IRSN, Institut de radioprotection et de sûreté nucléaire

Search our site :


Contact us :

En Fr

Enhancing Nuclear Safety


Research program

DISVER program

Last update in July 2011



Extending from 2008 to 2012, the main objective of the DISVER program (vertical dispersion) is to validate, by means of sea campaigns, MARS model computations. Jointly developed by IFREMER and IRSN, this three-dimensional hydrodynamic model simulates the dispersion of pollutants soluble in sea water.


The Disver project (vertical dispersion) has been conducted since 2008 by the radioecology laboratory (LRC) of IRSN in collaboration with IFREMER. Its aim is to validate the latest developments of a three-dimensional hydrodynamic dispersion model of a plume of pollutant. Initially developed by Ifremer and adapted to the needs of IRSN, the MARS 3D model enables the dispersion of liquid discharges during the minutes and hours that follow the discharge to be simulated, from the sea bottom to the surface, as a function of currents, tides and winds.

The dispersion of pollutants in the Channel may be simulated using 2D(1) models, by vertically integrating the velocities and the concentrations of the pollutant. Such use of 2D models is only valid if the plume is located at more than one kilometre from the release point and more than one hour after the discharge. In addition, it is only possible under certain conditions: density of the discharge close to that of sea water, vertically homogeneous concentrations of discharges and no thermal or haline stratification of the mass of water.

The MARS 3D model overcomes these constraints and represents the water column by discretising it into horizontal sections characterised by sigma levels (constant fraction of the total layer). This vertical discretisation makes it possible to represent vertical gradients of velocity and concentration of the pollutant in the water, which result from parameters such as the rugosity coefficient, the diffusion coefficient and the turbulence of the tracer.

In order to validate this model, sampling campaigns are organised, which make it possible to carry out concentration measurements of the pollutant in layers of predefined thickness. These in situ concentration measurements include the influence of all parameters and enable their adjustment to be verified.

The first experimental results show moreover a large spatial and temporal variability of the concentration gradients of radionuclides. They imply that the precision of the model needs to be improved to an elementary computing mesh of less than 30 metres. Consequently, it is necessary to represent turbulent diffusion at this scale, a fundamental challenge that needs to be met in order to correctly represent the evolution of the plume. Turbulent diffusion is in fact responsible for the rapid spreading out of the plume (more than 100 metres wide, 300 metres from the emission point), the formation of several distinct lobes and the temporal variability of the structures observed (concentrations varying by a factor of one hundred within ten minutes).


Experimental campaigns


The validation of these parameters is based on the exploitation of severalEn savoir plus sur les analyses de la campagne DisVer10 sea water sampling campaigns conducted in the open sea, off the La Hague reprocessing plant in France. Areva-NC is a partner in this program. The LRC uses the tritium(2)present in liquid discharges as dispersion tracer.

Three sampling campaigns have been planned. After the 2009 campaign, which enabled the methodology to be validated, the Disver10 campaign, conducted in October 2010, had the aim of studying the dispersion of tritium near to (between 100 m and 2 km) the emission point, when the plume is not yet vertically homogeneous. The Disver11 campaign, conducted in April 2011, enabled the study zone to be extended up to 10 km, in a region where depths and bathymetric gradients are important (from 30 to 100 metres depth). The 2010 and 2011 campaigns enabled 5,600 and 13,400 samples over 10 vertical levels to be collected, from the surface to a depth of 30 metres, using a sampling system developed by the LRC. The information collected at these distances from the outfall enables the dispersion parameters to be refined so as to better predict the dispersion of discharges, whether chronic or accidental, in seas with strong tides, as is the case in the geographic zone of La Hague.

Find out more about the 2010 and 2011 campaigns


A new system for sampling at depth, in a class of its own


A dynamic system for taking samples at depth has been developed by the LRC team for continuously sampling at several depths simultaneously while sailing at a normal speed of 5 to 10 knots (i.e. 5.1 m/s). The sampling line is maintained immersed down to a depth of 50 metres by means of a 70 kg dynamic ballast acting as a 1500 kg weight: the Dynalest. It is particularly adapted to vertical sampling at depth near to coastlines, in zones with strong currents such as in the sea off La Hague. It has been tested under actual conditions during four campaigns in 2008, 2009, 2010 and 2011. During the first two campaigns, it proved to be sufficiently reliable so as not to require surveillance during sampling.

The Dynalest (©IRSN)

Le Dynalest

A sampling line is provided with ten pipes immersed at a fixed depth, each pipe being associated with a specific level: every 2 metres for sampling close to the emission point and every 5 metres in the deepest zones. Depth sensors continuously record the actual immersion depth.

15 ml samples are taken simultaneously by several pumps and collected by an automated unit, which records the flow rates and sampling times. The automaton was developed specifically for this study jointly with the Ecole d'Ingénieurs de Cherbourg. Its versatility has made it possible to use it to perform high frequency monitoring in sea water or in atmospheric water vapour.

During campaigns at sea, more than 1200 samples are collected every hour, i.e. one every 30 seconds over 10 levels. Conditioning for measurement by liquid scintillation is carried out on-board, with the adjustment of volumes and the addition of scintillating liquid.

Once the measurements have been acquired, the collected data needs to be processed as a function of the output of the pumps, the speed of the vessel and the depths recorded so as to correspond to the sampling times acquired during the campaigns, the times and positions in longitude, latitude and depth where the sea water has entered the different pipes. This processing is indispensable, because the transit time of the water in the pipes can exceed 5 minutes, which is comparable to the duration of the crossing of the plume during a transect (see DisVer10 news).


The sampling table (© IRSN)

(1) : The area of application of 2D models on the continental platform of north west Europe extends between two and one thousand kilometres from the discharge point, for simulation times extending from one hour to several years. Such models have made it possible in particular to reconstitute the dispersion of radionuclide discharges in the Channel and at the scale of the continental platform since 1985. They have been validated by around 16,000 measurements of soluble radiotracers (3H, 125Sb, 99Tc) carried out on samples with a volume of 120 litres during 60 oceanographic campaigns conducted by the LRC between 1988 and 2006. 
(2): Tritium was chosen as tracer on account of its perfect conservativity in solution in sea water in the form HTO and its simplicity of sampling and measurement by liquid scintillation.


Send to a friend

The information you provide in this page are single use only and will not be saved.
* Required fields

Recipient's email:*  

Sign with your name:* 

Type your email address:*   

Add a message :

Do you want to receive a copy of this email?