Computational sensitivity studies
The effective doses received during 1 year for three age groups and four population groups were calculated using PREDO for 63Ni (Figure 1a), 106Ru (Figure 1b) and 125Sb (Figure 1c). The definitions of the age categories and population groups are given in [11]. As previously stated, the values are calculated assuming a hypothetical continuous release of 1 Bq/year for 100 years and are thus only intended for intercomparison. For all studied population groups and elements, the calculated dose was higher for lower age, which is expected due to the age dependence of the used dose coefficients [1] caused by differences in biokinetics between different age groups [18]. In general, the fishing family received the highest dose, for most studied age categories and elements, with some exceptions for nickel. For ruthenium, the calculated dose is significantly lower for the vegetarian family compared with the other families, and the dose for the other families is also almost directly proportional to the ruthenium CF, indicating that the dose from radioisotopes of ruthenium arises almost exclusively by ingestion. For antimony, the difference between the vegetarian and other families is less pronounced, whereas for nickel, the CF is less important for the calculated dose, indicating more exposure through other paths than the marine food chain. At the currently implemented value of the nickel CF, 1000 L/kg, the dose received by the farmer child is almost 3 times higher than that of the vegetarian child, so the ingestion exposure pathway is still dominant, indicating that for nickel, as well as for ruthenium and antimony, the CF is essential for calculation of radiation dose to all nonvegetarian population groups.
Element concentrations in seawater samples
Results of the ICPMS analysis of the concentrations of nickel, ruthenium and antimony in the seawater samples are listed in Table 2. The measured nickel concentration is almost an order of magnitude larger than concentrations previously found in Kattegat and the Baltic Sea (5 – 14 nM) [19], which could be due to contamination during sampling or storage [20]. Data on ruthenium and antimony concentrations in Kattegat or the Baltic Sea could not be found. The ruthenium concentration in heavily trafficked waters in the Mediterranean Sea has been measured to be 0.5 nM [21] which is much lower than our measured values. However, ruthenium pollution occurs to an important extent through wear of ruthenium containing alloys which are frequently used in highwear applications [21], resulting in small rutheniumcontaining particles which may have been included in our analysis. The antimony concentration in the Baltic Sea has been measured to be 0.3 – 0.8 nM [22] and 1.7 nM in the North Atlantic [23] which is similar to our measured data. The added amount of nickel in the experiment is small compared with the measured baseline concentration, whereas for ruthenium and antimony, the added element concentrations are about an order of magnitude larger than the measured concentrations.
Table 2: Concentrations of Ni, Ru and Sb. First the added radionuclide and stable carrier concentrations (Note the different prefixes for radionuclide and stable carrier concentrations), then the measured concentrations in the seawater samples, measured using ICPMS. Given uncertainties for the measurements are one standard deviation.
Element

Nickel

Ruthenium

Antimony

Radionuclide concentration [kBq/L]

36

1000

5.4

Radionuclide concentration [pM]

280

210

0.42

Added stable carrier concentration [nM]

2

160

53

Measured element concentration at Anholt E [nM]

84±5

43±2

3.3±0.8

Measured element concentration at Karlsödjupet [nM]

48±7

13±1

<0.5*

Total element concentration (Anholt samples) [nM]

86

200

56

Total element concentration (Karlsödjupet samples) [nM]

50

170

53

* The concentration of antimony at the Karlsödjupet sampling station was below the detection limit of 0.5 nM.

Concentration factors
A CF was calculated individually for each sample according to equation 1, where a phytoplankton dry weight of 23 pg was established and used in the calculation. For each element and water origin, a mean value and standard deviation was calculated from the three CF thus obtained. These values are listed in Table 3. For nickel, the mean activity measured on the control filters was less than 1% of that caught on filters with phytoplankton. For ruthenium and antimony, the corresponding number was 25% and 21% respectively.
Table 3: Concentration factors (CF) in P. Tricornutum for nickel, ruthenium and antimony. CFs and uncertainties (one standard deviation) are given in terms of dry weight (dw). Relative standard deviations (RSD) are given in %. CFs are also given in L/kg fresh weight (fw) for comparison with literature values below, where the conversion factor of 0.18 dry weight to fresh weight recommended by IAEA [5] has been used.
Element

Water origin

CF (dw) [L/kg]

CF (fw) [L/kg]

RSD

Nickel

Anholt E

6 400 ±

1 900

1 200 ±

300

30%


Karlsödjupet

6 100 ±

800

1 100 ±

140

13%

Ruthenium

Anholt E

15 000 ±

11 000

2 700 ±

2 000

75%


Karlsödjupet

20 000 ±

8 000

3 600 ±

1 400

40%

Antimony

Anholt E

190 ±

150

34 ±

27

81%


Karlsödjupet

890

*

160*



* No standard deviation is given because one of the triplicate samples was excluded.

Nickel
The difference between the CFs calculated for the water samples from Anholt E and Karlsödjupet respectively is not statistically significant (the difference is smaller than one standard deviation). Both values (1 200 L/kg and 1 100 L/kg fresh weight, respectively) are also, within one standard deviation, similar to the geometric mean of the values listed by IAEA TECDOC 479 [5] (570 ± 740 L/kg fresh weight).
Ruthenium
The calculated CFs for ruthenium are also, within one standard deviation, identical for the two water sampling sites. The values, 2 700 L/kg and 3 600 l/kg fresh weight, for Anholt E and Karlsödjupet respectively, are both within one standard deviation from the mean value given by IAEA TECDOC 479 [5] (6 700 ± 8 500 L/kg fresh weight). It should be mentioned that the very high value recommended in IAEA TECDOC 422 [4], which is currently implemented in PREDO, refers to Lowman et al. [9] who in turn refer to Slowey et al. [24] who explicitly mention that their data, established for Caribbean ecosystems, are preliminary.
Antimony
Finally, the CF values for antimony calculated from our measurement data are quite widely spread. For one of the samples with water from Karlsödjupet, the measured activity in the phytoplankton fraction was an order of magnitude larger than for the other two samples, although the phytoplankton concentration was similar. We assume that this was due to some foreign particle(s) entering the culture, onto which much nickel was adsorbed. This sample was therefore excluded from the analysis. The mean value for Anholt E (34 L/kg fresh weight) has an 81% RSD and the mean value for Karlsödjupet (160 L/kg fresh weight) is more than two standard deviations larger, however, given that no standard deviation could be calculated for the Karlsödjupet value, it is not possible to determine if the difference is statistically significant. No value is given in IAEA TECDOC 479 [5], while the recommended value in IAEA TECDOC 422 [4] (1000 L/kg) is the same as in IAEA TECDOC 211 [3], for which no original source is stated. Furthermore, no information is given on how the value was obtained or whether it refers to phytoplankton fresh or dry weight. Thus, our values probably represent the first published CF measurement data for antimony in phytoplankton.