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.
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.
C.V. Srinivas, R. Venkatesan / Atmospheric Environment 39 (2005) 1497–1511 1501
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;
C.V. Srinivas, R. Venkatesan / Atmospheric Environment 39 (2005) 1497–15111502
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.
C.V. Srinivas, R. Venkatesan / Atmospheric Environment 39 (2005) 1497–1511 1503
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
ARTICLE IN PRESSC.V. Srinivas, R. Venkatesan / Atmospheric Environment 39 (2005) 1497–15111504
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
Grids in W-E direction
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
Grids in W-E direction
06 hr LST 08 hr LSTStable (E)
Grids in W-E direction
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.
C.V. Srinivas, R. Venkatesan / Atmospheric Environment 39 (2005) 1497–1511 1505
ARTICLE IN PRESSC.V. Srinivas, R. Venkatesan / Atmospheric Environment 39 (2005) 1497–15111506
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-
Inhalation 0.0185 -do-
Exposure 7.256E�6 -do-
08:00 E Cloud gamma 0.1304 -do-
Inhalation 0.020 -do-
Exposure 8.22E�6 -do-
13:00 Fumigation Cloud gamma 0.0861 -do-
Inhalation 0.0116 -do-
Exposure 3.28E�6 -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'
Grids in W-E direction
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.
C.V. Srinivas, R. Venkatesan / Atmospheric Environment 39 (2005) 1497–1511 1507
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
Grids in W-E direction
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.
C.V. Srinivas, R. Venkatesan / Atmospheric Environment 39 (2005) 1497–15111508
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
Grids in W-E direction
5 10 15 20 25 30 35 40 45 50
Grids in W-E direction
Grids in W-E direction
5 10 15 20 25 30 35 40 45 50
Grids in W-E direction5 10 15 20 25 30 35 40 45 50
Grids in W-E direction
5 10 15 20 25 30 35 40 45 50
Grids in W-E direction
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.
C.V. Srinivas, R. Venkatesan / Atmospheric Environment 39 (2005) 1497–1511 1509
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
C.V. Srinivas, R. Venkatesan / Atmospheric Environment 39 (2005) 1497–15111510
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.