- Email: [email protected]
Relevance of Radiocaesium Interception Potential (RIP) on a worldwide scale to assess soil vulnerability to 137Cs contamination
Journal of Environmental Radioactivity 104 (2012) 87e93
Contents lists available at SciVerse ScienceDirect
Journal of Environmental Radioactivity journal homepage: www.elsevier.com/locate/jenvrad
Relevance of Radiocaesium Interception Potential (RIP) on a worldwide scale to assess soil vulnerability to 137Cs contamination Louis Vandebroeka, b, May Van Heesa, *, Bruno Delvauxb, Otto Spaargarenc, Yves Thirya,1 a
SCKCEN, Belgian Nuclear Research Center, Foundation of Public Utility, Biosphere Impact Studies, Boeretang 200, 2400 Mol, Belgium Université Catholique de Louvain, Unité des Sciences du Sol, Place Croix du Sud 2/10, 1348 Louvain-la-Neuve, Belgium c International Soil Ressource and Information Center, Duivendaal 9, 6701 AR Wageningen, The Netherlands b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 12 November 2009 Received in revised form 19 August 2011 Accepted 2 September 2011 Available online 2 October 2011
The extent of radiocaesium retention in soil is important to quantify the risk of further foodchain contamination. The Radiocaesium Interception Potential (RIP e Cremers et al., 1988, Nature 335, 247e249) is an intrinsic soil parameter which can be used to categorize soils or minerals in terms of their capacity to selectively adsorb radiocaesium. In this study, we measured RIP for a large soil collection (88 soil samples) representative of major FAO soil reference groups on a worldwide scale and tested the possibility to predict the RIP on the basis of other easily accessible or measurable soil data. We also compared RIP values with those obtained from separate chemical extraction experiments. The range of measured RIP values (1.8e13300 mmol kg1) was shown to include nearly all possible cases of agricultural soil contamination. Only Podzols, Andosols and Ferralsols were clearly characterized by a very low RIP (<2000 mmol kg1). On a worldwide scale, RIP was in fact slightly related to soil reference type or other simple major physicochemical parameters such as clay percentage or organic matter. Conversely our results indicated a link between the RIP and radiocaesium extractability across very different soils. We showed that, with the proposed scale of RIP values, a simple acid extraction method can provide an operational result highly predictive of potential RIP despite very contrasting soil properties. The RIP could be estimated from the empirical equation: RIP ¼ (31.701 * log(AER) þ 58.886)2 where AER is the fraction of acid-extractable radiocaesium. Ó 2011 Elsevier Ltd. All rights reserved.
Keywords: 137 Cs Radiocaesium Interception Potential Soil taxonomy Acid extraction Soil collection Micaceous clay
1. Introduction Following an accidental atmospheric release, radiocaesium is of great environmental and public health concern, due to its possible long transport, its long half-live, its high solubility in water and its biogeochemical behavior similar to that of K, a major nutrient for plants and animals. After deposition, radiocaesium is retained in the topsoil because of its selective adsorption on soil particles, largely ruled by weathered micaceous clays (Maes et al., 1998). Vertical migration rates are usually very low in agricultural mineral soils (Almgren and Isaksson, 2006). Even in soils of semi-natural areas, with low clay and high organic matter content, the net 137 Cs export from the rooting zone was determined to be less than 0.007% per year (Tikhomirov et al., 1993). In organic-rich forest soils
* Corresponding author. Tel.: þ32 14 33 21 05; fax: þ32 14 31 50 21. E-mail address: [email protected] (M. Van Hees). 1 Present address: ANDRA, Agence nationale pour la gestion des déchets radioactifs, Direction scientifique, Service transferts, 1-7 rue Jean Monnet, 92298 Châtenay-Malabry Cedex, France. 0265-931X/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jenvrad.2011.09.002
in particular, 137Cs persistence in the surface layers can be a source of long-lasting recycling by the vegetation (Kruyts et al., 2000; Goor and Thiry, 2004; Thiry et al., 2000). In general, the extent of 137Cs retention in agricultural and forest soils plays a key role in understanding the risk of foodchain contamination. Various experimental approaches were used to quantify such retention, yielding a variety of conceptual coefficients. One of them, the Radiocaesium Interception Potential (RIP e Cremers et al., 1988) has the particularity to be an intrinsic soil parameter but it must be determined under standardized experimental conditions. The measurement procedure involves soil chemical equilibration and its labeling with a sufficient amount of radiocaesium and thus necessitates adapted facilities to manipulate artificial radioactivity, which can discourage the use of RIP as a standard indicator at international level. Only a few European laboratories have produced most RIP values. Moreover, while the RIP method was frequently used to characterize the 137Cs adsorption capacity of various temperate soils containing micas (Waegeneers et al., 1999; Delvaux et al., 2000; Gil-García et al., 2008), tropical and sub-tropical soils were largely overlooked. However, Joussein et al. (2004) have shown that
88
L. Vandebroek et al. / Journal of Environmental Radioactivity 104 (2012) 87e93
Table 1 Selected characteristics of the 88 soils including mean RIP values (n ¼ 3) and mean fraction of total radiocaesium that is extractable in 0.1 M HCl and 1 M NH4OAc (n ¼ 3). Ref. Group
Land
Texture fractions Sand
Silt
Ca %
pH Clay
Water
KCl
Ca cmolc kg
% Andosol Andosol Andosol Andosol Andosol Andosol Andosol Andosol Andosol Andosol Arenosol Calcisol Calcisol Calcisol Calcisol Calcisol Calcisol Cambisol Cambisol Cambisol Cambisol Cambisol Cambisol Cambisol Cambisol Cambisol Cambisol Chernozem Chernozem Chernozem Chernozem Chernozem Chernozem Ferralsol Ferralsol Ferralsol Ferralsol Ferralsol Ferralsol Ferralsol Ferralsol Ferralsol Fluvisol Fluvisol Fluvisol Fluvisol Fluvisol Gleysol Gleysol Gleysol Gleysol Greyzem Kastanozem Kastanozem Luvisol Luvisol Luvisol Luvisol Luvisol Luvisol Luvisol Nitisol Nitisol Nitisol Nitisol Nitisol Nitisol Nitisol Phaeozem Podzol Podzol
C. Rica Ecuador Ecuador Ecuador Ecuador Ecuador Indonesia Indonesia Japan Kenya China Spain Turkey Turkey Turkey Turkey Turkey China China China China China China China China Italy Italy China Germany Peru Peru Romania USA Brazil Brazil China China Cuba Cuba Indonesia Indonesia Indonesia Japan Malaysia Nicaragua Peru Poland Cuba Cuba Thailand Thailand China USA USA China China Italy Kenya Nicaragua Spain Turkey Brazil Cameroon Columbia Italy Kenya Kenya Kenya China Belgium Malaysia
13.2 61.4 58.6 50.6 57.2 31.0 29.5 41.2 31.0 19.2 95.9 25.7 34.4 27.6 6.7 10.4 ND 21.0 12.3 15.7 32.5 45.7 36.1 31.8 0.8 12.2 ND 18.1 8.3 38.9 15.6 15.0 ND 16.0 36.0 10.5 74.1 23.9 10.3 52.2 28.6 6.0 56.1 98.2 37.6 52.1 41.7 6.3 5.0 28.5 1.1 7.8 34.6 ND 5.6 19.3 7.9 71.1 20.4 62.8 1.1 7.2 18.1 42.4 7.0 7.0 6.2 5.8 8.0 68.0 97.2
Exchangeable cations
10.4 29.4 34.6 34.7 38.9 52.6 46.4 33.6 38.8 54.8 1.4 21.3 26.1 27.5 29.7 33.3 ND 53.0 43.1 54.9 23.4 17.9 30.5 16.8 24.6 30.7 ND 46.3 75.0 36.2 52.5 56.3 ND 27.0 22.0 12.4 6.7 10.6 9.5 13.0 9.0 8.9 34.9 0.7 38.6 38.0 38.6 29.1 20.4 52.4 25.6 66.6 33.1 ND 61.9 59.2 25.1 4.0 36.0 26.1 38.2 37.5 28.3 15.5 21.8 36.2 23.6 19.3 54.2 16.4 2.6
76.5 9.3 6.8 14.7 4.0 16.4 24.2 25.2 30.2 26.1 2.6 53.0 39.5 44.9 63.6 56.3 65.7 26.0 44.5 29.5 44.2 36.2 33.5 51.3 74.9 57.1 75.5 35.5 16.7 24.9 31.9 28.7 24.6 57.0 42.0 77.0 19.2 65.6 80.1 34.6 62.3 84.0 9.0 1.1 23.8 10.0 19.7 64.7 74.4 19.1 73.3 25.6 32.3 23.1 32.4 21.6 67.2 24.8 43.6 11.1 60.7 55.4 53.6 42.1 71.3 56.8 70.0 74.8 37.9 15.6 0.2
4.9 5.7 6.7 6.4 6.6 5.2 5.6 5.0 5.5 6.7 8.9 7.9 8.2 8.1 8.1 8.0 7.5 8.3 8.2 7.4 4.8 4.3 3.9 6.2 4.7 6.8 6.3 8.0 8.1 8.1 8.3 7.7 5.9 4.5 4.0 4.3 4.8 6.7 5.2 4.2 4.7 3.5 6.6 8.3 7.3 6.0 5.1 7.7 6.6 5.8 5.0 6.7 7.0 8.1 5.6 6.5 6.2 6.7 6.8 5.7 8.0 6.2 6.4 4.8 5.6 5.4 5.8 6.4 6.5 6.2 4.9
4.3 4.8 5.4 4.3 5.7 4.6 5.2 4.5 4.6 6.1 8.2 7.1 7.3 7.3 7.0 7.3 6.6 7.1 7.1 6.5 4.1 3.8 3.6 4.8 4.0 5.9 5.6 7.3 7.1 7.8 7.4 7.0 4.8 4.2 4.0 4.1 4.1 5.6 4.2 3.7 4.2 3.5 5.7 8.1 6.7 5.5 4.6 6.8 5.1 5.3 4.1 6.3 6.1 ND 4.9 5.8 5.4 5.8 5.4 4.3 6.6 5.2 5.8 4.0 4.9 4.5 4.9 5.4 5.5 5.7 3.6
6.88 1.11 2.48 0.43 1.00 3.40 7.72 4.15 5.80 9.99 0.03 0.88 0.67 0.77 0.67 1.06 2.97 0.32 1.04 0.68 0.53 2.91 1.57 1.41 1.74 1.70 4.15 2.01 1.32 3.95 2.11 2.51 2.11 3.33 2.58 1.48 1.50 1.56 1.49 2.67 3.52 6.87 1.11 0.16 1.72 2.08 1.26 1.02 1.43 0.84 1.46 1.77 1.33 0.88 11.31 1.29 2.11 0.93 2.35 0.93 0.73 2.84 3.00 2.57 3.69 4.87 2.33 6.54 1.67 2.49 1.34
0.2 3.8 5.8 4.8 4.1 4.9 20.4 3.1 8.5 33.0 20.0 16.7 29.5 19.7 35.4 24.3 ND 49.8 49.0 20.3 0.8 0.8 0.0 10.2 2.3 20.9 ND 43.0 12.2 41.0 47.2 20.0 ND 0.6 0.2 0.2 0.6 8.0 5.8 1.8 3.0 0.2 7.1 7.3 37.9 14.9 12.7 28.9 29.9 7.1 15.1 20.8 15.0 ND 29.7 16.0 14.9 2.4 22.8 3.3 56.9 10.4 9.3 0.0 14.0 ND 7.9 12.8 23.7 ND 0.8
Mg
CECb
RIPc mmol kg1
K
Cs extractions
HCl
1
0.4 1.3 1.0 2.8 1.9 0.6 4.2 0.2 1.8 4.1 1.0 1.3 3.7 3.1 3.1 2.2 ND 1.0 2.4 2.4 0.3 0.3 0.0 2.7 0.3 1.3 ND 2.8 0.7 21.3 13.2 2.9 ND 0.4 0.1 0.3 0.3 2.7 1.5 0.5 0.7 0.2 2.7 0.8 2.4 0.0 3.1 12.4 12.2 1.3 3.9 2.4 4.5 ND 5.2 2.7 3.2 0.9 5.2 1.1 8.9 2.2 4.0 0.1 4.5 ND 1.9 8.1 6.6 ND 0.3
137
NH4OAc
% of total activity 0.2 0.1 0.7 1.1 0.4 0.3 2.7 0.2 0.8 1.5 0.2 0.9 2.6 0.9 0.7 1.8 ND 0.3 0.5 0.6 0.1 0.3 0.1 0.7 0.3 0.4 ND 0.2 0.3 0.9 1.7 0.3 ND 0.3 0.1 0.0 0.2 0.5 0.3 0.2 0.2 0.2 0.9 0.0 1.1 0.4 0.3 0.2 0.2 0.0 0.6 0.1 1.7 ND 1.2 0.2 1.2 0.4 0.9 0.2 1.9 1.1 0.4 0.2 1.8 ND 2.1 2.9 0.6 ND 0.0
23.3 5.9 10.0 10.2 6.7 12.4 36.7 20.7 48.6 70.7 2.1 20.0 25.3 27.5 29.9 17.2 50.5 22.2 23.0 19.2 5.1 10.5 8.0 13.2 10.0 25.6 30.5 30.2 16.4 24.9 39.1 28.9 20.0 6.7 7.3 5.9 3.5 12.0 13.2 9.5 11.8 25.2 14.9 2.2 33.8 12.1 16.2 38.3 46.3 11.5 30.2 20.4 20.2 14.0 51.0 16.1 25.7 6.7 29.4 11.1 83.8 17.7 22.7 14.8 25.0 ND 25.7 30.9 25.6 10.0 5.2
177 938 1020 484 94 97 450 208 1630 1320 1680 4340 12100 4220 9630 9090 9220 10500 12900 5310 945 1290 1370 4410 1680 6770 8860 5770 5820 5100 8590 6790 4510 26 203 217 245 1410 835 509 87 180 2720 150 9620 1730 6530 2110 504 4460 10100 4760 8390 5720 8050 5020 6750 1470 13 300 2270 7970 1550 539 324 4920 4370 4430 3510 6080 944 1.8
22.8 6.19 19.8 3.95 55.7 51.4 16.2 31.4 2.02 10.1 1.21 0.12 0.07 0.16 0.05 0.21 0.43 0.05 0.04 0.15 9.87 6.21 1.49 0.57 2.88 0.46 0.47 0.15 0.36 0.26 0.12 0.11 0.41 48.9 23.3 22.2 29.3 3.28 3.03 6.07 16.0 4.12 2.23 13.6 0.10 1.52 0.60 2.05 5.05 0.89 0.28 0.22 0.39 0.14 1.52 0.20 0.54 2.80 0.20 0.74 0.15 14.8 6.15 23.7 0.98 3.26 1.09 2.21 0.12 3.07 87.6
50.3 44.4 41.8 24.4 74.2 52.6 68.7 80.1 55.1 70.5 34.0 14.9 15.9 13.3 11.0 36.3 21.1 5.32 4.59 5.28 61.2 37.7 15.7 25.3 39.1 18.7 19.5 3.66 10.3 5.87 7.21 4.47 15.1 62.0 60.5 63.2 54.1 66.8 55.6 16.1 48.7 21.4 29.6 38.5 3.96 6.48 7.82 23.5 30.0 13.1 11.7 4.96 18.0 13.3 17.2 3.24 23.5 53.8 11.7 12.5 8.80 56.0 32.5 60.9 30.0 57.4 46.2 43.4 3.83 32.6 94.7
L. Vandebroek et al. / Journal of Environmental Radioactivity 104 (2012) 87e93
89
Table 1 (continued ) Ref. Group
Land
Texture fractions Sand
Silt
Ca %
pH Clay
Water
KCl
Ca
Malaysia Peru Peru USA Zambia Poland China Columbia Turkey Zambia Italy Italy Spain Spain Uruguay Uruguay Zambia
Arithmetic mean Geometric mean Minimum Maximum a b c
Mg
CECb
RIPc mmol kg1
K
137
Cs extractions
HCl
cmolc kg1
% Podzol Podzol Podzol Podzol Podzol Podzoluvisol Regosol Regosol Regosol Regosol Vertisol Vertisol Vertisol Vertisol Vertisol Vertisol Vertisol
Exchangeable cations
NH4OAc
% of total activity
96.7 91.9 84.1 92.0 98.0 5.4 13.6 75.1 10.3 82.6 6.1 5.5 20.9 15.0 4.3 6.6 29.5
2.5 8.0 11.0 5.0 2.0 85.4 70.7 13.8 51.1 8.0 33.7 25.4 26.5 28.9 49.9 58.5 21.5
0.8 0.0 5.1 3.0 0.0 9.1 15.7 11.1 38.6 9.3 60.3 69.2 52.7 56.1 45.8 34.9 48.9
4.6 4.5 4.0 5.0 4.4 7.0 8.6 4.1 7.7 5.5 5.4 6.3 7.8 7.6 7.3 5.3 7.3
4.2 3.8 3.5 3.6 3.5 6.3 8.3 3.7 7.0 4.3 4.5 5.1 6.3 6.3 6.6 4.5 5.8
1.28 0.63 6.05 0.27 1.90 1.22 0.73 1.60 3.07 0.90 1.94 2.60 0.46 0.94 2.38 3.72 1.88
1.5 0.2 0.4 ND 1.9 9.7 40.7 0.4 52.3 1.1 10.5 15.1 34.8 45.0 39.2 12.2 17.0
0.3 0.0 0.0 ND 0.5 0.6 3.6 0.1 2.0 0.4 2.0 2.9 8.2 7.5 3.9 3.8 10.2
0.1 0.0 0.0 ND 0.1 0.1 2.7 0.0 2.8 0.2 0.6 1.4 1.0 0.9 1.2 0.3 0.4
5.0 0.9 29.1 1.1 9.0 10.0 7.6 5.1 ND 4.8 20.3 34.7 44.4 67.0 43.2 36.0 29.4
13 38 12 62 7.4 3470 6820 1140 9650 744 6310 8210 3050 2100 3400 1820 2120
48.1 45.7 56.8 19.9 81.1 1.28 0.24 2.98 0.12 31.1 0.75 0.40 0.32 0.68 0.29 0.36 0.49
40.7 29.3 29.0 10.9 91.6 15.0 31.4 24.6 8.36 48.8 25.1 20.8 9.59 9.13 7.63 2.25 6.26
32.3 22.5 0.8 98.2
30.7 29.0 0.7 85.4
37.4 34.8 0.0 84.0
6.3 6.3 3.5 8.9
5.4 5.4 3.5 8.3
2.30 1.69 0.03 11.31
15.7 12.2 0.0 56.9
2.9 2.0 0.0 21.3
0.7 0.4 0.0 2.9
21.5 19.6 0.9 83.8
3730 2200 1.8 13300
9.87 1.39 0.04 87.6
30.0 23.9 2.25 94.7
C is the organic carbon content. CEC is the cation exchange capacity. RIP is the Radiocaesium Interception Potentia.
African volcanic soils devoid of micas could also selectively retain radiocaesium and thus have a RIP. In this study, we analyzed the soil parameters controlling radiocaesium retention for a large variety of agricultural substrates representative of major FAO soil reference groups including (sub)-tropical soils and other soils devoid of micas. The general objective was twofold: (1) further document the relevance of RIP as an indicator of 137Cs retention parameter in soils on a worldwide scale and (2) test the possibility to predict this parameter starting from easily accessible or measurable soil criteria such as soil type or texture, or from a simple Cs extraction using a batch method. 2. Materials and methods 2.1. A worldwide soil collection A collection of 88 surface soil samples from the International Soil Reference and Information Center (ISRIC, Wageningen) was used to measure RIP (see Table 1). Only the surface horizon was taken into account because of the low vertical migration usually observed for radiocaesium in mineral soils (Almgren and Isaksson, 2006). These soils were chosen to be representative of major soil types in agriculture on a worldwide scale covering different climatic zones. The classification of each soil sample into reference soil groups was achieved according to the World Reference Base for Soil Resource (FAO/ISRIC/ISSS, 1998). All these soils were analyzed by ISRIC for major parameters: pHwater, pHKCl, organic C content, texture, cation exchange capacity (CEC) and exchangeable bases as reported in Batjes (1995) based on classical methods of soil analyses. In brief pHwater and pHKCl were measured in suspension using soil:water and soil:KCl 1 M ratios of 1:2.5. The organic C content was determined following the Walkley-Black method. The clay (0e2 mm) and silt (2e50 mm) fractions were collected by gravimetric sedimentation (pipette method) after separation of the sand (50e2000 mm) fraction from the fine earth (<2 mm) by dispersion and wet sieving. The reference method based on ammonium acetate (1 M NH4OAc, pH7) was used to measure the soil CEC and to remove the exchangeable bases (Ca, Mg, K).
2.2. RIP measurements RIP was measured in a well defined ionic scenario according to a procedure adapted from Wauters et al. (1996a). One g of each soil sample (3 replicates) was introduced in a dialysis bag (Dialysis Tubing-Visking code DTV12000.06.000, size 3-20/32; Medicell International Ltd, UK) (5 mL) and equilibrated with a 5 104 M KCl e 0.1 M CaCl2 equilibration solution (Potassium Adsorption Ratio (PAR) ¼ 0.05 mmol1/2 L1/2). The solution of the container with the dialysis bag was changed 10 times during 7 days. During equilibration, the containers were agitated for 2 h each 12 h. Each dialysis bag was then transferred in a new container filled with 95 mL of the KCl e CaCl2 equilibration solution labeled with carrier-free 137CsCl. These bags were again agitated for 2 h each 12 h. After 5 days, aliquots of 20 mL were taken and 137Cs activity of the equilibrium solution was measured by gamma-counting and a Kd determined. Each measurement was determined in triplicate. The RIP (mmol.kg1) is defined by the relation: þ RIP ¼ KC * [FES] ¼ KCs D * [K ], where [FES] is the capacity of Frayed Edge Sites with high adsorption selectivity for Csþ, KC is the 137CseK selectivity coefficient on these sites (considering an average KC of 137 Cs solideliquid distri1000 (Wauters et al., 1996a)), KCs D is the bution coefficient and [Kþ] is the concentration of K in solution (Cremers et al., 1988). 2.3. Chemical extractions The contamination of a larger sub-sample (about 10 g) of the 88 soils was also carried out separately by mixing the soil aliquot with an adequate amount of deionized water labeled with carrier-free 137 CsCl. The volume of solution was adapted to reach soil moisture content close to field capacity and allow further hand mixing. The time elapsed since soil contamination is known to increase radiocaesium fixation. Wetting-drying cycles have been shown to simulate the 137Cs ageing in soil (Noordijk et al., 1992; Gommers et al., 2005) resulting in increase of Cs sorption and fixation in soil and micaceous clay (Degryse et al., 2004; Vandenhove et al., 2005). In order to quantify 137Cs retention in the long term after
L. Vandebroek et al. / Journal of Environmental Radioactivity 104 (2012) 87e93
2.4. Statistics Pearson correlation coefficients were calculated between RIP and the other soil properties without any data transformation. Data transformations concern only the relationship between RIP values and results from the chemical extractions. The statistical analysis was done with the software Minitab 15. 3. Results and discussion 3.1. RIP variability and reference soil groups The soil collection involved a large set of soil groups with very diverse properties and was found to cover a very large range of adsorption selectivity for radiocaesium in terms of RIP (Table 1). In fact, the measured RIP values extended over four orders of magnitude. These RIP values ranged from 1.8 mmol kg1 in a Malaysian humic podzol to 13 300 mmol kg1 in a Nicaraguan ferric luvisol. In comparison with RIP measured in previous studies (Table 2), the measured values are to the best of our knowledge the smallest and the highest RIP values ever determined for the soils. Our maximum value of RIP (13 300 mmol kg1) was even higher than the RIP measured for a pure illite (Silver Hill, Montana e 12 600 mmol g1 e Wauters, 1994) which was found to be characterized by an especially strong cesium retention (Delvaux et al., 2001). Fig. 1 represents the confidence interval (a ¼ 0.05) and maxima and minima of RIP for each major reference soil group represented in our collection. Only 3 groups of soil were characterized by a very low RIP: Podzols, Andosols and Ferralsols with all values <2000 mmol kg1. For the other soils, our results showed uniformity in RIP ranges indicating that in most cases, reference soil
Table 2 RIP range in the literature (adapted and completed from Kruyts, 2002). Source
Number of samples
O.M. [%]
Clay [%]
RIP [mmol kg1]
Cremers et al., 1990 Sweeck et al., 1990 Valcke, 1993 Smolders et al., 1997 Waegeneers et al., 1999 Rigol et al., 1999 Sanchez et al., 1999 Yera et al., 1999 Delvaux et al., 2000 Sanchez et al., 2002 Gil-García et al., 2008 This study
8 12 33 30 88 4 23 5 47 53 30 88
23e88 0.5e8.6 7e97 2e28 2e28 69e97 12.6e96.5 1e4.5 0.5e96 1.9e96.5 0.2e9.4 0e19.5
3e23 0.5e33 e 0.5e36 0.5e40 e 2e57.6 4.9e16.4 00.7e66 0.5e93.8 6.3e52.4 0e84
65e2450 67e4890 7e9199 54e5861 50e11 200 7e520 5e6545 443e2732 13e4861 5e6545 179e7000 1.8e13 343
14000 12000 10000 8000 6000 4000 2000
er tis ol
so l
l
V
go
zo Re
so l
Po d
iti
so l
N
ys ol
vi Lu
so l
so l
vi
le G
Fl u
m ze er no
Ch
Fe ra l
l
l
so bi m
lc
iso
Ca
Ca
os ol
0
nd
soil contamination, each soil sample was submitted to 10 dryingewetting cycles. At each cycle deionized water was added to reach soil moisture content close to 150% of field capacity. After 2e3 h the soil was transferred to a ventilated oven (40 C) and let dry overnight. The total 137Cs content of the soil was measured at the end of the treatment. Afterwards the fraction of acid extractable 137 Cs was determined following 137Cs desorption with HCl 101 M. In parallel, the fraction of NH4-exchangeable 137Cs was also determined following 137Cs desorption with ammonium acetate 1 M buffered at pH 7. These 2 extractions were carried out by using a batch system during 24 h in a ratio 18 mL extraction solution per g of soil. The 137Cs content in different samples was measured by gamma spectrometry (NaI detectoreMinaxi g Packard 5000 series). The counting time was adjusted in order to maintain the counting error below 3%. Each extraction was carried out in triplicate.
A
90
Fig. 1. Confidence interval and maxima and minima of RIP [mmol.kg1] for major reference soil groups. Bars are 95th percentile confidence bounds and lines indicate maxima and minima.
groups can hardly be used as a criterion of classification for 137Cs sorption capacity of a soil. In fact, the FAO international soil classification was primarily based on agronomical criteria and not specifically on mineralogy which mainly controls the sorption capacity of trace Cs in different soils (Maes et al., 1998; Joussein et al., 2004). The specific behaviour of Podzols, Andosols and Ferralsols could be explained by the particular mineralogical assemblage characterizing those reference soil groups: surface horizons of Podzols are characterized by an accumulation of organic matter on a quartzitic substrate; Andosols are characterized by a predominance of amorphous minerals; and Ferralsols are characterized by an association of kaolinite and iron and aluminium oxy-hydroxides. All those soil components are deprived of selective adsorptions sites for singly charged positive alkali ions including Csþ (FAO/ ISRIC/ISSS, 1998). In many soils, most adsorption sites presenting a high selectivity for trace Csþ are “Frayed Edge Sites” (FES); they are localized at the edge of micaceous minerals and especially in the interlayer space formed by the juxtaposition of hydrated and non-hydrated zones. The pool of FES may be considered as the fraction of the CEC that is selective for trace Cs and was estimated from the relation RIP ¼ FES * KC considering an average KC of 1000 (Wauters et al., 1996a). From previous studies (Brouwer et al., 1983; Cremers et al., 1988), the percentage of FES in the CEC was reported to be less than 2%. In this study, FES/CEC values covered three orders of magnitude and ranged from 0.003% in a Podzol to 8.97% in a Regosol, confirming that any possible site-specific cases and related ranges of RIP are most likely included within the proposed scale of RIP. 3.2. RIP variability and major physicochemical parameters of soil Calculated Pearson correlation coefficients between RIP and soil characteristics are shown in Table 3. RIP shows a positive highly significant correlation (P-Value < 0.001) with: the pH (water and KCl), the CEC of the soil, the percentage of silt, the exchangeable Ca, K and the sum of exchangeable cations. The RIP was also positively correlated (P-Value < 0.05) with the percentage of clay and the exchangeable Mg. There was a highly significant negative correlation between the RIP and the percentage of sand. However, no Table 3 Correlation coefficients (N ¼ 88) between RIP and major soil characteristics.
R
R
pH water
pH KCl
CEC
%C
Sum cations
0.569 ***
0.542 ***
0.374 ***
0.109 NS
0.636 ***
% clay
% silt
% sand
Ca exch.
Mg exch.
K exch
0.239 *
0.458 ***
0.475 ***
0.679 ***
0.268 *
0.467 ***
*,*** Significant at the P 0.05 and 0.001 levels, respectively.
L. Vandebroek et al. / Journal of Environmental Radioactivity 104 (2012) 87e93
3.3. Relation between RIP and chemical extraction of
137
Cs
0
2000
4000
6000
8000
10000
12000
14000
RIP [mmol kg ]
Fig. 3. Relation between RIP [mmol.kg1] and fraction of total radiocaesium that is extractable in ammonium acetate.
displace radiocaesium readily from selective sites. In previous mechanistic studies mainly focusing on micaceous clay minerals and agricultural soils containing micas, NHþ 4 was shown to be 4e7 times more selective than Kþ for Csþ specific adsorption sites (Sweeck et al., 1990; Wauters et al., 1996b). Because of its small size, Hþ can also compete with Kþ for interlayer adsorption sites. For different Belgian soils, Sweeck (1996) demonstrated that Hþ exhibits efficiency comparable to Kþ in displacing radiocaesium. This findings in addition to the lower concentration of Hþ in the acid extraction solution compared to that of NHþ 4 explain the lower acid extraction efficiency generally observed in our study. More surprising was the high diverging extraction efficiency among soils having equal affinity for Cs sorption, especially for NH4-extraction (Fig. 3). For soils with very low RIP, it is possible that the diversity in both extraction fractions points to important but varying adsorption of Csþ on non-selective sites. For other soils characterized by higher RIP and vermiculitic properties, NHþ 4 loading during NHþ 4 extraction might have introduced an artifact by inducing Cs entrapment within interlayer spaces of micaceous particles. Indeed, although the NHþ 4 is generally considered as competitive with other cations for interlayer adsorption sites, the collapse-inducing effect of the vermiculite in presence of macroquantities of NHþ 4 and consecutive ions entrapment were shown to hinder radiocaesium extraction (Jacobs, 1963; Thiry et al., 1996). On the other hand, since trace Csþ is strongly and selectively adsorbed against Kþ in the selective sites, the intrinsic factors governing Kþ release can also greatly affect the desorption of trace Csþ from different minerals (Delvaux et al., 2000; Thiry et al., 2005). As such, just as for Kþ, the release of trace Csþ from soils exposed to acid extraction will also depend on the susceptibility of the minerals to weathering and thus on their structural properties. Based on the data presented here and similar observations reported by Sweeck (1996) for Belgian soils, it appears that different Cs sorption reversibility for a particular RIP range largely reflect significant but non-reducible structural differences which exist between the pools of highly selective sites among different soils or even within same soil groups. In those conditions, Hþ-extraction seems however more suitable to reflect changes in Cs-sorption 140
100 90 80 70 60 50 40 30 20 10 0
2.5
120
2.0
100
1.5
SQRT(RIP)
AER [% ]
Fig. 2 presents the fraction of acid-extractable radiocaesium (AER) as a function of RIP values for the different soils considered. The corresponding desorption yields of radiocaesium, using ammonium acetate 1 N as a desorbing agent, are shown in Fig. 3. For both extractants, the desorption yields decreased with increasing RIP, the decrease being particularly important in acid extraction for RIP values <2000 mmol kg1. In general, the ammonium extraction was more efficient than the acid extraction, with the exception of only 4 soils with very low RIP. For the other soils presenting RIP values <2000 mmol kg1, the NH4-extraction yield was 1.1e28 higher than the acid extraction efficiency. For RIP > 2000 mmol kg1, the differences between extraction yields þ increased with a NHþ 4 /H extraction yield ratio ranging from 11 to 229. A pronounced diversity in desorption behaviour was observed in soils presenting nevertheless identical RIP values. Those differences were particularly high for NH4-extraction. Soil contamination with radiocaesium as simulated in this study relates to micro-quantities of the element. At such trace loading, the selectivity patterns of radiocaesium with respect to other competitive cations in soils will be confined essentially to adsorption site groups with high affinity for radiocaesium. The relatively high extraction yields obtained with NH4 for numerous soils including those with a high RIP confirmed that NHþ 4 is capable to
100 90 80 70 60 50 40 30 20 10 0
NH OAc Exch. Cs [%] .
significant correlation was obtained between the RIP and the percentage of organic C. The R values for correlation to RIP decreased in the sequence: Ca exch. > Sum of cations > pH water > pH KCl > % sand > K exch. > % silt > CEC soil > Mg exch. > % clay > % C and the R values were in general quite low (Table 3). Different multi regression analyses between 137Cs adsorption and major chemical soil properties have been developed and give reasonable predictions on a local scale (e.g. Waegeneers et al., 1999). The ranking of soil properties we obtained was however notably different from that obtained by Waegeneers et al. (1999) for Belgian soils. Moreover, the low values of correlation coefficients we measured for each soil parameter illustrated the difficulty to make the link between RIP and classical major soil properties on a worldwide scale. Surprisingly, certain parameters which were empirically found to control 137Cs adsorption by soils such as clay percentage and organic matter showed the lowest correlation with RIP. The geological origin of soils and especially their mineralogy are confirmed to be the most important intrinsic soil properties which control Cs retention by agricultural soils. A good example is Ferralsols which are clay-rich soils but with a very low fixation capacity of kaolinite as the predominant clay. In agreement with Waegeneers et al. (1999) our observations illustrate also that, when considering large spatial scale in particular, attention should focus on the extrapolation of statistical relationships from one soil group or one region to respective other categories.
91
1.0 0.5 0.0 2000
4000
6000
8000
10000
12000
14000
y = -31.701x + 58.886
80
R = 0.819
60 40 20
0
2000
4000
6000
8000
10000
12000
14000
RIP [mmol kg ]
0 -2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
Log(AER)
Fig. 2. Relation between RIP [mmol.kg1] and fraction of total radiocaesium that is acid extractable (AER, %).
Fig. 4. Plot of the square root of RIP as a function of log (AER).
2.5
92
L. Vandebroek et al. / Journal of Environmental Radioactivity 104 (2012) 87e93 14000
RIP predicted
12000 10000 8000 6000 4000 2000 0 0
2000
4000
6000
8000
10000
12000
14000
RIP observed
Fig. 5. Predicted versus observed RIP (mmol kg1) according to the function shown in Fig. 4.
reversibility compared to NHþ 4 -extraction since extraction yields were less diverging for soils with similar RIP. 3.4. Possibility to predict the RIP from a simple extraction Results shown in Fig. 2 imply the possible existence of a relationship between RIP and the fraction of acid extractable radiocaesium (AER). Different transformations to normalize at once the RIP and the AER were tested but we observed effectively a highly significant (P-Value < 0.001) linear relation only between the square root of RIP and the logarithm of AER. This empirical relationship, shown in Fig. 4, was characterized by a correlation coefficient of R ¼ 0.905 (N ¼ 88). A highly significant (P-Value <0.001) correlation between the square root of RIP and the logarithm of ammonium extractable radiocaesium was also observed, but these two parameters were less well correlated. The correspondent correlation coefficient R was 0.613. Accordingly, we kept the AER to test the RIP-estimating capacity of a simple batch extraction. According to the results of the linear regression, the RIP could be estimated from the equation: RIP ¼ (31.701 * log(AER) þ 58.886)2. To our knowledge, the comparison between a simple acid extraction and RIP was never done before for a large set of soil types. The observed empirical equation could help laboratories with no possibility to measure RIP to make the link with a RIP scale and further rationale on radiocaesium behavior in soils (Gil-García et al., 2009) based on a simple acid extraction and activity counting of the 137Cs still measurable in many soils. Comparison of measured and modeled RIP values resulted in a determination coefficient R2 of 0.736 (N ¼ 88). These results are shown in Fig. 5. We defined the error percentage (% error) as the ratio (RIPpredicted RIPobserved)/ RIPobserved * 100. A positive % error means an over-estimation of the RIP while a negative % error resulted in an underestimation of the RIP. As shown in Fig. 6, the largest error percentages were observed in the range of RIP < 2000 mmol g1 where the % error goes from 87% to þ764% while it goes from 65% to þ123% for the range of RIP > 2000 mmol g1. This finding could be explained by the fact that below a RIP of 2000 mmol g1, the quantities of 800 700 600
% error
500 400 300 200 100 0 -100 0
2000
4000
6000
8000
10000
12000
14000
-200 RIP observed
Fig. 6. Evolution of the % error in function of RIP (mmol kg1). First and third quartile in dotted-line.
radiocaesium desorbed by the HCl solution were higher and more variable than for RIP above 2000 mmol g1 as discussed above. The range of error percentages which characterize 50% of the predicted RIP values may be defined by the interval between the first and the third quartile. This range of % error goes from 29 to þ67. When considering only the RIP values below 2000 mmol g1, the interval of % error is increased from 21 to þ128%. However, this interval is reduced between 31 and þ26% for the RIP above 2000 mmol g1 which involves 53% of soils of the collection and a large majority of the tested soil groups with most agronomical interest. That would mean that deducing a reliable RIP value whatever the soil properties could be achieved through a simple acid extraction. Ideally those results should be confirmed by an independent data base. 4. Conclusion Until now, there was no statistical information on the relationship between RIP values and soil type or properties at a worldwide scale. The RIP values we measured (1.8e13300 mmol kg1) included almost all possible cases of agricultural soils contamination including non-temperate zones, and especially the emerging countries where data are scarce or missing. Faced with the difficulty in identifying which soil data are relevant for standard setting regarding radiocaesium retention in soils, RIP was shown to be an intrinsic soil parameter (Cremers et al., 1988). In this study, only Podzols, Andosols and Ferralsols were characterized by a very low RIP (<2000 mmol kg1) most likely due to particular mineral assemblages deprived of micaceous clay. Large and uniform RIP ranges for other reference soil groups indicated that FAO soil classification can hardly be used as a tool of categorization for 137Cs mobility in soil. The difficulty to make the link between RIP and classical major soil properties on a worldwide scale was also illustrated. Clay percentage or organic C by example showed the lowest correlation with RIP. Our results showed however that a simple acid extraction could be used as an operational test to estimate the RIP of any soil and make the link with the proposed scale of RIP values. An interesting next step would imply to test the relationship between RIP, AER and 137Cs transfer to plant. Acknowledgments The authors are grateful to Xavier Draye (UCL, Belgium) and Hildegarde Vandenhove (SCKCEN, Belgium) for advice and useful discussions. The authors also thank Ad Van Oostrum (ISRIC, The Netherlands) for his help in preparing the soil collection and Achim Albrecht (Andra, France) for his recommendations in improving the manuscript. References Almgren, S., Isaksson, M., 2006. Vertical migration studies of 137Cs from nuclear weapons fallout and the Chernobyl accident. J. Environ. Radioactivity 91, 90e102. Batjes, N.H., 1995. A Homogenized Soil Data File for Global Environnemental Research: A Subset of FAO, ISRIC and NRCS Profiles (Version 1.0). Working Paper and Preprint95/10b. International Soil Reference and Information Center, Wageningen. Brouwer, A., Bayens, A., Maes, A., Cremers, A., 1983. Cesium and rubidium equilibria in illite clay. J. Phys. Chem. 87, 1213e1219. Cremers, A., Elsen, A., De Preter, P., Maes, A., 1988. Quantitative analysis of radiocesium retention in soils. Nature 335, 247e249. Cremers, A., Elsen, A., Valcke, E., Wauters, J., Sandalls, F., Gaudern, S., 1990. The sensitivity of upland soils to radiocaesium contamination. In: Desmet, G., Nassimbeni, P., Belli, M. (Eds.), Transfer of Radionuclides in Natural and Seminatural Environments. Elsevier Applied Science, London, pp. 238e248. Degryse, F., Smolders, E., Cremers, A., 2004. Enhanced sorption and fixation of radiocaesium in K-bentonites submitted to wetting-drying cycles. Eur. J. Soil Sci. 55, 513e522.
L. Vandebroek et al. / Journal of Environmental Radioactivity 104 (2012) 87e93 Delvaux, B., Kruyts, N., Cremers, A., 2000. Rhizospheric mobilization of radiocaesium in soils. Environ. Sci. Tech. 34, 1489e1493. Delvaux, B., Kruyts, N., Maes, E., Smolders, E., et al., 2001. Fate of radiocaesium in soil and rhizopshere. In: Gobran (Ed.), Trace Elements in the Rhizosphere. CRC Press, Bocca Raton, pp. 61e91. Food and Agricultural Organisation, International Soil Reference and Information Center, International Society of Soil Science, 1998. World Reference Base for Soil Resources Rome, Italy. Gil-García, C.J., Rigol, A., Rauret, G., Vidal, M., 2008. Radionuclide sorptiondesorption pattern in soils from Spain. Appl. Radiat. Isotope. 66, 126e138. Gil-García, C.J., Rigol, A., Vidal, M., 2009. New best estimates for radionuclides solidliquid distribution coefficients in soils, part 1: radiostrontium and radiocaesium. J. Environ. Radioactivity 100, 690e696. Gommers, A., Thiry, Y., Delvaux, B., 2005. Rhizospheric mobilization and plant uptake of radiocaesium from weathered micas: I. influence of potassium depletion. J. Environ. Qual. 34, 2167e2173. Goor, F., Thiry, Y., 2004. Processes, dynamics and modelling of radiocaesium cycling in a chronosequence of Chernobyl-contaminated Scots pine (Pinus sylvestris L.) plantations. Sci. Tot. Environ. 325, 163e180. Jacobs, D.G., 1963. . The effect of collapse-inducing cations on the cesium sorption properties of hydrobiotite. In: Int. Clay Conference, Stockholm, Sweden. A Pergamon Press BookeThe Macmillan Company, New York, pp. 12e16. Joussein, E., Kruyts, N., Righi, D., Petit, S., Delvaux, B., 2004. Specific retention of radiocesium in volcanic ash soils devoid of micaceous clay minerals. Soil Sci. Soc. Am. J. 68, 313e319. Kruyts, N., Thiry, Y., Delvaux, B., 2000. Respective horizon contributions to Cesium137 soil-to-plant transfer: a rhizospheric experimental approach. J. Environ. Qual. 29, 1180e1185. Kruyts, N., 2002. Influence de la nature des constituents du sol sur la mobilization rhizosphérique du radiocésium dans quelques sols forestiers. PhD Thesis, Université Catholique de Louvain, Belgium. Maes, E., Delvaux, B., Thiry, Y., 1998. Fixation of radiocaesium in an acid brown forest soil. Eur. J. Soil Sci. 49, 133e140. Noordijk, H., van Bergeijk, K.E., Lembrechts, J., Frissel, M.J., 1992. Impact of ageing and weather conditions on soil-to-plant transfer of radiocesium and radiostrontium. J. Environ. Radioactivity 15, 277e286. Rigol, A., Roig, M., Vidal, M., Rauret, G., 1999. Sequential extractions for the study of radiocaesium and radiostrontium dynamics in mineral and organic soils from Western Europe and Chernobyl areas. Environ. Sci. Tech. 33, 887e895. Sanchez, A.L., Wright, S.M., Smolders, E., Naylor, C., Stevens, P.A., Kennedy, V.H., Dodd, B.A., Singleton, D.L., Barnett, C.L., 1999. High plant uptake of radiocesium from organic soils due to Cs mobility and low soil K content. Environ. Sci. Tech. 33, 2752e2757. Sanchez, A.L., Smolders, E., Van den Brande, K., Merckx, R., Wright, S.M., Naylor, C., 2002. Predictions of in situ solid/liquid distribution of radiocaesium in soils. J. Environ. Radioactivity 63, 35e47.
93
Smolders, E., Van den Brande, K., Merckx, R., 1997. Concentrations of Cs-137 and K in soil solution predict the plant availability of Cs-137 in soils. Environ. Sci. Tech. 31, 3432e3438. Sweeck, L., Wauters, J., Valcke, E., Cremers, A., 1990. The specific interception potential of soils for radiocaesium. In: Desmet, G., Nassimbeni, P., Belli, M. (Eds.), Transfer of Radionuclides in Natural and Semi-natural Environments. Elsevier Applied Science, London, pp. 249e258. Sweeck, L., 1996. Soil-chemical availability of radiocesium in mineral soils. PhD thesis, Katholieke Universiteit Leuven, Belgium. Tikhomirov, F.A., Shcheglov, A.I., Sidorov, V.P., 1993. Forests and forestry: radiation protection measures with special reference to the Chernobyl accident zone. Sci. Tot. Environ. 137, 289e305. Thiry, Y., Vandecasteele, C.M., Delvaux, B., 1996. Ability of specimen vermiculitic minerals to fix radiocaesium: effect of the chemical environment. In: Gerzabek, M. (Ed.), Ten Years Terrestrial Radioecological Research Following the Chernobyl Accident, International Symposium on Radioecology, pp. 22e24. April 1996, Vienna, Austria. Mitt. d. Österr. Bodenkundl. Ges., H.53, S.93-100. Thiry, Y., Kruytz, N., Delvaux, B., 2000. Respective horizon contributions to Cesium137 soil-to-plant transfer: a pot experiment approach. J. Environ. Qual. 29, 1194e1199. Thiry, Y., Gommers, A., Iserentant, A., Delvaux, B., 2005. Rhizospheric mobilization and plant uptake of radiocaesium from weathered micas: II. influence of mineral alterability. J. Environ. Qual. 34, 2174e2180. Valcke, E., 1993. The behaviour dynamics of radiocesium and radiostrontium in soil rich in organic matter. PhD Thesis, Katholieke Universiteit Leuven. Vandenhove, H., Cremers, A., Smolders, E., Van Hees, M., 2005. Effect of K and bentonite additions on Cs-transfer to ryegrass. J. Environ. Radioactivity 81, 233e253. Waegeneers, N., Smolders, E., Merckx, R., 1999. A statistical approach for estimating the radiocaesium interception potential of soils. J. Environ. Qual. 28, 1005e1011. Wauters, J., 1994. Radiocesium in aquatic sediments: sorption, remobilization and fixation. PhD Thesis, Katholieke Universiteit Leuven, Belgium. Wauters, J., Elsen, A., Cremers, A., Kolopnev, A.V., Bulgakov, A.A., Comans, R.N.J., 1996a. Prediction of solid-liquid distribution coefficients of radiocaesium in soils and sediments. I. a simplified procedure for the solid phase characterization. Appl. Geochem. 11, 589e594. Wauters, J., Vidal, M., Elsen, A., Cremers, A., 1996b. Prediction of solid-liquid distribution coefficients of radiocaesium in soils and sediments. II. A new procedure for solid phase speciation of radiocaesium. Appl. Geochem. 11, 595e599. Yera, T.S., Vallejo, V.R., Valcke, E., Colle, C., Forstel, H., Millan, R., Jouglet, H., 1999. Cs137 and Sr-90 root uptake prediction under close-to-real controlled conditions. J. Environ. Radioactivity 45, 191e217.
Contents lists available at SciVerse ScienceDirect
Journal of Environmental Radioactivity journal homepage: www.elsevier.com/locate/jenvrad
Relevance of Radiocaesium Interception Potential (RIP) on a worldwide scale to assess soil vulnerability to 137Cs contamination Louis Vandebroeka, b, May Van Heesa, *, Bruno Delvauxb, Otto Spaargarenc, Yves Thirya,1 a
SCKCEN, Belgian Nuclear Research Center, Foundation of Public Utility, Biosphere Impact Studies, Boeretang 200, 2400 Mol, Belgium Université Catholique de Louvain, Unité des Sciences du Sol, Place Croix du Sud 2/10, 1348 Louvain-la-Neuve, Belgium c International Soil Ressource and Information Center, Duivendaal 9, 6701 AR Wageningen, The Netherlands b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 12 November 2009 Received in revised form 19 August 2011 Accepted 2 September 2011 Available online 2 October 2011
The extent of radiocaesium retention in soil is important to quantify the risk of further foodchain contamination. The Radiocaesium Interception Potential (RIP e Cremers et al., 1988, Nature 335, 247e249) is an intrinsic soil parameter which can be used to categorize soils or minerals in terms of their capacity to selectively adsorb radiocaesium. In this study, we measured RIP for a large soil collection (88 soil samples) representative of major FAO soil reference groups on a worldwide scale and tested the possibility to predict the RIP on the basis of other easily accessible or measurable soil data. We also compared RIP values with those obtained from separate chemical extraction experiments. The range of measured RIP values (1.8e13300 mmol kg1) was shown to include nearly all possible cases of agricultural soil contamination. Only Podzols, Andosols and Ferralsols were clearly characterized by a very low RIP (<2000 mmol kg1). On a worldwide scale, RIP was in fact slightly related to soil reference type or other simple major physicochemical parameters such as clay percentage or organic matter. Conversely our results indicated a link between the RIP and radiocaesium extractability across very different soils. We showed that, with the proposed scale of RIP values, a simple acid extraction method can provide an operational result highly predictive of potential RIP despite very contrasting soil properties. The RIP could be estimated from the empirical equation: RIP ¼ (31.701 * log(AER) þ 58.886)2 where AER is the fraction of acid-extractable radiocaesium. Ó 2011 Elsevier Ltd. All rights reserved.
Keywords: 137 Cs Radiocaesium Interception Potential Soil taxonomy Acid extraction Soil collection Micaceous clay
1. Introduction Following an accidental atmospheric release, radiocaesium is of great environmental and public health concern, due to its possible long transport, its long half-live, its high solubility in water and its biogeochemical behavior similar to that of K, a major nutrient for plants and animals. After deposition, radiocaesium is retained in the topsoil because of its selective adsorption on soil particles, largely ruled by weathered micaceous clays (Maes et al., 1998). Vertical migration rates are usually very low in agricultural mineral soils (Almgren and Isaksson, 2006). Even in soils of semi-natural areas, with low clay and high organic matter content, the net 137 Cs export from the rooting zone was determined to be less than 0.007% per year (Tikhomirov et al., 1993). In organic-rich forest soils
* Corresponding author. Tel.: þ32 14 33 21 05; fax: þ32 14 31 50 21. E-mail address: [email protected] (M. Van Hees). 1 Present address: ANDRA, Agence nationale pour la gestion des déchets radioactifs, Direction scientifique, Service transferts, 1-7 rue Jean Monnet, 92298 Châtenay-Malabry Cedex, France. 0265-931X/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jenvrad.2011.09.002
in particular, 137Cs persistence in the surface layers can be a source of long-lasting recycling by the vegetation (Kruyts et al., 2000; Goor and Thiry, 2004; Thiry et al., 2000). In general, the extent of 137Cs retention in agricultural and forest soils plays a key role in understanding the risk of foodchain contamination. Various experimental approaches were used to quantify such retention, yielding a variety of conceptual coefficients. One of them, the Radiocaesium Interception Potential (RIP e Cremers et al., 1988) has the particularity to be an intrinsic soil parameter but it must be determined under standardized experimental conditions. The measurement procedure involves soil chemical equilibration and its labeling with a sufficient amount of radiocaesium and thus necessitates adapted facilities to manipulate artificial radioactivity, which can discourage the use of RIP as a standard indicator at international level. Only a few European laboratories have produced most RIP values. Moreover, while the RIP method was frequently used to characterize the 137Cs adsorption capacity of various temperate soils containing micas (Waegeneers et al., 1999; Delvaux et al., 2000; Gil-García et al., 2008), tropical and sub-tropical soils were largely overlooked. However, Joussein et al. (2004) have shown that
88
L. Vandebroek et al. / Journal of Environmental Radioactivity 104 (2012) 87e93
Table 1 Selected characteristics of the 88 soils including mean RIP values (n ¼ 3) and mean fraction of total radiocaesium that is extractable in 0.1 M HCl and 1 M NH4OAc (n ¼ 3). Ref. Group
Land
Texture fractions Sand
Silt
Ca %
pH Clay
Water
KCl
Ca cmolc kg
% Andosol Andosol Andosol Andosol Andosol Andosol Andosol Andosol Andosol Andosol Arenosol Calcisol Calcisol Calcisol Calcisol Calcisol Calcisol Cambisol Cambisol Cambisol Cambisol Cambisol Cambisol Cambisol Cambisol Cambisol Cambisol Chernozem Chernozem Chernozem Chernozem Chernozem Chernozem Ferralsol Ferralsol Ferralsol Ferralsol Ferralsol Ferralsol Ferralsol Ferralsol Ferralsol Fluvisol Fluvisol Fluvisol Fluvisol Fluvisol Gleysol Gleysol Gleysol Gleysol Greyzem Kastanozem Kastanozem Luvisol Luvisol Luvisol Luvisol Luvisol Luvisol Luvisol Nitisol Nitisol Nitisol Nitisol Nitisol Nitisol Nitisol Phaeozem Podzol Podzol
C. Rica Ecuador Ecuador Ecuador Ecuador Ecuador Indonesia Indonesia Japan Kenya China Spain Turkey Turkey Turkey Turkey Turkey China China China China China China China China Italy Italy China Germany Peru Peru Romania USA Brazil Brazil China China Cuba Cuba Indonesia Indonesia Indonesia Japan Malaysia Nicaragua Peru Poland Cuba Cuba Thailand Thailand China USA USA China China Italy Kenya Nicaragua Spain Turkey Brazil Cameroon Columbia Italy Kenya Kenya Kenya China Belgium Malaysia
13.2 61.4 58.6 50.6 57.2 31.0 29.5 41.2 31.0 19.2 95.9 25.7 34.4 27.6 6.7 10.4 ND 21.0 12.3 15.7 32.5 45.7 36.1 31.8 0.8 12.2 ND 18.1 8.3 38.9 15.6 15.0 ND 16.0 36.0 10.5 74.1 23.9 10.3 52.2 28.6 6.0 56.1 98.2 37.6 52.1 41.7 6.3 5.0 28.5 1.1 7.8 34.6 ND 5.6 19.3 7.9 71.1 20.4 62.8 1.1 7.2 18.1 42.4 7.0 7.0 6.2 5.8 8.0 68.0 97.2
Exchangeable cations
10.4 29.4 34.6 34.7 38.9 52.6 46.4 33.6 38.8 54.8 1.4 21.3 26.1 27.5 29.7 33.3 ND 53.0 43.1 54.9 23.4 17.9 30.5 16.8 24.6 30.7 ND 46.3 75.0 36.2 52.5 56.3 ND 27.0 22.0 12.4 6.7 10.6 9.5 13.0 9.0 8.9 34.9 0.7 38.6 38.0 38.6 29.1 20.4 52.4 25.6 66.6 33.1 ND 61.9 59.2 25.1 4.0 36.0 26.1 38.2 37.5 28.3 15.5 21.8 36.2 23.6 19.3 54.2 16.4 2.6
76.5 9.3 6.8 14.7 4.0 16.4 24.2 25.2 30.2 26.1 2.6 53.0 39.5 44.9 63.6 56.3 65.7 26.0 44.5 29.5 44.2 36.2 33.5 51.3 74.9 57.1 75.5 35.5 16.7 24.9 31.9 28.7 24.6 57.0 42.0 77.0 19.2 65.6 80.1 34.6 62.3 84.0 9.0 1.1 23.8 10.0 19.7 64.7 74.4 19.1 73.3 25.6 32.3 23.1 32.4 21.6 67.2 24.8 43.6 11.1 60.7 55.4 53.6 42.1 71.3 56.8 70.0 74.8 37.9 15.6 0.2
4.9 5.7 6.7 6.4 6.6 5.2 5.6 5.0 5.5 6.7 8.9 7.9 8.2 8.1 8.1 8.0 7.5 8.3 8.2 7.4 4.8 4.3 3.9 6.2 4.7 6.8 6.3 8.0 8.1 8.1 8.3 7.7 5.9 4.5 4.0 4.3 4.8 6.7 5.2 4.2 4.7 3.5 6.6 8.3 7.3 6.0 5.1 7.7 6.6 5.8 5.0 6.7 7.0 8.1 5.6 6.5 6.2 6.7 6.8 5.7 8.0 6.2 6.4 4.8 5.6 5.4 5.8 6.4 6.5 6.2 4.9
4.3 4.8 5.4 4.3 5.7 4.6 5.2 4.5 4.6 6.1 8.2 7.1 7.3 7.3 7.0 7.3 6.6 7.1 7.1 6.5 4.1 3.8 3.6 4.8 4.0 5.9 5.6 7.3 7.1 7.8 7.4 7.0 4.8 4.2 4.0 4.1 4.1 5.6 4.2 3.7 4.2 3.5 5.7 8.1 6.7 5.5 4.6 6.8 5.1 5.3 4.1 6.3 6.1 ND 4.9 5.8 5.4 5.8 5.4 4.3 6.6 5.2 5.8 4.0 4.9 4.5 4.9 5.4 5.5 5.7 3.6
6.88 1.11 2.48 0.43 1.00 3.40 7.72 4.15 5.80 9.99 0.03 0.88 0.67 0.77 0.67 1.06 2.97 0.32 1.04 0.68 0.53 2.91 1.57 1.41 1.74 1.70 4.15 2.01 1.32 3.95 2.11 2.51 2.11 3.33 2.58 1.48 1.50 1.56 1.49 2.67 3.52 6.87 1.11 0.16 1.72 2.08 1.26 1.02 1.43 0.84 1.46 1.77 1.33 0.88 11.31 1.29 2.11 0.93 2.35 0.93 0.73 2.84 3.00 2.57 3.69 4.87 2.33 6.54 1.67 2.49 1.34
0.2 3.8 5.8 4.8 4.1 4.9 20.4 3.1 8.5 33.0 20.0 16.7 29.5 19.7 35.4 24.3 ND 49.8 49.0 20.3 0.8 0.8 0.0 10.2 2.3 20.9 ND 43.0 12.2 41.0 47.2 20.0 ND 0.6 0.2 0.2 0.6 8.0 5.8 1.8 3.0 0.2 7.1 7.3 37.9 14.9 12.7 28.9 29.9 7.1 15.1 20.8 15.0 ND 29.7 16.0 14.9 2.4 22.8 3.3 56.9 10.4 9.3 0.0 14.0 ND 7.9 12.8 23.7 ND 0.8
Mg
CECb
RIPc mmol kg1
K
Cs extractions
HCl
1
0.4 1.3 1.0 2.8 1.9 0.6 4.2 0.2 1.8 4.1 1.0 1.3 3.7 3.1 3.1 2.2 ND 1.0 2.4 2.4 0.3 0.3 0.0 2.7 0.3 1.3 ND 2.8 0.7 21.3 13.2 2.9 ND 0.4 0.1 0.3 0.3 2.7 1.5 0.5 0.7 0.2 2.7 0.8 2.4 0.0 3.1 12.4 12.2 1.3 3.9 2.4 4.5 ND 5.2 2.7 3.2 0.9 5.2 1.1 8.9 2.2 4.0 0.1 4.5 ND 1.9 8.1 6.6 ND 0.3
137
NH4OAc
% of total activity 0.2 0.1 0.7 1.1 0.4 0.3 2.7 0.2 0.8 1.5 0.2 0.9 2.6 0.9 0.7 1.8 ND 0.3 0.5 0.6 0.1 0.3 0.1 0.7 0.3 0.4 ND 0.2 0.3 0.9 1.7 0.3 ND 0.3 0.1 0.0 0.2 0.5 0.3 0.2 0.2 0.2 0.9 0.0 1.1 0.4 0.3 0.2 0.2 0.0 0.6 0.1 1.7 ND 1.2 0.2 1.2 0.4 0.9 0.2 1.9 1.1 0.4 0.2 1.8 ND 2.1 2.9 0.6 ND 0.0
23.3 5.9 10.0 10.2 6.7 12.4 36.7 20.7 48.6 70.7 2.1 20.0 25.3 27.5 29.9 17.2 50.5 22.2 23.0 19.2 5.1 10.5 8.0 13.2 10.0 25.6 30.5 30.2 16.4 24.9 39.1 28.9 20.0 6.7 7.3 5.9 3.5 12.0 13.2 9.5 11.8 25.2 14.9 2.2 33.8 12.1 16.2 38.3 46.3 11.5 30.2 20.4 20.2 14.0 51.0 16.1 25.7 6.7 29.4 11.1 83.8 17.7 22.7 14.8 25.0 ND 25.7 30.9 25.6 10.0 5.2
177 938 1020 484 94 97 450 208 1630 1320 1680 4340 12100 4220 9630 9090 9220 10500 12900 5310 945 1290 1370 4410 1680 6770 8860 5770 5820 5100 8590 6790 4510 26 203 217 245 1410 835 509 87 180 2720 150 9620 1730 6530 2110 504 4460 10100 4760 8390 5720 8050 5020 6750 1470 13 300 2270 7970 1550 539 324 4920 4370 4430 3510 6080 944 1.8
22.8 6.19 19.8 3.95 55.7 51.4 16.2 31.4 2.02 10.1 1.21 0.12 0.07 0.16 0.05 0.21 0.43 0.05 0.04 0.15 9.87 6.21 1.49 0.57 2.88 0.46 0.47 0.15 0.36 0.26 0.12 0.11 0.41 48.9 23.3 22.2 29.3 3.28 3.03 6.07 16.0 4.12 2.23 13.6 0.10 1.52 0.60 2.05 5.05 0.89 0.28 0.22 0.39 0.14 1.52 0.20 0.54 2.80 0.20 0.74 0.15 14.8 6.15 23.7 0.98 3.26 1.09 2.21 0.12 3.07 87.6
50.3 44.4 41.8 24.4 74.2 52.6 68.7 80.1 55.1 70.5 34.0 14.9 15.9 13.3 11.0 36.3 21.1 5.32 4.59 5.28 61.2 37.7 15.7 25.3 39.1 18.7 19.5 3.66 10.3 5.87 7.21 4.47 15.1 62.0 60.5 63.2 54.1 66.8 55.6 16.1 48.7 21.4 29.6 38.5 3.96 6.48 7.82 23.5 30.0 13.1 11.7 4.96 18.0 13.3 17.2 3.24 23.5 53.8 11.7 12.5 8.80 56.0 32.5 60.9 30.0 57.4 46.2 43.4 3.83 32.6 94.7
L. Vandebroek et al. / Journal of Environmental Radioactivity 104 (2012) 87e93
89
Table 1 (continued ) Ref. Group
Land
Texture fractions Sand
Silt
Ca %
pH Clay
Water
KCl
Ca
Malaysia Peru Peru USA Zambia Poland China Columbia Turkey Zambia Italy Italy Spain Spain Uruguay Uruguay Zambia
Arithmetic mean Geometric mean Minimum Maximum a b c
Mg
CECb
RIPc mmol kg1
K
137
Cs extractions
HCl
cmolc kg1
% Podzol Podzol Podzol Podzol Podzol Podzoluvisol Regosol Regosol Regosol Regosol Vertisol Vertisol Vertisol Vertisol Vertisol Vertisol Vertisol
Exchangeable cations
NH4OAc
% of total activity
96.7 91.9 84.1 92.0 98.0 5.4 13.6 75.1 10.3 82.6 6.1 5.5 20.9 15.0 4.3 6.6 29.5
2.5 8.0 11.0 5.0 2.0 85.4 70.7 13.8 51.1 8.0 33.7 25.4 26.5 28.9 49.9 58.5 21.5
0.8 0.0 5.1 3.0 0.0 9.1 15.7 11.1 38.6 9.3 60.3 69.2 52.7 56.1 45.8 34.9 48.9
4.6 4.5 4.0 5.0 4.4 7.0 8.6 4.1 7.7 5.5 5.4 6.3 7.8 7.6 7.3 5.3 7.3
4.2 3.8 3.5 3.6 3.5 6.3 8.3 3.7 7.0 4.3 4.5 5.1 6.3 6.3 6.6 4.5 5.8
1.28 0.63 6.05 0.27 1.90 1.22 0.73 1.60 3.07 0.90 1.94 2.60 0.46 0.94 2.38 3.72 1.88
1.5 0.2 0.4 ND 1.9 9.7 40.7 0.4 52.3 1.1 10.5 15.1 34.8 45.0 39.2 12.2 17.0
0.3 0.0 0.0 ND 0.5 0.6 3.6 0.1 2.0 0.4 2.0 2.9 8.2 7.5 3.9 3.8 10.2
0.1 0.0 0.0 ND 0.1 0.1 2.7 0.0 2.8 0.2 0.6 1.4 1.0 0.9 1.2 0.3 0.4
5.0 0.9 29.1 1.1 9.0 10.0 7.6 5.1 ND 4.8 20.3 34.7 44.4 67.0 43.2 36.0 29.4
13 38 12 62 7.4 3470 6820 1140 9650 744 6310 8210 3050 2100 3400 1820 2120
48.1 45.7 56.8 19.9 81.1 1.28 0.24 2.98 0.12 31.1 0.75 0.40 0.32 0.68 0.29 0.36 0.49
40.7 29.3 29.0 10.9 91.6 15.0 31.4 24.6 8.36 48.8 25.1 20.8 9.59 9.13 7.63 2.25 6.26
32.3 22.5 0.8 98.2
30.7 29.0 0.7 85.4
37.4 34.8 0.0 84.0
6.3 6.3 3.5 8.9
5.4 5.4 3.5 8.3
2.30 1.69 0.03 11.31
15.7 12.2 0.0 56.9
2.9 2.0 0.0 21.3
0.7 0.4 0.0 2.9
21.5 19.6 0.9 83.8
3730 2200 1.8 13300
9.87 1.39 0.04 87.6
30.0 23.9 2.25 94.7
C is the organic carbon content. CEC is the cation exchange capacity. RIP is the Radiocaesium Interception Potentia.
African volcanic soils devoid of micas could also selectively retain radiocaesium and thus have a RIP. In this study, we analyzed the soil parameters controlling radiocaesium retention for a large variety of agricultural substrates representative of major FAO soil reference groups including (sub)-tropical soils and other soils devoid of micas. The general objective was twofold: (1) further document the relevance of RIP as an indicator of 137Cs retention parameter in soils on a worldwide scale and (2) test the possibility to predict this parameter starting from easily accessible or measurable soil criteria such as soil type or texture, or from a simple Cs extraction using a batch method. 2. Materials and methods 2.1. A worldwide soil collection A collection of 88 surface soil samples from the International Soil Reference and Information Center (ISRIC, Wageningen) was used to measure RIP (see Table 1). Only the surface horizon was taken into account because of the low vertical migration usually observed for radiocaesium in mineral soils (Almgren and Isaksson, 2006). These soils were chosen to be representative of major soil types in agriculture on a worldwide scale covering different climatic zones. The classification of each soil sample into reference soil groups was achieved according to the World Reference Base for Soil Resource (FAO/ISRIC/ISSS, 1998). All these soils were analyzed by ISRIC for major parameters: pHwater, pHKCl, organic C content, texture, cation exchange capacity (CEC) and exchangeable bases as reported in Batjes (1995) based on classical methods of soil analyses. In brief pHwater and pHKCl were measured in suspension using soil:water and soil:KCl 1 M ratios of 1:2.5. The organic C content was determined following the Walkley-Black method. The clay (0e2 mm) and silt (2e50 mm) fractions were collected by gravimetric sedimentation (pipette method) after separation of the sand (50e2000 mm) fraction from the fine earth (<2 mm) by dispersion and wet sieving. The reference method based on ammonium acetate (1 M NH4OAc, pH7) was used to measure the soil CEC and to remove the exchangeable bases (Ca, Mg, K).
2.2. RIP measurements RIP was measured in a well defined ionic scenario according to a procedure adapted from Wauters et al. (1996a). One g of each soil sample (3 replicates) was introduced in a dialysis bag (Dialysis Tubing-Visking code DTV12000.06.000, size 3-20/32; Medicell International Ltd, UK) (5 mL) and equilibrated with a 5 104 M KCl e 0.1 M CaCl2 equilibration solution (Potassium Adsorption Ratio (PAR) ¼ 0.05 mmol1/2 L1/2). The solution of the container with the dialysis bag was changed 10 times during 7 days. During equilibration, the containers were agitated for 2 h each 12 h. Each dialysis bag was then transferred in a new container filled with 95 mL of the KCl e CaCl2 equilibration solution labeled with carrier-free 137CsCl. These bags were again agitated for 2 h each 12 h. After 5 days, aliquots of 20 mL were taken and 137Cs activity of the equilibrium solution was measured by gamma-counting and a Kd determined. Each measurement was determined in triplicate. The RIP (mmol.kg1) is defined by the relation: þ RIP ¼ KC * [FES] ¼ KCs D * [K ], where [FES] is the capacity of Frayed Edge Sites with high adsorption selectivity for Csþ, KC is the 137CseK selectivity coefficient on these sites (considering an average KC of 137 Cs solideliquid distri1000 (Wauters et al., 1996a)), KCs D is the bution coefficient and [Kþ] is the concentration of K in solution (Cremers et al., 1988). 2.3. Chemical extractions The contamination of a larger sub-sample (about 10 g) of the 88 soils was also carried out separately by mixing the soil aliquot with an adequate amount of deionized water labeled with carrier-free 137 CsCl. The volume of solution was adapted to reach soil moisture content close to field capacity and allow further hand mixing. The time elapsed since soil contamination is known to increase radiocaesium fixation. Wetting-drying cycles have been shown to simulate the 137Cs ageing in soil (Noordijk et al., 1992; Gommers et al., 2005) resulting in increase of Cs sorption and fixation in soil and micaceous clay (Degryse et al., 2004; Vandenhove et al., 2005). In order to quantify 137Cs retention in the long term after
L. Vandebroek et al. / Journal of Environmental Radioactivity 104 (2012) 87e93
2.4. Statistics Pearson correlation coefficients were calculated between RIP and the other soil properties without any data transformation. Data transformations concern only the relationship between RIP values and results from the chemical extractions. The statistical analysis was done with the software Minitab 15. 3. Results and discussion 3.1. RIP variability and reference soil groups The soil collection involved a large set of soil groups with very diverse properties and was found to cover a very large range of adsorption selectivity for radiocaesium in terms of RIP (Table 1). In fact, the measured RIP values extended over four orders of magnitude. These RIP values ranged from 1.8 mmol kg1 in a Malaysian humic podzol to 13 300 mmol kg1 in a Nicaraguan ferric luvisol. In comparison with RIP measured in previous studies (Table 2), the measured values are to the best of our knowledge the smallest and the highest RIP values ever determined for the soils. Our maximum value of RIP (13 300 mmol kg1) was even higher than the RIP measured for a pure illite (Silver Hill, Montana e 12 600 mmol g1 e Wauters, 1994) which was found to be characterized by an especially strong cesium retention (Delvaux et al., 2001). Fig. 1 represents the confidence interval (a ¼ 0.05) and maxima and minima of RIP for each major reference soil group represented in our collection. Only 3 groups of soil were characterized by a very low RIP: Podzols, Andosols and Ferralsols with all values <2000 mmol kg1. For the other soils, our results showed uniformity in RIP ranges indicating that in most cases, reference soil
Table 2 RIP range in the literature (adapted and completed from Kruyts, 2002). Source
Number of samples
O.M. [%]
Clay [%]
RIP [mmol kg1]
Cremers et al., 1990 Sweeck et al., 1990 Valcke, 1993 Smolders et al., 1997 Waegeneers et al., 1999 Rigol et al., 1999 Sanchez et al., 1999 Yera et al., 1999 Delvaux et al., 2000 Sanchez et al., 2002 Gil-García et al., 2008 This study
8 12 33 30 88 4 23 5 47 53 30 88
23e88 0.5e8.6 7e97 2e28 2e28 69e97 12.6e96.5 1e4.5 0.5e96 1.9e96.5 0.2e9.4 0e19.5
3e23 0.5e33 e 0.5e36 0.5e40 e 2e57.6 4.9e16.4 00.7e66 0.5e93.8 6.3e52.4 0e84
65e2450 67e4890 7e9199 54e5861 50e11 200 7e520 5e6545 443e2732 13e4861 5e6545 179e7000 1.8e13 343
14000 12000 10000 8000 6000 4000 2000
er tis ol
so l
l
V
go
zo Re
so l
Po d
iti
so l
N
ys ol
vi Lu
so l
so l
vi
le G
Fl u
m ze er no
Ch
Fe ra l
l
l
so bi m
lc
iso
Ca
Ca
os ol
0
nd
soil contamination, each soil sample was submitted to 10 dryingewetting cycles. At each cycle deionized water was added to reach soil moisture content close to 150% of field capacity. After 2e3 h the soil was transferred to a ventilated oven (40 C) and let dry overnight. The total 137Cs content of the soil was measured at the end of the treatment. Afterwards the fraction of acid extractable 137 Cs was determined following 137Cs desorption with HCl 101 M. In parallel, the fraction of NH4-exchangeable 137Cs was also determined following 137Cs desorption with ammonium acetate 1 M buffered at pH 7. These 2 extractions were carried out by using a batch system during 24 h in a ratio 18 mL extraction solution per g of soil. The 137Cs content in different samples was measured by gamma spectrometry (NaI detectoreMinaxi g Packard 5000 series). The counting time was adjusted in order to maintain the counting error below 3%. Each extraction was carried out in triplicate.
A
90
Fig. 1. Confidence interval and maxima and minima of RIP [mmol.kg1] for major reference soil groups. Bars are 95th percentile confidence bounds and lines indicate maxima and minima.
groups can hardly be used as a criterion of classification for 137Cs sorption capacity of a soil. In fact, the FAO international soil classification was primarily based on agronomical criteria and not specifically on mineralogy which mainly controls the sorption capacity of trace Cs in different soils (Maes et al., 1998; Joussein et al., 2004). The specific behaviour of Podzols, Andosols and Ferralsols could be explained by the particular mineralogical assemblage characterizing those reference soil groups: surface horizons of Podzols are characterized by an accumulation of organic matter on a quartzitic substrate; Andosols are characterized by a predominance of amorphous minerals; and Ferralsols are characterized by an association of kaolinite and iron and aluminium oxy-hydroxides. All those soil components are deprived of selective adsorptions sites for singly charged positive alkali ions including Csþ (FAO/ ISRIC/ISSS, 1998). In many soils, most adsorption sites presenting a high selectivity for trace Csþ are “Frayed Edge Sites” (FES); they are localized at the edge of micaceous minerals and especially in the interlayer space formed by the juxtaposition of hydrated and non-hydrated zones. The pool of FES may be considered as the fraction of the CEC that is selective for trace Cs and was estimated from the relation RIP ¼ FES * KC considering an average KC of 1000 (Wauters et al., 1996a). From previous studies (Brouwer et al., 1983; Cremers et al., 1988), the percentage of FES in the CEC was reported to be less than 2%. In this study, FES/CEC values covered three orders of magnitude and ranged from 0.003% in a Podzol to 8.97% in a Regosol, confirming that any possible site-specific cases and related ranges of RIP are most likely included within the proposed scale of RIP. 3.2. RIP variability and major physicochemical parameters of soil Calculated Pearson correlation coefficients between RIP and soil characteristics are shown in Table 3. RIP shows a positive highly significant correlation (P-Value < 0.001) with: the pH (water and KCl), the CEC of the soil, the percentage of silt, the exchangeable Ca, K and the sum of exchangeable cations. The RIP was also positively correlated (P-Value < 0.05) with the percentage of clay and the exchangeable Mg. There was a highly significant negative correlation between the RIP and the percentage of sand. However, no Table 3 Correlation coefficients (N ¼ 88) between RIP and major soil characteristics.
R
R
pH water
pH KCl
CEC
%C
Sum cations
0.569 ***
0.542 ***
0.374 ***
0.109 NS
0.636 ***
% clay
% silt
% sand
Ca exch.
Mg exch.
K exch
0.239 *
0.458 ***
0.475 ***
0.679 ***
0.268 *
0.467 ***
*,*** Significant at the P 0.05 and 0.001 levels, respectively.
L. Vandebroek et al. / Journal of Environmental Radioactivity 104 (2012) 87e93
3.3. Relation between RIP and chemical extraction of
137
Cs
0
2000
4000
6000
8000
10000
12000
14000
RIP [mmol kg ]
Fig. 3. Relation between RIP [mmol.kg1] and fraction of total radiocaesium that is extractable in ammonium acetate.
displace radiocaesium readily from selective sites. In previous mechanistic studies mainly focusing on micaceous clay minerals and agricultural soils containing micas, NHþ 4 was shown to be 4e7 times more selective than Kþ for Csþ specific adsorption sites (Sweeck et al., 1990; Wauters et al., 1996b). Because of its small size, Hþ can also compete with Kþ for interlayer adsorption sites. For different Belgian soils, Sweeck (1996) demonstrated that Hþ exhibits efficiency comparable to Kþ in displacing radiocaesium. This findings in addition to the lower concentration of Hþ in the acid extraction solution compared to that of NHþ 4 explain the lower acid extraction efficiency generally observed in our study. More surprising was the high diverging extraction efficiency among soils having equal affinity for Cs sorption, especially for NH4-extraction (Fig. 3). For soils with very low RIP, it is possible that the diversity in both extraction fractions points to important but varying adsorption of Csþ on non-selective sites. For other soils characterized by higher RIP and vermiculitic properties, NHþ 4 loading during NHþ 4 extraction might have introduced an artifact by inducing Cs entrapment within interlayer spaces of micaceous particles. Indeed, although the NHþ 4 is generally considered as competitive with other cations for interlayer adsorption sites, the collapse-inducing effect of the vermiculite in presence of macroquantities of NHþ 4 and consecutive ions entrapment were shown to hinder radiocaesium extraction (Jacobs, 1963; Thiry et al., 1996). On the other hand, since trace Csþ is strongly and selectively adsorbed against Kþ in the selective sites, the intrinsic factors governing Kþ release can also greatly affect the desorption of trace Csþ from different minerals (Delvaux et al., 2000; Thiry et al., 2005). As such, just as for Kþ, the release of trace Csþ from soils exposed to acid extraction will also depend on the susceptibility of the minerals to weathering and thus on their structural properties. Based on the data presented here and similar observations reported by Sweeck (1996) for Belgian soils, it appears that different Cs sorption reversibility for a particular RIP range largely reflect significant but non-reducible structural differences which exist between the pools of highly selective sites among different soils or even within same soil groups. In those conditions, Hþ-extraction seems however more suitable to reflect changes in Cs-sorption 140
100 90 80 70 60 50 40 30 20 10 0
2.5
120
2.0
100
1.5
SQRT(RIP)
AER [% ]
Fig. 2 presents the fraction of acid-extractable radiocaesium (AER) as a function of RIP values for the different soils considered. The corresponding desorption yields of radiocaesium, using ammonium acetate 1 N as a desorbing agent, are shown in Fig. 3. For both extractants, the desorption yields decreased with increasing RIP, the decrease being particularly important in acid extraction for RIP values <2000 mmol kg1. In general, the ammonium extraction was more efficient than the acid extraction, with the exception of only 4 soils with very low RIP. For the other soils presenting RIP values <2000 mmol kg1, the NH4-extraction yield was 1.1e28 higher than the acid extraction efficiency. For RIP > 2000 mmol kg1, the differences between extraction yields þ increased with a NHþ 4 /H extraction yield ratio ranging from 11 to 229. A pronounced diversity in desorption behaviour was observed in soils presenting nevertheless identical RIP values. Those differences were particularly high for NH4-extraction. Soil contamination with radiocaesium as simulated in this study relates to micro-quantities of the element. At such trace loading, the selectivity patterns of radiocaesium with respect to other competitive cations in soils will be confined essentially to adsorption site groups with high affinity for radiocaesium. The relatively high extraction yields obtained with NH4 for numerous soils including those with a high RIP confirmed that NHþ 4 is capable to
100 90 80 70 60 50 40 30 20 10 0
NH OAc Exch. Cs [%] .
significant correlation was obtained between the RIP and the percentage of organic C. The R values for correlation to RIP decreased in the sequence: Ca exch. > Sum of cations > pH water > pH KCl > % sand > K exch. > % silt > CEC soil > Mg exch. > % clay > % C and the R values were in general quite low (Table 3). Different multi regression analyses between 137Cs adsorption and major chemical soil properties have been developed and give reasonable predictions on a local scale (e.g. Waegeneers et al., 1999). The ranking of soil properties we obtained was however notably different from that obtained by Waegeneers et al. (1999) for Belgian soils. Moreover, the low values of correlation coefficients we measured for each soil parameter illustrated the difficulty to make the link between RIP and classical major soil properties on a worldwide scale. Surprisingly, certain parameters which were empirically found to control 137Cs adsorption by soils such as clay percentage and organic matter showed the lowest correlation with RIP. The geological origin of soils and especially their mineralogy are confirmed to be the most important intrinsic soil properties which control Cs retention by agricultural soils. A good example is Ferralsols which are clay-rich soils but with a very low fixation capacity of kaolinite as the predominant clay. In agreement with Waegeneers et al. (1999) our observations illustrate also that, when considering large spatial scale in particular, attention should focus on the extrapolation of statistical relationships from one soil group or one region to respective other categories.
91
1.0 0.5 0.0 2000
4000
6000
8000
10000
12000
14000
y = -31.701x + 58.886
80
R = 0.819
60 40 20
0
2000
4000
6000
8000
10000
12000
14000
RIP [mmol kg ]
0 -2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
Log(AER)
Fig. 2. Relation between RIP [mmol.kg1] and fraction of total radiocaesium that is acid extractable (AER, %).
Fig. 4. Plot of the square root of RIP as a function of log (AER).
2.5
92
L. Vandebroek et al. / Journal of Environmental Radioactivity 104 (2012) 87e93 14000
RIP predicted
12000 10000 8000 6000 4000 2000 0 0
2000
4000
6000
8000
10000
12000
14000
RIP observed
Fig. 5. Predicted versus observed RIP (mmol kg1) according to the function shown in Fig. 4.
reversibility compared to NHþ 4 -extraction since extraction yields were less diverging for soils with similar RIP. 3.4. Possibility to predict the RIP from a simple extraction Results shown in Fig. 2 imply the possible existence of a relationship between RIP and the fraction of acid extractable radiocaesium (AER). Different transformations to normalize at once the RIP and the AER were tested but we observed effectively a highly significant (P-Value < 0.001) linear relation only between the square root of RIP and the logarithm of AER. This empirical relationship, shown in Fig. 4, was characterized by a correlation coefficient of R ¼ 0.905 (N ¼ 88). A highly significant (P-Value <0.001) correlation between the square root of RIP and the logarithm of ammonium extractable radiocaesium was also observed, but these two parameters were less well correlated. The correspondent correlation coefficient R was 0.613. Accordingly, we kept the AER to test the RIP-estimating capacity of a simple batch extraction. According to the results of the linear regression, the RIP could be estimated from the equation: RIP ¼ (31.701 * log(AER) þ 58.886)2. To our knowledge, the comparison between a simple acid extraction and RIP was never done before for a large set of soil types. The observed empirical equation could help laboratories with no possibility to measure RIP to make the link with a RIP scale and further rationale on radiocaesium behavior in soils (Gil-García et al., 2009) based on a simple acid extraction and activity counting of the 137Cs still measurable in many soils. Comparison of measured and modeled RIP values resulted in a determination coefficient R2 of 0.736 (N ¼ 88). These results are shown in Fig. 5. We defined the error percentage (% error) as the ratio (RIPpredicted RIPobserved)/ RIPobserved * 100. A positive % error means an over-estimation of the RIP while a negative % error resulted in an underestimation of the RIP. As shown in Fig. 6, the largest error percentages were observed in the range of RIP < 2000 mmol g1 where the % error goes from 87% to þ764% while it goes from 65% to þ123% for the range of RIP > 2000 mmol g1. This finding could be explained by the fact that below a RIP of 2000 mmol g1, the quantities of 800 700 600
% error
500 400 300 200 100 0 -100 0
2000
4000
6000
8000
10000
12000
14000
-200 RIP observed
Fig. 6. Evolution of the % error in function of RIP (mmol kg1). First and third quartile in dotted-line.
radiocaesium desorbed by the HCl solution were higher and more variable than for RIP above 2000 mmol g1 as discussed above. The range of error percentages which characterize 50% of the predicted RIP values may be defined by the interval between the first and the third quartile. This range of % error goes from 29 to þ67. When considering only the RIP values below 2000 mmol g1, the interval of % error is increased from 21 to þ128%. However, this interval is reduced between 31 and þ26% for the RIP above 2000 mmol g1 which involves 53% of soils of the collection and a large majority of the tested soil groups with most agronomical interest. That would mean that deducing a reliable RIP value whatever the soil properties could be achieved through a simple acid extraction. Ideally those results should be confirmed by an independent data base. 4. Conclusion Until now, there was no statistical information on the relationship between RIP values and soil type or properties at a worldwide scale. The RIP values we measured (1.8e13300 mmol kg1) included almost all possible cases of agricultural soils contamination including non-temperate zones, and especially the emerging countries where data are scarce or missing. Faced with the difficulty in identifying which soil data are relevant for standard setting regarding radiocaesium retention in soils, RIP was shown to be an intrinsic soil parameter (Cremers et al., 1988). In this study, only Podzols, Andosols and Ferralsols were characterized by a very low RIP (<2000 mmol kg1) most likely due to particular mineral assemblages deprived of micaceous clay. Large and uniform RIP ranges for other reference soil groups indicated that FAO soil classification can hardly be used as a tool of categorization for 137Cs mobility in soil. The difficulty to make the link between RIP and classical major soil properties on a worldwide scale was also illustrated. Clay percentage or organic C by example showed the lowest correlation with RIP. Our results showed however that a simple acid extraction could be used as an operational test to estimate the RIP of any soil and make the link with the proposed scale of RIP values. An interesting next step would imply to test the relationship between RIP, AER and 137Cs transfer to plant. Acknowledgments The authors are grateful to Xavier Draye (UCL, Belgium) and Hildegarde Vandenhove (SCKCEN, Belgium) for advice and useful discussions. The authors also thank Ad Van Oostrum (ISRIC, The Netherlands) for his help in preparing the soil collection and Achim Albrecht (Andra, France) for his recommendations in improving the manuscript. References Almgren, S., Isaksson, M., 2006. Vertical migration studies of 137Cs from nuclear weapons fallout and the Chernobyl accident. J. Environ. Radioactivity 91, 90e102. Batjes, N.H., 1995. A Homogenized Soil Data File for Global Environnemental Research: A Subset of FAO, ISRIC and NRCS Profiles (Version 1.0). Working Paper and Preprint95/10b. International Soil Reference and Information Center, Wageningen. Brouwer, A., Bayens, A., Maes, A., Cremers, A., 1983. Cesium and rubidium equilibria in illite clay. J. Phys. Chem. 87, 1213e1219. Cremers, A., Elsen, A., De Preter, P., Maes, A., 1988. Quantitative analysis of radiocesium retention in soils. Nature 335, 247e249. Cremers, A., Elsen, A., Valcke, E., Wauters, J., Sandalls, F., Gaudern, S., 1990. The sensitivity of upland soils to radiocaesium contamination. In: Desmet, G., Nassimbeni, P., Belli, M. (Eds.), Transfer of Radionuclides in Natural and Seminatural Environments. Elsevier Applied Science, London, pp. 238e248. Degryse, F., Smolders, E., Cremers, A., 2004. Enhanced sorption and fixation of radiocaesium in K-bentonites submitted to wetting-drying cycles. Eur. J. Soil Sci. 55, 513e522.
L. Vandebroek et al. / Journal of Environmental Radioactivity 104 (2012) 87e93 Delvaux, B., Kruyts, N., Cremers, A., 2000. Rhizospheric mobilization of radiocaesium in soils. Environ. Sci. Tech. 34, 1489e1493. Delvaux, B., Kruyts, N., Maes, E., Smolders, E., et al., 2001. Fate of radiocaesium in soil and rhizopshere. In: Gobran (Ed.), Trace Elements in the Rhizosphere. CRC Press, Bocca Raton, pp. 61e91. Food and Agricultural Organisation, International Soil Reference and Information Center, International Society of Soil Science, 1998. World Reference Base for Soil Resources Rome, Italy. Gil-García, C.J., Rigol, A., Rauret, G., Vidal, M., 2008. Radionuclide sorptiondesorption pattern in soils from Spain. Appl. Radiat. Isotope. 66, 126e138. Gil-García, C.J., Rigol, A., Vidal, M., 2009. New best estimates for radionuclides solidliquid distribution coefficients in soils, part 1: radiostrontium and radiocaesium. J. Environ. Radioactivity 100, 690e696. Gommers, A., Thiry, Y., Delvaux, B., 2005. Rhizospheric mobilization and plant uptake of radiocaesium from weathered micas: I. influence of potassium depletion. J. Environ. Qual. 34, 2167e2173. Goor, F., Thiry, Y., 2004. Processes, dynamics and modelling of radiocaesium cycling in a chronosequence of Chernobyl-contaminated Scots pine (Pinus sylvestris L.) plantations. Sci. Tot. Environ. 325, 163e180. Jacobs, D.G., 1963. . The effect of collapse-inducing cations on the cesium sorption properties of hydrobiotite. In: Int. Clay Conference, Stockholm, Sweden. A Pergamon Press BookeThe Macmillan Company, New York, pp. 12e16. Joussein, E., Kruyts, N., Righi, D., Petit, S., Delvaux, B., 2004. Specific retention of radiocesium in volcanic ash soils devoid of micaceous clay minerals. Soil Sci. Soc. Am. J. 68, 313e319. Kruyts, N., Thiry, Y., Delvaux, B., 2000. Respective horizon contributions to Cesium137 soil-to-plant transfer: a rhizospheric experimental approach. J. Environ. Qual. 29, 1180e1185. Kruyts, N., 2002. Influence de la nature des constituents du sol sur la mobilization rhizosphérique du radiocésium dans quelques sols forestiers. PhD Thesis, Université Catholique de Louvain, Belgium. Maes, E., Delvaux, B., Thiry, Y., 1998. Fixation of radiocaesium in an acid brown forest soil. Eur. J. Soil Sci. 49, 133e140. Noordijk, H., van Bergeijk, K.E., Lembrechts, J., Frissel, M.J., 1992. Impact of ageing and weather conditions on soil-to-plant transfer of radiocesium and radiostrontium. J. Environ. Radioactivity 15, 277e286. Rigol, A., Roig, M., Vidal, M., Rauret, G., 1999. Sequential extractions for the study of radiocaesium and radiostrontium dynamics in mineral and organic soils from Western Europe and Chernobyl areas. Environ. Sci. Tech. 33, 887e895. Sanchez, A.L., Wright, S.M., Smolders, E., Naylor, C., Stevens, P.A., Kennedy, V.H., Dodd, B.A., Singleton, D.L., Barnett, C.L., 1999. High plant uptake of radiocesium from organic soils due to Cs mobility and low soil K content. Environ. Sci. Tech. 33, 2752e2757. Sanchez, A.L., Smolders, E., Van den Brande, K., Merckx, R., Wright, S.M., Naylor, C., 2002. Predictions of in situ solid/liquid distribution of radiocaesium in soils. J. Environ. Radioactivity 63, 35e47.
93
Smolders, E., Van den Brande, K., Merckx, R., 1997. Concentrations of Cs-137 and K in soil solution predict the plant availability of Cs-137 in soils. Environ. Sci. Tech. 31, 3432e3438. Sweeck, L., Wauters, J., Valcke, E., Cremers, A., 1990. The specific interception potential of soils for radiocaesium. In: Desmet, G., Nassimbeni, P., Belli, M. (Eds.), Transfer of Radionuclides in Natural and Semi-natural Environments. Elsevier Applied Science, London, pp. 249e258. Sweeck, L., 1996. Soil-chemical availability of radiocesium in mineral soils. PhD thesis, Katholieke Universiteit Leuven, Belgium. Tikhomirov, F.A., Shcheglov, A.I., Sidorov, V.P., 1993. Forests and forestry: radiation protection measures with special reference to the Chernobyl accident zone. Sci. Tot. Environ. 137, 289e305. Thiry, Y., Vandecasteele, C.M., Delvaux, B., 1996. Ability of specimen vermiculitic minerals to fix radiocaesium: effect of the chemical environment. In: Gerzabek, M. (Ed.), Ten Years Terrestrial Radioecological Research Following the Chernobyl Accident, International Symposium on Radioecology, pp. 22e24. April 1996, Vienna, Austria. Mitt. d. Österr. Bodenkundl. Ges., H.53, S.93-100. Thiry, Y., Kruytz, N., Delvaux, B., 2000. Respective horizon contributions to Cesium137 soil-to-plant transfer: a pot experiment approach. J. Environ. Qual. 29, 1194e1199. Thiry, Y., Gommers, A., Iserentant, A., Delvaux, B., 2005. Rhizospheric mobilization and plant uptake of radiocaesium from weathered micas: II. influence of mineral alterability. J. Environ. Qual. 34, 2174e2180. Valcke, E., 1993. The behaviour dynamics of radiocesium and radiostrontium in soil rich in organic matter. PhD Thesis, Katholieke Universiteit Leuven. Vandenhove, H., Cremers, A., Smolders, E., Van Hees, M., 2005. Effect of K and bentonite additions on Cs-transfer to ryegrass. J. Environ. Radioactivity 81, 233e253. Waegeneers, N., Smolders, E., Merckx, R., 1999. A statistical approach for estimating the radiocaesium interception potential of soils. J. Environ. Qual. 28, 1005e1011. Wauters, J., 1994. Radiocesium in aquatic sediments: sorption, remobilization and fixation. PhD Thesis, Katholieke Universiteit Leuven, Belgium. Wauters, J., Elsen, A., Cremers, A., Kolopnev, A.V., Bulgakov, A.A., Comans, R.N.J., 1996a. Prediction of solid-liquid distribution coefficients of radiocaesium in soils and sediments. I. a simplified procedure for the solid phase characterization. Appl. Geochem. 11, 589e594. Wauters, J., Vidal, M., Elsen, A., Cremers, A., 1996b. Prediction of solid-liquid distribution coefficients of radiocaesium in soils and sediments. II. A new procedure for solid phase speciation of radiocaesium. Appl. Geochem. 11, 595e599. Yera, T.S., Vallejo, V.R., Valcke, E., Colle, C., Forstel, H., Millan, R., Jouglet, H., 1999. Cs137 and Sr-90 root uptake prediction under close-to-real controlled conditions. J. Environ. Radioactivity 45, 191e217.