Atmospheric

Dispersion

R. Bhattacharya
Member of IAEA’s Expert Mission to Malayasia
Email: rbhattacharya@[Link]
1
 Contents
• Dispersion Model
– Fundamental concepts
– Classification of models

• Gaussian plume model

– Derivation and limitations

• Other models

• Examples

• Exposure pathways and

radiological impact
assessment

2
What is an atmospheric dispersion model?

An atmospheric dispersion model is a:

• mathematical simulation of the physics and chemistry
governing the transport, dispersion and transformation
of pollutants in the atmosphere

• means of estimating downwind air pollution

concentrations, given information about the pollutant
emissions and nature of the atmosphere.

3
 Atmospheric dispersion model and
Authorisation of NPP

Industrial facilities are required to obtain

permits to emit into the atmosphere and to
demonstrate their compliance with regulations

In the process of applying for permits,

dispersion models are generally used to assess
the impact of point source emission
4
 Emission Source
Meteorology
Information

Topography Receptors

Dispersion Model

Atmospheric Concentration at Receptors

5
 Meteorological
Model

Emission Chemical
Model Model
Model
types

Source
Receptor
Dispersion
Model
Model
6
 • Estimates temporal and spatial emission
Emission Model rates based on activity level, emission rate
per unit of activity and meteorology

Meteorological • Describes transport, dispersion, vertical

mixing and moisture in time and space
Model

• Describes transformation of directly

Chemical emitted particles and gases to secondary
particles and gases; also estimates the
Model equilibrium between gas and particles for
volatile species

7
 Source • Uses the outputs from the previous models
to estimate concentrations measured at
receptors; includes mathematical
Dispersion simulations of transport, dispersion, vertical
mixing, deposition and chemical models to
Model represent transformation.

Receptor • Infers contributions from different primary

source emissions or precursors from
Model multivariate measurements taken at one
ore more receptor sites.

8
 Classifications of Models
 Developed for a number of pollutant types and
time periods

◦ Short-term models – for a few hours to a few days;

worst case episode conditions
◦ Long-term models – to predict seasonal or annual
average concentrations; health effects due to exposure

 Classified by
◦ Non-reactive models – pollutants such as SO2 and CO,
FPNG, Ar-41 etc
◦ Reactive models – pollutants such as O3, NO2, etc.
9
 Classifications of Models
 Classified by coordinate system used
◦ Grid-based
 Region divided into an array of cells
 Used to determine compliance with NAAQS
◦ Trajectory
 Follow plume as it moves downwind

 Classified by level of sophistication

◦ Screening: simple estimation use preset, worst-case
meteorological conditions to provide conservative
estimates.
◦ Refined: more detailed treatment of physical and
chemical atmospheric processes; require more detailed
and precise input data.

10
 Data Inputs: Emission Sources

• Source types

• Emissions of each pollutant

• Location and Height of emission sources

• Stack Characteristics
11
 Source Types

Point sources
(stacks)

 Line sources
(roads)

 Area sources
(treatment
ponds)

 Volume
sources
(buildings)

12
Atmospheric dispersion modeling
procedure

13
• Dispersion models
– Box model
– Gaussian plume model
– Gaussian puff model
– Complex numerical models
• ‘now-casting’ (diagnostic)
• 'forecasting' (prognostic)

14
 Modeling Procedure
 The modeling procedures can be categorized into four generic
classes:
 Gaussian
 Numerical
 statistical or empirical and
 physical

 The emphasis is on Gaussian-plume type models for continuous

releases, which are at the core of most regulatory models

 Gaussian models are the most widely used techniques for estimating
the impact of nonreactive pollutants

15
 Gaussian Dispersion Models
 Most widely used
 Based on the assumption
◦ plume spread results primarily by molecular diffusion
◦ horizontal and vertical pollutant concentrations in the plume are
normally distributed (double Gaussian distribution)
X
 Plume spread and shape vary in response to meteorological
conditions Z
Q

u
Y
H

16
 Atmospheric Modelling

Gaussian Plume Model (GPM):

The three-dimensional diffusion equation for atmospheric diffusion
is:
𝝏𝝌 𝝏 𝝏𝝌 𝝏 𝝏𝝌 𝝏 𝝏𝝌
= 𝑲 + 𝑲 + 𝑲
𝝏𝒕 𝝏𝒙 𝒙 𝝏𝒙 𝝏𝒚 𝒚 𝝏𝒚 𝝏𝒛 𝒛 𝝏𝒛

Where,
x, y and z = the downwind, cross-wind and vertical directions,
respectively
Kx , Ky , Kz = Diffusivity in x-, y- and z- directions

17
 Atmospheric Modelling
Assuming continuity of mass, solution to the diffusion equation with
varying initial and boundary conditions, yields Gaussian Distribution of
concentration, χ.
and, the standard deviation of the Gaussian distribution is given by:

σ2 = 2Kt

Schematic representation of a Gaussian Plume 18

 The Diffusion Equation and the Gaussian
Plume Model

The mass rate of diffusion Nx of a gaseous

species in the x-direction at some cross-
sectional area A is given by the expression
 DxC  
N x   A  
 x 

Where;
Nx is the mass transfer per unit time
Dx is the mass diffusivity, area/time, in the x direction
C is the concentration in mass per unit volume
A is the cross- sectional area in the x-direction

19
Schematic for the Development of
Gaussian Plume Model

DxC 
N x  dy dz N x  dx
x
Bulk Motion Out
(bulk motion in) = CU dy dz

20
 DxC 
N x  dy dz N x dx
x
(bulk motion) =
CU dy dz Bulk Motion Out

 DxC 
N x  dy dz
x

DxC    DxC  
N x  dx  dy dz   dy dz  dx
x x  x  
  DxC  
N x  dx  N x   
x  x   dx dy dz
dy dz

rate in (bulk motion)  CU dy dz

rate out (bulk motion )  CU dy dz  CU dy dz dx
x

net rate (bulk motion )   CU dx dy dz
x

C
rate of change within dx dy dz  dx dy dz
t

21
 C    DxC     DyC     DzC 
  CU        
t x x  x  y  y  z  z 

Where;
x = along- wind coordinate measured in wind direction from the source
y = cross-wind coordinate direction
z = vertical coordinate measured from the ground
C(x,y,z) = mean concentration of diffusing substance at a point (x,y,z) [kg/m3]
Dy,Dz = mass diffusivity in the direction of the y- and z- axes [m2/s]
U = mean wind velocity along the x-axis [m/s]

C CU 
 Time rate of change and advection of the cloud by the mean wind
t x

  DxC   Turbulent diffusion of material relative to the center of the pollutant cloud.
 , etc. ( the cloud will expand over time due to these terms.)
x  x 

22
 Assumptions
 Mass transfer due to bulk motion in x-direction far out shadows the
contribution due to mass diffusion. That is the second term on the right
side of Equation is far smaller than the first term and may be dropped
from the equation
 We are primarily interested in the steady-state solution
C to the dispersion
of the pollutants in the atmosphere. Hence the t quantity is zero
 Even though the wind speed does vary in the three coordinate directions,
the variation is relatively small. Therefore it is appropriate to assume
that the wind speed U is constant
 The major transport direction due to the wind is chosen to lie along the
x-axis
 Dx, Dy and Dz are constant

23
 C    DxC     DyC     DzC 
  CU        
t x x  x  y  y  z  z 

 C    2C    2C 
U   Dy 2   Dz 2 
 x   y   z 

The general solution to this second-order partial differential equation is


  y   z  U 
2 2

C  Kx exp  
1
    

  Dy   Dz  4 x 
Where K is an arbitrary constant whose value is determined by the boundary
conditions.
The rate of transfer of pollutant through any vertical plane downwind from the source is a
constant in steady state, and this constant must equal the emission rate of the source, Q.

Q   UC dy dz
Generally the limits of integration on dy are minus to plus infinity and for a point source at
Elevation H above the ground level the limits of integration on z are taken
  tofrom


 
  y2 z2  U 
Q    KUx exp  
1
   dy dz
     Dy Dz  4x 
24
After integrating

Q
K
4 Dy Dz 
1/ 2

Where Q is the strength of the emission source, mass emitted per unit time

Q   y2 z2  U 
C ( x, y , z )  exp     
4xDy Dz 
1/ 2
  Dy Dz  4 x 

Gaussian parameters

x x
 y  2 Dy and  z  2 Dz
U U

25
The general equation to calculate the steady state concentration
of an air contaminant in the ambient air resulting from a point
source is given by:

Q   y2  
   z  H 2    z  H 2  

C ( x, y, z )  exp    exp    exp   
2U  y z   2 2     
2 2
  y    2 z   2 z 

Where;
c(x,y,z) = mean concentration of diffusing substance at a point (x,y,z) [kg/m3]
x = downwind distance [m],
y = crosswind distance [m],
z = vertical distance above ground [m],
Q = contaminant emission rate [mass/s],
= lateral dispersion coefficient function [m],
=y vertical dispersion coefficient function [m],
 =z mean wind velocity in downwind direction [m/s],
U
H = effective stack height [m].

26
Plume Boundary

27
Plume Dispersion by Gaussian Distribution and Coordinate System

28
 Effect of
atmospheric
stability on
plumes:
Gaussian Parameters
The lateral dispersion coefficient function and, the vertical dispersion
coefficient functions depend on the downwind distance and the atmospheric
stability class. These coefficients in meters can be obtained using Pasquill-
Gifford-Turner estimates shown in the equations below

where ,
s = an integer [1-6] representing the atmospheric stability shown in Table 1
kx,x = empirical constants, values for each of the stability class can be obtained from Green et al. (1960)

30
Atmospheric Stability Classes

Pasquill Stability categories (1995)

31
Dispersion Coefficients: Horizontal

32
Dispersion Coefficients: Vertical

33
 Gaussian Dispersion Equation
If the emission source is at ground level with no
effective plume rise then
  2 2 
C x, y, z  
Q 1 y z 
exp   2 
 y  z u  2   y  z 
2

Plume Rise
• H is the sum of the physical stack height
and plume rise. H  h h
plume rise actual stack

34
 Plume Rise
Buoyant plume: Initial buoyancy >> initial momentum
Forced plume: Initial buoyancy ~ initial momentum
Jet: Initial buoyancy << initial momentum
 For neutral and unstable atmospheric conditions, buoyant rise can be
calculated by
21.425F 0.75
hplumerise  (F  55 m 4 / s 3 )
u
38.71F 0.6
hplume rise  (F  55 m 4 / s 3 )
u
Vs: Stack exit velocity, m/s
where buoyancy flux is d: top inside stack diameter, m
Ts: stack gas temperature, K
F  gVs d 2 (Ts  Ta ) / 4TS Ta: ambient temperature, K
g: gravity, 9.8 m/s2

35
 Wark & Warner, “Air Pollution: Its Origin & Control”

 y 2    z  H    z  H 2  

2
C x, y, z  
Q
exp  2  exp    exp   
2 y  z u  2 y  
  2 z 
2
 2 z  
2

36
Ground level concentration
Q  y2   H2 
C exp  2  exp  2 
 y  z u  2 y   2 z 
37
Maximum Ground Level Concentration
Under moderately stable to near neutral conditions,
 y  k1 z
The ground level concentration at the center line is
 H2 
C x,0,0 
Q
exp  2 
k1 z u
2
 2 z 
The maximum occurs at
H
dC / d z  0  z 
2
Once z is determined, x can be known and subsequently C.

C x,0,0  exp  1  0.1171

Q Q
 y  z u  y zu

38
 Atmospheric Modelling

• Restrictions for GPM

– In practical terms, Gaussian Plume Model should
not be applied under conditions of:

1. Low wind speed

2. Complex terrain
3. Spatial and temporal changes in meteorological
parameters
4. deposition and transformation (e.g. radiological
decay) within the plume during travel

39
Limitations of simplified models

40
 Atmospheric Modelling
Limitations of GPM:
Sl. No. Conditions Accuracy
1 1. Ideal Conditions: 10% to 20%

Near field (<1.0 km) short averaging times (min to hour),

flat terrain, steady meteorology, surface source

same as above for elevated source 20% to 40%

2 2. Real World Applications: Factor of two

Meteorological parameters reasonably well known and

steady with no exceptional circumstances
3 3. Exceptional Circumstances: Poorer than a factor of
two; may be as poor
Building wakes, buoyant plumes, varied surfaces such as as factor of ten.
forests, cities, shorelines, rough terrain, extreme stable
and unstable conditions, distances greater than 10-20 km

Site-specific field studies or physical modelling, or numerical modelling using complex models
may be required to understand the airflow and the resultant diffusion in practical situations. 41
 Atmospheric Modelling

 Theoretically, the Gaussian Plume Model (GPM) is only valid

when certain basic assumptions are completely met.

 In reality, some of these basic assumptions are never met.

42
 Atmospheric Modelling

 Three of the most common scenarios, where

modification of the basic GPM would be necessary,
are:
1. Complex terrain
2. Building wake effect
3. Sea breeze fumigation

43
 Atmospheric Modelling

Complex Terrain
-Complex terrain, such as hills,
valleys etc, influences the path and
diffusion of a pollutant plume.

-Depending upon meteorological

condition, the same terrain can
affect a plume in different ways.

44
 Atmospheric Modelling

Building Wake Effect

Complex airflow condition
around buildings in the plume
pathway 45
 Atmospheric Modelling
Sea Breeze Fumigation
-Cold marine air advects
over land during
daytime

-The lower layer of the

air mass gets
progressively heated up

-In consequence,
boundary layer develops
over land region (highly
turbulent) known as
'Thermal Internal
Boundary Layer' (TIBL)
46
 Atmospheric Modelling

• Empirically based adjustments to the basic

Gaussian plume formation, allow it to be
applied in many situations, where
theoretically it should not be applied.

47
 Atmospheric Modelling
Comparative evaluation of dispersion models and their application

Model Type Input data needed Application Accuracy Remarks

Box Model Meteorologic Vertical average wind Area sources, Gives uniform concentration Generally
al speed, volume of distributed sources, in domain, hence poor used
model model domain, Mixing long range plume for point source near as screening
height trajectory modeling field application model
Gaussian Combined Surface wind speed, Point, area, volume Gives concentration Widely used
Plume meteorology direction, insolation, source estimates within an
Model and diffusion cloud cover order of magnitude for
model continuous releases
over homogeneous
terrain
Gaussian Dispersion Surface wind speed, Dispersion under Better than Gaussian Used also in
puff model direction, insolation, time varying plume model for time mesoscale
model cloud cover meteorological varying meteorology. models
Conditions, continuous Not satisfactory under
short term releases strong wind shear
under emergency
situations.
Particle Dispersion Atmospheric stability, Dispersion over Good for complex Used also in
trajectory model wind and turbulence complex terrain terrain mesoscale
model data from prognostic models
model 48
 Other Models
• Numerical Modelling
– Diagnostic modeling
– Prognostic modeling
• Meterological model to obtain space-time evolution of
meteorological parameters
• Particle trajectory model-to obtain concentration
estimates using the meteorological model values as
inputs

49
 Example-1: Routine release scenario of NPP

 Suppose Argon-41 is routinely released at a rate

of 1Bq/sec from a stack of height 100m from a
reactor. Find out the centreline Ground Level
concentration at 1km from the stack, maximum
GLC and downwind distance of its occurrence,
given wind speed at 10m is 4m/sec and stability
class is B-C. (note: σz = H/√2 )

50
 U100 = u10 (100/10)0.11 = 5.15 m/sec

 σz and σy from graphs for B and C category are

 σz = 100m (B cat)and 60 m (C cat)
 σy = 140m (B cat)and 100 m (C cat)

 Using Q = 1 Bq/sec, h = 100m, u = 5.15 m/s in

Q  H 
2
C exp  2 
 y z u  2 z 

51
 GLC (1000, 0,0) = 2.654 E-6 Bq/m3 for B category
 and 2.568 E-6 Bq/m3 for C category

 GLC (B-C class) = average of the two values = 2.611 E -6 Bq/m3

 Now, σz = H/√2 = 100/1.14 = 71m

 From the graph, the corresponding X max = 820 m

 GLC(max) = (2Q/eлuHeff2) (σz/ σy)

 σy for B-C category for 820m estimated from graph is 105m

 Therefore, GLC (max) = 3.08 E-6 Bq/m3 which will occur at 820m
downwind

52
 Example-2: Accidental release scenario of NPP

 Suppose 1 Bq/sec of I-131 is released at ground level from a reactor

building whose cross sectional area is 1600 m2. Find out downwind,
centreline GLC at 1 km downwind and deposited activity for the
worst meteorological conditions ( i.e F category with surface level
speed of 2m/sec).

 Assume deposition velocity of 0.1m/s

 Formulae used:
GLC = Q/[u(л σy σz + Cw.A) where Cw = 0.5

53
 From the graph,
 σy and σy for F category at 1000m is 38m and 15m
respectively
 u = 2m/s
 Q = 1 Bq/sec
 A = 1600 m2

 GLC = 1.93 E -4 Bq/m3

Deposited Activity = GLC x deposition velocity

 = 1.93 E -4 Bq/m3 x 0.1 m/sec = 1.93 E-5 Bq/m2.

 In actual practice, GLC should be corrected for plume decay

and then used for estimation of deposited activity.

54
Typical Exposure Pathways

55
 RADIONUCLIDE
RELEASE

ATMOSPHERIC SURFACE WATER GROUNDWATER

TRANSPORT TRANSPORT TRANSPORT

TERRESTRIAL
PATHWAY

REFERENCE MAN
- EXPOSURE

INTERNAL &
EXTERNAL
DOSIMETRY
Typical components
of Radiological HEALTH EFFECTS
Impact Assessment
IMPACT ASSESSMENT
----------
56
REGULATORY STANDARDS
 Exposure pathways

Behaviour of effluents released to the atmosphere 57

Radiation exposure from an atmospheric releases could result in:

• External exposure
– Direct radiation from plume (immersion)
– Deposited radio nuclides on surfaces (ground shine)

• Internal Exposure
– Inhalation of radionuclides in air
– Ingestion of foods, contaminated by radio nuclides

58
59
 Exposure pathways and dose assessment
Accumulation of contaminants in human and other biological tissues plays a
key role in contributing to overall dose and risk resulting from environmental
releases of radionuclides.

60
Evaluation of physical and biological transport processes is essential for
environmental dose and risk assessment.

Processes related to Terrestrial food chain pathways:

 Direct deposition on to plant surfaces
 Reduction of radionuclide concentration from surfaces of vegetation
 Deposition on soil
 Reduction of radionuclide concentration in the soil surface
 Uptake from soil by edible portions of vegetation and the implicit
assumption of inadvertent soil ingestion
 Intake of radionuclides by animals and transfer to milk and meat
 Ingestion of milk / meat by human

Radionuclides discharged into the aquatic environment are also passed along the
aquatic food chain and may eventually reach humans.
61
 During normal operating R*

conditions, Radionuclides R*
R*
are released through high
Stack after treatment R*

R*

R*

62
The radionuclides may get
deposited on soil and R
vegetation surfaces by wet * R
*
and dry deposition

R R
R * *
R
R *
*
*

Wet
Dry
Deposition
Deposition

63
 Dry
Deposition

Radionuclides may
directly get deposited on
grass surfaces by dry
deposition
64
Foliar Dry
Deposition

Or get deposited on the

leaves of plants or trees
65
 Root
Uptake

66
Assimilation And then get incorporated
in leaf and other edible
parts of the plant.

67
Animal Radionuclides enter to
Uptake grazing animals through
grass and finally appear in
their milk and meat

68
 Ingestion
by Human

Enters to human body

through intake of plant
and animal products

69
 Exposure pathways and dose assessment

Once the concentrations in various environmental

media are estimated, dose to human can be computed.

 Safety Report Series no. 19 of IAEA describes the procedure

to be followed for individual dose calculation.

 Basic equations for calculation of following doses from

airborne radionuclides are presented in following slides:
 Immersion dose
 Inhalation dose
 Ingestion dose
 Dose from ground deposition

70
 Exposure pathways and dose assessment
The annual effective dose from immersion in the atmospheric discharge plume
Eim (Sv/a) is given by

Eim = CA Dfim Of
Where,
CA = annual average concentration of nuclide ‘i’ in air (Bq/m3)
DFim = effective dose coefficient for immersion (Sv/a per Bq/m),
Of = fraction of the year for which the hypothetical critical group member
is exposed to this particular pathway.
---------------------------------------------------------------------------
The annual effective dose from inhalation Einh (Sv/a) is:

Einh = CA Rinh Dfinh

Where,
Rinh = inhalation rate (m3/a),
DFinh = inhalation dose coefficient (Sv/Bq)
71
 Exposure pathways and dose assessment
The ingestion doses for infants and adults are calculated using the following general
equation:
Eing, p = Cp,i Hp DFing
where
Eing, p = annual effective dose from consumption of nuclide ‘i ‘ in foodstuff ‘p ‘
(Sv/a),
Cp,i = concentration of radionuclide ‘i’ in foodstuff ‘p’ at the time of consumption
(Bq/kg),
Hp = consumption rate for foodstuff ‘p’ (kg/a),
DFing = dose coefficient for ingestion of radionuclide ‘i’ (Sv/Bq).
--------------------------------------------------------------------------------------------
The annual effective dose from ground deposition Egr (Sv/a) is given by

Egr = Cgr DFgr Of

DFgr = dose coefficient for exposure to ground deposits (Sv/a per Bq/m2),
Cgr = deposition density of radionuclide ‘i’ (Bq/m2).

Cgr is obtained from the ground deposition rate. 72

 Exposure pathways and dose assessment

•Recommended values of different dose

coefficients are given the Safety Report
Series no. 19 of IAEA .

•Subsequent to the dose assessment,

epidemiologic analyses are to be carried
out to evaluate potential health effects of
human exposure to radioactive materials.

73
Thank you

74