Modelling the dispersion of radionuclides in rivers (PhD)

Since more than half a century, nuclear wastewater is discharged into Belgian rivers, resulting in an accumulation of radionuclides in the riverbeds of the rivers and reaches downstream. Due to magnification, measurable radioactive levels in sediments and aquatic biota (benthos, aquatic plants and fish) may be obtained, posing a potential hazard for the environment and humans.

Because sediment transport and sediment-radionuclide interactions are not well understood, it is common practice to use simple, purely empirical models to calculate the sediment-based radionuclide transport. While such models are valuable as screening tool (e.g. for routine releases), they are usually not appropriate to be used in case of accidental releases or as a diagnostic and pro-active environmental protection tool in case of site restoration.

As case study, the Molse Nete, a river of the Rupel basin receiving liquid discharges from nuclear activities, is selected. This project can be considered as  a first step in the development of a aquatic model for the whole Scheldt basin of which the Rupel basin is a subbasin. 

The main objective of the proposal is to adapt and improve existing river models to forecast the time- and space-dependent radioactivity level in the water column and river bed sediment by taking into account the removal processes and remobilization processes.

In this context, it is important to evaluate the influence of environmental factors on the distribution coefficient (Kd). The Kd is the most important parameter describing the radionuclide sediment interactions. Classical hydrological models use the Kd to model the radionuclide-sediment interaction process, without considering the underlying environmental factors. This generally results in predictions with large uncertainty ranges. The innovative aspect of this project is that we want to reduce the uncertainty on the sediment-radionuclide interaction process by increasing the understanding of the environmental factors which are important in  the radionuclide-sediment interaction process and hence to obtain a more mechanistic derived Kd value instead of a purely empirical derived Kd.

·         Experimental work

As mentioned, the distribution coefficient (Kd) is the most important parameter describing the radionuclide sediment interactions. The Kd is influenced by many factors (e.g. pH, sorption capacity, sediment composition, ionic composition of the water column). To examine the effect of these environmental factors on the adsorption/desorption kinetics of the radionuclides on river sediments, each factor will be varied within natural occuring ranges. For the sediment composition, this means that samples will be taken not only in the Molse Nete but also on other locations of the Rupel Basin.

Laboratory experiments will be carried out to assess the adsorption/desorption kinetics under varying conditions for  3 radionuclides ( ¹³⁷Cs, ⁶⁰Co (which can also be used as tracer for stable Co) and ¹³³Ba (as tracer of stable Ba and as analogue for  ²²⁶Ra)) on selected sediment samples of the Rupel basin.

A series of measurement campaigns will be organised aiming at the determination of river characteristics and physico-chemical characteristics of the Molse Nete. Measurements will take place at minimum 3 locations of the Molse Nete.

These laboratory and field experiments will be complemented with historical measurents.

The current river water quality models for transport of radionuclides work with distribution coefficients (Kd) to take the interactions of radionuclides with particulate matter into account. This simple approach has the advantage that only one, in situ parameter, has to be determined, but assumes an immediate and irreversible adsorption under equilibrium conditions (which is not always the case) and does not take the underlying processes into account. In the framework of the project, a theoretical structure will be developed in which the uncertainties in the model resulting from the simplified approach will be decreased by a better characterisation of the radionuclide- sediment interaction, taking into account the influence of variables such as pH, ionic strength, competitive elements in sediments and water, mineralogy, kinetic adsorption and desorption effects.

·         Numerical modelling

The numerical modelling includes a hydrologic model and the development of a river model for the Molse Nete. 

In the framework of previous studies, the VUB has developed a hydrologic model for the subbasins of the Nete. This model will be applied on the Molse Nete.

The river model consists of a hydraulic component, a sediment component and a water quality component. Concerning this model, the long term objectives - the modelling of the whole basin of the Scheldt - will be taken into account. While the modelling of the Molse Nete can be performed by a one dimensional approach, a 2- or 3-dimensional model is required for the estuarine part of the Scheldt river.

While within this project, new theoretical models for radionuclides will be developed and need to be integrated in a software package, the development of a completely new river model is not realistic and we opt for the use of open source software. The selected software packages are EFDC and WASP. EFDC includes a complete hydrodynamic module in 1D, 2D or 3D and a sediment transport module that accounts for deposition and resuspension of cohesive and non-cohesive sediment. It also includes modules for bedload transport, for river bank erosion and a sediment bed module. EFDC provides information about the flows and the sediments to the water quality software WASP. Besides modules for eutrofication, heath and salinity, WASP also contains modules for toxic substances (including metals, which shown similar behaviour as radionuclides in rivers.

In the framework of the project, WASP will be adapted in order to enable it to model radionuclides and an EFDC-WASP model will be developed for the Molse Nete.