Journal of Environmental Radioactivity 134 (2014) 128e135
Contents lists available at ScienceDirect
Journal of Environmental Radioactivity journal homepage: www.elsevier.com/locate/jenvrad
Sub-daily periodic radon signals in a confined radon system G. Steinitz*, O. Piatibratova 1, P. Kotlarsky Geological Survey of Israel (GSI), 30 Malkhei Israel, Jerusalem 95501, Israel
a r t i c l e i n f o
a b s t r a c t
Article history: Received 10 December 2013 Received in revised form 8 March 2014 Accepted 15 March 2014 Available online 3 April 2014
Signals from radon in air enclosed in a tight canister are recorded by five gamma detectors located around the horizontal plane and along the vertical axis. At steady state conditions (diffusion input ¼ radon decay) the primary variation is of daily radon (DR) signals with amplitudes of around 20 e25%. The DR signal, with a rounded form, is characterized by periodicities of 24-, 12- and 8-h (i.e. 1, 2 & 3 CPD). Similar DR variation patterns occur in the east and west sensors whereas inverse DR patterns are recorded by the north and south sensors. Short term (ST) signals, having saw tooth form and periods of 2 e3 h (frequencies in the range of 9e12 CPD) are observed at all five sensors and are superimposed on the DR signals with relative amplitudes of around 20%. They exhibit differing forms and phase at the different sensors, located at different directions around the canister. The latter is similar to the spatial manifestation of form and phase of the DR signal in such experiments, indicating a communality of the driving mechanism. At this stage a geophysical explanation cannot be presented for the ST signals. In this respect a peculiar observation is that their extraordinary occurrence coincides in time with the Tohoku Earthquake (Mw ¼ 9.0; 11 March 2011). Ó 2014 Elsevier Ltd. All rights reserved.
Keywords: Radon signals Confined systems Geodynamics Tohoku earthquake
1. Introduction Radon (222Rn) is a radioactive inert gas formed by disintegration from 226Ra as part of the 238U decay series. The combination of its noble gas character and its radioactive decay make it a unique ultratrace component for tracking temporally varying natural processes. In the geological environments it occurs at varying concentrations and shows large and complex temporal variation patterns which exhibit periodic and non-periodic signals of annual to daily scale. The issue of its exceptional behavior was addressed in numerous works in the last decades reporting on observations from diverse environs and situations. Many different environmental, geodynamical and physical processes have been suggested as influencing the variability of radon. Despite the efforts of the scientific community, the nature of the physical processes driving the temporal patterns observed in 222Rn time series remains elusive and interpretation of the observed phenomena on a physical basis is not straightforward. Suggested processes such as exhalation, diffusion, advection, and transport in porous media, stack effect, atmospheric influence (variation of pressure, temperature, and humidity) and solar tides sometimes even oppose each other. * Corresponding author. Tel.: þ972 2 5314241. E-mail addresses: [email protected] (G. Steinitz), [email protected] (O. Piatibratova), [email protected] (P. Kotlarsky). 1 Tel.: þ972 2 5314214. http://dx.doi.org/10.1016/j.jenvrad.2014.03.012 0265-931X/Ó 2014 Elsevier Ltd. All rights reserved.
Advance in the investigation of the radon system in air was achieved by the radon research group at the Geological Survey of Israel (GSI). Extensive monitoring measurements conducted since 1990 in subsurface geological sites are complemented in recent years by simulation experiments in the laboratory. In these radon signals in an enhanced confined mode (ECM) are investigated (Steinitz et al., 2011). In these experiments a relatively high level of radon is maintained inside a confined volume by diffusion (via tube) from a connected radon source. The nuclear radiation from the radon (and progeny) in the confined volume is monitored mainly with alpha sensors (inside volume) and gamma sensors (inside and outside the volume). Under such conditions where radon only diffuses into a closed volume it is expected that stable and uniform nuclear radiation will be observed determined by a balance between the diffusion controlled supply and the decay rate of radon (half-life of 3.82 days). In difference with the expected, nuclear radiation from the radon (progeny) shows (Steinitz et al., 2011, 2013: a) temporal variations (signals) spanning annual to daily scale; b) directionality of the nuclear radiation reflected as inverse signal patterns in the eastewest versus northesouth directions; c) patterns, periodicities and their characteristics which are similar to those observed on radon in the geological environment. Previous observations derived from the subsurface geological environment indicated the unique temporal variations of nuclear radiation from radon (Steinitz et al., 2007; Steinitz and Piatibratova,
G. Steinitz et al. / Journal of Environmental Radioactivity 134 (2014) 128e135
129
Fig. 1. The experiment comprised of a 3.53 L cylindrical SS canister and five gamma sensors in vertical orientation. 222Rn diffuses from a commercial source to the experimental volume via a tube and valves. All components of the system e radon source and air e are enclosed within leak tight SS components (canisters, tube) which are connected using UHV flanges and valves. Four sensors (PM-11) are placed in four orientations (g-E, g-N, g-W, g-S) next to the middle of the canister. A further sensor (Scionix) is placed in the middle of the top of the canister (g-T). Lead shielding minimized the effect of radiation from the source lowered environmental radiation. The laboratory wall is oriented EeW.
2010a,b), and raised the suggestion that an unrecognized geophysical driver influences the variations, primarily in the diurnal band. The outcome of the first experiments using the ECM configuration substantiated this assertion and led to the further suggestion that the prominent diurnal periodicity was associated with solar tide (Steinitz et al., 2011). Reanalysis of the long experimental radon time series by Sturrock et al. (2012) corroborated the results from Steinitz et al. (2011) and provided further evidence for the solar influence by identifying solar rotational frequencies (which are independent of Earth) in time series of the nuclear radiation and by showing a 24-h modulation of the gamma radiation of the radon system in the annual and in the solar rotational frequency bands. In conformity with the view in previous works Sturrock et al. (2012) suggested that the decay is influenced by solar radiation and solar neutrinos were considered as a possible particle involved. Systematic radon signals of sub-daily duration have been observed in the geological environment (Steinitz and Piatibratova, 2010a,b). In this contribution an occurrence of sub-daily shortterm (ST) signals in a simulation experiment are described. 2. Methods Simulation experiments of radon signals (Steinitz et al., 2011, 2013) are performed using the Enhanced Confined Mode (ECM) principle, which is based on: 1. A tight container in which a high level of radon in air is maintained. 2. A radon source (commercial e RaCl2) connected with a tube (and valves) supplies radon in a diffusion regime. 3. Nuclear radon detectors e alpha as internal or directly connected to the container volume, and/or gamma detectors which are either inserted into the volume (large volume) or placed outside next to the container. Our field and laboratory experience has shown that similar results in monitoring temporal variation of radon are obtained using alpha and gamma detectors. Furthermore, Zafrir et al. (2011) have shown that gamma sensors are advantageous due primarily to the higher count rates. 4. When using the gamma radiation a lead shield is added to minimize the effect of environmental gamma radiation and the eventual influence of the source on the detectors. The radiation (a, g) from the radon in the system is expected: a) to be proportional to concentration in the container air; b) to rise
once the source is connected due to diffusion from the source, and c) to level-off once the rate of diffusion equals the rate of decay of radon. Experiments conducted using different configuration showed that in contradiction with the expected large temporal variations are encountered (Steinitz et al., 2011, 2013). The phenomenology of these variations and the characteristics of the signals resemble those encountered in the natural geological environs. The specific experiment dealt with here is placed on a table, north of the eastewest laboratory wall (Fig. 1). The experimental work is based on an enhanced confined module (ECM) in which an enhanced level of radon is maintained inside a confined volume by diffusion (via tube) from an attached commercial radon source. The nuclear radiation from the radon in the confined volume is monitored, basically, with gamma sensors placed outside the confined volume. The setup used (Fig. 1) is the setup used in EXP #1 described by Steinitz et al. (2011). It consists of a 3.53 L cylindrical (Ø ¼ 15 cm) stainless steel (SS) canister, ultra-high vacuum tested, and equipped with a UHV valve. Radon is let in (diffusion) from a commercial 222Rn source (103.2 kBq), placed within a leak tight SS enclosure, via a UHV valve on the source and a 60-cm-long SS pipe, resulting in a radon level in the order of 8400 kBq/m3. In this specific experiment the atmosphere inside the ECM was of argon. Air was first evacuated from the canister and from the radon source to around 5 102 mbar. After evacuation the source valve was closed and argon was introduced to around 1 atm (controlled with a Pirani gauge). Once the system was isolated diffusion of radon was enabled by opening the source valve. Detection of temporal variation of radiation is achieved by utilizing gamma detectors with NaI(Tl) scintillation detectors. The utilization of gamma detectors for monitoring radon in air has been described in detail by Zafrir et al. (2011). Five gamma detectors were placed around the canister. Four sensors (PM-11; Rotem Industries Inc.) are based on 2 200 crystals tuned to the energy range of 50e3000 keV. These gamma sensors (PM-11) are set vertically around the central horizontal plane of the canister. They are placed, relative to the canister, according to the global directions. The orientation of these sensors is estimated to deviate not more than 5 from the true geographic directions. To allow comparison among them these sensors were adjusted (inter-calibrated) during the experiment using the radon in the canister. This was done by placing each sensor at position north and recording the radiation from radon in the canister (relative adjustment: North ¼ 1, East ¼ 1.65, West ¼ 1.8, South ¼ 1.78). A further gamma sensor (Scionix Ltd.) is based on a 36 76 mm crystal and is tuned to the energy range of 475e3000 keV. It was placed on top of the canister, in a vertical position along the axis of the canister. The
130
G. Steinitz et al. / Journal of Environmental Radioactivity 134 (2014) 128e135
experimental setup is enclosed (bottom, sides) by lead shielding (5 cm) to minimize the influence of gamma radiation from the source and from the environment. Data acquisition was with a datalogger (CR-800; Campbell Inc.) which also supplied the stabilized power to the detectors by a continuously loaded 12 V supply battery. Collected data (1 min intervals; timing precision better than 30 s) was integrated to 15min intervals. Time is shown on a decimal-day scale relative to 1.1.1992 (T ¼ 0 which is the time base of the radon measurements at the GSI). 3. Results The experimental system was used in this configuration in a series of experiments conducted from May 2009 to June 2012. The overall pattern of variation of radon during the experiment (16/1/ 2011e11/4/2011; Fig. 2) is similar to other experiments obtained with this setup (Steinitz et al., 2011, 2013). Once activated by enabling diffusion from the source buildup of radon occurs inside the SS canister, lasting around 15 days. Daily radon (DR) signals accompany the buildup. The DR signal is observed by the five sensors as discordant signals the amplitude of which is proportional to the attained level of radon. Leveling-off the signals occurs once a balance is attained between the diffusion input and the nuclear decay of radon. The plateau is dominated by large variations (20e25%) which are composed of DR (primary) and multi-day (MD) signals. The DR signal is characterized by periodicities of 24-, 12- and 8-h (i.e. 1, 2 & 3 CPD). These variations are recorded as signals of different pattern and phase at the different sensors. Once the system is isolated from the source a systematic decrease occurs which follows the nuclear decay scheme of radon. Discordant DR signals are also superimposed on the decay pattern with amplitudes which are proportional to the remaining radon level. Fig. 3 shows a typical 10-day long example of the variation pattern, dominated by the DR signal, at four primary directions along the horizontal plane of the canister and along the vertical axis. Similar daily variation occurs at the opposite sensors (east & west; north & south) while inverse patterns are observed among these two perpendicular directions. The DR pattern at the vertical sensor similar to the pattern in the northesouth direction. These features described and discussed in detail by Steinitz et al. (2013)
Fig. 2. Overall pattern of the gamma radiation from radon in the experiment, shown using hourly averages of the 15-min count rate. Time is given as a decimal day relative to 1.1.1992 (¼ Day 0). The experiment follows three stages: buildup of radon in the canister following the opening of the source (Day 6955), leveling off (around Day 6972) of the signal owing to attaining a balance between the diffusion and decay rates, and decay of radon following the isolation (Day 7019.54) of the system from the source.
Fig. 3. Detail (10 days) of the measured gamma radiation pattern (counts/15-min; hour averages) at the five sensors around the horizontal plane of the canister and along its vertical axis. The variation is dominated by DR signals with are non-similar among the sensors. Concordant patterns occur in the east and west directions and also among the north and south directions. Inverse patterns exist among the two concordant sets. The variation pattern along the vertical axis is similar to that in the northesouth direction. See text.
are attributed to directionality linked to global orientation of the temporally varying gamma radiation from radon in air at confined conditions. Besides to the DR signal a further variation pattern occurs in the time span of Days 7009e7012. The temporal pattern of the gamma radiation from radon at the five sensors of the experiment, at a resolution of 15-min, in the 10-day interval around the ST signals is shown in Fig. 4. Short term (ST) signals lasting 2e3 h are superimposed on the DR signal in this interval. The inverse parallel patterns of the DR signal in the eastewest and northesouth directions are maintained in this interval. The ST signals are recorded by all five sensors. Similar signals are not encountered in further and previous experiments performed using this configuration conducted from May 2009 to June 2012, nor in other experiments. Fig. 5 shows a five day detail around these ST signals using the measured 1-min count rate. At this resolution the ST signals are clearly resolved and manifest a periodicity in the range of 2e3 h. The strength of the ST signal is amplified during two-three intervals in Days 7009e7022.5. Different phase and amplitude are observed among the different sensors. The details of the ST signal and its relation to the DR signal are resolved by applying a 7.5-h sliding average to decompose the measured signal into the DR and ST components. Fig. 6 demonstrates both communalities as well as differences among the two signals. The DR signal (smoothed component) shows a rounded pattern. The variation of the ST signal (residual component) is composed of distinct sharp (saw-tooth) forms. The DR signal demonstrates directionality manifested as inverse mutual variation in the eastewest versus the northesouth directions, and the DR signal along the vertical axis is varying in tandem with the northe south direction. This pattern is in conformity with its pattern at other times. Conversely, the ST signal also shows directionality, but which is different. Mutual co-variation occurs in the directions eastewestenorth and an inverse pattern in the south direction. The variation along the vertical axis tends to be similar to the easte westenorth trend.
G. Steinitz et al. / Journal of Environmental Radioactivity 134 (2014) 128e135
Fig. 4. The measured gamma radiation pattern using 15-min averages, in the days around the TE event at the five sensors around the horizontal plane of the canister and along its vertical axis. The variation is composed of DR signals and also Short-Term (ST) signals which occur during three days around the TE, Compare with Fig. 3.
Fig. 5. Five day detail of the measured count rate (1-min) around the interval showing the ST signals. The ST signals, which are well defined at this time resolution, are superimposed at all sensors on the DR signal. The relative amplitude of the ST signal attains up to 20% of the amplitude of the DR signal. The two subintervals with prominent ST signals occur in the interval of Days 7009e7012 are shown in detail in Fig. 6.
The systematic patterns of the variations of the DR and ST signals indicate that directionality of the driving system is playing a role for both signal types. The systematic difference in this directionality among them indicates a different relative geometrical configuration of the driver of each signal type. This implies that the drivers must be attributed to sources at different locations, suggesting in turn that different sources are involved. In both cases the source affecting the radon signals is of a remote nature. In the case of the DR signal the driver is attributed to a component in solar irradiation (i.e. of extraterrestrial origin; op. cit.). Taking into account the observed difference of the directionality leads to the eventual conclusion that in the case of the ST signal the source is bound to Earth. Further insight is obtained by analysis in the frequency domain. The ST signal is examined by applying a FFT filter on the de-trended
131
and normalized time series, and reconstructing the signal in the range of 9e12 CPD (assuming periodicity in the range of 2e3 h). Fig. 7 shows the filtered signal during 50 days (left) and in a detail of 10 days (right). The reconstructed signals demonstrate similar patterns at the five sensors, with clear enlarged relative amplitudes in two pulses in the time interval lasting some 48 h in Days 7009e 7012. In a further step of analysis pairs of reconstructed time series at opposite directions are multiplied. The assumption is that calculating the product of normalized measurements among time series will lead to the amplification of co-varying variations while reducing, in parallel, unrelated signals. Fig. 8-A shows the result of the multiplication of the filtered time series of the two pairs of opposite sensors (east & west; north & south). The signals around in the interval of Days 7009e7012 is clearly enhanced in each of the two perpendicular directions. Further enhancing of the intense signal is obtained by forming a combined product time series of the measurements in the horizontal plane (Fig. 8-B; upper plot). This plot demonstrates the extreme reduction obtained for the background variation relative to the signal in Days 7009e7011 by using measurements in the multi-sensor configuration. The outcome is further supported by the significant co-varying temporal pattern also observed in the independent measurements by the sensor along the vertical axis of the canister (Fig. 8-B, lower plot). Sub-daily ST radon signals with a period in the range of 9e12 h have not been reported from natural environs nor have they been observed in other simulation experiments we performed. In this sense they stand as a solitary observation which cannot be associated, at this stage, with a known recurring phenomenon. Notwithstanding, the occurrence of these signals concurs with a particular outstanding geophysical event e the Tohoku earthquake. The Tohoku earthquake (TE) is a magnitude 9.0 (Mw) earthquake that occurred some 70 km east of Tohoku (Japan) at an estimated depth of 32 km, at 05:46 (UTC) on 11 March 2011. The time interval of the appearance of the ST signal coincides with the occurrence of the Tohoku earthquake (TE). The ST signals are observable from around 10 h prior to the TE event and last some 48-h after TE. This is demonstrated in Figs. 9e11. Fig. 9 shows time in two methods: a) in decimal days, as in the previous figures, in the upper x-axis, and b) relative to the TE event in the lower x-axis. The primary TE event and a reported fore-shock are indicated as vertical lines. The plots show a detailed comparison of the combined FFT reconstructed signal in the horizontal plane and the reconstructed signal of the sensor along the vertical axis of the canister. The enhancement of the signal/noise ratio as described in Figs. 8 and 9 permits improving the observations by enlarging the y-axis, as shown in Fig. 10. The two reconstructed patterns, which are derived from measurements which are geometrically independent, show co-varying response in the temporal dimension. An association is indicated with the two major seismic events e the TE and the earlier foreshock. In the case of the TE the onset of the radon signal in the 9e12 CPD band precedes the event by some 10 h in both horizontal and vertical dimension of the nuclear radiation from radon in the canister. This pre-TE signal is clearly observable relative to the filtered background. The intense primary ST signal occurs from TE þ 6 h and lasts to around TE þ 48 h. The response of the ST pattern to the main event (TE) is significantly larger compared to the response to the foreshock, occurring at TE2.13 days. In the case of the foreshock the indication in the ST radon signal occurs from 3 h after the foreshock and lasts to around 10 h after this event. The measured signal recorded at 1-min resolution and the 9e12 CPD reconstructed signal in a 1 day interval around TE are compared in Fig. 11. The time relations among the patterns confirm that the ST signal, which is significantly observed in the lower plot, was indeed initiated already 6e9 h prior to the TE event.
132
G. Steinitz et al. / Journal of Environmental Radioactivity 134 (2014) 128e135
Fig. 6. The decomposed DR signal (left) and ST signal (right), using 1-min measurements. The decomposition was obtained by applying a 7.5 h sliding average. The ST signal is shown in detail for two time intervals in the interval 7009e7012 (see Fig. 5). The E&W versus N&S directional pattern of the DR signal is maintained. The directional pattern of the decomposed ST signal is different. See text.
Fig. 7. The FFT filtered and reconstructed signal in the frequency range of 9e12 CPD for the four sensors around the horizontal plane of the canister and the vertical sensor along the axis of the canister. Left e 50 day sequence. Right e 10 day detail around the interval showing the ST signals.
G. Steinitz et al. / Journal of Environmental Radioactivity 134 (2014) 128e135
133
2e-3
1e-3
7005
7007
7009
7011
7013
0
2
4
5.0e-3
gE*gW*gN*gS
g-TOP
Day since 1.1.1992
EW*NS g-TOP Foreshock (Mw = 7.2) TE (Mw = 9)
2.5e-3 0
-1e-3
-2e-3
0.0
-3e-3
-4
-2
Time relative to TE (day) Fig. 9. A 5-day detail of Fig. 8-B showing the temporal relation between the FFT 9e12 CPD reconstructed signals and the TE seismic event. The lower x-axis shows time relative to the main TE event (at t ¼ 0) and the pre-shock (at t ¼ 2.15 days). The combined FFT reconstructed signal in the horizontal plane is shown in the lower plot (black); the upper plot (red) shows the reconstructed signal of the sensor along the vertical axis of the canister. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 8. A e Composite time series formed by the multiplication of reconstructed time series shown in Fig. 7. The two time series represent the result in the eastewest and northesouth direction. Enhanced signals occur around in the interval of the ST signals (see text). B e The variation in the horizontal plane: Composite time series (lower plot) formed by the multiplication of the product in the two principal directions (shown in A); a concordant pattern is exhibited (upper plot) by the filtered 9e12 CPD signal at the sensor along the vertical axis of the canister. The pattern of the 9e12 CPD signal is further enhanced in the interval of the ST signals.
communality in the characteristics of the driving mechanism. For the DR signals it was previously shown in such experiments that they due neither to mechanical nor to local environmental influences on the experimental system, but rather to an unidentified remote driver. A similar conclusion is reached for the ST signals e i.e. that a remote driver is also involved in their generation. In the case of the DR signal it was recently suggested that a component of solar irradiance is influencing the pattern of nuclear radiation from radon in air (Sturrock et al., 2012). Taking into account both similarity and differences between the ST and DR signal in terms of their global orientation patterns it is suggested that the ST signal is
4. Discussion Radon signals spanning from annual to daily scale are frequent in the air of the subsurface geological environment, and differing views are offered as to the nature of the processes driving them. Recently it was demonstrated that similar radon signals can be simulated in the laboratory using the enhanced confined mode (ECM) configuration. Utilizing an ECM system with a multi-sensor configuration around it the described experiment demonstrates, for the first time, that short term (ST) radon signals with periods of 2e3 h (9e12 CPD) also occur. These signals are due to temporal variation of the gamma radiation evolving from radon (progeny) inside the canister. Based on the accumulated experience the occurrence of ST signals is rare compared to the longer term AR, MD and DR radon signals. The ST signal clearly observed in the measured data in this experiment using a multi-sensor configuration is also well characterized in the reconstructed signal obtained by filtering and combining in the frequency domain. Systematic spatial inhomogeneity of the ST signals is reflected as differing forms and phase at different directions around the canister, clearly indicating that the ST signal is generated in the canister due to an external influence, probably of a remote nature. This feature is similar to the spatial manifestation of form and phase of the DR signal in such experiments, indicating a
Fig. 10. Comparison of the 9e12 CPD FFT reconstructed signals and their temporal relation to the primary seismic event TE (at t ¼ 0) and the pre-shock (t ¼ 2.15 days). An enlarged y-axis is used, compared to the one in Fig. 9. The plots portray the combined reconstructed signal in the horizontal plane and the signal along the signal perpendicular to it e along the vertical axis of the canister. A 9e12 CPD variation occurs around 6 h prior to the TE event. An enlarged y-axis is used, compared to the one in Fig. 9.
134
G. Steinitz et al. / Journal of Environmental Radioactivity 134 (2014) 128e135
pre-shock. These ST signals appear from around 10 h prior to the TE and are maintained to around 48 h after the TE. The relatively intense manifestation of this phenomenon occurs after TE. If this association between TE event and the ST radon signals is indeed due to a geophysical connection then the following implications are indicated: 1. The overall progression of the TE (Mw ¼ 9) event is also associated with a non-mechanical geodynamic process which is reflected by nuclear radiation features of radon in air. This process is modifying (superimposed) the driver of the primary DR signal of the radon system. 2. The influencing process is operating at a global scale, but at this stage it remains open whether the effect is transferred to the experimental setup via the solid earth and/or the atmosphere. 3. Detection of such phenomena is possible in radon systems of specific configuration the parameters of which are so far undetermined. 4. The possibility is raised that a pre-cursor of the TE has been detected.
Fig. 11. Comparison, in the two day interval around TE, of the measured signals (upper plot) with the signal (lower plot) reconstructed in the frequency range of 9e12 CPD. The time scale, relative to the TE event, is the same in both plots.
driven by a remote source which in this case is located on Earth, in variance with the DR signal. The fact that similar short term signals have not been encountered in other experiments or field locations may be due to several factors. The experience gained shows that patterns of radiation from radon in air are influenced by geometrical parameters (canister dimensions, location, and orientation) which so far are unresolved, and that the sensitivity for the detection of DR signals is strongly related to the overall radon level. The following implications are indicated for the processes underlying the temporal and spatial variation of the nuclear radiation from radon in air: 1. Results obtained by the GSI group indicate that nuclear radiation from radon (and progeny) inside a confined volume of air varies spatially and temporally at time scales from annually to daily. The geophysical drivers of these signals are unclear at this stage. The new experimental results demonstrate that the same radon system is responding to further geophysical drivers operating at a time scale of 2e3 h. 2. The results further demonstrate the potential of investigations utilizing enhanced radon levels within confined volumes for detection of a new type of time varying geophysical phenomena. 3. Performing ECM experiments utilizing relatively high radon levels leads to improved signal to noise conditions and thereby allows detection and investigation of ST signals in the radon signals. 4. If causality is indeed underlying the high concordance in time between the occurrence of the ST signal in the radon system and the Tohoku earthquake, which is an extreme geodynamic event, then a new research track is indicated. In the temporal dimension the exceptional ST signals occur in association with the Tohoku earthquake and possibly also with its
Due to the large effect of disastrous earthquakes the issue of their prediction is often dealt with. A wide spectrum of physical phenomena for earthquake prediction is suggested and is extensively debated in the literature. For a recent review see Jordan et al. (2011). Very different views exist including ones maintaining that such events cannot be predicted (e.g. Geller, 1997; Geller et al., 1997). Radon in the natural environment, either in air or in water, is one of the leading candidates suggested as a proxy of seismogenic activity (Hartmann and Levy, 2005; Immè and Morelli, 2012). As is the case of other proxies a clear-cut situation has not been demonstrated. Furthermore, the lack of a sound physical and geophysical frame for the exceptional and complex patterns of radon variation in nature stresses the fact that this assertion is not established. The present contribution demonstrates, in agreement with other observations we recently presented, that the fundamental physical and geophysical properties are different from those assumed so far. Only a new understanding of the prosperities of the radon system may facilitate the application in earthquake prediction. This work shows that other physical drivers underlie its variation. Furthermore it demonstrates, again, the potential in investigating the geophysical aspects influencing radon system and their implication for using it as an indicator of geophysical phenomena and processes. Acknowledgments This investigation was supported by the Geological Survey of Israel. References Geller, R.J., 1997. Earthquake prediction: a critical review. Geophys. J. Int. 131 (3), 425e450. http://dx.doi.org/10.1111/j.1365-246X.1997.tb06588. Geller, R.J., Jackson, D.D., Kagan, Y.Y., Mulargia, F., 1997. Earthquakes cannot be predicted. Science 275, 1616. http://dx.doi.org/10.1126/science.275.5306.1616. Hartmann, J., Levy, J.K., 2005. Hydrogeological and gasgeochemical earthquake precursors e a review for application. Nat. Hazards 34, 279e304. Immè, G., Morelli, D., 2012. Radon as an earthquake precursor (Chapter 7). In: D’Amico, S. (Ed.), Earthquake Research and Analysis e Statistical Studies, Observations and Planning, pp. 143e160. Jordan, T.H., Chen, Y.-T., Gasparini, T., Madariaga, P., Main, I., Marzocchi, W., Papadopoolos, G., Sobolev, G., Yamaoka, K., Zschau, J., 2011. Operational earthquake forecasting: state of knowledge and guidelines for utilization. Ann. Geophys. 54 (4), 315e391. http://dx.doi.org/10.4401/ag-5350. Steinitz, G., Piatibratova, O., Barbosa, S.M., 2007. Radon daily signals in the Elat Granite, southern Arava, Israel. J. Geophys. Res. 112, B10211. http://dx.doi.org/ 10.1029/2006JB004817.
G. Steinitz et al. / Journal of Environmental Radioactivity 134 (2014) 128e135 Steinitz, G., Piatibratova, O., 2010a. Radon signals in the Gavnunim intrusion, Makhtesh Ramon. Isr. Geophys. J. Int. 180, 651e665. Steinitz, G., Piatibratova, O., 2010b. Radon signals at the roded site, southern Israel. Solid Earth 1, 99e109. http://dx.doi.org/10.5194/se-1-99-2010. Steinitz, G., Piatibratova, O., Kotlarsky, P., 2011. Possible effect of solar tides on radon signals. J. Environ. Radioact. 102, 749e765. http://dx.doi.org/10.1016/ j.jenvrad.2011.04.002. Steinitz, G., Kotlarsky, P., Piatibratova, O., 2013. Anomalous non-isotropic temporal variation of gamma-radiation from radon (progeny) within air in confined conditions. Geophys. J. Int. 193, 1110e1118. http://dx.doi.org/10.1093/gji/gg057.
135
Sturrock, P.A., Steinitz, G., Fischbach, E., Javorsek, D., Jenkins, J.H., 2012. Analysis of gamma radiation from a radon Source: indications of a solar influence. Astropart. Phys. 35/1, 18e25. Zafrir, H., Haquin, G., Malik, U., Barbosa, S.M., Piatibratova, O., Steinitz, G., 2011. Gamma versus alpha sensors for Rn-222 long-term monitoring in geological environments. Radiat. Meas. 46, 611e620. http://dx.doi.org/10.1016/ j.radmeas.2011.04.027.
Contents lists available at ScienceDirect
Journal of Environmental Radioactivity journal homepage: www.elsevier.com/locate/jenvrad
Sub-daily periodic radon signals in a confined radon system G. Steinitz*, O. Piatibratova 1, P. Kotlarsky Geological Survey of Israel (GSI), 30 Malkhei Israel, Jerusalem 95501, Israel
a r t i c l e i n f o
a b s t r a c t
Article history: Received 10 December 2013 Received in revised form 8 March 2014 Accepted 15 March 2014 Available online 3 April 2014
Signals from radon in air enclosed in a tight canister are recorded by five gamma detectors located around the horizontal plane and along the vertical axis. At steady state conditions (diffusion input ¼ radon decay) the primary variation is of daily radon (DR) signals with amplitudes of around 20 e25%. The DR signal, with a rounded form, is characterized by periodicities of 24-, 12- and 8-h (i.e. 1, 2 & 3 CPD). Similar DR variation patterns occur in the east and west sensors whereas inverse DR patterns are recorded by the north and south sensors. Short term (ST) signals, having saw tooth form and periods of 2 e3 h (frequencies in the range of 9e12 CPD) are observed at all five sensors and are superimposed on the DR signals with relative amplitudes of around 20%. They exhibit differing forms and phase at the different sensors, located at different directions around the canister. The latter is similar to the spatial manifestation of form and phase of the DR signal in such experiments, indicating a communality of the driving mechanism. At this stage a geophysical explanation cannot be presented for the ST signals. In this respect a peculiar observation is that their extraordinary occurrence coincides in time with the Tohoku Earthquake (Mw ¼ 9.0; 11 March 2011). Ó 2014 Elsevier Ltd. All rights reserved.
Keywords: Radon signals Confined systems Geodynamics Tohoku earthquake
1. Introduction Radon (222Rn) is a radioactive inert gas formed by disintegration from 226Ra as part of the 238U decay series. The combination of its noble gas character and its radioactive decay make it a unique ultratrace component for tracking temporally varying natural processes. In the geological environments it occurs at varying concentrations and shows large and complex temporal variation patterns which exhibit periodic and non-periodic signals of annual to daily scale. The issue of its exceptional behavior was addressed in numerous works in the last decades reporting on observations from diverse environs and situations. Many different environmental, geodynamical and physical processes have been suggested as influencing the variability of radon. Despite the efforts of the scientific community, the nature of the physical processes driving the temporal patterns observed in 222Rn time series remains elusive and interpretation of the observed phenomena on a physical basis is not straightforward. Suggested processes such as exhalation, diffusion, advection, and transport in porous media, stack effect, atmospheric influence (variation of pressure, temperature, and humidity) and solar tides sometimes even oppose each other. * Corresponding author. Tel.: þ972 2 5314241. E-mail addresses: [email protected] (G. Steinitz), [email protected] (O. Piatibratova), [email protected] (P. Kotlarsky). 1 Tel.: þ972 2 5314214. http://dx.doi.org/10.1016/j.jenvrad.2014.03.012 0265-931X/Ó 2014 Elsevier Ltd. All rights reserved.
Advance in the investigation of the radon system in air was achieved by the radon research group at the Geological Survey of Israel (GSI). Extensive monitoring measurements conducted since 1990 in subsurface geological sites are complemented in recent years by simulation experiments in the laboratory. In these radon signals in an enhanced confined mode (ECM) are investigated (Steinitz et al., 2011). In these experiments a relatively high level of radon is maintained inside a confined volume by diffusion (via tube) from a connected radon source. The nuclear radiation from the radon (and progeny) in the confined volume is monitored mainly with alpha sensors (inside volume) and gamma sensors (inside and outside the volume). Under such conditions where radon only diffuses into a closed volume it is expected that stable and uniform nuclear radiation will be observed determined by a balance between the diffusion controlled supply and the decay rate of radon (half-life of 3.82 days). In difference with the expected, nuclear radiation from the radon (progeny) shows (Steinitz et al., 2011, 2013: a) temporal variations (signals) spanning annual to daily scale; b) directionality of the nuclear radiation reflected as inverse signal patterns in the eastewest versus northesouth directions; c) patterns, periodicities and their characteristics which are similar to those observed on radon in the geological environment. Previous observations derived from the subsurface geological environment indicated the unique temporal variations of nuclear radiation from radon (Steinitz et al., 2007; Steinitz and Piatibratova,
G. Steinitz et al. / Journal of Environmental Radioactivity 134 (2014) 128e135
129
Fig. 1. The experiment comprised of a 3.53 L cylindrical SS canister and five gamma sensors in vertical orientation. 222Rn diffuses from a commercial source to the experimental volume via a tube and valves. All components of the system e radon source and air e are enclosed within leak tight SS components (canisters, tube) which are connected using UHV flanges and valves. Four sensors (PM-11) are placed in four orientations (g-E, g-N, g-W, g-S) next to the middle of the canister. A further sensor (Scionix) is placed in the middle of the top of the canister (g-T). Lead shielding minimized the effect of radiation from the source lowered environmental radiation. The laboratory wall is oriented EeW.
2010a,b), and raised the suggestion that an unrecognized geophysical driver influences the variations, primarily in the diurnal band. The outcome of the first experiments using the ECM configuration substantiated this assertion and led to the further suggestion that the prominent diurnal periodicity was associated with solar tide (Steinitz et al., 2011). Reanalysis of the long experimental radon time series by Sturrock et al. (2012) corroborated the results from Steinitz et al. (2011) and provided further evidence for the solar influence by identifying solar rotational frequencies (which are independent of Earth) in time series of the nuclear radiation and by showing a 24-h modulation of the gamma radiation of the radon system in the annual and in the solar rotational frequency bands. In conformity with the view in previous works Sturrock et al. (2012) suggested that the decay is influenced by solar radiation and solar neutrinos were considered as a possible particle involved. Systematic radon signals of sub-daily duration have been observed in the geological environment (Steinitz and Piatibratova, 2010a,b). In this contribution an occurrence of sub-daily shortterm (ST) signals in a simulation experiment are described. 2. Methods Simulation experiments of radon signals (Steinitz et al., 2011, 2013) are performed using the Enhanced Confined Mode (ECM) principle, which is based on: 1. A tight container in which a high level of radon in air is maintained. 2. A radon source (commercial e RaCl2) connected with a tube (and valves) supplies radon in a diffusion regime. 3. Nuclear radon detectors e alpha as internal or directly connected to the container volume, and/or gamma detectors which are either inserted into the volume (large volume) or placed outside next to the container. Our field and laboratory experience has shown that similar results in monitoring temporal variation of radon are obtained using alpha and gamma detectors. Furthermore, Zafrir et al. (2011) have shown that gamma sensors are advantageous due primarily to the higher count rates. 4. When using the gamma radiation a lead shield is added to minimize the effect of environmental gamma radiation and the eventual influence of the source on the detectors. The radiation (a, g) from the radon in the system is expected: a) to be proportional to concentration in the container air; b) to rise
once the source is connected due to diffusion from the source, and c) to level-off once the rate of diffusion equals the rate of decay of radon. Experiments conducted using different configuration showed that in contradiction with the expected large temporal variations are encountered (Steinitz et al., 2011, 2013). The phenomenology of these variations and the characteristics of the signals resemble those encountered in the natural geological environs. The specific experiment dealt with here is placed on a table, north of the eastewest laboratory wall (Fig. 1). The experimental work is based on an enhanced confined module (ECM) in which an enhanced level of radon is maintained inside a confined volume by diffusion (via tube) from an attached commercial radon source. The nuclear radiation from the radon in the confined volume is monitored, basically, with gamma sensors placed outside the confined volume. The setup used (Fig. 1) is the setup used in EXP #1 described by Steinitz et al. (2011). It consists of a 3.53 L cylindrical (Ø ¼ 15 cm) stainless steel (SS) canister, ultra-high vacuum tested, and equipped with a UHV valve. Radon is let in (diffusion) from a commercial 222Rn source (103.2 kBq), placed within a leak tight SS enclosure, via a UHV valve on the source and a 60-cm-long SS pipe, resulting in a radon level in the order of 8400 kBq/m3. In this specific experiment the atmosphere inside the ECM was of argon. Air was first evacuated from the canister and from the radon source to around 5 102 mbar. After evacuation the source valve was closed and argon was introduced to around 1 atm (controlled with a Pirani gauge). Once the system was isolated diffusion of radon was enabled by opening the source valve. Detection of temporal variation of radiation is achieved by utilizing gamma detectors with NaI(Tl) scintillation detectors. The utilization of gamma detectors for monitoring radon in air has been described in detail by Zafrir et al. (2011). Five gamma detectors were placed around the canister. Four sensors (PM-11; Rotem Industries Inc.) are based on 2 200 crystals tuned to the energy range of 50e3000 keV. These gamma sensors (PM-11) are set vertically around the central horizontal plane of the canister. They are placed, relative to the canister, according to the global directions. The orientation of these sensors is estimated to deviate not more than 5 from the true geographic directions. To allow comparison among them these sensors were adjusted (inter-calibrated) during the experiment using the radon in the canister. This was done by placing each sensor at position north and recording the radiation from radon in the canister (relative adjustment: North ¼ 1, East ¼ 1.65, West ¼ 1.8, South ¼ 1.78). A further gamma sensor (Scionix Ltd.) is based on a 36 76 mm crystal and is tuned to the energy range of 475e3000 keV. It was placed on top of the canister, in a vertical position along the axis of the canister. The
130
G. Steinitz et al. / Journal of Environmental Radioactivity 134 (2014) 128e135
experimental setup is enclosed (bottom, sides) by lead shielding (5 cm) to minimize the influence of gamma radiation from the source and from the environment. Data acquisition was with a datalogger (CR-800; Campbell Inc.) which also supplied the stabilized power to the detectors by a continuously loaded 12 V supply battery. Collected data (1 min intervals; timing precision better than 30 s) was integrated to 15min intervals. Time is shown on a decimal-day scale relative to 1.1.1992 (T ¼ 0 which is the time base of the radon measurements at the GSI). 3. Results The experimental system was used in this configuration in a series of experiments conducted from May 2009 to June 2012. The overall pattern of variation of radon during the experiment (16/1/ 2011e11/4/2011; Fig. 2) is similar to other experiments obtained with this setup (Steinitz et al., 2011, 2013). Once activated by enabling diffusion from the source buildup of radon occurs inside the SS canister, lasting around 15 days. Daily radon (DR) signals accompany the buildup. The DR signal is observed by the five sensors as discordant signals the amplitude of which is proportional to the attained level of radon. Leveling-off the signals occurs once a balance is attained between the diffusion input and the nuclear decay of radon. The plateau is dominated by large variations (20e25%) which are composed of DR (primary) and multi-day (MD) signals. The DR signal is characterized by periodicities of 24-, 12- and 8-h (i.e. 1, 2 & 3 CPD). These variations are recorded as signals of different pattern and phase at the different sensors. Once the system is isolated from the source a systematic decrease occurs which follows the nuclear decay scheme of radon. Discordant DR signals are also superimposed on the decay pattern with amplitudes which are proportional to the remaining radon level. Fig. 3 shows a typical 10-day long example of the variation pattern, dominated by the DR signal, at four primary directions along the horizontal plane of the canister and along the vertical axis. Similar daily variation occurs at the opposite sensors (east & west; north & south) while inverse patterns are observed among these two perpendicular directions. The DR pattern at the vertical sensor similar to the pattern in the northesouth direction. These features described and discussed in detail by Steinitz et al. (2013)
Fig. 2. Overall pattern of the gamma radiation from radon in the experiment, shown using hourly averages of the 15-min count rate. Time is given as a decimal day relative to 1.1.1992 (¼ Day 0). The experiment follows three stages: buildup of radon in the canister following the opening of the source (Day 6955), leveling off (around Day 6972) of the signal owing to attaining a balance between the diffusion and decay rates, and decay of radon following the isolation (Day 7019.54) of the system from the source.
Fig. 3. Detail (10 days) of the measured gamma radiation pattern (counts/15-min; hour averages) at the five sensors around the horizontal plane of the canister and along its vertical axis. The variation is dominated by DR signals with are non-similar among the sensors. Concordant patterns occur in the east and west directions and also among the north and south directions. Inverse patterns exist among the two concordant sets. The variation pattern along the vertical axis is similar to that in the northesouth direction. See text.
are attributed to directionality linked to global orientation of the temporally varying gamma radiation from radon in air at confined conditions. Besides to the DR signal a further variation pattern occurs in the time span of Days 7009e7012. The temporal pattern of the gamma radiation from radon at the five sensors of the experiment, at a resolution of 15-min, in the 10-day interval around the ST signals is shown in Fig. 4. Short term (ST) signals lasting 2e3 h are superimposed on the DR signal in this interval. The inverse parallel patterns of the DR signal in the eastewest and northesouth directions are maintained in this interval. The ST signals are recorded by all five sensors. Similar signals are not encountered in further and previous experiments performed using this configuration conducted from May 2009 to June 2012, nor in other experiments. Fig. 5 shows a five day detail around these ST signals using the measured 1-min count rate. At this resolution the ST signals are clearly resolved and manifest a periodicity in the range of 2e3 h. The strength of the ST signal is amplified during two-three intervals in Days 7009e7022.5. Different phase and amplitude are observed among the different sensors. The details of the ST signal and its relation to the DR signal are resolved by applying a 7.5-h sliding average to decompose the measured signal into the DR and ST components. Fig. 6 demonstrates both communalities as well as differences among the two signals. The DR signal (smoothed component) shows a rounded pattern. The variation of the ST signal (residual component) is composed of distinct sharp (saw-tooth) forms. The DR signal demonstrates directionality manifested as inverse mutual variation in the eastewest versus the northesouth directions, and the DR signal along the vertical axis is varying in tandem with the northe south direction. This pattern is in conformity with its pattern at other times. Conversely, the ST signal also shows directionality, but which is different. Mutual co-variation occurs in the directions eastewestenorth and an inverse pattern in the south direction. The variation along the vertical axis tends to be similar to the easte westenorth trend.
G. Steinitz et al. / Journal of Environmental Radioactivity 134 (2014) 128e135
Fig. 4. The measured gamma radiation pattern using 15-min averages, in the days around the TE event at the five sensors around the horizontal plane of the canister and along its vertical axis. The variation is composed of DR signals and also Short-Term (ST) signals which occur during three days around the TE, Compare with Fig. 3.
Fig. 5. Five day detail of the measured count rate (1-min) around the interval showing the ST signals. The ST signals, which are well defined at this time resolution, are superimposed at all sensors on the DR signal. The relative amplitude of the ST signal attains up to 20% of the amplitude of the DR signal. The two subintervals with prominent ST signals occur in the interval of Days 7009e7012 are shown in detail in Fig. 6.
The systematic patterns of the variations of the DR and ST signals indicate that directionality of the driving system is playing a role for both signal types. The systematic difference in this directionality among them indicates a different relative geometrical configuration of the driver of each signal type. This implies that the drivers must be attributed to sources at different locations, suggesting in turn that different sources are involved. In both cases the source affecting the radon signals is of a remote nature. In the case of the DR signal the driver is attributed to a component in solar irradiation (i.e. of extraterrestrial origin; op. cit.). Taking into account the observed difference of the directionality leads to the eventual conclusion that in the case of the ST signal the source is bound to Earth. Further insight is obtained by analysis in the frequency domain. The ST signal is examined by applying a FFT filter on the de-trended
131
and normalized time series, and reconstructing the signal in the range of 9e12 CPD (assuming periodicity in the range of 2e3 h). Fig. 7 shows the filtered signal during 50 days (left) and in a detail of 10 days (right). The reconstructed signals demonstrate similar patterns at the five sensors, with clear enlarged relative amplitudes in two pulses in the time interval lasting some 48 h in Days 7009e 7012. In a further step of analysis pairs of reconstructed time series at opposite directions are multiplied. The assumption is that calculating the product of normalized measurements among time series will lead to the amplification of co-varying variations while reducing, in parallel, unrelated signals. Fig. 8-A shows the result of the multiplication of the filtered time series of the two pairs of opposite sensors (east & west; north & south). The signals around in the interval of Days 7009e7012 is clearly enhanced in each of the two perpendicular directions. Further enhancing of the intense signal is obtained by forming a combined product time series of the measurements in the horizontal plane (Fig. 8-B; upper plot). This plot demonstrates the extreme reduction obtained for the background variation relative to the signal in Days 7009e7011 by using measurements in the multi-sensor configuration. The outcome is further supported by the significant co-varying temporal pattern also observed in the independent measurements by the sensor along the vertical axis of the canister (Fig. 8-B, lower plot). Sub-daily ST radon signals with a period in the range of 9e12 h have not been reported from natural environs nor have they been observed in other simulation experiments we performed. In this sense they stand as a solitary observation which cannot be associated, at this stage, with a known recurring phenomenon. Notwithstanding, the occurrence of these signals concurs with a particular outstanding geophysical event e the Tohoku earthquake. The Tohoku earthquake (TE) is a magnitude 9.0 (Mw) earthquake that occurred some 70 km east of Tohoku (Japan) at an estimated depth of 32 km, at 05:46 (UTC) on 11 March 2011. The time interval of the appearance of the ST signal coincides with the occurrence of the Tohoku earthquake (TE). The ST signals are observable from around 10 h prior to the TE event and last some 48-h after TE. This is demonstrated in Figs. 9e11. Fig. 9 shows time in two methods: a) in decimal days, as in the previous figures, in the upper x-axis, and b) relative to the TE event in the lower x-axis. The primary TE event and a reported fore-shock are indicated as vertical lines. The plots show a detailed comparison of the combined FFT reconstructed signal in the horizontal plane and the reconstructed signal of the sensor along the vertical axis of the canister. The enhancement of the signal/noise ratio as described in Figs. 8 and 9 permits improving the observations by enlarging the y-axis, as shown in Fig. 10. The two reconstructed patterns, which are derived from measurements which are geometrically independent, show co-varying response in the temporal dimension. An association is indicated with the two major seismic events e the TE and the earlier foreshock. In the case of the TE the onset of the radon signal in the 9e12 CPD band precedes the event by some 10 h in both horizontal and vertical dimension of the nuclear radiation from radon in the canister. This pre-TE signal is clearly observable relative to the filtered background. The intense primary ST signal occurs from TE þ 6 h and lasts to around TE þ 48 h. The response of the ST pattern to the main event (TE) is significantly larger compared to the response to the foreshock, occurring at TE2.13 days. In the case of the foreshock the indication in the ST radon signal occurs from 3 h after the foreshock and lasts to around 10 h after this event. The measured signal recorded at 1-min resolution and the 9e12 CPD reconstructed signal in a 1 day interval around TE are compared in Fig. 11. The time relations among the patterns confirm that the ST signal, which is significantly observed in the lower plot, was indeed initiated already 6e9 h prior to the TE event.
132
G. Steinitz et al. / Journal of Environmental Radioactivity 134 (2014) 128e135
Fig. 6. The decomposed DR signal (left) and ST signal (right), using 1-min measurements. The decomposition was obtained by applying a 7.5 h sliding average. The ST signal is shown in detail for two time intervals in the interval 7009e7012 (see Fig. 5). The E&W versus N&S directional pattern of the DR signal is maintained. The directional pattern of the decomposed ST signal is different. See text.
Fig. 7. The FFT filtered and reconstructed signal in the frequency range of 9e12 CPD for the four sensors around the horizontal plane of the canister and the vertical sensor along the axis of the canister. Left e 50 day sequence. Right e 10 day detail around the interval showing the ST signals.
G. Steinitz et al. / Journal of Environmental Radioactivity 134 (2014) 128e135
133
2e-3
1e-3
7005
7007
7009
7011
7013
0
2
4
5.0e-3
gE*gW*gN*gS
g-TOP
Day since 1.1.1992
EW*NS g-TOP Foreshock (Mw = 7.2) TE (Mw = 9)
2.5e-3 0
-1e-3
-2e-3
0.0
-3e-3
-4
-2
Time relative to TE (day) Fig. 9. A 5-day detail of Fig. 8-B showing the temporal relation between the FFT 9e12 CPD reconstructed signals and the TE seismic event. The lower x-axis shows time relative to the main TE event (at t ¼ 0) and the pre-shock (at t ¼ 2.15 days). The combined FFT reconstructed signal in the horizontal plane is shown in the lower plot (black); the upper plot (red) shows the reconstructed signal of the sensor along the vertical axis of the canister. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 8. A e Composite time series formed by the multiplication of reconstructed time series shown in Fig. 7. The two time series represent the result in the eastewest and northesouth direction. Enhanced signals occur around in the interval of the ST signals (see text). B e The variation in the horizontal plane: Composite time series (lower plot) formed by the multiplication of the product in the two principal directions (shown in A); a concordant pattern is exhibited (upper plot) by the filtered 9e12 CPD signal at the sensor along the vertical axis of the canister. The pattern of the 9e12 CPD signal is further enhanced in the interval of the ST signals.
communality in the characteristics of the driving mechanism. For the DR signals it was previously shown in such experiments that they due neither to mechanical nor to local environmental influences on the experimental system, but rather to an unidentified remote driver. A similar conclusion is reached for the ST signals e i.e. that a remote driver is also involved in their generation. In the case of the DR signal it was recently suggested that a component of solar irradiance is influencing the pattern of nuclear radiation from radon in air (Sturrock et al., 2012). Taking into account both similarity and differences between the ST and DR signal in terms of their global orientation patterns it is suggested that the ST signal is
4. Discussion Radon signals spanning from annual to daily scale are frequent in the air of the subsurface geological environment, and differing views are offered as to the nature of the processes driving them. Recently it was demonstrated that similar radon signals can be simulated in the laboratory using the enhanced confined mode (ECM) configuration. Utilizing an ECM system with a multi-sensor configuration around it the described experiment demonstrates, for the first time, that short term (ST) radon signals with periods of 2e3 h (9e12 CPD) also occur. These signals are due to temporal variation of the gamma radiation evolving from radon (progeny) inside the canister. Based on the accumulated experience the occurrence of ST signals is rare compared to the longer term AR, MD and DR radon signals. The ST signal clearly observed in the measured data in this experiment using a multi-sensor configuration is also well characterized in the reconstructed signal obtained by filtering and combining in the frequency domain. Systematic spatial inhomogeneity of the ST signals is reflected as differing forms and phase at different directions around the canister, clearly indicating that the ST signal is generated in the canister due to an external influence, probably of a remote nature. This feature is similar to the spatial manifestation of form and phase of the DR signal in such experiments, indicating a
Fig. 10. Comparison of the 9e12 CPD FFT reconstructed signals and their temporal relation to the primary seismic event TE (at t ¼ 0) and the pre-shock (t ¼ 2.15 days). An enlarged y-axis is used, compared to the one in Fig. 9. The plots portray the combined reconstructed signal in the horizontal plane and the signal along the signal perpendicular to it e along the vertical axis of the canister. A 9e12 CPD variation occurs around 6 h prior to the TE event. An enlarged y-axis is used, compared to the one in Fig. 9.
134
G. Steinitz et al. / Journal of Environmental Radioactivity 134 (2014) 128e135
pre-shock. These ST signals appear from around 10 h prior to the TE and are maintained to around 48 h after the TE. The relatively intense manifestation of this phenomenon occurs after TE. If this association between TE event and the ST radon signals is indeed due to a geophysical connection then the following implications are indicated: 1. The overall progression of the TE (Mw ¼ 9) event is also associated with a non-mechanical geodynamic process which is reflected by nuclear radiation features of radon in air. This process is modifying (superimposed) the driver of the primary DR signal of the radon system. 2. The influencing process is operating at a global scale, but at this stage it remains open whether the effect is transferred to the experimental setup via the solid earth and/or the atmosphere. 3. Detection of such phenomena is possible in radon systems of specific configuration the parameters of which are so far undetermined. 4. The possibility is raised that a pre-cursor of the TE has been detected.
Fig. 11. Comparison, in the two day interval around TE, of the measured signals (upper plot) with the signal (lower plot) reconstructed in the frequency range of 9e12 CPD. The time scale, relative to the TE event, is the same in both plots.
driven by a remote source which in this case is located on Earth, in variance with the DR signal. The fact that similar short term signals have not been encountered in other experiments or field locations may be due to several factors. The experience gained shows that patterns of radiation from radon in air are influenced by geometrical parameters (canister dimensions, location, and orientation) which so far are unresolved, and that the sensitivity for the detection of DR signals is strongly related to the overall radon level. The following implications are indicated for the processes underlying the temporal and spatial variation of the nuclear radiation from radon in air: 1. Results obtained by the GSI group indicate that nuclear radiation from radon (and progeny) inside a confined volume of air varies spatially and temporally at time scales from annually to daily. The geophysical drivers of these signals are unclear at this stage. The new experimental results demonstrate that the same radon system is responding to further geophysical drivers operating at a time scale of 2e3 h. 2. The results further demonstrate the potential of investigations utilizing enhanced radon levels within confined volumes for detection of a new type of time varying geophysical phenomena. 3. Performing ECM experiments utilizing relatively high radon levels leads to improved signal to noise conditions and thereby allows detection and investigation of ST signals in the radon signals. 4. If causality is indeed underlying the high concordance in time between the occurrence of the ST signal in the radon system and the Tohoku earthquake, which is an extreme geodynamic event, then a new research track is indicated. In the temporal dimension the exceptional ST signals occur in association with the Tohoku earthquake and possibly also with its
Due to the large effect of disastrous earthquakes the issue of their prediction is often dealt with. A wide spectrum of physical phenomena for earthquake prediction is suggested and is extensively debated in the literature. For a recent review see Jordan et al. (2011). Very different views exist including ones maintaining that such events cannot be predicted (e.g. Geller, 1997; Geller et al., 1997). Radon in the natural environment, either in air or in water, is one of the leading candidates suggested as a proxy of seismogenic activity (Hartmann and Levy, 2005; Immè and Morelli, 2012). As is the case of other proxies a clear-cut situation has not been demonstrated. Furthermore, the lack of a sound physical and geophysical frame for the exceptional and complex patterns of radon variation in nature stresses the fact that this assertion is not established. The present contribution demonstrates, in agreement with other observations we recently presented, that the fundamental physical and geophysical properties are different from those assumed so far. Only a new understanding of the prosperities of the radon system may facilitate the application in earthquake prediction. This work shows that other physical drivers underlie its variation. Furthermore it demonstrates, again, the potential in investigating the geophysical aspects influencing radon system and their implication for using it as an indicator of geophysical phenomena and processes. Acknowledgments This investigation was supported by the Geological Survey of Israel. References Geller, R.J., 1997. Earthquake prediction: a critical review. Geophys. J. Int. 131 (3), 425e450. http://dx.doi.org/10.1111/j.1365-246X.1997.tb06588. Geller, R.J., Jackson, D.D., Kagan, Y.Y., Mulargia, F., 1997. Earthquakes cannot be predicted. Science 275, 1616. http://dx.doi.org/10.1126/science.275.5306.1616. Hartmann, J., Levy, J.K., 2005. Hydrogeological and gasgeochemical earthquake precursors e a review for application. Nat. Hazards 34, 279e304. Immè, G., Morelli, D., 2012. Radon as an earthquake precursor (Chapter 7). In: D’Amico, S. (Ed.), Earthquake Research and Analysis e Statistical Studies, Observations and Planning, pp. 143e160. Jordan, T.H., Chen, Y.-T., Gasparini, T., Madariaga, P., Main, I., Marzocchi, W., Papadopoolos, G., Sobolev, G., Yamaoka, K., Zschau, J., 2011. Operational earthquake forecasting: state of knowledge and guidelines for utilization. Ann. Geophys. 54 (4), 315e391. http://dx.doi.org/10.4401/ag-5350. Steinitz, G., Piatibratova, O., Barbosa, S.M., 2007. Radon daily signals in the Elat Granite, southern Arava, Israel. J. Geophys. Res. 112, B10211. http://dx.doi.org/ 10.1029/2006JB004817.
G. Steinitz et al. / Journal of Environmental Radioactivity 134 (2014) 128e135 Steinitz, G., Piatibratova, O., 2010a. Radon signals in the Gavnunim intrusion, Makhtesh Ramon. Isr. Geophys. J. Int. 180, 651e665. Steinitz, G., Piatibratova, O., 2010b. Radon signals at the roded site, southern Israel. Solid Earth 1, 99e109. http://dx.doi.org/10.5194/se-1-99-2010. Steinitz, G., Piatibratova, O., Kotlarsky, P., 2011. Possible effect of solar tides on radon signals. J. Environ. Radioact. 102, 749e765. http://dx.doi.org/10.1016/ j.jenvrad.2011.04.002. Steinitz, G., Kotlarsky, P., Piatibratova, O., 2013. Anomalous non-isotropic temporal variation of gamma-radiation from radon (progeny) within air in confined conditions. Geophys. J. Int. 193, 1110e1118. http://dx.doi.org/10.1093/gji/gg057.
135
Sturrock, P.A., Steinitz, G., Fischbach, E., Javorsek, D., Jenkins, J.H., 2012. Analysis of gamma radiation from a radon Source: indications of a solar influence. Astropart. Phys. 35/1, 18e25. Zafrir, H., Haquin, G., Malik, U., Barbosa, S.M., Piatibratova, O., Steinitz, G., 2011. Gamma versus alpha sensors for Rn-222 long-term monitoring in geological environments. Radiat. Meas. 46, 611e620. http://dx.doi.org/10.1016/ j.radmeas.2011.04.027.
