Probabilistic Prognosis of Environmental Radioactivity Concentrations due to Radioisotopes Discharged to Water Bodies from Nuclear Power Plants.

Due to their very low values, the complexity of comparing the contribution of nuclear power plants (NPPs) to environmental radioactivity with modeled values is recognized. In order to compare probabilistic prognosis of radioactivity concentrations with environmental measurement values, an exercise was performed using public data of radioactive routine discharges from three representative Spanish nuclear power plants. Specifically, data on liquid discharges from three Spanish NPPs: Almaraz, Vandellós II, and Ascó to three different aquatic bodies (river, lake, and coast) were used. Results modelled using generic conservative models together with Monte Carlo techniques used for uncertainties propagation were compared with values of radioactivity concentrations in the environment measured in the surroundings of these NPPs. Probability distribution functions were inferred for the source term, used as an input to the model to estimate the radioactivity concentrations in the environment due to discharges to the water bodies. Radioactivity concentrations measured in bottom sediments were used in the exercise due to their accumulation properties. Of all the radioisotopes measured in the environmental monitoring programs around the NPPs, only Cs-137, Sr-90, and Co-60 had positive values greater than their respective detection limits. Of those, Sr-90 and Cs-137 are easily measured in the environment, but significant contribution from the radioactive fall-out due to nuclear explosions in the atmosphere exists, and therefore their values cannot be attributed to the NPPs. On the contrary, Co-60 is especially useful as an indicator of the radioactive discharges from NPPs because its presence in the environment can solely be attributed to the impact of the closer nuclear facilities. All the modelled values for Co-60 showed a reasonable correspondence with measured environmental data in all cases, being conservative in two of them. The more conservative predictions obtained with the models were the activity concentrations in the sediments of a lake (Almaraz) where, on average, values two times higher were obtained. For the case of rivers (Ascó), calculated results were adequately conservative-up to 3.4 times on average. However, the results for coasts (Vandellos II) were in the same range as the environmental measurements, obtaining predictions that are only-at maximum-1.1 times higher than measured values. Only for this specific case of coasts could it be established that the models are not conservative enough, although the results, on average, are relatively close to the real values.