AT9900006 Plant uptake of radionuclides in lysimeter experiments

M.H. Gerzabek F. Strebl B. Temmel

June 1998

OEFZS—4820

SEIBERSDORF 30-20 / OEFZS-4820 June 1998

Plant uptake of radionuclides in lysimeter experiments

In: Environmental Pollution 99 (1998) 93-103

M.H. Gerzabek, F. Strebl, B. Temmel

Department of Environmental Research Division of Life Sciences ENVIRONMENTAL POLLUTION

ELSEVIER Environmental Pollution 99 (1998) 93-103

Plant uptake of radionuclides in lysimeter experiments

M.H. Gerzabek*, F. Strebl, B. Temmel Austrian Research Centre Seibersdorf, Division of Life Sciences, A-2444 Seibersdorf Austria Received 20 June 1997; accepted 15 October 1997

Abstract

The results of seven years lysimeter experiments to determine the uptake of 60 Co, 137Cs and 226 Ra into agricultural crops (endive, maize, wheat, mustard, sugarbeet, potato, Faba bean, rye grass) are described. The lysimeter consists of twelve monolithic soil profiles (four soil types and three replicates) and is located in Seibersdorf/Austria, a region with a pannonian climate (pronounced differences between hot and semi-arid summers and humid winter conditions, annual mean of precipitation: 517 mm, mean annual temperature: 9.8°C). Besides soil-to-plant transfer factors (TF), fluxes were calculated taking into account biomass production and growth time. Total median values of TF’s (dry matter basis) for the three radionuclides decreased from 226 Ra (0.068 kg kg" 1) to ,37Cs (0.043 kg kg" 1) and 60 Co (0.018 kg kg" 1); flux values exhibited the same ranking. The varying physical and chemical proper­ ties of the four experimental soils resulted in statistically significant differences in transfer factors or fluxes between the investigated soils for l37Cs and 226 Ra, but not for 60 Co. Differences in transfer between plant species and plant parts are distinct, with grami­ naceous species showing, on average, TF values 5.8 and 15 times lower than dicotyledonous species for 137Cs and 60 Co, respectively. This pattern was not found for 226 Ra. It can be concluded that 137Cs transfer is heavily influenced by soil characteristics, whilst the plant-specific factors are the main source of TF variability for 60 Co. The variability of 226 Ra transfer originates both from soil properties and plant species behaviour. © 1998 Elsevier Science Ltd. All rights reserved.

Keywords: Caesium; Cobalt; Fluxes; Plant uptake; Radium

1. Introduction waste of money and foodstuff, while underestimations must be avoided to meet radiological standards. Critics In recent years a growing interest in the evaluation of of the TF-concept demand a more realistic description fluxes of nutrients as well as contaminants through eco­ of soil-plant relationships, one that takes into account systems has been expressed in many fields of environ­ physiological mechanisms of nutrient and radionuclide mental research, including radioecology (Frissel and uptake. This approach might reduce the extremely high Pennders, 1983). Dealing with dynamic phenomena like variability of TF values for different crops (Smolders bioaccumulation requires the introduction of major and Merckx, 1993). driving processes into descriptions of radionuclide The present paper presents soil-plant transfer data transfer (Desmet et al., 1991). for 60 Co, 137Cs and 226 Ra from seven years of lysimeter Many radiological food-chain transport models as experiments. The calculation of traditional transfer fac­ well as biospheric assessment models are based on the tors was supplemented by a new approach that takes traditional transfer factor concept, where soil-plant into account dry matter production of vegetation and relationships are described by concentration ratios the exposure time of the crops to nutrients and radio­ assuming first order linearity. Radiological databases nuclides from artificially contaminated soils. The results like those published by the International Atomic Energy from both approaches are compared and correlations Agency (1994) intend to cover a broad range of soil with soil characteristics are calculated. types and plants. Model predictions that form the basis for decision making, e.g. in the calculation of dose limits 2. Materials and methods for human consumption of agricultural products, need to be as precise as possible; overestimation leads to a 2.1. Soils

The selection of soil types for the lysimeter experi­ * Corresponding author. E-mail: [email protected] . ments was based on their relative abundance in Austrian

0269-7491/98/$ 19.00 © 1998 Elsevier Science Ltd. All rights reserved. PII: S0269-7491 (97)00167-X 94 M.H. Gerzabek et al.j Environmental Pollution 99 (1998) 93-103 regions suitable for geological radioactive waste dis­ and powdered with an agate mill. The upper 20 cm posal. The lysimeter facility consists of 12 soil monoliths of each lysimeter profile were mixed thoroughly with from four sites prepared in triplicate (Soil I: Eutric the contaminated soil aliquots, then placed on top of Cambisol, a slightly alkaline, calcareous soil with a high the monolithic subsoils in the appropriate lysimeter silt content and a low amount of gravel; Soil II: Dystric pots. Introduced activities for each lysimeter (1 m2 sur­ Cambisol on fine colluvium with a medium content of face area) amounted to 588 ± 30.8 kBq 137Cs, coarse gravel; Soil III: Dystric Cambisol on crystalline 687.9 ± 40.7 kBq 60 Co, and 79.0 ±8.5 kBq 226 Ra. Added rock, extremely high amount of coarse stones, with low activities were monitored before and after mixing; the pH-value and high extractable potassium contents; Soil contamination procedure yielded standard deviations of IV: Dystric Gleysol already drained with a medium to less than 10%, proving homogeneity within pots and high fraction of gravel). The size of the monoliths is between the different lysimeters. 'The radionuclide 1x1x0.75 m. Soil characteristics are listed in Table 1 concentrations originally present in the soil profiles and described in more detail in Gerzabek et al. (1996). (depositions from Chernobyl or nuclear weapon fallout Soil analysis was performed according to Austrian in the case of mCs and the natural 226 Ra) were deter­ standard methods (Blum et al., 1996). Descriptions of mined in composite samples from the sampling area of lysimeter construction and maintenance have been pre­ the soil monoliths. Results of these determinations are viously described (Gerzabek, 1990). presented in Table 1.

2.2. Radionuclide contamination 2.3. Gammaspectrometric measurements

After contaminating 1 kg surface soil from each lysi­ Gammaspectrometric determinations were carried meter monolith with a solution containing a mixture of out using a multichannel analyser system with high the chosen radionuclides, these aliquots were air dried purity germanium detectors (30% relative efficiency) in

Table 1 Soil characteristics of lysimeter soils in 0-20 cm depth

Parameter Soil I Soil II Soil III Soil IV pH (CaCl2) 7.5 4.6 5.1 5.9 pH (H20) 8.1 5.1 5.3 6.2 Sand (%) 17 48 61 24 Silt (%) 65 43 27 62 Clay (%) 18 9 12 14 Caolinite (% of clay) 8 8 23 35 Illite (% of clay) 69 59 42 28 Humus (%) 1.5 2.4 2.6 5.2 CaCOj (%) 18.3

< d.l. = below detection limit. a Calciumacetate-lactate or double-lactate extraction. M.H. Gerzabek et al.jEnvironmental Pollution 99 (1998) 93-103 95

Table 2 ments were carried out in Marinelli beakers with air­ Growth time and biomass production of lysimeter crops of different tight lids after a storage time of at least three weeks to years allow the necessary equilibrium for the daughter Species Mean yield Growth Year nuclides of 226 Ra. (g dry time matter m-2) (days) 2.6. Calculations Endivia sp. 34 160 1990 (endive leaves) 2.6.1. Transfer factor Zea mays 2370 153 1991 Transfer factor (TF) values were determined on a (maize straw) mass basis: the activity concentration of plant samples Zea mays 1158 80 1991 (maize corn) (dry matter d.m.) was related to the radionuclide activ­ Triticum aestivum 424 252 1991/ ityconcentration in the first 20 cm of soil and calculated (wheat straw) 1992 as follows: TF = (Bq kg -1 plant d.m.) / (Bq kg"" 1 soil Triticum aestivum 326 50 1991/ d.m.). (wheat grain) 1992 The radioactive decay of 60 Co and 137Cs was taken Sinapis sp. 308 70 1992 (mustard leaves + stems) into account. Radionuclide migration during the seven- Beta saccharifera 213 178 1993 year period could not be introduced into our calcula­ (sugarbeet leaves) tions because sampling with corers would have Beta saccharifera 259 148 1993 destroyed the lysimeter monoliths. However, from the (sugarbeet root) experiences after the Chernobyl fallout in Austria Solanum tuberosum 106 73 1994 (potato leaves) (Meisel et ah, 1991) and from our measurements of Solanum tuberosum 96 22 1994 radionuclide leaching losses in the lysimeters (Gerzabek (potato tubers) et ah, 1996) only a minor impact on soil-to-plant trans­ Vicia faba 398 100 1995 fer is expected. (Faba bean shoots) A time-dependent increase of radionuclide fixation in Vicia faba 271 42 1995 (Faba beans kernels) soils is significant for 137Cs Chernobyl data (Frissel, Lolium perenne 296 197 1996 1992), data from experiments with artificial 137Cs con­ (rye grass) tamination seem to show a small effect. Correction fac­ tors calculated by Frissel (1992) from the IUR (International Union of Radioecologists) databank heavy shielding consisting of 10 cm lead and 5 cm cop­ before the introduction of TF derived in field studies per. The gamma lines used for radionuclide determina­ after the Chernobyl accident are 0.92, 0.86, and 0.80 for tion were 661.6keV for 137Cs, as well as 1173.2 and the first, second and third year following contamina­ 1332.5 keV for 60 Co. 226 Ra was determined both using tion, corresponding values for 90Sr (comparable to the gamma line at 185 keV and by determination of the 226 Ra) are even less important (0.94, 0.89, and 0.84). daughter nuclides 214Bi and 214Pb (609 and 351 keV). The dynamics of radionuclide concentration in leachate water of the lysimeter experiments gives no evidence for 2.4. Vegetation samples higher mobility of the applied radionuclides in the first years after contamination (Gerzabek et ah, 1996). The crop species cultivated on the lysimeter pots were Therefore, a correction of TF values for time-lag effects changed every year (Table 2). Sowing and harvesting was not applied in this study. dates were those used on farms in the vicinity of Sei- bersdorf. Other agricultural practices, such as the use of 2.6.2. Fluxes pesticides and fertiliser were restricted to a minimum, as For a more realistic description of the mass transfer far as possible. The lysimeter station at Seibersdorf/ of radionuclides from soil to plant material we calcu­ Lower Austria lies in a region with a Pannonian climate lated ‘fluxes’, taking into account biomass production with continental semi-arid and hot summers, and humid and exposure time of the crops investigated. The idea winter conditions (mean annual temperature: 9.8°C; behind this simple approach is to allow recalculation of mean annual precipitation: 517 mm). In dry periods a transfer factors using some additional easily available minimum of irrigation was applied in order to avoid information like yields and growth time. In this way, cracking of surface soil and plant drought stress. some sources of bias, e.g. dilution effects due to different biomass production, could be excluded. From investi­ 2.5. Sample preparation gations of perennial plants (e.g. spruce-twig buds) it is well known that 137Cs decreases during spring time due After biomass determination, harvested material was to biomass buildup and dry matter production (Molzahn dried at 70°C, milled and ashed if necessary. Measure­ and Assmann-Wertmuller, 1993). Nimis et ah (1988) 96 M.H. Gerzabek et al.lEnvironmental Pollution 99 (1998) 93-103

stress the great influence and dilution effect of cellulose All datasets followed a log-normal distribution (see content on the 137Cs concentration in plant samples, as below). Therefore, to obtain unbiased estimates, median radiocaesium, similar to potassium, is dissolved in values (x) and the average absolute deviation (AAD, cytoplasmatic and apoplasmatic solutions, while plant Sachs, 1993) from median were calculated. fiber contains less activity. The principal assumptions of this concept are: AAD = ^ x Y] | Xj - x | (2)

• the radionuclide concentration of vegetation material is related to the total amount of radio­ nuclides within the soil compartment (upper 20 cm Accordingly, further data analysis was carried out by of soil), which is thought to be the main rooting use of non-parametric methods calculating the con­ zone of the plants growing in the lysimeter; servative Spearman rank correlation coefficients and • the relationship between radionuclide uptake and Kruskal-Wallis ANOVA by ranks (Sachs, 1992). dry matter production is set linear.

Under these assumptions, /sp (flux soil-plant, [m2 g~] 3. Results and discussion d-1]) is calculated as follows: 3.1. Comparison of investigated radionuclides

/sp — p x y ^ £ or ^p1 =/sp x Pt x kt x As (1) Results of both soil-to-plant-transfer models (TF values and fluxes) are listed in Tables 3 and 4. They follow log-normal distribution functions, which is typi­ where Ap represents the total activity of the radionuclide cal for transfer factor datasets (Konshin, 1992). With in the biomass (Bq m-2) at harvesting time (t); P the respect to distribution, there seems to be no substantial amount of plant material ((g m~2) dry matter), pro­ difference between transfer factors and fluxes. Fig. 1 duced; in V the growth time in days (d); and As the total shows an example for 137Cs. activity of radionuclides (Bq mr2) in the rooting zone The total median values of TF data for the three (uppermost 20 cm of profile). radionuclides decreased from 226 Ra (0.068 ±0.098) to In order to determine fluxes, growth time was recor­ ,37Cs (0.043 ±0.083), with the lowest values being found ded between appearance and harvesting date of plants; for 60 Co (0.018 ±0.040); flux values exhibited the same in the case of fruits, the appearance of flowers was ranking: 32±28xlO"10 (226 Ra), 21 ±42x10-'° (137Cs), defined as the starting point of biomass production. For and 8.6±26x 10_1 ° m2 g_1 d_1 (60 Co). subsurface plant-parts (sugar beet, potato tubers) lit­ For further comparison, data were grouped by the erature data were used to define the starting point of four different soil types and the 13 different plant species tuber and storage-root development (Geisler, 1988, and plant parts, respectively. The varying physical and 1983). Harvested biomass was determined both on fresh chemical properties of the four experimental soils resul­ and dry matter basis. Mean results of dry matter pro­ ted in statistically significant differences (tested by duction are presented in Table 2. Kruskal-Wallis ANOVA by ranks, see Table 5) of transfer factors or fluxes for 137Cs and 226 Ra. For 60 Co 2.6.3.Statistics data differences were not significant. Grouping by soils Statistical tests were carried out with Microsoft leads to absolute deviations from the median of ±110- EXCEL 5.0 and STATISTICA for Windows software. 120% for 137Cs transfer data; the values for 60 Co show

Table 3 Median values of transfer factors and fluxes (± average absolute deviation from median) for four different soil types

Total median Soil I Soil II Soil III Soil IV

Transfer factor TF (kg kg ') '"Cs 0.043 ±0.083 0.022 ±0.026 0.121 ±0.123 0.007 ±0.016 0.096 ±0.103 ™Ra 0.068 ±0.098 0.023 ±0.072 0.118 ± 0.154 0.093 ±0.091 0.052 ±0.057 60 Co 0.018±0.040 0.014±0.032 0.039 ±0.068 0.023 ±0.042 0.014±0.017 Flux (m2 g-1 d~’)x 10-10 ':?Cs 21.0 ±42.2 10.2 ±10.6 50.6 ±53.8 2.06 ±7.34 54.3 ±73.6 ™Ra 32.2 ±27.5 11.6 ± 11.6 48.9 ±37.0 34.0 ±21.5 34.2 ±25.0 60 Co 8.60 ±25.8 6.35 ± 19.3 17.9 ±39.2 6.62 ±25.2 8.83 ±17.2

n = 13, median of total: n = 149. M.H. Table 4

Median values of transfer factors and fluxes for different agricultural crops Gerzabek Endive Mustard Potato Potato Maize Maize V. faba V. faba Rye Wheat Wheat Sugar beet Sugar beet leaves tuber straw grains leaves beans grass straw grains leaves root

et

Transfer factor TF (kg kg ” ') al.jEnvironmental 0.047 0.163 0.156 0.089 0.041 0.005 0.098 0.050 0.027 0.021 0.009 0.312 0.046 ±0.059 ±0.227 ±0.075 ±0.048 ±0.066 ±0.013 ±0.048 ±0.044 ±0.014 ±0.027 ±0.012 ±0.252 ±0.049 ™Ra 0.141 0.063 0.197 0.021 0.030 < d.l. 0.064 0.006 0.147 0.255 0.024 0.327 0.095 ±0.104 ±0.027 ±0.109 ±0.013 ±0.021 ±0.035 ±0.005 ±0.064 ±0.140 ±0.016 ±0.221 ±0.046 60 Co 0.023 0.020 0.073 0.014 0.004 0.0004 0.061 0.217 0.010 0.002 0.001 0.056 0.025

±0.027 ±0.013 ±0.046 ±0.006 ±0.001 ±0.0003 ±0.034 ±0.101 ±0.010 ±0.001 ±0.001 ±0.036 ±0.016 Pollution Flux (m2 g” 1 d—1)x 10—10 16.25 127.9 88.85 96.73 15.40 3.662 37.00 53.45 11.80 4.725 10.79 84.65 11.67

± 17.86 ±153.4 ±50.88 ±61.15 ± 19.89 ±6.923 ±27.06 ±54.74 ±6.258 ±5.148 ±11.56 ±70.42 ± 16.66 99

^Ra 40.85 37.95 126.80 24.64 8.40

±6.37 ±6.02 ±22.75 ±5.06 ±0.206 ±0.115 ±11.32 ±77.28 ±4.50 ±0.152 ±0.361 ±6.88 ±4.09 93-103 n—12; ± average absolute deviation from median. < d.l. = below detection limit.

3 98 M.H. Gerzabek et al.j Environmental Pollution 99 (1998) 93-103

100 T 100 T Cs-TF-all soils Cs-Rux - all soils

Fig. 1. Histograms of frequency distribution of l37Cs TF and flux data.

Table 5 Results of Kruskal-Wallis ANOVA by ranks

Soil 137Cs Sum of ranks 226 Ra Sum of ranks 60 Co Sum of ranks

TF flux TF flux TF flux

I 1928.0 1950.5 2169.5 1807.5 2714.5 2567.5 II 4185.5 3914.0 3564.0 3698.0 3399.5 3379.5 III 1329.5 1323.5 2935.5 2721.5 2682.5 2499.0 IV 3732.0 3987.0 2506.0 2948.0 2378.5 2729.0 H (« = 149, d.f. = 3) 68.41 66.92 16.20 24.20 6.78 • 4.22 P = 0.000 0.000 0.001 0.000 0.079 0.239 n: number of observations; d.f.: degrees of freedom; p: significance; H: test figure.

,37Cs TF value (kg/kg) ■ III Dystric Cambisol 0.7 □ I calcareous Cambisol a__ _ m 0.6 □ IV drained Gleysol 0.5 HII Dystric Cambisol 0.4 0.3 |: 0.2 jj FI r 0.1 h 0 wheat maize tye wheat sugar potato maize endivie faba potato faba must- beet gram gram grass straw beet straw leaves leaves beans ard leaves

137Cs flux(m2 g"1 d"1) x 10'10 ■ III Dystric Cambisol 400 □ I calcareous Cambisol □ IV drained G ley sol 300 11II Dystric Cambisol 200 100 0 maize wheat wheat sugar rye endivie maize beet potato potato faba faba must- grain grain straw beet grass straw leaves leaves leaves beans ard

Fig. 2. Comparison of ,37Cs TF and flux data for different crops grown on four soil types. M.H. Gerzabek et al.j Environmental Pollution 99 (1998) 93-103 99

2 2 6 Ra TF value (kg / kg) ■ m Dystric Cambisol 0.8 MI calcareous Cambisol 0.7 11IV drained Gleysol 0.6 ■ II Dystric Cambisol 0.5 0.4 0.3 0.2 0.1 0 maize faba maize potato wheat faba must- sugar endivie wheat potato rye beet grain beans straw grain leaves ard beet straw leaves grass leaves

226 Ra flux(m2 g'1 d'1) x 10'10 ■ III Dystric Cambisol Hi calcareous Cambisol 50

40

maize faba maize wheat sugar faba potato endivie must- wheat rye beet potato grain beans straw grain beet leaves ard straw grass leaves leaves

Fig. 3. Comparison of 226 Ra TF and flux data for different crops grown on four soil types.

60 Co TF value (kg / kg) ■ III Dystric Cambisol 0.4 HI I calcareous Cambisol - || IV drained Gleysol 0.3 ■ II Dy stric Cambisol -

0.2 0.1

0 maize wheat wheat maize potato tye must- sugar endivie potato faba beet faba grain grain straw straw grass ard beet leaves leaves leaves beans

60 Co flux(m2 g'1 d"1) x 10"10 ■ III Dystric Cambisol gl calcareous Cambisol ||IV drained Gleysol gn DystricCambisol xlO g

maize wheat maize wheat sugar rye potato endivie must- beet faba potato faba grain straw straw grain beet grass ard leaves leaves leaves beans

Fig. 4. Comparison of 60 Co TF and flux data for different crops grown on four soil types. 100 M.H. Gerzabek et al.jEnvironmental Pollution 99 (1998) 93-103

Table 6 radionuclides (Andersen, 1967; Frissel et al., 1990; Ger­ Spearman correlation coefficients of physical/chemical soil character­ zabek et al., 1990). In the present study, radium exhib­ istics with soil-plant transfer-factors and fluxes for 137Cs, 226 Ra and ited the highest plant availability followed by caesium 60 Co and cobalt. It is interesting to note that this behaviour "7Cs ™Ra 60 Co does not match with the mobility of these elements as TF flux TF flux TF flux reflected by their leaching losses. Gerzabek et al. (1996) reported the following ranking of leaching losses in the TF,37Cs 0.932 0.329 0.413 0.532 0.543 respective lysimeters: cobalt > radium > caesium. The TF60 Co 0.517 0.356 0.390 0.945 TF226 Ra 0.886 explanation for this phenomenon is based on the differ­ ences between the radionuclides with respect to selective pH value -0.287 -0.317 -0.387 % Sand 0.285 0.318 plant uptake. Caesium, with its plant physiological % Silt -0.285 -0.318 similarity to potassium (Collander, 1941), obviously has % Clay -0.287 -0.317 -0.387 a much higher affinity to the uptake sites of the plant % Humus roots than cobalt, an element not considered essential mg K20 100 g"la -0.650 -0.603 mg P205 100g_la -0.526 -0.571 for most plants (Mengel, 1984). % Ca -0.285 -0.318 All soil-plant transfer data were tested against % Mg each other by Spearman correlation. The significant % K -0.650 -0.603 (/><0.05) correlation values are listed in Table 6. We % Fe 0.585 0.549 0.362 Ca exch -0.276 observed significant positive correlations between Mg exch -0.276 transfer factors for Co and Cs and between Cs, Co and K exch -0.650 -0.603 Ra. For all investigated radionuclides, fluxes and trans­ Na exch -0.276 fer factors were highly correlated (r between 0.886 and Water 0.945, see Table 6). conductivity Electrical -0.276 conductivity 3.1.1. Caesium Fe (EDTA) Differences between the investigated plant species are Mn (EDTA) -0.276 distinct, the TF values decrease from 0.631 in sugarbeet Cu (EDTA) 0.591 0.625 Zn (EDTA) -0.419 -0.326 leaves on soil II to 0.001 in wheat and maize grains % Elite grown on soil III (see Fig. 2). Graminaceous species % Caolinite show TF values on average 5.8 times lower than dicoty­ % Illite %clay/100 ledonous species. Higher TF values of 137Cs into % Caol. %clay/100 vegetative plant parts (versus produced seeds) were n= 52; r{ta = 0.05):0.274; r(ta = 0.01):0.356. found for maize (8.2-fold), Faba bean (2.0-fold) and a Plant available fraction, calciumacetate-lactate or double-lactate winter wheat (2.3-fold). The difference between below ­ extraction. ground organs and shoots amounts to 1.75-fold higher TF values for potato leaves and 6.8-fold values variations of approximately ± 180% for TF values and for sugar beet leaves. Considering flux data, for three an extremely high variation for flux data of 262%. For species the ratio turns to the opposite direction with 226 Ra transfer factors grouped by the four soil types, we higher 137Cs fluxes into edible plant parts than into found variations of ± 120-160%, while for 226 Ra-flux leaves for wheat (1:0.4), faba bean (1:0.7) and potato the variation was comparatively low with only ±75% (1:0.9). For maize (1:4.2) and sugar beet (1:7.2) the of the median (see Table 3). ratios of 137Cs flux are more similar to the ratios of TF The variability of TF values and fluxes within one values. Comparing the average transfer factors (see plant species was ± 100-110% for 137Cs, but only 50- Table 4) with the ‘best estimates’ of the Handbook of 60% for 60 Co and 226 Ra (see Table 4 and Figs. 2-4). Parameter Values for the Prediction of Radionuclide These findings underline that 137Cs transfer is heavily Transfer in Temperate Environments (IAEA, 1994) shows influenced by soil characteristics; for 60 Co the plant that our findings are generally in a good agreement and specific factors are the main source of TF variability, as lie in the same order of magnitude. Transfer factors into ANOVA between soils was not significant. The varia­ wheat grain and potato tubers in our experiment are bility of 226 Ra transfer stems both from soil properties almost identical to the ‘best estimates’. The lowest soil- and plant species behaviour. All radionuclides investi­ to-plant TF for caesium was observed for maize grains, gated exhibited the highest average TF value on soil II which is supported by earlier findings based on Cher­ (Table 3). This Dystric Cambisol has the lowest clay nobyl fallout studies in Austria (Gerzabek et al., 1990). content and pH value and by far the lowest cation Additionally these findings indicate that time lag effects exchange capacity in the plough layer of all studied soil of TF-values in the present experiment at least are profiles, which explains the high plant availability of the small. After endive, grown in 1990, maize and wheat M.H. Gerzabek et al./Environmental Pollution 99 (1998) 93-103 101 were cultivated in 1991 and 1992 (see Table 2). A sig­ corresponding seeds of Faba bean, maize and wheat. nificant higher mobility of 137Cs would be expected only The same ratio (1:10) was found for the comparison of in the first years after contamination due to insufficient edible and non edible plant parts of potato, while sugar binding of artificially introduced radionuclides to the beet showed only a 4-fold TF to leaves in relation to TF soil matrix. into storage roots (see Table 4). The nearly 10-fold The generally higher accumulation of radiocaesium in accumulation of radium in wheat straw in comparison dicotyledonous species versus graminaceous species is with the grains reproduces findings from Mordvedt also supported by previous results (Gerzabek et al., (1986), who refers to literature data reporting Ra con­ 1990). Considering the Cs fluxes, differences of transfer centrations in wheat straw to be 5 to 16 times higher factors between leaves (straw) and harvested plant parts than in grains. This finding reflects the chemical are only apparent to some extent due to the uptake similarity of radium with the essential plant nutrient (translocation) rates (Table 4, Fig. 2). The lower trans­ calcium, which is mainly transported in the xylem sap. fer factors into maize grain and sugar beet root com­ Therefore, in storage tissues supplied by phloem like pared with straw and leaves, respectively, can be roots and seeds, radium concentrations are compara­ explained both by lower flux rates and a shorter growth tively low (Mordvedt, 1994). period of the harvested plant part. In all other cases Calculated transfer factors for 226 Ra are in agreement (wheat, Faba beans, potato) we observed slightly higher with findings from the Glen Radium Extraction sites or similar Cs-flux rates for both investigated plant parts. (Kirchmann et al., 1973) and are within errors of the Thus, the shorter growing period explains solely the ‘best estimates’ of the IAEA-handbook (IAEA, 1994). lower transfer factors into the storage organs of these Transfer factors reported by Vasconcellos et al. (1987) crops. The distribution of radiocaesium within the plant from a uranium mining and milling area in Brazil (P090S is similar to that of potassium. It is well known that de Caldas Plateau) tend to be lower than our values. potassium concentrations in roots are about half the For 226 Ra the analysis revealed even more significant shoot concentrations (Flowers and Lauchli, 1983), and correlations between soil-to-plant transfer and soil potassium contents of cereal straw exceeds the K-con- characteristics than for 137Cs: while the TF values tents of grains by a factor of 1.7 (rye) to 4.6 (oat) showed significant negative correlations with the pH (Geisler, 1988), which reflects the results of our TF value, the silt and clay content of the topsoil, the total investigations. Ca content, the exchangeable fraction of Ca, Mg and For ,37Cs TF values, significant negative correlations Na, the EDTA-extractable fraction of Mn and the elec­ were found with pH, the clay content, ‘plant available ’ trical conductivity, a positive relation was observed for potassium and phosphorus, total as well as exchange­ the sand content. 226 Ra flux data were negatively corre­ able potassium content, and the amount of EDTA- lated with the pH, the silt and clay content, and the extractable zinc (Table 6). Positive relations were found total amount of Ca and positively correlated with the with the iron content of the soil and the amount of sand content and Fetot. The pH-effect is due to lower EDTA-extractable copper. 137Cs flux data yielded simi­ availability (Hewamanna et al., 1988). Soil calcium, lar results, without showing the correlation with pH and based on its chemical similarity to radium, is considered clay content. The highest correlation coefficients were to be important in controlling Ra availability to plants, obtained for all soil parameters describing potassium which explains the effect of soil Ca on plant uptake of (Ktot, Kexch, K20Cal,dl)- This close relation of caesium radium. Accordingly, the only calcareous experimental uptake to the potassium status of the soil has been pre­ soil I of this study (Eutric Cambisol) showed lowest TF viously described by Jackson et al. (1965). From our values for almost all plant species (see Fig. 3). In field data we could deduce a significant effect of clay content, studies, however, Hewamanna et al. (1988) and Vas­ but not of clay mineral types, as suggested by Valcke et concellos et al. (1987) failed to detect a significant al. (1996), on the soil-to-plant transfer of 137Cs. These influence of Ca in areas of high natural radioactivity. authors proposed an improvement of 137Cs transfer models by consideration of the amount of illitic clays 3.1.3. Cobalt with a high number of Frayed Edge Sites in the experi­ Although cobalt is not an essential element for plants mental soils. Increasing phosphorus supply seems to (Mengel, 1984), cobalt uptake is highly dependent on diminish Cs-transfer, probably due to a dilution effect plant species (Fig. 3). The comparison of the three caused by enhanced biomass production. monocotyledonous species with the other species showed extremely low TF-values into grasses (approxi­ 3.1.2. Radium mately 15-fold lower TF). Uptake of cobalt into roots is Transfer factors of 226 Ra showed no difference considered to be a passive process leading to an accu­ between graminaceous and dicotyledonous species, mulation in the root tissue. Only a small part is trans­ whereas the transfer to vegetative plant parts turned out located into the shoot (Coughtrey and Thorne, 1983). to be one order of magnitude higher than the TF to the This picture could not be found for TF-data of sugar 102 M.H. Gerzabek et al./Environmental Pollution 99 (1998) 93-103 beet from this study (see Table 4). Excluding Faba bean work. They are indebted to Dr K. Muck and Dr F. Steger, pods, the transfer to edible plant parts was approxi­ both of the Austrian Research Centre Seibersdorf, for mately 5-fold lower than to non edible tissues. The most gammaspectrometric measurements, to the Institute for striking result was the high accumulation of cobalt in Soil Research of the University for Agricultural Sci­ Faba bean pods (Fig. 4), which was also reported from ences, Vienna, for the clay characterisation and to Dr earlier investigations (Coughtrey and Thorne, 1983; Michael Stachowitsch/Vienna for linguistic help. Gerzabek et al., 1994). Cobalt is a component of vita­ min B12, which plays a key role in the build-up of leghaemoglobin, an essential enzyme of nitrogen fixa­ References tion in leguminous metabolism (Mengel, 1984). The soil-to-plant transfer of cobalt was not sig­ Andersen, A.J., 1967. Investigations on the plant uptake of fission nificantly correlated with any of the investigated soil products from contaminated soils. I. Influence of plant species and soil types on the uptake of radioactive strontium and caesium. Riso properties (Table 6), although some differences between Report 170. soils were observed (Table 3, Fig. 4). The Dystric Gley- Blum, W.E.H., Spiegel, H., Wenzel, W.W., 1996. Bodenzus- sol (soil IV) in most cases exhibited the lowest transfer tandsinventur, 2nd ed. Bundesministerium fur Land- und Forst- factors and fluxes. Apart from this, the remaining three wirtschaft, Vienna. investigated soils could be grouped by their pH, show­ Collander, R., 1941. Selective absorption of cations by higher plants. Plant Physiology 6, 691-720. ing decreasing transfer factors and fluxes with increas­ Coughtrey, P.J., Thorne, M.C., 1983. Radionuclide Distribution and ing pH (Table 1); this is supported by results obtained Transport in Terrestrial and Aquatic Ecosystems, vol. 2. Balkema, by the evaluation of the IUR (International Union of Rotterdam. Radioecologists) databank (Frissel et al., 1990). Klessa Desmet, G.M., Van Loon, L.R., Howard, B.J., 1991. Chemical spe- et al. (1989) showed in a long-term field experiment that ciation and bioavailability of elements in the environment and their relevance to radioecology. The Science of the Total Environment Co concentration in herbage is inverse logarithmically 100, 105-124. related to soil pH. Flowers, T.J., Lauchli, A., 1983. Sodium versus potassium: Substitu ­ tion and Compartmentation. In: Lauchli, A., Bieleski, R.L. (Eds.), Inorganic Plant Nutrition, Part B. Springer Verlag, Berlin, pp. 651 — 4. Conclusions 681. Frissel, M.J., Pennders, R., 1983. Models for the accumulation and migration of 90Sr, 137Cs, 239/240pu ancj 24i^m jn the; upper layer of In comparison with the literature, the present data set soil. In: Coughtrey, P.J., Bell, J.N.B., Roberts, T.M. (Eds.), Eco­ shows that in general soil-plant transfer of 137Cs, 226 Ra, logical Aspects of Radionuclide Release. Special Publication No. 3 and 60 Co in a Pannonian climate does not differ of the British Ecological Society, pp. 63-72. distinctly from other temperate environments. 137Cs- Frissel, M.J., Noordijk, H., Van Bergeijk, K.E., 1990. The impact of extreme environmental conditions, as occurring in natural ecosys­ transfer is heavily influenced by soil characteristics; for tems, on the soil-to-plant transfer of radionuclides. In: Desmet, G., 60 Co the plant specific factors are the main source of Nassimbeni, P., Belli, M. (Eds.), Transfer of Radionuclides in Nat­ variability; for the 226 Ra soil-to-plant transfer soil ural and Semi-natural Environments. Elsevier, London, pp. 40-47. characteristics and plant factors seem to be equally Frissel, M.J., 1992. An update of the recommended soil-to-plant important. Data calculated using the traditional transfer transfer factors of 90Sr, ,37Cs and transuranics. VIIIth Report of the Working Group Soil-to-plant Transfer Factors. IUR Pub R-9212-02. factor concept and using fluxes in principle yielded Geisler, G., 1988. Pflanzenbau —Bin Lehrbuch— Biologische Grund- similar information when comparing different soils and lagen und Technik der Pflanzenproduktion, 2nd Edition Paul Parey, plant species with respect to plant uptake of caesium, Berlin, Hamburg. cobalt and radium. Correlations between soil param­ Geisler, G., 1983. Ertragsphysiologie von Kulturarten des gemaBigten eters, and TF and fluxes, respectively, were also similar. Klimas. Paul Parey, Berlin, Hamburg. Gerzabek, M.H., 1990. Bine einfache Vorrichtung zur Entnahme The potential of the simple flux values cannot be monolithischer Bodenkorper. Die Bodenkultur 41, 283-288. conclusively evaluated at the present time. Advantages Gerzabek, M.H., Horak, O., Miick, K., 1990. Cs-137 soil to plant can be envisaged both for estimating plant uptake of transfer studies and their implications on parameters used in the radionuclides under distinctly different climatic condi­ Austrian version of ECOSYS. In: Desmet, G., Nassimbeni, P., Belli, tions and the corresponding growth times, and for M. (Eds.), Transfer of Radionuclides in Natural and Semi-natural Environments. Elsevier, London, pp. 611-618. modelling radionuclide concentrations in pastures. The Gerzabek, M.H., Mohammad, S.A., Miick, K., Horak, O., 1994. latter issue is currently being investigated in the lysi- 60 Co, 63 Ni and 94Nb Soil-to-Plant transfer in pot experiments. meter by maintaining a continuous rye grass cover. Journal of Environmental Radioactivity 25, 205-212. Gerzabek, M.H., Miick, K., Steger, F., Algader, S.M., 1996. Die Auswaschung von 60 Co, 137Cs und 226 Ra im Lysimeterversuch. Die Bodenkultur 47, 71-80. Acknowledgements Hewamanna, R., Samarakoon, C.M., Karunaratne, P.A.V.N., 1988. Concentration and chemical distribution of radium in plants The authors thank the Austrian Ministry for Science, from monazite-bearing soils. Environmental and Experimental Arts and Traffic for providing funds for the present Botany 28, 137-143. M.H. Gerzabek et al.lEnvironmental Pollution 99 (1998) 93-103 103

International Atomic Energy Agency (IAEA), 1994. Handbook of Mordvedt, J.J., 1986. Effects of calcium silicate slag application on Parameter Values for the Prediction of Radionuclide Transfer in radium-226 concentrations in plant tissues. Communications in Soil Temperate Environments. Technical Report Series No. 364, IAEA, Science and Plant Analysis 17, 75-84. Vienna. Mordvedt, J.J., 1994. Plant and soil relationships of uranium and Jackson, W.A., Craig, D., Lugo, H.M., 1965. Effects of various thorium decay series radionuclides—A review. Journal of Enviro- cations on cesium uptake from soils and clay suspensions. Soil Sci­ mental. Quality 23 (4), 643-650. ence 99, 345-353. Nimis, P.L., Giovani, C., Padovani, R., 1988. On the ways of expres­ Kirchmann, R., Lafontaine, A., Cantieon, G., Boulenger, R., 1973. sing radiocesium contamination in plants for radioecological Etude du cycle biologique pare ouru par la radioactivity. SCK. research. Studia Geobotanica 8, 3-12. CEN-Report BLG-477. Sachs, L., 1992. Angewandte Statistik, 7th revised Edition. Springer Klessa, D.A., Dixon, J., Voss, R.C., 1989. Soil and agronomic factors Publishers, Berlin, Heidelberg, New York. influencing the cobalt content of herbage. Research and Develop­ Sachs, L., 1993. Statistische Methoden. 7th extended Edition. Springer ment in Agriculture 6, 25-35. Publishers, Berlin, Heidelberg. Konshin, O.V., 1992. Transfer of 137Cs from soil to grass—analysis Smolders, E., Merckx, R., 1993. Some principles behind the selection of possible sources of uncertainty. Health Physics 63 (3), 307- of crops to minimise radionuclide uptake from soil. The Science of 315. the Total Environment 137, 135-146. Meisel, S., Gerzabek, M.H., Muller, H.K., 1991. Influence of plowing Valcke, E., Vandecasteele, C.M., Delvaux, B., 1996. The use of on the depth distribution of various radionuclides in the soil. Zeits- mineral and organic adsorbents as countermeasures in con­ chrift f. Pflanzenernahrung und Bodenkunde 154, 211-215. taminated soils: a soil chemical approach. Mitteilungen der Oster- Mengel, K., 1984. Emahrung und Stoffwechsel der Pflanze. Gustav reichischen Bodenkundlichen Gesellschaft 53, 85-92. Fischer Verlag, Stuttgart. Vasconcellos, L.M.H., Amaral, E.C.S., Vianna, M.E., Penna Franca, Molzahn, D., Assmann-Wertmiiller, U., 1993. Caesium radioactivity E., 1987. Uptake of 226 Ra and 2,0Pb by food crops cultivated in a in serveral selected species of honey. The Science of the Total region of high natural radioactivity in Brazil. Journal of Environ­ Environment 130/131, 95-108. mental Radioactivity 5, 287-302. Als Manuskript vervielfaltigt Fur diesen Bericht behalten wir uns alle Rechte vor. OEFZS-Berichte ISSN 0253-5270 Herausgeber, Verleger, Redaktion, Hersteller: Osterreichisches Forschungszentrum Seibersdorf Ges.m.b.H. A-2444 Seibersdorf, Austria Telefon 02254-780-0, Fax 02254-74060 Email [email protected] Server http://www.arcs.ac.at/ SEIBERSDORF