In this paper we present an atmospheric dispersion scenario for a proposed nuclear power plant in Pakistan involving the hypothetical accidental release of radionuclides. For this, a concept involving a Lagrangian stochastic particle model (LSPM) coupled with an Eulerian regional atmospheric modelling system (RAMS) is used. The atmospheric turbulent dispersion of radionuclides (represented by non-buoyant particles/neutral traces) in the LSPM is modelled by applying non-homogeneous turbulence conditions. The mean wind velocities governed by the topography of the region and the surface fluxes of momentum and heat are calculated by the RAMS code. A moving least squares (MLS) technique is introduced to calculate the concentration of radionuclides at ground level. The numerically calculated vertical profiles of wind velocity and temperature are compared with observed data. The results obtained demonstrate that in regions of complex terrain it is not sufficient to model the atmospheric dispersion of particles using a straight-line Gaussian plume model, and that by utilising a Lagrangian stochastic particle model and regional atmospheric modelling system a much more realistic estimation of the dispersion in such a hypothetical scenario was ascertained. The particle dispersion results for a 12 h ground release show that a triangular area of about 400 km(2) situated in the north-west quadrant of release is under radiological threat. The particle distribution shows that the use of a Gaussian plume model (GPM) in such situations will yield quite misleading results.

No abstract

The dispersion and concentration of particles (fluid elements) that are continuously released into a neutral planetary boundary layer is presented. The velocity fluctuations of the particles are generated using a Markov chain–Monte Carlo (MCMC) process at random time intervals with a one-step memory. The local mean concentration of the particles is calculated by using a fully Lagrangian method, which performs an efficient near-neighbor search and employs a smoothing kernel for eliminating the statistical noise. The predicted vertical and transversal root-mean-square of the particles’ deviation from their mean position [()1/2 and ()1/2] for an elevated continuous release source in a neutral atmosphere are compared with empirical parameters like the Pasquill–Gifford σz and σy. The numerical predictions of the particle concentration are compared with a Gaussian model and field measurement data on the ground concentration obtained during the Green Glow Program. The comparison between the numerical predictions and the field data shows that the MCMC model can successfully predict the particle dispersion and concentration in a neutral atmosphere.

No abstract