Dispersion of Radioactive Buoyant Plume in the Convective Boundary Layer Michael Momeni Hoshang: Momeni & Associates Radiological Monitoring and Assessment Associates 37 Cheshire Court Chatham, Illinois 62629 mailto:[email protected] INTRODUCTION Estimation of radiological impacts from an explosive release or from a nuclear reactor accident is a coupled temporal and spatial estimation of the quantity of the releases, dispersive forces, depletion processes, and exposure pathways. An accidental release from a nuclear reactor could have thermal and kinetic energies resulting in buoyancy. Additionally, nocturnal releases could penetration into the atmospheric inversion cap. Large-scale convective updrafts and downdrafts might lead to looping of the plume, resulting in high ground level concentrations of radioactivity at distances far from the release point. This summary report presents the structure of procedures for estimation of airborne and ground-deposited radionuclides from a buoyant radioactive plume using a variable-trajectory puff superposition model [1]. DISTRIBUTION The joint probability density function (P) for distribution of a buoyant plume in a convective boundary layer often has been defined using independent distributions in the horizontal Py (r) and vertical (zenith) Pz(r) directions for a plume travel distance r(x,y,z). The ensemble average concentration C(r,i) for radionuclide (i) is:

standard deviations of the turbulence in the y and z-directions, s(r) is the radial distance from the center of the puff. The summation index (n) is truncated to –4 to + 4 range. The effective release height (h e) is he = hs + hr The parameters, h s and h r ,are the release height and the plume rise.

CALCULATION The concentrations are calculated from Q(r,i) using representative meteorological parameters. Previously [2] in this conference, a Database Grid System, 40 by 40 grid, each cell 10 km in length, and Dosimetric Grid, were defined. The convective boundary layer and the upper level are each subdivided up to five layers. The default, a three-layer atmospheric structure, assumes surface, mixed, and non-turbulent layers. The surface layer has a shallow (10 m) thickness; it rapidly adjusts to the condition of the ground surface. The mixed layer has a uniform distribution of radionuclides within the puff and it extends to the base of the boundary layer. Due to the non-turbulent structure of the upper layer, the material would not mix by turbulence.

C(r,i) = Q(r)(2π σ2 y (r)) – 1 Py (r) Pz(r) Py (r) = exp[ - s ( r)y (2 σ y (r)) ] 2

2

-1

Pz(r) = 2(2π σ2 z(r)) – 0. 5 ∑n exp [(h e + 2n zm)2 (2σ2 z(r))-1 ] for an elevation (z) satisfying 0 ¼ z ¼ zm ,where zm is the thickness of the convective boundary layer height. Q(r,i) is the activity of the radionuclide (i) in the puff after reduction for depletion processes (wet and dry deposition) and radioactive decay during transit, σy and σz are the

For each period, at each grid point, and in each layer, representative values for each meteorological parameter (wind speed, wind direction, standard deviation of the atmospheric turbulence in the horizontal and vertical directions, the ambient temperature, and ground roughness) and precipitation are defined. The representative input values are extracted from data downloaded from 11 surface stations and 4 upper atmosphere meteorological stations. Thus, a puff within each grid’s boundaries would be subjected to the representative values

ProcDaD

ProcMet

Measured Data ProcData

Calculation

Optimization

Output Data

Data Processing

GIS

Figure 1. Parallel structure of ProcDaD

for the physical surface structures and meteorological parameters. In order to minimize discontinuity at the grid boundaries, the average value of each parameter is estimated from the values in the adjacent grid points. Puffs are advected during each time step ∆t from location r (t) to r (t + ∆t) according to Lagrangian trajectory function. The distance for the puff travel is calculated from the wind speed U(r) for the transport-advection period of ∆t (960 seconds) assuming a homogeneous steady-state condition during each advection period. The total concentration for each radionuclide is calculated from superposition of contributions from each of the puffs. The parallel structure of the computation ProcDaD [1] allows dynamic collection of all preprocessing data (ProcMet, Measured Data) and creation of the grid database prior to calculation (Figure 1). A ProcDaD query of the databases for time-dependent inputs allows a

dynamic exchange of information. The results of the computation are stored into an output database for further data processing, display (GIS) and hardcopy production. DISCUSSION Q(r,i) often is not known. Thus, Q(r,i) has to be calculated using an optimization procedure [3] for the ratio C(r,i) / Q(r,i). The above model may be conservative for z ¼ zm, but, it could under estimate airborne concentrations for releases above the boundary layer cap. Pz(r) accounts for zero flux both at the ground and at zm. A partial plume penetration at zm may reduce Q(r,i); but for releases above zm , it would contribute to the concentrations when z ¼ zm.

ACKNOWLEDGEMENT The author is grateful to Sheryl Roethlinger, Phillip Wilson, and Thomas Bellinger from Illinois Department of Nuclear Safety for the discussions and suggestions during development of this on going project

Priorities, Proc American Nuclear Society Conference, Spokane, Washington, September 17 (2000). 2.

M.H. MOMENI, ” Assessment of Regional Radiological Transport and Impacts,” Proc American Nuclear Society: International ANS/ENS 2003, New Orleans, Louisiana, November10 (003).

3.

M.H. MOMENI, S. L. SODERDAHL, K.A. FOSTER, “Optimizing Dose Assessment Model Predictions Using Key Radionuclide Indicators,” Proc American Nuclear Society Conference, TRANSO 17-1-560, 77, p. 316 (1997).

REFERENCES 1.

M.H. MOMENI, “ProcDaD, a Tool for Assessment of Doses and Induced Health Effects: Analysis of Radiological Impacts Following a Nuclear Reactor Accident,” Radiation Protection for Our National