Introduction

Radon (222Rn) in the environment has been extensively studied to evaluate dose levels due to inhaling the radioactive gas into the body, to estimate its flux from the ground surface to the atmosphere (exhalation) as a tracer of air movement in the lower atmosphere, and to predict seismic activity in tectonically active areas (Dörr et al. 1983; Zahorowski et al. 2004; Vaupotič et al. 2010). Another important and basic aspect exists in the study of 222Rn in soil air to elucidate mechanisms of its migration in soil and transportation to the ground surface (Nazaroff 1992; Neznal and Neznal 2005; Fujiyoshi et al. 2010, 2013). Soil radon monitoring has shown that 222Rn activity concentration varies to a great extent depending on geological, meteorological and hydrological factors, and that its migration through the soil and rock is controlled not only by diffusion but also by convection (Yakovleva 2005; Perrier and Girault 2013). Monitoring soil radon for several years since 2004 gave the results that a major factor affecting soil 222Rn concentration was soil temperature under high atmospheric pressure region in three seasons, except for winter on a campus forest site of the Hokkaido University, Japan (Fujiyoshi et al. 2005, 2006, 2010). In contrast, lower 222Rn concentration with small variability appeared in soil under persisting snow in winter months from December to March. They further detected a small amount of 222Rn releasing from the ground surface to the overlying snowpack with a mean flux density of 0.4 mBq m−2 s−1 at their observation site in Hokkaido (Fujiyoshi et al. 2013).

It is widely recognized that both liquid and gaseous water movements are fundamental factors controlling many processes in soil. Soil water dynamics are strongly linked to temperature variations and then biological activities. These processes, complicated due to interrelations among controlling factors, have not been clarified thoroughly (Wells et al. 2007; Bittelli et al. 2008). Understanding radon transportation in soil is useful for evaluating soil air movement in the surface soil layers, because radon is chemically inert and radioactive.

A site concerned in this study was on the campus forest of the Fukushima University, Fukushima, Japan. Large amounts of radionuclides such as 131I (T 1/2 = 8.04d), 134Cs (T 1/2 = 2.06y) and 137Cs (T 1/2 = 30.17y) were released and deposited in the environment due to the accident of the Fukushima Daiichi Nuclear Power Plant in March 2011 as a consequence of a big earthquake (Great East Japan Earthquake) of magnitude 9.0 and following a 15 m tsunami. Total amounts of 131I, 134Cs and 137Cs released to the atmosphere were estimated by several institutes and universities to be approximately 200, 20 and 20 PBq over March 12–31, 2011, respectively (Science Council of Japan 2012). The majority of the contaminated area (about 70 %) is covered by forests in which a large portion of deposited radiocesium existed in the canopies of coniferous forests, whereas fallen leaves on the ground surface contained most of the radiocesium in deciduous forests as of September, 2011 (Hashimoto et al. 2012). Much effort has been devoted to remove deposited fallout radionuclides in forest areas so far; however, a large part still remains without decontamination. Our previous results showed that most radiocesium deposited on fallen leaves of deciduous trees in March 2011 rapidly moved to the organic layer of the soil surface in November of this year (Hao et al. 2013). The above fact probably resulted from intensive (micro) biological activities in the organic layer in summer. Decomposing organic matter releases CO2 and other gaseous components into pore spaces in soil. It is therefore important to evaluate potential effects of soil air (and water) movement on the fate of deposited radiocesium in the surface layer of forest soils.

The primary purpose of the present study is to elucidate basic behavior of soil radon (222Rn) including activity level, variability and transportation mechanism(s) in forest soil in Fukushima contaminated with fallout nuclides. Special concern is to clarify effects of two typhoons passing over the observation site on soil radon transportation.

Experimental procedure

Description of the site

Figure 1 shows a location map of our observation site (37.68457N, 140.45347E), a campus forest of the Fukushima University, Fukushima Prefecture, Japan. The site (about 200 m a.s.l.) belongs to the Fukushima Basin which is surrounded by the Azuma Mountain Range in the west and Abukuma Highland in the east. The original sloping terrain in this area was modified to construct a new university building in the 1980s (Fukushima Prefecture 1982).

Fig. 1
figure 1

Location map and instrumentation of radon probes in soil on the campus forest of Fukushima University, Fukushima, Japan

Full size image

It is a typical basin-specific climate of high temperature and high humidity in summer with annual mean temperature and precipitation, 12.8 °C and 1,105 mm, respectively. Annual mean of maximum snow depth in winter is about 8 cm.

There is a great variety of vegetation on the campus of the Fukushima University (Kurosawa et al. 2010). For example, tree species growing at the site are: oaks (Quercus serrate, Quercus acutissima), pine tree (pinus densiflora), chestnut tree (Castanea crenata) and Japanese laurel (Aucuba japonica).

Soil properties

Several soil properties were measured including porosity, humidity, soil organic matter and pH. Porosity (and humidity) was determined by measuring the weight of a sample in a container of known mass: (1) in the field (bulk weight), (2) after filling it with pure water (water-filled weight) and (3) after complete drying at 100 °C for more than 24 h (dried weight). The amount of soil organic matter was estimated as a difference of a sample weight before and after heating at 500 °C for 2 hours. Soil pH was measured in situ with a pH meter (HI 99121, HANNA INSTRUMENTS, USA), in which the electrode was inserted into the soil at a depth of about 10 cm (Fujiyoshi and Sawamura 2004).

Soil radon (222Rn) measurement

Activity concentration of 222Rn in soil air was monitored at three soil depths with radon probes (VDG, Algade, France), in which silicon detectors of 450 mm2 in the total detection area count alpha particles in the 0.7–6.1 MeV energy range, emitted from 222Rn radioactive transformations. The radon probes in PVC housing tubes (0.5, 0.8 and 1.2 m in length) were buried in soil at 0.3, 0.6 and 1.0 m depths on August 21, 2013. Detection efficiency was 51.6, 51.8 and 45.8 Bq m−3/(impulse h−1) at 0.3, 0.6 and 1.0 m, respectively. Probes were located close to each other in the ground, which could be assumed to be the same place of a local environment. A data logger stored hourly data on 222Rn activity concentration, barometric pressure and temperature at each depth.

Monitoring differential barometric pressure

Differences in barometric pressure on the surface and at three depths were measured once an hour in the casing tubes of 222Rn probes. The whole system prepared on demand by a private company (North One Co., Ltd., Sapporo, Japan) consists of a micro-barometric sensor (JP208, Yokogawa Electric, Japan) and a data logger (KADEC21-U4-C, North One Co., Ltd., Japan).

Gamma spectrometry

Activity concentration of several environmental radionuclides (40K, 134Cs, 137Cs, 210Pb) as well as 226Ra, the parent nuclide of 222Rn, in soil was determined by gamma spectrometry with a HPGe detection system (SEIKO EG&G, Japan). Standard reference materials (IAEA 327 and IAEA 444) were used to evaluate activity concentrations of individual radionuclides from counts obtained with the same geometry under identical operating conditions. Energy and efficiency calibrations were periodically carried out, as well as checking the background. Details of the measurements were described in Fujiyoshi et al. (2010, 2013).

Results and discussion

Figure 1 shows the monitoring point on the campus of the Fukushima University in Fukushima City, Japan, where the areas were contaminated with fallout radionuclides due to the accident of Fukushima Daiichi Nuclear Power Plant on Mar. 11 2011. As of 2013, most campus sites except for the forest parts have already been decontaminated. University staff members have regularly measured and reported radiation dose rate at 1 m height above the ground at many selected points on the campus (http://www.fukushima-u.ac.jp/guidance/top/fukudai-housyasen.html/).

Table 1 summarizes some of the soil properties including activity concentration of environmental radionuclides (40K, 210Pb, 226Ra) in our test site where the original hilly landscape was modified to construct a new university building more than 30 years ago. This fact reflects relatively homogeneous distribution of 40K and 226Ra with depth as summarized in the table. Depth distribution profiles of 210Pb, a 222Rn progeny, showed a small surface enrichment, suggesting atmospheric lead deposition on the forest floor since 1981 when the construction of university buildings was finished. It should be noted here that much higher concentration of 210Pb is usually observed in the surface layer of forest soils undisturbed for more than 100 years (Fujiyoshi and Sawamura 2004).

Table 1 Summary of soil properties (humidity, pH, porosity, activity concentrations of 40K, 210Pb and 226Ra) on the campus forest of Fukushima University, Fukushima, Japan
Full size table

Figure 2 shows the depth distribution profiles of radiocesium (134Cs and 137Cs) activity concentration in soil, in which all the values in the figure were calculated on August 21, 2013. Two and a half years after the accident, activity concentration of short-lived 134Cs (T 1/2 = 2.07 y) was about a half of that of 137Cs (T 1/2 = 30.17 y). Activity concentration of radiocesium present within the surface portion of soil (depth of <5 cm) decreased exponentially with soil depth, thus suggesting no natural and/or anthropogenic intervention since then (Fujiyoshi et al. 2011).

Fig. 2
figure 2

Depth distribution profiles of radiocesium (134Cs and 137Cs) in soil collected on the test site on August 21 2013. Activity concentration of both nuclides was corrected to the values on the time of soil sampling (August 21 2013)

Full size image

Soil radon (222Rn) monitoring started in August, 2013 on the observation site of known basic information on geology, climate and soil properties, already described in the site description in the “Experimental procedure” and also in Table 1 Materials and methods). Figure 3 shows time series changes in 222Rn activity concentration in soil air at different soil depths (0.3, 0.6, and 1.0 m) at the observation point, together with atmospheric pressure from August 21 to December 6, 2013. Soil radon level varied to a great extent depending on various factors, including meteorological and soil parameters. Different soil radon levels appeared at different soil depths, indicating equivalent 222Rn concentrations (222Rneq) to be 7.5, 14 and 23 kBq m−3 at 0.3, 0.6 and 1.0 m in depth, respectively (Dörr and Münnich 1990). Plotting 222Rneq values against soil depth gave an infinite 222Rn concentration (222Rn ∞ ) as 53.6 kBq m−3 at our observation site (Dörr and Münnich 1990). Using this value of 222Rn ∞ (53.6 kBq m−3), soil density (ρ), 226Ra activity concentration in soil (226 Ra) and total porosity (p) shown in Table 1, emanation coefficient of 222Rn (ε) was estimated with the following equation:

$$ {}^{222}\text{Rn}_{\infty } = {}^{226}Ra \cdot \varepsilon \cdot \rho \cdot (1 - p)/p $$
(1)

Values of ε obtained at different depths were then used for evaluating 222Rn generation rate (v s ) from the source with the decay constant of 222Rn (λ Rn) as follows:

$$ v_{s} (Bq \, m^{ - 3} \, s^{ - 1} ) = \varepsilon \cdot \rho \cdot \, {}^{226}Ra \cdot \lambda_{Rn} \cdot (1 - p)/p $$
(2)

Figure 4 gives the v s values at different soil depths giving a minimum value (0.64 Bq m−3 s−1) at a depth of 0.2 m, and increasing values up to 2.2 Bq m−3 s−1 down to a depth of about 0.5 m. The results may suggest that it takes several hours for 222Rn to be in an equilibrium state in soil.

Fig. 3
figure 3

Time series plots of 222Rn activity concentration at different depths (0.3, 0.6 and 1.0 m) and of atmospheric pressure from August 21 to December 6 in 2013. Two typhoons struck the test site on September 16 (Typhoon No. 16) and October 16 (Typhoon No. 23) in 2013. Upward (E1, E2, E3, E4, E5, E6, and E7) and downward (E8 and E9) changes in 222Rn concentration with time were observed at all the depths, in which E denotes to an abbreviation of event

Full size image
Fig. 4
figure 4

Change in 222Rn generation rate (Bq m−3 s−1) with soil depth calculated from sets of observed data on 226Ra activity concentration and porosity of the soil

Full size image

Now, it should be noted that two big typhoons (Typhoons No. 18 and No. 26) passed over the site in mid-September and in mid-October, 2013, respectively, as shown in Fig. 3. Atmospheric pressure decreased drastically, and it then recovered within a short period of time depending on the moving speed of a typhoon. In contrast, soil 222Rn activity concentration decreased slowly to a bottom value (~4 kBq m−3) at all the depths (0.3, 0.6 and 1.0 m) during the typhoon periods. It is probably because supplying rate of 222Rn from the parent 226Ra in soil was too low to catch up with the concentration in a steady-state level in this period.

There is a lack of information in the literature concerning effects of typhoon on soil radon concentration. It may be the only one that Richon et al. (2003) monitored soil radon (222Rn) at Taal volcano in Philippines from 1993 to 1996 to investigate possible relationship between 222Rn and an earthquake. They concluded that a 222Rn anomaly (extremely high 222Rn level) appeared 22 days before the M 7.1 Mindoro earthquake in 1994, being a precursor of the quake, not resulting from typhoon Teresa passing a few days before. According to them, the only proof for the above conclusion was that 2022Rn level was not affected so seriously by another super typhoon (Angela) striking the island just one year later. Findings of theirs were obviously different from ours in the present study, in which soil radon concentration was affected greatly by the passing typhoon.

To elucidate the mechanism of 222Rn transportation in soil air, nine periods of time (E1–E9) depicted in Fig. 3 were selected, in which upward (E1–E7) and downward (E8 and E9) changes in 222Rn concentration with time appeared at all three depths. Here, it was assumed that increasing 222Rn concentration after the Typhoon No. 18 (Event 3) in the figure should be a diffusion-controlled process due to a 222Rn concentration gradient between shallow and deeper portions of the soil under a recovering high-pressure region in the atmosphere. This assumption leads to obtaining effective diffusion coefficient (D e ) of 222Rn with the equation below, where 222 Rn, F and x denote to 222Rn concentration (Bq m−3), 222Rn flux (kBq m−2 s−1) obtained at E3 in Fig. 3 and soil depth (m), respectively. A differential part (d222Rn/dx) in Eq. (3) was evaluated using an equilibrium concentration of 222Rn at each depth.

$$ F = D_{e} \left( {\frac{{d{}^{222}Rn}}{dx}} \right) $$
(3)

Mean D e value thus obtained as 5 × 10−6 m2 s−1 is reasonable by considering homogeneous and well-drained soil in our test site (Nazaroff 1992; Sakoda et al. 2011). Diffusion-controlled 222Rn flux in soil air was therefore estimated to be about 1 × 10−1 Bq m−2 s−1 during the observation period from August to September, 2013. It is much higher than that obtained in a forest site of Sapporo under thick snowpack in winter (Fujiyoshi et al. 2013).

Transportation mechanism of 222Rn in soil air was considered with a set of data on 222Rn flux at each period of time (E1–E9). As shown in Fig. 5, 222Rn in soil air moves by a diffusion-controlled mechanism in most of the periods shown in Fig. 3. However, there are two cases that mass flow of soil air controls 222Rn transport as in cases E8 and E9 in which an approaching low-pressure front caused strong upward movement of air from the soil to the atmosphere resulting in a lack of 222Rn in soil air during the typhoon periods. The difference in the 222Rn flux between downward (E8) and recovering (E3) time affected by the Typhoon No. 18 (Sept 16 2013) is probably due to mass flow of soil air in the former.

Fig. 5
figure 5

Flux of 222Rn as a function of soil depth obtained on nine events shown in Fig. 3. Most of the events except for E1, E8 and E9 followed diffusion-controlled 222Rn transportation in soil air. Mass flow of soil air was predominant in the 222Rn transportation only in the cases of E8 and E9, when extremely low atmospheric pressure front was passing over the observation site

Full size image

After the typhoon No. 26 (October 16, 2013) passed on, 222Rn concentration was not recovered up to the previous level at each depth (Fig. 3). Small up and down changes in radon level appeared for about 10 days after this typhoon. Such a behavior of 222Rn in soil air is supposed to be associated with changing meteorological conditions and soil properties in autumn season when atmospheric temperature gradually decreases. Figure 6a shows consequences of 222Rn concentration at three depths and of precipitation (mm h−1). As shown in the figure, it rained several times ranging from 1 to 9 mm h−1, which clearly affected subsequent soil radon level to a great extent. Temperature in soil air (and also on the ground surface) did not show clear diurnal variability in this period (Fig. 6b). Barometric pressure at different depths showed small differences, especially when the surface portion (0.3 m) is compared with the other two (0.6 and 1.0 m), as shown in Fig. 6c. However, it is clear in Fig. 6d that barometric pressure difference (P g –P s ) between ground surface (P g ) and soil air (P s ) differed in time series profile at a depth of 0.3 m if compared with other depths. As depicted from Figs. 6a and b, properties of surface soil was directly affected by meteorological conditions, such as air movement and intensity and amount of precipitation. Frequent precipitation in autumn season gave some retarding effect on 222Rn concentration at different depths, in which a deeper layer of the soil responded much slower to changing meteorological conditions (Fig. 6a). Dynamic behavior of 222Rn in surface soil found in this study suggests that further investigation should be necessary for evaluating effect of humidity on 222Rn variability in soil air. This may be further required to elucidate effects of water (and water vapor) transportation on the fate of fallout nuclides deposited on the forest floor. Monitoring has still been continued, aimed at evaluating water (water vapor) movement in soil as well soil 222Rn concentration at three depths in the present test site.

Fig. 6
figure 6

a, b, c, d Time series plots of 222Rn concentration with precipitation (a), 222Rn concentration with soil temperature (b), barometric pressure (c) and difference in barometric pressure on the ground surface from those in soil air at different depths (d) obtained from October 15 to 30, when Typhoon No. 23 had passed over the test site. In a, variability of 222Rn concentrations with time did not show similar patterns at individual depths. It is clear in d that pressure difference between the ground surface and soil air was quite different only at the surface portion (0.3 m in depth) of the soil

Full size image

Conclusions

Soil radon (222Rn) was monitored from mid-summer to early winter in 2013 at three depths in a forest site contaminated with fallout radionuclides derived from the accident of the Fukushima Daiichi Nuclear Power Plant in 2011. Different 222Rn levels appeared depending on soil depths, which gave an equivalent 222Rn concentration (222 Rn eq ) of 7.5, 14 and 23 Bq m−3 at the depths of 0.3, 0.6 and 0.7 m, respectively. Two big typhoons passing over the observation site showed a great effect on soil radon variability, in which upward transportation of 222Rn associated with passed typhoon was governed by diffusion-controlled mechanism. Effective diffusion coefficient of 222Rn (D e ) was easily evaluated using time series 222Rn data at three different soil depths during a typhoon event in September, 2013. This is important for investigating dynamic behavior of 222Rn in soil air with an “in situ” diffusion coefficient. Diffusion-controlled 222Rn flux was obtained in selected time regions using this effective diffusion coefficient. Precipitation and subsequent change in soil humidity also affected 222Rn level to a great extent, depending on soil depth, which remains to be investigated thoroughly. The authors have been further studying dynamic behavior of gaseous and water components in soil using 222Rn as a radiotracer for tracking the fate of deposited radiocesium on the forest floor.