocity mantle under the Basin and Range Province, actually starting underneath the Moho (at the depth of 50–60 km). The lower boundary of the low-velocity mantle is located at the depth of 130–160 km. Thus, there are grounds to conclude that the change in the slope of the curve showing dependence of δ from the depth is related to the deep structure of the medium. The abrupt changes of δ and n are associated with the velocity discontinuities of the medium. The high values of δ and n, which are characteristic of the upper part of the profile, indicate the high degree of heterogeneity of the medium, which is also confirmed by the low velocities of seismic waves in the area under study [Wagner et al., 2012]. The reduction of parameters  δ and n in the middle and lower parts of the profile suggests a more homogeneous structure of the medium at larger
depths.

As a result of the study of the characteristics of seismic wave’s attenuation in the lithosphere and the upper mantle of the northern part of the Basin and Range Province, it is established that the effective seismic quality factor QC is highly dependent on the frequency in the range of 0.5–16.0 Hz. The empirical relationships of Q(f) for various lapse time windows are obtained; it is shown that increasing the lapse time window causes the values of the effective seismic quality factor to increase, which may be interpreted as reduction of attenuation with depth. By comparing the depth variations of the attenuation coefficient and the frequency parameter against the velocity structure, it is shown that there is a distinct change in attenuation of seismic waves at the velocity discontinuities in the northern part of the Basin and Range Province.