| dc.description.abstract | Anthropogenic radionuclides are introduced to the global marine environment from a range of different sources, such as global fallout from atmospheric weapon tests and nuclear accidents as well as operational or accidental discharges from nuclear installations. Transportation of radionuclides in the marine environment depends on the physico-chemical properties and different oceanic processes. Long-lived radionuclides like 129I and 236U are commonly utilized as tracers of ocean currents due to conservative behavior and the relatively well-documented discharges from nuclear installations. Discharges of 129I and 236U from Sellafield and La Hague reprocessing plants can be used as model input in ocean models to predict ocean transport.
These oceanographic models are used to study ocean processes, including climate changes such as ocean acidification and ice melting, and predict potential consequences of various contaminant releases to the marine environment. Regarding nuclear events, these models can be valuable pre-accident to simulate different scenarios to predict the transport and fate of radionuclides, while also contributing to emergency decision-making. Models can also be utilized post-accident to assess consequences and future measures. However, an oceanographic model needs calibration and validation before being used as an emergency preparedness tool.
In the present work, a hydrodynamic ocean model in combination with an ocean transport model was utilized to simulate the ocean transport of Lagrangian particles representing 129I and 236U discharges released from Sellafield and La Hague.The ocean transport model is an open-source trajectory model called OpenDrift, developed by the Norwegian Meteorological Institute (MET). The model simulations aimed to improve the source term by investigating the source contribution of 129I and 236U originating from Sellafield and La Hague and to validate the transport model by comparing model estimations to observations.
To validate OpenDrift, an extensive literature search of existing 129I and 236U observations within the model domain was carried out. The search resulted in observations in the North Sea, Irish Sea and Barents Sea which could be compared to model estimations. The literature search revealed a lack of 129I and 236U observations in the Norwegian Sea. Some of the identified data gaps were filled by providing seawater samples from various sampling stations in the Norwegian Sea through collaboration with the Institute of Marine Research (IMR). These samples were then analyzed using accelerator mass spectrometry (AMS) through collaboration with the Czech Technical University (CTU) and the University of Seville (USEV).
The preparation of the seawater samples was conducted at the Norwegian University of Life Sciences (NMBU) in Ås and at the Czech Technical University (CTU) in Prague. The preparations consisted of radiochemical separations and aimed to reduce the sample volume and obtain a high purity and yield of the analyte for measurements with mass spectrometric techniques. The samples were prepared into a fine homogenous powder and shipped to Centro Nacional de Aceleradores (CNA) in Seville for accelerator mass spectrometer (AMS) measurements. Since the AMS measurements yielded 129I/236U ratios,
127I was analyzed using inductively coupled plasma spectrometry (ICP-MS) at NMBU, enabling the conversion of the AMS results to 129I concentrations (atoms/l). The measured concentrations of 129I and existing literature data of 129I and 236U from different locations and sampling dates were compared to model outputs to validate OpenDrift.
Model simulations and observations provided 129I and 236U concentrations as well as 129I/236U atom ratios at nine different locations, including the Irish Sea, English Channel, North Sea (south, east and north), Norwegian Sea, Barents Sea, Fram Strait and Komsomolets. The 129I concentrations (atoms/l) measured with AMS from samples in the Norwegian Sea ranged from 0.63 to 35.0 x 10^9 atoms/l. The 129I/236U ratios tended to be more comparable to the model outputs further away from the discharge sources when water masses are better mixed and local currents play a lesser role in the transport. The model estimations aimed to improve the source terms of 129I and 236U by comparing literature and unpublished data with observations, and separate source contributions from Sellafield and La Hague.
General trend observations indicated a higher 129I contribution from La Hague and a higher 236U contribution from Sellafield in most model locations, reflecting the reprocessing discharges. Based on comparisons of model outputs with field observation data, an alleged substantial retention of 236U in the Irish Sea could not be confirmed. This interpretation is based on 1) the model underestimated 236U concentrations and 2) the estimated 129I/236U ratios were relatively comparable to corresponding observations, varying within a factor ranging between 1 and 38.Whereas the opposite outcomes would be expected in case of substantial retention.
Analysis of Lagrangian model particle ages revealed transportation time differences between particles released from Sellafield and La Hague. Between 2003 and 2023, the average particle ages from Sellafield (6.98 and 7.53 years) were older than for La Hague (3.54 and 4.12 years) in the Norwegian and Barents Seas. In the Irish Sea, however, the model particles released from La Hague were over three times older, implying longer transportation time. In most parts of the North Sea, particles from
Sellafield were older, meaning La Hague particles are transported faster due to the strong coastal currents along the northwestern European coast into the Barents Sea and the Arctic Ocean.
The ocean model simulations provided valuable information regarding source contributions, transportation pathways and model particle ages. The validation of the ocean model provided relatively comparable ratios, the most favorable validations of 129I/236U ratios being further away from the discharge sources. More refinements should be considered in the simulations to achieve more accurate predictions from the ocean model. The speciation of radionuclides, the background signal from global fallout and nuclear reprocessing discharges pre-1990 should be implemented in the model for future research. | |
