Exchangeability of ¹³⁷Cs and K in soils of agricultural fields after decontamination in the eastern coastal area of Fukushima

ABSTRACT

To examine the exchangeability of ¹³⁷Cs and K in soils of agricultural fields after decontamination, 173 soil samples were collected in 2016 from the plowed layer (0–15 cm depth) of decontaminated agricultural fields in Tomioka town, Fukushima prefecture, Japan. We investigated total ¹³⁷Cs, exchangeable ¹³⁷Cs, exchangeable K, and nonexchangeable K content for these soils. The total ¹³⁷Cs content in soils was, on average, 1,200 Bq kg⁻¹ (value range: 20–4,400 Bq kg⁻¹), which was an 80% decrease from the values determined before the decontamination. This result indicated that the decontamination process considerably reduced the total ¹³⁷Cs content in the plowed layer. The exchangeable K content in soils was on average 172 mg kg⁻¹ (value range: 27.9 – 743 mg kg⁻¹). More than 80% of soils showed the exchangeable K content lower than the recommended threshold value for paddy soils needed in order to reduce the ¹³⁷Cs transfer from soil to plant (i.e. 200 mg kg⁻¹). The nonexchangeable K content showed a wide range of values (value range: 98.8–1,280 mg kg⁻¹) and no significant correlation with the exchangeable K content. Because nonexchangeable K is released more slowly but is comparable to exchangeable K as a reservoir of phytoavailable K, more frequent K application is recommended for soils with lower contents of both exchangeable and nonexchangeable K. The total ¹³⁷Cs content in soils showed a significant positive correlation with the exchangeable K content in soils (p < 0.01), which means that decontamination efforts intended to reduce the total ¹³⁷Cs content in soils might also lead to a reduction in the exchangeable K content. In conclusion, the application of K fertilizer is strongly recommended in the decontaminated agricultural fields because the decontamination causes a considerable decrease in exchangeable K content along with a reduction in total ¹³⁷Cs content.

KEY WORDS:

• Decontamination
• Fukushima
• potassium
• radiocesium
• soil

1. Introduction

The Great East Japan Earthquake and the subsequent tsunami on 11 March 2011 caused TEPCO’s (Tokyo Electric Power Company Holdings, Inc.) Fukushima Daiichi Nuclear Power Plant (FDNPP) accident. Radioactive fission products, including ¹³¹I, ¹³⁴Cs, and ¹³⁷Cs, were released into the environment (Evangeliou et al. 2013), and soils over a wide area of Fukushima and the neighboring prefectures were contaminated by these radionuclides (Christoudias and Lelieveld 2013). Of all these radionuclides, ¹³⁷Cs has been a major concern in terms of contamination even though seven years have passed since the 2011 accident, because the amount released was very large (1.3 × 10¹⁶ Bq) (Chino et al. 2011) and it has a longer half-life (30.2 y) than ¹³¹I (8.04 d) or ¹³⁴Cs (2.06 y). Because the maximum standard limit of radiocesium in food was set to 100 Bq kg⁻¹ (Ministry of Health, Labor and Welfare 2012), it was necessary to reduce radiocesium in the food produced from these contaminated agricultural fields. Hence, by 2018, as much as 2.4 × 10⁴ km² in Fukushima and the neighboring prefectures had been decontaminated to reduce the total radiocesium in agricultural soils derived from the accident, although the decontaminated area did not include the ‘difficult-to-return’ zones (Ministry of the Environment 2018).

After the FDNPP accident, Tomioka town, located approximately 10 km south of FDNPP, was designated as a residence restriction area. The evacuation order for this town was lifted on 1 April 2017, except for the difficult-to-return zone in the northeastern part of town. The decontamination of agricultural fields in this town involved the following steps: (1) removing the contaminated surface soil layer (0–5 cm depth); (2) dressing typically with uncontaminated granitic sand (0–5 cm thickness) in place of the removed soil; and (3) mixing the fresh soil with the subsurface soil (5–15 cm depth). After the decontamination, the total radiocesium (¹³⁷Cs + ¹³⁴Cs) content in the plowed layer (0–15 cm depth) had been reduced from 9,100 Bq kg⁻¹ to 1,260 Bq kg⁻¹ in a representative sample of the fields (n = 17) (Tomioka town 2016). The lower ¹³⁷Cs content in soils directly contributes to a smaller external radiation exposure during agricultural activity in the fields (Satoh et al. 2014), whereas it does not necessarily reduce the ¹³⁷Cs transfer from soil to plant (Saito et al. 2012) if the soils have high exchangeability of ¹³⁷Cs or low exchangeability of K. The ¹³⁷Cs transfer becomes smaller if the soils have a lower exchangeable ¹³⁷Cs content (Kondo et al. 2015); however, the ¹³⁷Cs transfer is also reduced under conditions of higher phytoavailable K content (Eguchi et al. 2015; Kato et al. 2015; Zhu and Smolders 2000). Phytoavailable K can be separated into exchangeable K (Ex-K) and nonexchangeable K (Nex-K) (Kitagawa et al. 2018). Ex-K, electrostatically adsorbed on the surface of minerals and organic matter, is readily available for uptake by plants. Although Nex-K is released more slowly and retained more strongly by minerals such as 2:1 phyllosilicates, it is also an important reservoir of phytoavailable K (Moritsuka et al. 2004). The uptake of radiocesium by plants was controlled by these phytoavailable K fractions, especially by the Ex-K content in soils. In spite of their importance, no previous study has elucidated this exchangeability of ¹³⁷Cs and K, or the levels of their exchangeable fractions, over a wide area of decontaminated agricultural fields.

The objective of this study was, therefore, to investigate the exchangeability of ¹³⁷Cs and K and to discuss the relationship between these exchangeabilities and other soil properties for soils over a wide area of the decontaminated agricultural fields in Fukushima.

2. Materials and methods

2.1. Soil samples

In total, 173 soil samples were collected on November 8–10, 2016 from the plowed layer (0–15 cm depth) of 173 agricultural fields distributed widely in Tomioka town, Fukushima prefecture, Japan (N37°20′, E141°00′) (Fig. 1). No fertilizer had been applied after decontamination when we collected the soil samples, i.e. the soil samples were in the post-decontamination state just after decontamination. The soils investigated were classified into five soil groups based on the classification of original soils described in the Soil-Inventory (National Agricultural and Food Research Organization): Muck soils, Andosols (Wet Andosols and Andosols), Lowland soils (Gray Lowland soils and Brown Lowland soils), Yellow soils, and Brown Forest soils (Cultivated soil classification committee 1995). The number of samples for Muck soils, Andosols, Lowland soils, Yellow soils, and Brown Forest soils was 22, 57, 64, 11, and 7, respectively, with an additional 12 unclassified samples. In each field, subsamples of soil were collected from five points (at the center and the four vertices of a square with a side length of 15 m) and mixed together. The hourly air dose rate (µSv h⁻¹) at a height of 1 m was also determined at the center of each field using a gamma-ray radiation detector (Dose e, Fuji Electric Co., Ltd., Japan). Three extraordinarily high air dose rates were detected in those fields with solar panel installations; these values of air dose rates were not used in this study owing to possibly unreliable measurements of the air dose rates due to electromagnetic malfunction. Soil samples were dried at 50°C for more than two days and then sieved to less than 2 mm before analysis.

Figure 1. Locations of the sampling sites in Tomioka town, Fukushima prefecture, Japan. The pin indicates the location of the Fukushima Daiichi Nuclear Power Plant (FDNPP) in Okuma town. This map is based on the blank map and aerial photograph published by Geospatial Information Authority of Japan (https://maps.gsi.go.jp/development/ichiran.html)

2.2. Determination of total and exchangeable ¹³⁷Cs content

A 20 cm³ volume of soil sample corresponding to approximately 20–40 g was filled in a polyethylene vial (20 cm³). The vial was set in an auto gamma counter with a NaI (Tl) detector (2480WIZARD², PerkinElmer, Waltham, Massachusetts, USA, Radioisotope Research Center of Kyoto Prefectural University) to determine the gamma activity (between 589 and 735 keV) in soils. The counting time was set to 1,800–7,200 s to decrease the statistical counting error to less than 1.7%. One of these 173 soil samples showed a total ¹³⁷Cs content that was less than the detection limit. As the resolution of the NaI detector is not high enough to differentiate the gamma ray of ¹³⁴Cs from that of ¹³⁷Cs, the results were calibrated with a germanium (Ge) detector. More specifically, approximately 60–80 g portions of four representative soil samples were filled into styrene U-8 vessels (100 mL), and then their respective gamma activities from ¹³⁷Cs and ¹³⁴Cs were analyzed using a Ge semiconductor detector connected to a multichannel analyzer system (GC-4018, Canberra, Japan, Radioisotope Research Center of Fukushima University). The counting time was set to 3,845–5,004 s to decrease the statistical counting error to less than 0.3%. To elucidate the mobility of ¹³⁷Cs with relatively high levels of contamination, the exchangeable ¹³⁷Cs content in soils with a total ¹³⁷Cs content of higher than 2,000 Bq kg⁻¹ was determined for 32 out of the 173 soil samples (Table 1). Although sample extraction in this way may cause some statistical bias in the following results, soils with a total ¹³⁷Cs content of lower than 2,000 Bq kg⁻¹ was removed in this experiment to ensure data reliability. Duplicates of 50 g of soil sample were shaken with 250 mL of 1 mol L⁻¹ ammonium acetate (CH₃COONH₄) for 24 h. The suspension was centrifuged, and the supernatant was collected by filtration. The duplicate supernatant solutions were combined into a 500 mL beaker after filtration through a 0.45 µm Millipore filter and heated gently until the volume of the supernatant decreased to approximately 20 mL. To determine the exchangeable radiocesium (¹³⁷Cs + ¹³⁴Cs) content by counting the gamma activity between 589 and 735 keV, approximately 20 mL of the concentrated supernatant in a polyethylene vial (20 mL) was analyzed using the auto gamma counter. The counting time was set to 1,800 s to decrease the statistical counting error to less than 2%. The activity ratio of ¹³⁷Cs/¹³⁴Cs emitted from FDNPP in March 2011 was reported to be approximately 1:1 (Komori et al. 2013). Considering the difference in their half-lives, this ratio was expected to be approximately 8:1 when we measured the exchangeable ¹³⁷Cs content in November 2017. The ¹³⁷Cs desorption ratio (%), i.e. the percentage of exchangeable ¹³⁷Cs content in the total ¹³⁷Cs content, was calculated by the following equation: [¹³⁷Cs desorption ratio (%)] = [exchangeable ¹³⁷Cs content in soils (Bq kg⁻¹)]/[total ¹³⁷Cs content in soils (Bq kg⁻¹)] × 100. Although Nex-¹³⁷Cs is as phytoextractable as Nex-K, we considered that it is much less phytoavailable due to high competition with K⁺ at the ionic channel on plant root.

Table 1. Selected physicochemical properties of the soils

		Total-^137Cs	Ex-^137Cs	^137Cs desorption ratio	AlO+1/2FeO	Total C	Sand	Silt	Clay
Sample No.	Soil group^a	Bq kg^−1	Bq kg^−1	%	g kg^−1	g kg^−1	%	%	%
1	Andosols	3200	350	11	5.87	15.8	-	-	-
14	Andosols	2300	60	2.5	8.51	23.5	53.0	28.0	19.0
15	Andosols	2500	210	8.4	7.13	13.2	56.8	27.2	16.0
23	Andosols	4400	120	2.7	7.40	17.0	53.8	30.7	15.5
29	Muck soils	3500	110	3.3	3.96	14.8	57.5	30.0	12.5
32	Muck soils	3700	240	6.5	5.76	26.8	40.7	36.3	23.0
33	Muck soils	2800	170	6.0	3.99	12.9	62.7	23.6	13.7
57	Lowland soils	3900	200	5.1	2.35	13.3	69.0	21.8	9.19
72	Andosols	2600	340	13	4.23	10.6	71.5	18.1	10.3
78	Andosols	2800	510	18	6.88	12.2	48.5	29.5	22.0
81	Lowland soils	2200	100	4.5	3.10	8.59	75.4	18.1	6.48
86	Lowland soils	2200	110	4.9	5.64	14.6	53.5	33.4	13.2
92	Lowland soils	3900	230	6.0	2.48	10.1	66.7	23.2	10.1
93	Lowland soils	2200	40	1.9	2.09	7.89	57.2	33.7	9.09
94	Lowland soils	2200	230	11	3.06	11.5	65.4	24.1	10.5
96	Lowland soils	3200	380	12	2.35	11.6	70.3	20.0	9.75
97	Andosols	3900	200	5.1	2.55	8.72	64.1	23.1	12.8
101	Andosols	2700	150	5.5	9.10	32.7	58.9	25.1	16.0
103	Andosols	2500	170	6.8	9.49	24.2	70.3	15.2	14.4
105	Andosols	3600	40	1.0	4.91	17.4	63.7	25.3	11.0
110	Lowland soils	2400	140	5.9	1.98	5.46	83.7	10.6	5.76
120	Lowland soils	4000	350	8.8	2.66	9.59	69.4	21.9	8.68
122	Lowland soils	3200	120	3.9	1.62	13.4	76.3	15.1	8.65
127	Lowland soils	4200	180	4.4	4.57	15.1	49.4	34.5	16.1
152	Not classified	2400	520	22	14.4	30.5	20.1	37.3	42.6
156	Lowland soils	2100	410	20	2.86	5.97	77.1	14.4	8.51
161	Lowland soils	2100	90	4.3	3.22	13.8	69.2	20.9	9.83
165	Muck soils	2800	270	10	2.74	14.0	65.0	24.5	10.5
169	Lowland soils	2100	270	13	1.57	7.38	70.5	21.7	7.80
172	Lowland soils	3200	120	3.7	5.19	13.1	49.9	34.7	15.4
173	Andosols	2500	410	17	5.38	13.1	62.5	34.2	3.33
174	Andosols	3900	500	13	7.77	26.2	-	-	-

^aClassification of cultivated soils in Japan (3rd approximation) (Cultivated soil classification committee 1995).

2.3. Determination of phytoavailable K content

The Ex-K content in soils was determined using the 1 mol L⁻¹ CH₃COONH₄ extraction method (Thomas 1982). More specifically, 5.0 g of the soil sample was shaken with 25 mL of 1 mol L⁻¹ CH₃COONH₄ for 30 min. The suspension was centrifuged, and the supernatant was collected by filtration. These processes were repeated three times. The approximately 75 mL of supernatant was brought up to a total volume of 100 mL with 1 mol L⁻¹ CH₃COONH₄, and the K content of this solution was determined by atomic absorption spectroscopy (AAS) (AA-6200, Shimadzu, Kyoto, Japan).

The Nex-K content in soils was determined using the 1 mol L⁻¹ hot nitric acid (HNO₃) extraction method (Helmke and Sparks 1996). More specifically, 2.5 g of the soil sample was heated gently with 25 mL of 1 mol L⁻¹ HNO₃ for 15 min after boiling began. After cooling for 5 min, the sample was filtered, and the extract volume was brought up to 100 mL with 0.1 mol L⁻¹ HNO₃. The K content of the extract solution was determined by AAS. As the K in the extract includes both Ex-K and Nex-K, the amount of Nex-K was calculated by subtracting the amount of Ex-K from the hot HNO₃ extractable K.

2.4. Determination of selected physicochemical properties

Selected physicochemical properties of soils with a total ¹³⁷Cs content of higher than 2,000 Bq kg⁻¹ were determined for 32 out of 173 soil samples (Table 1) to elucidate the mobility of ¹³⁷Cs with relatively high levels of contamination. Oxalate acid extractable aluminum (Al_o) and iron (Fe_o) (McKeague and Day 1966) were determined using inductively coupled plasma optical emission spectroscopy (ICP-OES) (iCAP 4400 Duo Full MFC, Thermo Fisher Scientific, Waltham, USA) after extraction with 0.2 mol L⁻¹ acid ammonium oxalate solution (pH 3) for 4 h in darkness followed by filtration through a 0.45 µm Millipore filter. The total carbon (C) content in soils, as an indicator of total organic matter content, was determined by the dry combustion method with an NC analyzer (NC-95A, Smika Chem. Anal. Service, Tokyo, Japan) and gas chromatography (GC-8A, Shimadzu, Kyoto, Japan) after fine grinding. The sand (2 mm to 20 µm), silt (20–2 µm), and clay (<2 µm) contents were determined by sieving and pipetting methods after organic matter removal by oxidation with hydrogen peroxide and ultrasonic dispersion.

3. Results and discussion

3.1. Changes in total ¹³⁷Cs content and air dose rate after decontamination

According to Tomioka town (Tomioka town 2016), the total radiocesium (¹³⁷Cs + ¹³⁴Cs) content for local agricultural soils before soil decontamination was reported to be, on average, 9,100 Bq kg⁻¹ (collected in May 2014). Rate of ¹³⁷Cs content in this value can be calculated to be 6,700 Bq kg⁻¹ since activity ratio of ¹³⁷Cs/¹³⁴Cs was reported to be approximately 1:1 just after the FDNPP accident (Komori et al. 2013). Furthermore, the activity concentration of ¹³⁷Cs during our sampling period was calculated to be 6,000 Bq kg⁻¹, based on the effective half-life of ¹³⁷Cs in plowed layer (16.9 y) (Komamura et al. 1999). In addition, the hourly air dose rate in the agricultural fields in Tomioka town before decontamination was, on average, 2.00 µSv h⁻¹ (Tomioka town 2016).

After decontamination, our results showed that the total ¹³⁷Cs content in the soils ranged from 20 to 4,400 Bq kg⁻¹, with median and arithmetic mean (±standard deviation) values of 940 and 1,200 ± 1,000 Bq kg⁻¹, respectively. Total ¹³⁷Cs content in the decontaminated soils may vary depending on both the difference in the amount of removed and dressed soils among fields and the original contamination. We considered the former is more important because of the significant positive correlation between the Ex-K content and the total ¹³⁷Cs content in soils (Fig. 5). The hourly air dose rate at a height of 1 m ranged from 0.10 to 0.94 µSv h⁻¹, with median and arithmetic mean (±standard deviation) value of 0.34 and 0.36 ± 0.16 µSv h⁻¹, respectively. Both an average value of the total ¹³⁷Cs content in soils (Fig. 2) and the hourly air dose rate (Fig. 3) were about 80% smaller than those surveyed before the decontamination (Tomioka town 2016), whereas very similar to those surveyed after decontamination (Tomioka town 2016). The reductions were much larger than those caused only by the natural decay of radiocesium (total ¹³⁷Cs content: 10% decrease). These results strongly support the effectiveness of decontamination for the decrease of the total ¹³⁷Cs content in soils and the hourly air dose rate in Fukushima.

Figure 2. Boxplots of the distribution of the total ¹³⁷Cs content in soils before and after decontamination. The quartiles and median value are shown as the boxes and the maximum and minimum values are shown as the whiskers. The total ¹³⁷Cs content (collected in May 2014 and February 2016) before and after decontamination are from Tomioka town (2016) and corrected for decay from the sampling date

Figure 3. Comparison of the hourly air dose rate at a height of 1 m before and after decontamination (bars indicate standard deviation). The dose rate (measured from 30 July 2013 to 9 May 2016) before and after decontamination are from Tomioka town (2016)

The total ¹³⁷Cs content in soils showed a significant positive correlation with the hourly air dose rate (p < 0.01) (Fig. 4). Fields with higher air dose rates than expected from the regression line were those located close to the border of forests or difficult-to-return zone where the decontamination (including removal of the surface soil layer) had been less thorough. Thus, the gamma radiation from these areas may have increased the air dose rate in the fields. On the contrary, a lower air dose rate than expected from the regression line was detected in a few of the fields, which is probably because of insufficient plowing after the fields had been covered with dressed soils. Except for these specific situations, the air dose rate in the agricultural fields was found to be mainly controlled by the total ¹³⁷Cs content in the surface soil layer, in addition to a certain background level. Kohyama et al. (2013) also showed a similar positive relationship between the total ¹³⁷Cs content in soils and the hourly air dose rate in fields plowed after the FDNPP accident even across different conditions of sampling date and decontamination.

Figure 4. Relationship between the total ¹³⁷Cs content in soils and the hourly air dose rate at a height of 1 m (n = 170) (** indicates P ≤ 0.01). The darker plots indicate fields located close to the border of forests or difficult-to-return zone

Assuming that one spends 8 h outside and 16 h at home (Ministry of the Environment 2011), the average hourly air dose rate in Tomioka town, i.e. 0.36 µSv h⁻¹, corresponds to an annual additional exposure dose of 1.9 mSv y⁻¹. In Japan, 10 mSv y⁻¹ is the criterion for lifting the evacuation order, and 1 mSv y⁻¹ is the criterion for designating zones for which a decontamination plan is to be formulated. To achieve the latter criterion, therefore, agricultural fields with a high air dose rate that were often located close to the border of forests or difficult-to-return zones should be carefully managed.

3.2. Evaluation of phytoavailable K

The Ex-K content in soils ranged from 27.9 to 743 mg kg⁻¹, with median and arithmetic mean (±standard deviation) values of 159 and 172 ± 74.2 mg kg⁻¹, respectively. More than 80% of the soil samples showed an Ex-K content of less than 200 mg kg⁻¹, i.e. less than the recommended threshold value needed to reduce ¹³⁷Cs transfer from soil to plant (Kato et al. 2015). The Ex-K content was positively correlated with the total ¹³⁷Cs content in soils (p < 0.01) (Fig. 5). This correlation was not expected because their origins and adsorption sites were supposed to be largely different. Ex-K in agricultural soils mainly originated from K input as fertilizer before the FDNPP accident, which would probably be weakly adsorbed on the surface of minerals and organic matter in an exchangeable form (Helmke and Sparks 1996). ¹³⁷Cs in agricultural soils in Fukushima was mainly derived from the FDNPP accident and was strongly bound partly in the interlayer of micaceous minerals. This correlation suggests that the mixing of dressed soils in the decontamination processes reduced not only the total ¹³⁷Cs content in soils but also the Ex-K content, possibly because of the lower contents of ¹³⁷Cs and Ex-K in the dressed soils than in the surface soil layer. Therefore, fertilization with K after decontamination to improve the level of Ex-K is strongly recommended.

Figure 5. Relationship between the exchangeable K content in soils and the total ¹³⁷Cs content in soils (n = 172) (** indicates P ≤ 0.01)

The Nex-K content in soils ranged from 98.8 to 1,280 mg kg⁻¹, with median and arithmetic mean (±standard deviation) values of 391 and 460 ± 266 mg kg⁻¹, respectively. The Nex-K content showed no significant correlation with the Ex-K content (Fig. 6). Because the Nex-K content showed a wide range of values, the potential to reserve Ex-K varied widely among the soils. This means that the determination of not only Ex-K but also Nex-K is important in efforts to design appropriate countermeasures to reduce the ¹³⁷Cs transfer. Further studies will be required to determine the recommended threshold value of Nex-K content in soils needed in order to reduce the transfer of ¹³⁷Cs from soil to plant. Kitagawa et al. (2018) showed that Andosols and Wet Andosols in agricultural fields in Japan tended to have lower Nex-K contents (110 and 132 mg kg⁻¹, respectively) than the other soil types. However, the Nex-K content of Andosols in this study was, on average, 536 mg kg⁻¹, which was much higher than that of the previous study. Dressed soils may be rich in K-bearing minerals such as micaceous minerals, which would account for the increased Nex-K content. Mixing with dressed soil may have changed other properties of the surface soils as well. As shown in Table 1, none of the soils that have been classified as Andosols satisfy the diagnostic criterion of andic properties (i.e. Al_o+1/2Fe_o contents > 20 g kg⁻¹) (Soil Survey Staff 2014). This discrepancy is also caused by the dilution effect of dressed soils on surface soils.

Figure 6. Relationship between the exchangeable K content in soils and the nonexchangeable K content in soils (n = 173)

3.3. Factors influencing the ¹³⁷Cs desorption ratio

The exchangeable ¹³⁷Cs content in soils ranged from 40 to 520 Bq kg⁻¹, with median and arithmetic mean (±standard deviation) values of 200 and 230 ± 140 Bq kg⁻¹, respectively. The ¹³⁷Cs desorption ratio ranged from 1.0% to 22%, with median and arithmetic mean (±standard deviation) values of 6.0% and 8.1% ± 5.3%, respectively. The low ¹³⁷Cs desorption ratio indicated that most of the ¹³⁷Cs was strongly adsorbed in the soils and was not readily available for uptake by plants. These values are similar to but lower than those reported for soils investigated soon after the accident in Fukushima (Saito et al. 2014). This difference in the desorption ratio corresponded reasonably well to the research that showed the desorption ratio of the added ¹³⁷Cs to decrease with time (Takeda et al. 2013).

A Spearman correlation analysis indicated that the ¹³⁷Cs desorption ratio was positively correlated with the exchangeable ¹³⁷Cs content (p < 0.01) and negatively correlated with the Nex-K content (p < 0.05), whereas it was not significantly correlated with the total C content or the Al_o + 1/2Fe_o content (Table 2). The positive correlation with the exchangeable ¹³⁷Cs content indicates that the variability in the ¹³⁷Cs desorption ratio was mainly controlled by the variability in the exchangeable ¹³⁷Cs content rather than the total ¹³⁷Cs content. Therefore, the exchangeable ¹³⁷Cs content, as the phytoavailable fraction, may be more useful in the estimation of ¹³⁷Cs transfer from soil to plant. The negative correlation with the Nex-K content (p < 0.05) may be an indirect relationship; both a lower desorption ratio of ¹³⁷Cs and a higher amount of K release from the nonexchangeable fraction may be directly related to the abundance of weathered micaceous minerals, because the interlayer of these minerals can be both a fixation site for ¹³⁷Cs (Saito et al. 2014; Nakao et al. 2014) and a reservoir of Nex-K (Fanning et al. 1989). Therefore, nonexchangeable K content can be a rough index to estimate the immobility or non-exchangeability of ¹³⁷Cs. However, the lack of any significant correlations of the ¹³⁷Cs desorption ratio with the total C content or the Al_o + 1/2Fe_o content was unexpected. Yamaguchi et al. (2017) and Nakao et al. (2015a) reported that the radiocesium interception potential (RIP), as an indicator of the capacity and selectivity for the sorption of radiocesium by soil (Cremers et al. 1988), showed a significant negative correlation with the total C content and andic soil properties (e.g. phosphate absorption capacity). These correlations would occur because the organic matter coating on micaceous minerals may inhibit ¹³⁷Cs sorption (Tashiro et al. 2018), and Andosols tend to have lesser amounts of micaceous minerals in comparison to the large content of non-crystalline minerals (Nakao et al. 2015b). The factors controlling ¹³⁷Cs desorption may be different from those controlling its selective adsorption (Vandebroek et al. 2012). Another possible reason is that the inclusion of dressed soils on surface soils largely modified soil characteristics. Further investigation into the relationship between RIP and ¹³⁷Cs desorption ratio will provide some important information for designing appropriate countermeasures to reduce the ¹³⁷Cs transfer from soil to plant.

Table 2. Correlation coefficients between the ¹³⁷Cs desorption ratio and selected physicochemical properties of the soils

	Total-^137Cs	Ex-^137Cs	Ex-K	Nex-K	AlO+1/2FeO	Total-C	Sand	Silt	Clay
^137Cs desorption ratio	−0.20	0.94	0.19	−0.42	0.10	−0.16	−0.09	−0.18	−0.09
Statistical significance	ns	**	ns	*	ns	ns	ns	ns	ns

**, *, and ns indicate statistical significance at the level of p < 0.01, p < 0.05, and “not significant’, respectively.

4. Conclusions

We investigated exchangeability of ¹³⁷Cs and K, which will affect ¹³⁷Cs transfer from soil to plant, i.e. total ¹³⁷Cs, exchangeable ¹³⁷Cs, exchangeable K, and nonexchangeable K content, for the soils collected from 173 decontaminated fields in Tomioka town in the eastern coastal area of Fukushima. We found that the decontamination largely reduced the exchangeable K content in soils as well as the total ¹³⁷Cs content in soils with low ¹³⁷Cs desorption ratios. In fact, more than 80% of the soil samples showed an exchangeable K content less than the recommended threshold value needed to reduce ¹³⁷Cs transfer from soil to plant (200 mg kg⁻¹). We also found that nonexchangeable K content can be a rough index to estimate the immobility of ¹³⁷Cs. In conclusion, K fertilization of decontaminated surface soils is strongly recommended, taking account of both exchangeable K and nonexchangeable K contents, because the decontamination caused a considerable decrease in the exchangeable K content.

Acknowledgments

The experiment on measurement of ¹³⁷Cs content was carried out in the Laboratory of Radioisotopes of Kyoto Prefectural University and Fukushima University. This work was financially supported by the JSPS KAKENHI (grant number 16H06188) and research grant from Tomioka town. We would like to thank Mr. Sho Ogasawara, Ms. Shiori Uno, Ms. Yuka Kitagawa, Mr. Ryo Ito, Ms. Sumika Kataoka, Ms. Mai Terashima, Ms. Mayu Tomita, Ms. Yuki Matsuyama, Ms. Honoka Ota, Mr. Rikuo Kitayama, Mr. Yuki Tanaka, Ms. Mina Hirose, Ms. Natsumi Noguchi, Ms. Sae Ito, Ms. Yurie Nishii, Ms. Fukiko Masai, and Ms. Nanami Murashita of Laboratory of Soil Chemistry, Kyoto Prefectural University for collecting soil samples and sample pretreatments.

Additional information

Funding

This work was supported by the research grant from Tomioka town;JSPS KAKENHI [16H06188].