A simulation study of dispersion of air borne radionuclides from a nuclear power plant under a...

ARTICLE IN PRESS

1352-2310/$ - se

doi:10.1016/j.at

�CorrespondE-mail addr

Atmospheric Environment 39 (2005) 1497–1511

www.elsevier.com/locate/atmosenv

A simulation study of dispersion of air borne radionuclidesfrom a nuclear power plant under a hypothetical accidental

scenario at a tropical coastal site

C.V. Srinivas�, R. Venkatesan

Radiological Impact Assessment Section, Radiological Safety Division, Safety Group, Indira Gandhi Centre for Atomic Research,

Kalpakkam, Tamil Nadu 603102, India

Received 3 July 2004; accepted 22 November 2004

Abstract

Meteorological condition in coastal regions is diurnally variable and spatially heterogeneous due to complex

topography, land–sea interface, etc. A wide range of dispersion conditions is possible on a given day in the coastal regions.

In case of inadvertent accidental situations, though unlikely, it would be necessary to examine the potentially severe case

among different dynamically occurring local atmospheric conditions for dispersion and its range of impact around a

nuclear power plant for safety analysis. In this context, dispersion of air borne radioactive effluents during a hypothetical

accidental scenario from a proposed prototype fast breeder reactor (PFBR) at an Indian coastal site, Kalpakkam, is

simulated using a 3-D meso-scale atmospheric model MM5 and a random walk particle dispersion model FLEXPART. A

simulation carried out for a typical summer day predicted the development of land–sea breeze circulation and thermal

internal boundary layer (TIBL) formation, which have been confirmed by meteorological observations. Analysis of dose

distribution shows that the maximum dose for releases from a 100m stack occurs at two places within 4km distance

during sea breeze/TIBL fumigation hours. Maximum dose also occurred during nighttime stable conditions. Results

indicate that, on the day of present study, the highest concentrations occurred during periods of TIBL fumigation rather

than during stable atmospheric conditions. Further, the area of impact (plume width at the surface) spreads up to a down

wind distance of 4 km during fumigation condition. Simulation over a range of 25km has shown turning of plume at the

incidence of sea breeze circulation and two different dispersion patterns across the sea breeze front. These results are

significant in comparison to the expected pattern shown by Gaussian plume model used for routine analysis.

r 2004 Elsevier Ltd. All rights reserved.

Keywords: Radionuclide dispersion; Meso-scale model; FLEXPART; Sea breeze; TIBL fumigation

1. Introduction

A prototype fast breeder reactor (PFBR) is being

constructed at Kalpakkam, an eastern coastal site of

India. Analysis of the environmental radiological impact

e front matter r 2004 Elsevier Ltd. All rights reserve

mosenv.2004.11.016

ing author. Fax: +914114 280235.

ess: [email protected] (C.V. Srinivas).

under design-based probable accidental releases for any

proposed nuclear plant is a regulatory requirement and

therefore, as a part of it, a study of atmospheric

dispersion of radionuclides and the consequent dose to

the public is carried out.

The meteorological condition at Kalpakkam, like any

coastal site, is non-stationary and non-homogeneous

due to thermally driven land–sea breeze circulation,

d.

www.elsevier.com/locate/atmosenv

ARTICLE IN PRESSC.V. Srinivas, R. Venkatesan / Atmospheric Environment 39 (2005) 1497–15111498

which gives rise to a variety of complex atmospheric

dispersion conditions. Strong gradients of heat flux and

temperature develop across the coast during clear sunny

days and lead to direct circulation from sea to land

during daytime and an opposite circulation from land to

sea during nighttime (Estoque, 1961). The advection of

cool humid air mass during sea breeze time changes the

properties of the hot dry air mass over land and a

thermal internal boundary layer (TIBL) forms near the

coast (Vugts and Businger, 1977). TIBL is a shallow

unstable layer and forms an upper limit for the mixing

height. The stable boundary layer over the TIBL inhibits

vertical mixing across TIBL top. While entrainment of

pollutants and reflectance by TIBL leads to fumigation,

the incidence of sea breeze leads to turning of the plume

(Lyons and Cole, 1976; Ludwig, 1983; National

Academy of Sciences, 1992). The onset, intensity,

duration and inland extent of sea breeze and the

horizontal and vertical growth properties of TIBL

assume significance in determining the pollution disper-

sion in coastal regions.

The conventional practice is to calculate the radio-

active dose to the public at the boundary of the site

(1.5 km) and design the plant in such a way that the

doses are well within the prescribed safety limit.

Calculation for releases through air route is done

assuming the Pasquill stability category F for the worst

atmospheric condition using a Gaussian plume/puff

model. While this gives the maximum dose estimate for

ground release, the maximum dose for stack release would

be during an unstable (Passquill category A) situation,

near the site boundary, if the same wind speed were used

as in the case of F category. The question is, whether the

conception of worst situation (F for ground release and A

for stack release) for a plain inland condition is applicable

for a coastal terrain where fumigation of the plume is

expected under TIBL development.

The regulatory procedures for assessing the radiolo-

gical consequences use the standard Gaussian plume

model (GPM) for estimating the plume concentration

(Shirvaikar, 1973; IAEA, 1980; Clarke and Macdonald,

1978). Here, radioactive dose at the site boundary alone

is considered important with an assumption that it

decreases exponentially further away for ground level

releases. So, calculations are usually made for stability

category F to find out the dose at the site boundary.

However, there is a short-term release of noble gases

envisaged through stack and the maximum dose due to

it would be during an unstable (Passquill category A)

situation, near the site boundary, if the same wind speed

were used as in the case of F category. Further, the

distance of maximum concentration for stack release

could occur beyond the site boundary even in highly

mixing category A. In an earlier study, analysis of the

observed radioactive Ar41 plume dose within the site at

Kalpakkam had shown the influence of TIBL fumigation

(Venkatesan et al., 2002). Dose pattern beyond the site

needs to be studied using a suitable meso-scale model.

Although GPM provides conservative estimates for

safe design purposes, it is nevertheless required to know

the radiological impact due to complex meteorological

condition at the coastal site and suitably incorporate the

knowledge in accident analysis in order to ensure that the

design provides adequate safety to environment beyond

the site boundary. A suitable atmospheric hydrodynamic

model is therefore chosen to simulate the spatial and

temporal structure of the coastal atmospheric boundary

layer wind field and followed by a particle dispersion

model to get the ground-level concentration more

realistically. A good number of meso-scale atmospheric

models with advanced diffusion techniques have been

reported in literature for emergency response pro-

grammes (Lyons et al., 1983; Fast et al., 1995; Camps

et al., 1997; Hanna, 1982; Zannetti, 1990 among others).

The present work aims to study the dose distribution

of air borne radioactive effluents released from PFBR

proposed at Kalpakkam coastal site in a design basis

accidental (DBA) situation for the ambient atmospheric

conditions on a given day and the differences that would

arise in the estimates when a standard GPM is applied for

the same. The modelling system consists of a high-

resolution 3-D meso-scale meteorological model (MM5)

coupled with a Lagrangian particle model (FLEXPART).

The simulation is selected on 24 May 2003, a typical

summer sea breeze day, when meteorological measure-

ments were carried out at Kalpakkam using a tethered

balloon, a mini sodar and a meteorological tower.

Dispersion calculation is made over two ranges, i.e., over

a local range of 6.25 km for dose distribution near the

reactor and in a range of 25 km in the meso-scale.

2. Study area

The study area Kalpakkam is located on the east

coast of India (Fig. 1) at 12.501N and 80.101E. The

terrain elevation gently rises to the west and has a linear

coastline running in NE–SW direction. Along the coast,

the site is about 80 km away from Chennai city in the

northern side and nearly 70 km from Pondicherry on the

southern side. The land use pattern indicates a

corresponding average roughness length of 0.3m except

near plant area and a sandy-clay loam soil type.

3. Simulation of atmospheric parameters

3.1. Brief description of atmospheric model

A non-hydrostatic, primitive-equation, finite-differ-

ence meso-scale atmospheric model MM5 (Anthes and

ARTICLE IN PRESS

Fig. 1. The three nested domains used for MM5 simulations (left-hand side), the innermost domain shows the study region of

Kalpakkam (right-hand side). Height contours are at 20m interval.

C.V. Srinivas, R. Venkatesan / Atmospheric Environment 39 (2005) 1497–1511 1499

Warner, 1978) is used to provide the winds, vertical

temperature and stability structure, mixing depths and

other parameters to atmospheric transport and diffusion

model. The prognostic region can be selected with

suitable horizontal and vertical resolution and with

nested domains. The model incorporates a variety of

schemes for physical processes of radiation, convection

and atmospheric boundary layer (Grell et al., 1994). For

the present study, MM5 is designed with three nested

domains with horizontal resolutions 18, 6 and 2 km,

respectively. The outer first and second domains cover

the Indian peninsula and southeast region whereas the

inner third domain covers an area of 40,000 km2 around

Kalpakkam between 79.151E, 11.811N and 80.811E,

13.361N (Fig. 1). All domains have 100� 100 grids and

contain 26 vertical s levels of which six levels

(corresponding to the pressure levels 990, 980, 970,

950, 925, 900 hPa) are chosen near to the surface. The

initial meteorological conditions for the model are

provided from United States National Center for

Environmental Prediction (NCEP) 11� 11 resolution

final analysis (FNL) data which contains the horizontal,

vertical winds, temperature, specific humidity, cloud

cover, geopotential heights, soil moisture, soil tempera-

ture, etc. The model is initialized at 00:00 h IST on 24

May 2003 and is integrated for 24 h, the lateral

boundary conditions are updated every 6 h. The Eta

PBL scheme based on Mellor–Yamada formulae (Janjic,

1990) is used for atmospheric boundary layer. It predicts

TKE and has local vertical mixing. It uses a land surface

model (LSM) for surface temperature prediction. Before

the LSM the scheme calculates the surface layer

exchange coefficients using similarity theory and after

the LSM it calculates the vertical fluxes with an implicit

diffusion scheme. The lower boundary values for terrain

height, albedo, moisture availability, emissivity, rough-

ness coefficient, thermal inertia, thermal diffusivity of

soil, etc. are specified from data on topography,

vegetation from USGS and soil types from FAO. For

the fine nest these data are incorporated at a horizontal

resolution of 0.9 km. Atmospheric radiation and micro-

physics for atmospheric moisture are dealt with simple

schemes (Dudhia, 1989). The lateral boundary condition

is Newtonian relaxation for the outermost domain and

time dependent for the inner domains.

3.2. Simulated atmospheric structure

The simulated surface wind field from outer domain

(Fig. 2) at 6 AM indicates presence of westerly to

northwesterly winds along the west coast, westerly winds

over central parts and southerly to southwesterly over

Bay of Bengal. In order to validate the simulated

parameters in this domain, observations at model grids

corresponding to operational weather stations falling in

this region (Chennai, Karaikal, Cochin, Trivandrum,

Mangalore, Bangalore, Hyderabad, Machilipatnam,

Raipur and Nagpur ) are used. Since the data is already

assimilated and analysed by the NCEP-FNL and is used

subsequently as a boundary condition in MM5,

comparison is made before the model adjusts the

boundary conditions. Fig. 3 shows a qualitative agree-

ment of the model with radiosonde data as an example

for Chennai, which is near the study site Kalpakkam.

The model atmosphere is less stable than the observed

ARTICLE IN PRESS

Fig. 2. Simulated wind flow at the surface over Indian

peninsula at 06:00 h IST on 24 May 2003.

C.V. Srinivas, R. Venkatesan / Atmospheric Environment 39 (2005) 1497–15111500

state up to 1 km height. The wind is slightly under-

predicted below 500m height and overpredicted further

above. Model relative humidity on this cloudless day is

roughly 10% less than the observations up to 750m

height (not shown). These differences are due to

time–volume averaging in the model and are not gross

deviations from the actual state. Statistical analysis of

the simulated versus observed meteorological para-

meters at standard pressure levels (1000, 925, 850, 700

and 600 hPa) gave a correlation coefficient and standard

deviation of (0.818, 2.059) for wind speed, (0.724, 49.3)

for wind direction, (0.90, 2.846) for potential tempera-

ture and (0.654, 18.05) for relative humidity, respectively

(Srinivas et al., 2004).

In the fine grid domain, a land sea breeze circulation

extending about 50 km from the coast is seen (Fig. 4),

which is not resolvable by the outer domain. Simulated

wind and surface fluxes are compared with experimental

observations at 10m level (Fig. 5) at Kalpakkam site.

Winds are southerly between 00 and 04 h; southwesterly

between 04 and 08 h, westerly between 08 and 12 h. Sea

breeze is found as southeasterly between 13 and 20 h.

While the trends of wind velocity and heat flux are

similar, the sea breeze onset is predicted delayed by 2 h

and there is also a remarkable deviation in their

magnitudes particularly during sea breeze onset hours.

Vertical profiles of wind measured by Doppler Mini-

Sodar up to 300m height at Kalpakkam have also

confirmed this deviation at 12 h IST in the model wind

profiles (Srinivas et al., 2004). The lag in sea breeze onset

could slightly alter the plume direction and concentra-

tion. A significantly lower surface sensible heat flux is

also found in the model (Fig. 5c) during daytime. The

differences could be attributed to the surface boundary

condition specified using USGS data for soils and land

use which needs to be replaced using locally generated

data. A modification of initial soil moisture would also

be necessary for treatment of turbulent fluxes and to

improve the simulation further. The atmospheric stabi-

lity is determined from the simulated boundary layer

variables by the correspondence between Pasquill

stability type, aerodynamic roughness (zo) and inverse

Obukhov length (1/L) (Golder, 1972) (Table 1) based on

the facts that various atmospheric parameters charac-

terizing turbulence are functions of 1/L in the surface

layer and 1/L is proportional to the Richardson number.

A spectrum of dispersion conditions is found from

simulated atmosphere on the chosen day for DBA

analysis. The simulated atmosphere is neutral (D) in the

night (19–03 h IST), stable (E) in the early morning and

morning hours (04–08 h IST) and unstable (C,B,A)

during most of the day. The surface wind speed varied

3–4m s�1 in the nighttime; 2m s�1 in the morning land

breeze time, between 8 and 11 h; and about 5.5m s�1 in

the sea breeze time, between 13 and 17 h. (Table 1). Prior

to the onset of sea breeze, an unstable boundary layer

exists over the land. Its temperature structure is altered

by the advection of cool and humid air mass by the sea

breeze occurrence leading to the formation of TIBL.

TIBL is a shallow unstable layer with a neutral or stable

layer aloft. The height of TIBL is important as it acts as

a ‘lid’ for vertical mixing of pollutants and leads to

higher ground-level concentrations and fumigation in

coastal regions.

Analysis of the simulated temperature profiles across

the coast after the onset of sea breeze indicates formation

of TIBL with a horizontal extent of about 40 km by

16:00 h (Fig. 6). Its vertical extent, as determined from the

base of inversion in the temperature profile (Arritt, 1987),

varies parabolically with downwind distance. From the

tether sonde data, the TIBL height is estimated to be

100m at 5 km away from coastline (Sivaramakrishnan

and Venkatesan, 2002). The predicted mixing height

identified from TKE profile falls at 200–300m at this time

indicating the TIBL adjacent to coast.

4. Dispersion simulation

The atmospheric dispersion is simulated using a

Lagrangian particle model FLEXPART (Stohl, 1999)

that is briefly described below.

4.1. Description of the model FLEXPART

FLEXPART simulates the transport, diffusion, dry

and wet depositions and radioactive decay of air

ARTICLE IN PRESS

300 305 310 315 3200

500

1000

1500

2000

2500

3000

3500

4000

4500

5000Chennai, Meenambakkam24-05-2003 06:00 LST

Radiosonde MM5

Hei

ght (

m)

Potential temperature (K)

2 4 6 8 10 12 14 16 18 20

500

1000

1500

2000

2500

3000

3500

4000

4500

5000Chennai24-05-2003 06:00 LST

Radiosonde MM5

Hei

ght (

m)

Wind speed (m/s)

0 30 60 90 120 150 180 210 240 270 300 3300

500

1000

1500

2000

2500

3000

3500

4000

4500

5000Chennai

24-05-2003 06:00 LST MM5 Radiosonde

Hei

ght (

m)

Wind direction

(A) (B)

(C)

Fig. 3. Simulated and observed vertical profiles of potential temperature (A), wind speed (B) and wind direction (C) at 06:00 h IST on

24 May 2003 from model 1st domain and corresponding to the grid at Chennai location.


pollutants from point, line or area sources by computing

trajectories of infinitesimally small air parcels using the

time-varying 3-D meteorological fields. Its advantage is

that the numerical accuracy is only limited by the

number of particles released and by the resolution of

meteorological input data and so gives more reliable

estimates even near the source unlike Eularian models.

For the local and medium range dispersion calculation,

it uses the time-varying 3-D meteorological and

boundary layer data predicted by meso-scale atmo-

spheric models. It also adopts the surface boundary

conditions such as terrain, land use, etc., commonly

ARTICLE IN PRESS

Fig. 4. Simulated surface wind (at 1001 hPa level) indicating sea

breeze circulation in the fine mesh domain at Kalpakkam at

16:00 h IST. Contours indicate the wind speed (m s�1).

0 2 4 6 8 10 12 14 16 18 20 22 240

2

4

6

8

10

12

14

Wind speed at 10 level MM5 Sonic Anemometer

Win

d sp

eed

(m/s

)

Hour

Win

d di

rect

ion

0 2 4 6 8 10 12-50

0

50

100

150

200

250

300

350

400

450

500

Sen

sibl

e H

eat F

lux

(Wat

ts/s

q.m

)

Hou

(A) (B)

(C)

Fig. 5. Diurnal variation of (A) wind speed, and (B) direction;


available from global data bank for treatment of

pollutant deposition. The motion of each fictitious

particle is given by

xðt þ DtÞ ¼ xðtÞ þ vðx; tÞDt; (1)

where t is time, Dt the time increment, x the position and

v the wind velocity. The wind consists of a grid scale

component, the turbulent component and the meso-

scale component, i.e., v ¼ v þ vt þ vm: Depending on thetime scale and distance range of diffusion, the velocity

fluctuation terms are treated using separate schemes. In

the present case, it is important to find the dispersion

effects in a local range of 6 km and a meso-scale range of

about 25 km around the PFBR. In the local short range,

turbulent eddies occurring in the PBL corresponding to

timescales of less than 1 h and short length scales are

important. For this, the particle velocity is treated as a

sum of two terms: auto-correlation term and a term of

random contribution due to turbulence. The turbulent

motion (vt) is parameterized by Langevin equation

dvti ¼ vtidt

tLvþ

2

tLv

� �1=2

svti doj : (2)

0 2 4 6 8 10 12 14 16 18 20 22 240

30

60

90

120

150

180

210

240

270

300

330

360

Wind direction at 10 m MM5 Sonic Anemometer

Hours

14 16 18 20 22 24

Wind direction at 10 m level flux from MM5 flux from Sonic Anemometer

r

(C) sensible heat flux from simulation and observations.

ARTICLE IN PRESS

Table 1

Boundary layer parameters simulated by MM5 over land near coast around PFBR release location for different characteristic times

Time (h)

IST

Inverse Obukhov

length 1/L (m�1)

Mixing height H

(m)

Friction velocity

u* (m s�1)

Convective velocity

w* (m s�1)

Wind speed

(m s�1)

Pasquill stability

category

04:00 0.0025 250 0.33 0.00 3.9 E

06:00 0.0035 550 0.42 0.00 3.3 E

08:00 0.0019 580 0.38 0.00 2.1 E

10:00 �0.031 615 0.48 1.96 3.8 C

11:00 �0.043 1700 0.52 2.91 3.3 B/A

13:00 �0.0239 815 0.57 2.86 5.5 C

15:00 �0.020 225 0.64 1.05 5.5 C

17:00 �0.0028 250 0.55 0.81 5.6 C

19:00 �0.0004 275 0.49 0.37 5.1 D

0100200300400500600700800

1000

2000

3000

4000

5000

TIBL

PBL

LandSea1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

1600 LST

Hei

ght (

m)

No. of Grids

Fig. 6. Simulated vertical profiles of potential temperature

across the coast in x–z plane at 16 h IST. The cross-section

extends vertically up to 4 km (600mb). TIBL formation is

indicated by the parabola.


Here, the first term represents the drift, the second term

the diffusion which are functions of position, turbulent

velocity and time, svti is the standard deviation of

turbulent wind, tLv is the Lagrangian time scale for

velocity auto-correlation, doj are incremental compo-

nents with mean zero and unit standard deviation. svti,tL at any time and position of the trajectory are

determined from boundary layer parameters h, L, w*,

zo, u* (mixing height, Monin–Obukhov length, con-

vective velocity scale, roughness length and friction

velocity) (Hanna, 1982) predicted by meteorological

model. For the meso-scale range the velocity fluctua-

tions are described by an independent Langevin

equation with a much larger time step and using the

grid scale variance in the wind field. The source (mass) is

considered to be made up of a large number of pseudo-

particles and concentrations are estimated as integrated

mass of individual particles in a physical cell volume

(kernel) according to their mass fraction. A parabolic

kernel method as described in Stohl (1999) is used for

estimating receptor point concentration. The model has

been validated with a number of tracer studies (Stohl

et al., 1998).

4.2. Calculation domain and dose estimation method

Radiation dose estimates are made at different

distances from the source considering site boundary at

1.5 km. The accident is assumed to occur at 00 h on 24

May 2003. The integration of MM5 model also starts at

this point and would take some spin up time for the

adjustment of the large-scale flow according to the local

topography, land use, etc. in the third domain. Thus, the

release at the initial time steps would follow the model

interpolated flow on to its vertical levels and subse-

quently gets adjusted to the local flow dynamics. This

may result in some loss of accuracy in dispersion

calculation initially using this release time for dispersion

simulation. However, during operational runs using

dynamic initialization, the question of model spin-up

does not arise, as the meteorological model would

remain in dynamical balance with observations. Two

cases of release are considered to occur during the

accident viz., from ground and a 100m stack. Although

the ground release through leakage is proportional to

the pressure fall within the containment building, the

release rate is considered uniformly spread over for 24 h.

Due to expected delay in closure of the ventilation ducts,

a stack release of fission product noble gases (FPNG) is

envisaged for the first 60 s. The pathways of exposure

considered for the calculation are cloud gamma dose by

arbitrarily distributed radio active cloud of FPNG,

inhalation dose due to particulates and vapours and

external exposure due to ground deposited activity. In

reality, these pathways would continue to deliver dose

for the entire period of exposure lasting the complete

passage of the plume. However, a period of 24 h


simulation alone is considered with a break up for

various stability categories occurring at different times

on the given day. Emission rates of different isotopes are

considered from internal estimates of the release of

radioactivity to the environment under hypothetical

Core Disruptive Accident for PFBR (Keshavamurthy

Table 2

Radionuclides that are released through the stack (time 60 s)

Radionuclide Quantity released (Bq)

I-131 1.38 E 14

I-132 1.83 E 14

I-133 2.37 E 14

I-134 2.36 E 14

I-135 2.08 E 14

CS-134 5.26 E 09

CS-137 1.57 E 12

KR-85m 3.16 E+15

KR-87 5.71 E+15

KR-88 6.90 E+15

KR-85 6.55 E+13

XE-133 3.56 E+16

XE-135 3.72 E+16

Table 3

Radionuclides released to the environment under CDA from

PFBR (ground release)

Radionuclide Half-life Release to the

environment (Bq)

Release rate

(Bq s�1)

I-131 8.05 d 1.61E+13 1.86E+08

I-132 2.3 h 2.14E+13 2.47E+08

I-133 20.8 h 2.77E+13 3.20E+08

I- 134 52.5min 2.76E+13 3.19E+08

I-135 6.5 h 2.43E+13 2.81E+08

Cs-134 754d 1.63E+10 1.89E+05

Cs-137 30 yr 4.86E+12 5.63E+07

Rb-88 18min 7.14E+12 8.263E+07

Ru-103 39 d 3.37E+12 3.90E+07

Ru-106 374d 1.48E+12 1.71E+07

Sr-89 50 d 2.87E+12 3.32E+07

Sr-90 29 yr 1.16E+11 1.34E+06

Ce-141 32 d 4.64E+12 5.37E+07

Ce-144 285d 2.31E+12 2.67E+07

Te-131m 30h 9.50E+11 1.099E+07

Te-132 3.2 d 1.10E+13 1.273E+08

Ba-140 12.8 d 6.39E+12 7.396E+07

Zr-95 64 d 3.99E+12 4.618E+07

La-140 1.7 d 4.62E+12 5.347E+07

Pu-239 24400 yr 1.67E+07 1.933E+02

Kr-85m 4.5 h 1.30E+15 15.05E+10

Kr-87 1.3 h 2.35E+15 2.72E+10

Kr-88 2.8 h 2.85E+15 3.29E+10

Kr-85 10.7 yr 2.71E+13 3.14E+08

Xe-133 5.2 d 1.47E+16 1.70E+11

Xe-135 9.1 h 1.53E+16 1.77E+11

et al., 2003) and are given in Tables 2 and 3 for stack and

ground releases, respectively. Dose rates are estimated

from simulated plume, ground-level concentration or

ground deposition using the standard procedures avail-

able from literature (Kai, 1984; IAEA, 1996; Till and

Meyer, 1983).

Dispersion simulation is done over two separate

ranges viz., (i) a fine mesh area of 12.5� 12.5 km2 with

0.25 km grid size, covering a radius 6.25 km around

PFBR and (ii) a coarse mesh area of 50� 50 km2 with

1.0 km grid size which covers a range of 25 km from

PFBR. All grids have vertically 10 levels with a

resolution of 50m. A total emission mass of 86.4

E+03Bq split into 100,000 computational particles is

released over a period of 24 h with an equivalent release

rate of 1 Bq s�1. The concentration is sampled every 60 s

and averaged for 1800 s. Dispersion model is initialized

at 00 h on 24 May 2003 and MM5 predicted meteor-

ological and turbulence parameters are given as input to

FLEXPART every hour. These parameters are held

constant during the hourly periods.

5. Results

5.1. Dose distribution in the local range

The dose to the public due to the accidental releases

comprises (i) cloud gamma from the noble gases, (ii)

inhalation of particulate and noble gases and (iii)

exposure from ground deposited particulate activity.

The first one is computed from the distribution of

concentration from the entire plume and the latter two

are based on the ground-level concentration and

deposition. These are computed independently for stack

and ground release cases to study the coastal boundary

layer effects under different conditions. They are

discussed in the following.

5.1.1. stack release

Dose value is computed for the first 60 s after release

at 00 h. However, it is also interesting to know the dose

pattern if the release takes place in different other

conditions on that day particularly during sea breeze

and TIBL-fumigation condition. So the dose due to

stack releases of 60 s is calculated for different stability

conditions to identify which is the worst meteorological

situation at a coastal site. From the inhalation dose

computed from simulated ground-level concentration at

different times (Fig. 7), it can be seen that the plume

trajectory follows the streamline in this local range and

the horizontal spread (lateral and along plume) varies

with time due to changes in stability and wind. The

plume lies in the northeast direction, over the coast,

between 00 h and 06 h IST (stable condition E). Then, it

ARTICLE IN PRESS

0.07

0.07

0.07

0.12

0.12

0.12

0.12

0.20

0.20.03

0.06

0.11

0.11

0.19

0.19

0.19

0.

20 25 30 35 40 45

20

25

30

35

40

45

50

0.311

0.262

0.09

Distance (mSv)

7.2 -- 12 4.3 --7.2 2.6 -- 4.3 1.5 --2.6 0.92 -- 1.5 0.55 -- 0.92 0.33 -- 0.55 0.20 -- 0.33 0.12 -- 0.20 0.07 -- 0.12

Stable (E)G

rids

in S

-N d

irect

ion

0.006

20 25 30 35 40 45

20

25

30

35

40

45

0.425

Grid

s in

S-N

dire

ctio

n

0.07

0.12

0.20

0.33

0.330.0019

0.0030.006

0.012

0.022

0.040.08

0.140.27

20 25 30 35 40 45

20

25

30

35

40

45

0.23

0.437

0.41

0.45

0.44

10 hr LSTUnstable (C)

Grid

s in

S-N

dire

ctio

n

Grids in W-E direction

0.07

0.07

0.07

0.07

0.07

0.07

0.07

0.07

0.07

0.07

0.07

0.07

0.07

0.07

0.07

0.07

0.07

0.07

0.07

0.07

0.12

0.12

0.120.12

0.20

0.330.003

0.006

0.006

0.006

0.006

0.012

0.012

0.012

0.012

0.012

0.012 0.012

0.021

0.021

0.021

0.021

0.021

0.04

0.04

0.04

0.04

0.040.04

0.04

0.04

0.07

0.07

0.07

0.07

0.07

0.13

0.24

15 20 25 30 35 40 4515

20

25

30

35

40

45

0.199

0.43

12 hr LSTUnstable (A)

Grid

s in

S-N

dire

ctio

n


0.07

0.07

0.07

0.07

0.07

0.07

0.07

0.07

0.12

0.12

0.20

0.20

0.33

10 15 20 25 30 35 4020

25

30

35

40

45Unstable (C-TIBL)

0.34

0.41

Grid

s in

S-N

dire

ctio

n

0.07

0.12

0.20

0.33

10 15 20 25 30 35 4020

25

30

35

40

45

1.2

1.34

0.89

15 hr LSTUnstable (C-TIBL)

Grid

s in

S-N

dire

ctio

n

Grids in W-E directionGrids in W-E direction

13 hr LST


06 hr LST 08 hr LSTStable (E)


0.070.120.021

0.07

0.07 0.0120.20

Fig. 7. Simulated inhalation dose distribution due to stack release for 60 s in 6.25 km range calculation domain. Figures from top left

to bottom right represent dose corresponding to 06, 08, 10, 12, 13, 15 h representing the conditions F, C, A, TIBL fumigation at 13, 15,

respectively. The release location is indicated by aK,K represents the site boundary at 1.5 km from release site. The contour intervals

are uniform for all the plots.



backs towards east over sea during land breeze hours

and gradually turns in the northwest direction in the

afternoon by the incidence of sea breeze.

According to the regulatory requirement, site bound-

ary doses are examined first. The pattern of dose at

different times through the three pathways is shown in

Table 4. In addition to the values at the site boundary,

occurrence of maximum doses at other distances is also

presented. The major contribution to the total dose

comes from cloud gamma rays. It is seen that the dose

maximum occurs within the site boundary for all

stability condition of A to E whereas the other two

pathways change their distance of dose maximum from

0.5 km for A to 4 km for E conditions. This is because

the cloud dose is computed for the entire 3-D plume

while the other two doses are computed based on the

ground-level concentration and deposition at the re-

ceptor. The pattern of dose distribution downwind is as

well known from GPM. The difference occurs when the

fumigation condition is simulated. During this condi-

tion, the cloud gamma dose maximum is also shifted

away from the site boundary. Secondly, the inhalation

and exposure dose values are the highest during

fumigation time among all other stability conditions

and occur at around 2.5 km downwind. For example,

the exposure dose is about two orders higher than the E

category value. This is in contrast to the belief that

stable category leads to the worst situation. The plume

pattern during fumigation time is very narrow and the

dose value remains almost steady away from site

boundary up to 4 km as indicated by the central contour

value. Thus, for stack releases of noble gases at the

Table 4

Simulated radioactive dose due to stack releases (60 s) along plume cen

is diagnosed on basis of Goulder’s relationship

Time (IST) Stability

category

Pathway D

(m

11:00 A Cloud gamma 0.7

Inhalation 0.1

Exposure 3.6

10:00 C Cloud gamma 2.5

Inhalation 0.1

Exposure 3.0

08:00 E Cloud gamma 2.1

Inhalation 0.0

Exposure 5.2

13:00 Fumigation Cloud gamma 1.5

Inhalation 0.3

Exposure 1.3

15:00 Fumigation Cloud gamma 3.2

Inhalation 0.8

Exposure 2.6

coastal site, TIBL-fumigation condition leads to severe

radiological hazard beyond site boundary while at all

other conditions, the maximum hazard lies within the

site boundary.

5.1.2. Ground release

The doses due to all the three pathways are computed

for ground releases for 24 h. The values decrease

monotonically with distance downwind for all the

pathways for all the stability conditions. The interest

of our investigation once again pertains to the identifi-

cation of condition leading to worst hazard. So a break-

up of the doses at different stability conditions in

different pathways is presented (Table 5). Here again,

cloud gamma is the chief contributor to the total dose.

The site boundary dose value, among the different

stabilities, is higher during the TIBL fumigation time

followed by stable (E) condition than at other times. For

example, the exposure dose due to deposition at the site

boundary is an order higher than its corresponding

value at other times. The dose during TIBL fumigation

and stable E conditions remains almost a steady value

away from site boundary. The distance of this steady

dose value spreads up to a distance of 5.0 km during

TIBL condition and to 2.5 km for E-condition as seen

from cloud gamma dose (eighth contour incremented

from the lowest value) (Fig. 8). The plume is rather

narrow with less lateral dispersion during fumigation

time unlike in the case of E-category, where it spreads

laterally more. The narrow plume width in the sea

breeze time is probably due to higher wind velocities

(Table 1). The conclusion is that for ground releases, the

terline from a domain with 6.5 km range. The stability category

ose at 1.5 km

Sv)

Maximum dose

Dmax (mSv)

Distance of

Dmax (km)

5 5.91 0.0�0.25

4 0.45 0.5

7E�05 6.12E�5 0.0–0.50

98 3.886 0.0–0.25

76 0.425 2.25–4

3E�05 4.77E�05 2.5–4.0

8 4.17 0.0–0.25

25 0.425 3.5

E�07 4.43E�05 3.75

9 3.4 0.25–0.50

4 0.40 1.75–2.0

7E–4 1.78E�4 2.5

3.42 2.5–4

9 1.34 2–4

6E�4 0.00156 2.75–5

ARTICLE IN PRESS

Table 5

Simulated radioactive dose due to ground releases (24 h) along plume centerline from a domain with 6.5 km range. The stability

category is diagnosed on basis of Goulder’s relationship

Time (IST) Stability category Pathway Dose at 1.5 km

(mSv)

Dose away from

1.5 km

11:00 A Cloud gamma 0.0474 Monotonous

decrease

Inhalation 0.0047 -do-

Exposure 6.32E�6 -do-

10:00 C Cloud gamma 0.1216 -do-



08:00 E Cloud gamma 0.1304 -do-



13:00 Fumigation Cloud gamma 0.0861 -do-



15:00 Fumigation Cloud gamma 0.1632 0.142 (2–3 km )

Inhalation 0.0275 0.0178 (2–3 km )

Exposure 1.413E�5 7.846E�7 (2–3 km)

1E-4

2.5E-4

6E-4

0.00150.0015

0.004

0.009

0.022

0.05

0.13

15 20 25 30 35 40 45 5010

15

20

25

30

35

40

45

Grid

s in

S-N

dire

ctio

n

1E-4

2.5E-4

6E-4

0. 0015

0.004

0.009

0.022

0.05

0.08

10 15 20 25 30 35 4015

20

25

30

35

40

45

0.33 - - 0.80 0.13 -- 0.33 0.05 - - 0.13 0.022 -- 0.05 0.009 -- 0.022 0.004 -- 0.009 0.0015 -- 0.004 6E-4 -- 0.0015 2.5E-4 -- 6E-4 1E-4 -- 2.5E-4

Fumigation at 15 hr LST

Grid

s in

S-N

dire

ctio

n

Stable cateory 'E'


Dose (mSv)

Grids in W-E direction(A) (B)

Fig. 8. Simulated cloud gamma dose due to ground releases in 6.25 km range calculation domain for (A) stable E and (B) TIBL

fumigation conditions. The circle represents the site boundary at 1.5 km from release site. Contour intervals are uniform for both plots.


distance of constant dose extends to more distance for

TIBL condition whereas it spreads to relatively wider

area confined to a narrow range in the case of stable

conditions. The integrated dose due to ground releases

for 24 h for cloud gamma, inhalation and exposure due

to deposition amounts to 0.641, 0.21, 5.4E�5mSv,

respectively, at 1.5 km yielding a total dose of 0.91mSv

due to ground releases at site boundary (Fig. 9, Table 5).

The trace of plume trajectory for ground, stack releases

shows that the dispersion on the chosen day is confined

to N–E quadrant of the study region, a major area of it

falls over the sea.

Thus, the total dose due to the stack and ground

releases considering TIBL fumigation period as the

worst condition for short-term stack release and a full

day realistic scenario including fumigation for ground

release is 22.82mSv at the release location and 5.01mSv

at site boundary, respectively.

5.2. Dose pattern in the meso-scale range

The simulated cloud gamma dose pattern for stack

release at different times is shown in Fig. 10 in the

ARTICLE IN PRESS

0.01000.0100

0.0214

0.0214

0.046

0.098

0.209

0.209

0.4470.447

0.956

0 5 10 15 20 25 30 35 40 45 500

5

10

15

20

25

30

35

40

45

50

Site boundary

Bay

of B

enga

l

Coa

stlin

eDose (mSv)

9.35 -- 20.0 4.37 -- 9.35 2.05 -- 4.37 0.956 -- 2.05 0.447 -- 0.956 0.209 -- 0.447 0.098 - - 0.209 0.046 -- 0.098 0.0214 -- 0.046 0.0100 -- 0.0214

Grid

s in

S-N

dire

ctio

n


Fig. 9. Total integrated dose due to ground releases for 24 h in

6.25 km range calculation domain (cloud gamma+inhalatio-

n+exposure due to deposition). The circle represents the site

boundary at 1.5 km from release site.


meso-scale range (25 km). Unlike the local range, here

the influence of spatial inhomogeneity in the wind field is

manifested significantly. Before sea breeze sets in, the

wind is weak westerly land breeze. So the plume is over

the sea till 12 h IST. There is a wider spread of plume

during highly unstable conditions just before sea breeze.

When sea breeze just begins to develop at 13 h the plume

turns near the source in northwest while the large part of

the earlier releases stay in the northeast direction (Fig.

10d). The distinctly different shapes of the plume on

either side of the sea breeze front are also seen from the

dose pattern. The plume is narrow in the region of sea

breeze influence and is wide spread much away further.

This is attributed to the enhanced turbulence at the sea

breeze frontal zone. The narrow region of the plume

progresses further inland in the wake of the sea breeze

front (Fig. 10e,f).

Further, it is noticed again that the cloud gamma dose

values at the central line of the plume during fumigation

hours (15 h) are about more than twice that of the values

simulated during F category. The innermost contours

corresponding to the value of 0.40–1.80mSv are

extending to large downwind distances like a tongue,

posing more hazard during fumigation hours than

during any other condition of the atmosphere. Similar

pattern is noticed in the case of inhalation and ground

exposure doses too.

In the case of ground release, dose through all the

pathways decreases with distance for all conditions in

general. However, for TIBL fumigation and stable E

conditions, the high dose value (the tongue!) extends up

to some distance viz., about 20 km for fumigation and

12 km for E-condition. As in the case of local range, the

dose value through all the pathways is higher for TIBL

fumigation condition in the meso-scale range also. For

example, the cloud gamma dose at 15 km distance is

about three times higher during fumigation hours than

the value during stable condition E. The total dose at

large distances due to ground release is however below

the value at the site boundary and is well within the

safety limit.

5.3. Comparison with GPM estimates

A comparison of the dose estimates for the PFBR-

CDA obtained from the present modelling system is

made with those computed from a standard GPM, and

the latter ignores the coastal effect. According to

regulatory requirements, the GPM estimates are made

conservatively assuming a uniform wind speed of 2m s�1

and worst condition (F-category) of the atmosphere for

all the 24 h. Whereas, in the present estimation using the

hydrodynamic model, a typical realistic situation in-

cluding fumigation condition is simulated. Examination

of the dose values given by the two systems (Table 6)

clearly shows that the total dose at site boundary given

by present system for the realistic situation is about eight

times less than the GPM value. While for ground

releases the estimates by present system are much lower

than GPM values, for stack releases they are higher in

the fumigation condition than the corresponding GPM

values for stable F condition. If the wind speed during

fumigation is assumed to be 2m s�1 as in the case of

GPM, the dose value would be further enhanced by a

factor of 2. The actual wind speed is 5.5m s�1 during the

sea breeze-fumigation hours. In addition, the area of

radiological impact is obviously different from that

computed by straight-line GPM.

6. Summary and conclusion

The meteorological condition at coastal locations is

complex due to land–sea interface. For emergency

accidental situations, though unlikely, it is necessary to

evaluate the worst meteorological condition for disper-

sion at the site leading to potential impact and its range

for safety analysis. The conventional regulatory ap-

proaches for assessment of environmental radiological

impact of nuclear power plants assume F category of

stability as the worst situation and often employ the

GPM in dose assessment. Present study gives an insight

into the validity of these assumptions and a realistic

picture of the dispersion pattern near a coastal site.

Analysis of the plume dispersion of radionuclides

released from ground and partially from a 100m stack

ARTICLE IN PRESS

1E-3

0.0021

0.004

0.009

0.020

0.04

0.09

0.19

5

10

15

20

25

30

35

40

45

50

Bay

of B

enga

l

Coa

stlin

e

15 km

10 km

5 km

1.5 km

0.85 - - 1.8 0.40 - - 0.85 0.19 - - 0.40 0.09 - - 0.19 0.04 - - 0.09 0.020 -- 0.04 0.009 -- 0.020 0.004 -- 0.009 0.0021 -- 0.004 1E-3 -- 0.0021

1E-3

0.0021

0.004

0.009

5

10

15

20

25

30

35

40

45

50

15 km

10 km

5 km

1.5 km

Grid

s in

S-N

dire

ctio

n

1E-3

1E-3

1E-3

0.0021

0.0021

0.004

0.009

0.020

0.04

0.09 0.19

0.85

5 10 15 20 25 30 35 40 45 50

5 10 15 20 25 30 35 40 45 50

5

10

15

20

25

30

35

40

45

50

15 km

10 km

5 km

1.5 km

Grid

s in

S-N

dire

ctio

n

1E-3

1E-3

1E-3

1E-3

1E-3

1E-3

1E-30.0021

0.00210.004

0.004

0.009

0.020

0.04

0.04

0.09 0.19

0.40

5

10

15

20

25

30

35

40

45

50

15 km

10 km

5 km

1.5 km

Grid

s in

S-N

dire

ctio

n

1E-3

1E-3

1E-3

1E-3

1E-3

0.00212

0.0045

0.0045

0.0095

0.0095

0.0200

0.0200

0.0200

0.042

0.090

0.190

0.402

5

10

15

20

25

30

35

40

45

50

15 km

10 km

5 km1.5 km

Grid

s in

S-N

dire

ctio

n

Grid

s in

S-N

dire

ctio

n

2.11E-17

1E-3

1E-3

1E-3

1E-3

1E-31E-3

1E-3

0.002060.00206

0.0043

0.0043

0. 0043

0.0043

0.0043

0.0043

0 .0088

0.0088

0.0088

0.0181

0.0181

0.037

0.077

0.159

0.329

0.329

5

10

15

20

25

30

35

40

45

50

15 km

10 km

5 km

1.5 km

Stable (F) at 06 hr LST

Dose(mSv)

Unstable 'A' at 12 hr LST


5 10 15 20 25 30 35 40 45 50



5 10 15 20 25 30 35 40 45 50

Grids in W-E direction5 10 15 20 25 30 35 40 45 50


5 10 15 20 25 30 35 40 45 50


Fumigation at 14 hr

Unstable (C) at 10 hr LST

Sea breeze incidence Unstable 'C'

Fumigation at 15 hr LST

Grid

s in

S-N

dire

ctio

n

Fig. 10. Simulated cloud gamma dose pattern for 60 s. Stack release in the meso-scale calculation range during (A) stable-E, (B)

unstable-C, (C) unstable-A, (D) sea breeze onset, and (E) TIBL fumigation. The area covered is 50� 50 km with 1 km grid distance, the

release location is at grid point (34,25). Circles in different colours indicate different ranges of distance from release location. Contour

values are uniform for all the plots.


under DBA situation of a nuclear power plant proposed

at Kalpakkam on the east coast of India has helped to

understand the said issue.

A simulation is carried out using an advanced

dispersion system consisting of a 3-D non-hydrostatic

meso-scale atmospheric model (MM5) coupled to a

ARTICLE IN PRESS

Table 6

Dose estimate at the site boundary distance 1.5 km (in mSv) for CDA using GPM and MM5+FLEXPART

Dose Type of release Dose computed with GPM Dose computed with

MM5+FLEXPART

system

Cloud gamma Ground release 20.9 0.62

Stack release 7.9 3.20

Inhalation Ground release 6.4 0.21

Stack release 3.4E�5 0.89

Deposited activity Ground release 0.68 5.4E�5

Stack release 51 2.67E�4

Total site boundary

dose for 24 h

36.0 5.0


Lagrangian particle dispersion model (FLEXPART). A

typical summer sea breeze day is chosen so that available

meteorological field experimental data are used to

validate the simulated sea breeze circulation and

formation of TIBL at the coastal site. Results indicate

that for both stack and ground release, the maximum

site boundary dose occurs during the TIBL fumigation

time in the sea breeze hours. Unlike the ground source,

the dose due to stack releases does not fall mono-

tonically with distance as would be expected when GPM

is used but becomes maximum beyond the site boundary

at 3–4 km from the release location. The distance range

and area of maximum radiological hazard (the radio-

active tongue) is enhanced by about two to three times

during TIBL-sea breeze period at the coastal site. As the

sea breeze duration is longer in the months of June, July,

August and September at Kalpakkam, the fumigation

effects would be quite significant.

Simulation in a meso-scale range of 25 km shows

turning of plume and two dispersion ranges across the

sea breeze front. The dose remains almost a steady value

during TIBL formation time up to 20 km. The study

confirms that the radioactive plume dose from cloud

shine, ground shine and inhalation pathways is two to

three times more during sea breeze-fumigation hours

than the dose computed for F stability category.

However, the dose is eight times less than that computed

using the conservative GPM for regulatory purpose and

safety is doubly ensured to be within the limit.

Meteorological and dispersion simulation as well as

tracer experiments for various days typical of different

seasons are needed to arrive at a quantitatively more

meaningful result. The meteorological aspects simulated

by MM5 have been evaluated as forecast errors will

affect the dispersion pattern. In the present simulation,

the onset of sea breeze is about 1–2 h late and the

predicted atmosphere is less stable than the observations

at 6 AM. Also a lower sensible heat flux is present in the

model in the forenoon. A modification in the initial soil

moisture would be necessary to improve the calculation

of surface turbulent fluxes. Also MM5 has several

boundary layer parameterization schemes. Each of them

would simulate the boundary layer flow in a different

manner depending on the surface schemes, vertical

diffusion scheme (local/non-local), turbulence closure,

etc. Although a higher order, non-local scheme is used

here for coastal conditions, a proper choice of PBL

scheme is needed in order to improve the near-surface

predictions and the TIBL effects into the dispersion

forecast. Also, application of locally generated data on

land-use and soils of the region would further improve

the meteorological prognosys in the local range (Hanna

and Yang, 2001).

Acknowledgements

The authors sincerely thank Dr. Baldev Raj, Director,

IGCAR, Dr. S. Govindarajan, Director, Safety Group,

and Dr. A. Natarajan, Head, Radiological Safety

Division for their encouragement in carrying out the

study. Authors gratefully acknowledge Andrew Stohl,

Gerhard Wotawa, Technical University of Munich,

Germany, and Dr. Petra Seibert, Vienna, for helpful

suggestions in implementation of their dispersion code

FLEXPART. Authors are also grateful for the anon-

ymous reviewers for useful suggestions in the improve-

ment of the manuscript.

References

Anthes, R.A., Warner, T.T., 1978. Development of hydro-

dynamic models suitable for air pollution and meso-

meteorological studies. Monthly Weather Review 106,

1045–1078.

Arritt, R.W., 1987. The effect of water surface temperature on

lake breezes and thermal internal boundary layers. Bound-

ary-Layer Meteorology 40, 101–125.

ARTICLE IN PRESSC.V. Srinivas, R. Venkatesan / Atmospheric Environment 39 (2005) 1497–1511 1511

Camps, J., Massons, J., Soler, M.R., Nickerson, E.C., 1997.

Pollutant transport in coastal areas with and without

background wind. Annales Geophysicae 15, 476–486.

Clarke, R.H., Macdonald, H.F., 1978. Radioactive releases

from nuclear installations: evaluation of accidental

atmospheric discharges. Progress in Nuclear Energy 2,

77–152.

Dudhia, J., 1989. Numerical study of convection observed

during winter monsoon experiment using a mesoscale two-

dimensional model. Journal of Atmospheric Science 46,

3077–3107.

Estoque, M.A., 1961. A theoretical investigation of the sea

breeze. Quarterly Journal of the Royal Meteorological

Society 87, 136–146.

Fast, J.D., Steen, B., Addis, R.P., 1995. Advanced atmospheric

modeling for emergency response. Journal of Applied

Meteorology 34 (3), 626–649.

Grell, G.A., Dudhia, J., Stauffer, D.R., 1994. A description of

the fifth-generation Penn State/ NCAR mesoscale model

(MM5). NCAR Technical Note, NCAR/TN-398+STR,

117pp.

Golder, D., 1972. Relations among stability parameters in the

surface layer. Boundary-layer Meteorology 3, 47–58.

Hanna, S.R., 1982. Applications in air pollution modeling. In:

F.T.M., Van Dop, H. (Ed.), Atmospheric Turbulence and

Air Pollution Modeling. D. Reidel Publishing Company,

Dordrecht, Holland.

Hanna, S.R., Yang, R., 2001. Evaluations of meso-scale

model’s simulations of near-surface winds, temperature

gradients and mixing depths. Journal of Applied Meteorol-

ogy 40 (6), 1095–1104.

IAEA., 1980. Atmospheric dispersion in nuclear power plant: a

Safety Guide. Safety series No. 50-SG-S3.

IAEA., 1996. International basic safety standards for protec-

tion against ionising radiation and for the safety of

radiation sources. IAEA Safety Series-115.

Janjic, Z.I., 1990. The step-mountain coordinate: physical

package. Monthly Weather Review 118, 1429–1443.

Kai, M., 1984. A Computer Code for Calculating a g-ExternalDose from a Randomly Distributed Radioactive Cloud.

JAERI-M 84-006.

Keshavamurthy, R., Indra, R., Om Pal Sing, A., 2003. Gaseous

radioactivity and its controlled discharge into environment

from PFBR. Bulletin of Radiation Protection 26 (1–2),

236–240.

Ludwig, F.L., 1983. A review of coastal zone meteorological

processes important to the modeling of air pollution.

Fourteenth Technical Meeting on Air Pollution and its

Application, Copenhagen.

Lyons, W.A., Cole, H.S., 1976. Photochemical oxidant trans-

port: mesoscale lake breeze and synoptic-scale aspects.

Journal of Applied Meteorology 15, 733–744.

Lyons, W.A., Keen, C.S., Schuh, J.A., 1983. Modeling

Mesoscale Diffusion and Transport processes for Releases

Within Coastal Zones During Land/Sea Breezes. US

Nuclear Regulatory Commission, NUREG/CR-3542, Wa-

shington, DC, 183pp.

Mellor, G.C., Yamada, T., 1982. Development of a turbulence

closure model for geophysical fluid problems. Reviews of

Geophysics 20, 851–875.

National Academy of Sciences, 1992. Coastal Meteorology: A

Review of the State of the Science. National Academy Press,

Washington, DC.

Shirvaikar, V.V., 1973. Guide book for stack parameters in

nuclear and ancillary chemical installation. BARC Report-

722.

Sivaramakrishnan, S., Venkatesan., R., 2002. Coastal Atmo-

spheric Boundary Layer Experiment (CABLE-2001) at

Kalpakkam. Project Report, Safety Research Institute,

Atomic Energy Regulatory Board, Mumbai.

Srinivas, C.V., Venkatesan, R., Bagavath Singh, A., Somayaji,

K.M., 2004. A real case simulation of the air-borne effluent

dispersion on a typical summer day under CDA scenario for

PFBR using an advanced meteorological and dispersion

model. Research Report, IGC–259, p. 82.

Stohl, A., 1999. The FLEXPART Particle Dispersion Model

Version 3.1. User Guide, Lehrstuhl fur Bioklimatologie und

Immissionsforschung, University of Munich, Am Hochan-

ger 13, 85354 Freising, Germany.

Stohl, A., Hittenberger, M., Wotawa, G., 1998. Validation of

the particle dispersion model FLEXPART against large

scale tracer experiment data. Atmospheric Environment 24,

4245–4264.

Venkatesan, R., Mathiyarasu, R., Somayaji, K.M., 2002. A

study of atmospheric dispersion of radionuclides at a coastal

site using a modified Gaussian model and a mesoscale sea

breeze model. Atmospheric Environment 36, 2933–2942.

Vugts, H.F., Businger, J.A., 1977. Air modification due to a

step change in surface temperature. Boundary-Layer

Meteorology 11, 295–305.

Zannetti, P., 1990. Air Pollution Modelling. Computational

Mechanics Publication, Boston, MA.

Till, J.E., Meyer, 1983. Radiological Assessment: A Textbook

on Environmental Dose Analysis NUREG-CR 5332.

Date post:	15-May-2023
Category:	Documents
Upload:	umich
View:	1 times
Download:	0 times

A simulation study of dispersion of air borne radionuclides from a nuclear power plant under a...

Documents