NRPB-W16: Non-technical summary

Non-technical summary

In the United Kingdom, the Food Standards Agency (FSA) has statutory responsibility for protection of the food supply from radioactive contamination. One potential source of such contamination is the routine discharge of radioactivity to air from licensed sites. FSA carries out routine measurements of activity concentrations in foods around such sites, but also relies on the use of mathematical models to estimate the likely impact of airborne discharges on the foodchain. These models are continually reviewed, and the present study forms part of this process. Air flows around buildings on a site have the potential to influence the local air concentrations and resulting deposition (building wake effects). A simple Gaussian plume model has been used in many regulatory applications, but it does not make any allowance for the effects of buildings near the release point on the air flow and consequent dispersion. Simple modifications of this model to allow for some building wake effects have also been used in earlier studies and regulatory applications. A more complex model has recently been developed; this describes the air flow around buildings and hence the effect on plume dispersion in more detail than earlier extensions to the simple Gaussian model.

The purpose of this study is to explore whether there are circumstances in which it is necessary to model the influence of building wakes explicitly in order to obtain reliable estimates of activity concentrations in food from continuous releases, and how to model dispersion from sites with several buildings. The study was intended to provide advice to FSA on the extent to which its current models are adequate, to indicate whether there are circumstances for which the explicit modelling of building wake effects is required, and, if so, to recommend an appropriate model for this.

In order to carry out this study, three atmospheric dispersion models were selected, namely:

• a simple Gaussian plume model with no allowance for buildings (the 'R91' model).
• an extension to the Gaussian plume model describing buildings (the 'R157' model).
• a more complex model that explicitly characterises air flows around buildings (the ADMS model).

Results from the models were compared, for a range of on-site building configurations and release locations. In addition, the extent to which details of the buildings on the site are required in the complex model was considered; results are presented for several different descriptions of the building complex and conclusions are drawn. In particular, results are presented for the models which show the sensitivity of the predicted air concentrations and depositions to:

• the ways in which particular features of a nuclear site are described in the models.
• the extent to which small buildings should be included in the calculations.
• whether the models are sensitive to the precise location of the release point relative to buildings.

The main findings of this project can be summarised as follows. The predictions of R91 and ADMS for air concentration within about 10 km of the release if buildings are not considered are generally within a factor of 2 to 3. This is similar to the difference which a group of UK dispersion modelling experts felt could be expected between observed concentrations and predictions of Gaussian plume models. Building effects only modify predicted concentrations by more than a factor of 2 at distances within less than about 1 km of the release point. For the Amersham site, where the buildings are small (less than 10 m tall) the effects are confined to within one or two hundred metres of the release point. At Dungeness, the effects extend to larger distances, but they are still confined to distances less than 1 km from the stack. It is reasonable to conclude that building effects in general are only important at distances of a few hundred metres from a site, and so buildings need not be considered in assessing doses at larger distances. Within this distance range, some predictions are sensitive to particular features of the model, and results of any calculations in this region should be treated with caution. ADMS is considered to be a more appropriate model than a simple Gaussian model if building effects must be considered as it treats a number of effects that cannot easily be considered in simpler models.

Last reviewed: 1 September 2009