EPA/600/A-97/091
For presentation at the American Meteorological Society
10th Joint Conference on the Applications of Air Pollution
Meteorology with the kVMA, Phoenix, kZ Jan. 11 - 16, 1998
7A.12 MODELS AND LABORATORY EXPERIMENTS OF BUOYANT PUFF DISPERSION
IN A CONVECTIVE BOUNDARY LAYER
J.C. Weil
CIRES, University of Colorado
Boulder, Colorado
W.H. Snyder
\
Mechanical Engineering Department, University of Surrey
Guildford, Surrey, England
R.E. Lawson, Jr.*
ASMD, ARL, National Oceanic and Atmospheric Administration
Research Triangle Park, North Carolina
R.S, Thompson
AMD, NERL, U.S. Environmental Protection Agency
Research Triangle Park, North Carolina
M.S. Shipman
Geophex, Ltd.
Raleigh, North Carolina
1. INTRODUCTION
Buoyant puffs or thermals are generated by the
sudden release of heat in the atmosphere, e.g., from
an explosion. Much is known about the rise and
spread of thermals in a nonturbulent environment
from both models and laboratory experiments
(e.g., Turner, 1979) and recent experiments have
added to our knowledge of thermals in a neutral
environment capped by a stable layer (Thompson
et al., 1998). However, there have been relatively
few studies of buoyant puff behavior in a turbulent
environment such as the convective boundary layer
(CBL). The latter is pertinent to a number of
atmospheric problems. In this paper, we present:
1) a simple model of buoyant pulf dispersion in the
CBL, 2) results from experiments on puff dispersion
in a laboratory convection tank, and 3) a brief
comparison of the two. The experiments—the
first on buoyant pufis in a convection tank—were
conducted at the U.S. EPA Fluid Modeling Facility.
* On assignment to NERL, U.S. Environmental
Protection Agency.
Corresponding author address: J.C. Weil, NCAR,
P.O. Box 3000, Boulder, CO 80307
This study is motivated by the need to dispose
of obsolete munitions and ordnance at Department
of Defense (DOD) and Department of Energy
(DOE) facilities. The most widely-used disposal
method is open burning (OB) and open detonation
(OD) in an earthen pit. Since OBOD generates
air pollutants, any facility using this method must
meet source permit requirements and demonstrate
a low risk to human health and the environment.
This requires an appropriate dispersion model
to estimate ambient air concentrations, dosage,
surface deposition, etc. In particular, estimates of
the peak ground-level concentrations (GLCs) are
required for averaging times ranging from a few
minutes to an hour.
In earlier work (Weil et al., 1996), we presented
an overview of a model being developed for OBOD
sources, which are unique in having instantanteous
or short-duration releases of buoyant material.
The model includes: 1) a uniform treatment of
dispersion as the release varies from instantaneous
to continuous, 2) puff and plume rise estimated
from entrainment models, 3) relative and total
dispersion based on similarity scaling concepts for
the planetary boundary layer (PBL), and 4) pre-
processed meteorological variables (surface heat
flux, PBL deoth. etc) for estimating mean winds

-------
and turbulence in the PBL. This paper focuses on
puff releases in the CBL because OBOD activities
are currently restricted to daytime convective
periods. However, the overall OBOD model
addresses all PBL types including stable conditions.
2. DISPERSION MODEL
The dispersion of a buoyant puff is a random
phenomenon owing to the stochastic nature of tur-
bulence in the PBL. This means that the concen-
tration observed at some downwind receptor is a
random variable and should be estimated statisti-
cally through a probability distribution. The dis-
tribution can be parameterized using an analytical
form such as a gamma probability density function
(p.d.f.) and requires two variables to character-
ize it—the ensemble-mean concentration C and the
root-mean-square (rms) concentration fluctuation
crxoy P ^ 2o\ 2o*J
k"« \ J
(i)
where Q is the pollutant mass released, U is the
mean wind speed, t is time, ht is the effective puff
height, ax and ay are the puff dispersion in the x
and y directions, crZj = Vjx/U, and zCj — wJx/U
with j = 1,2. The Aj, WJ, and 
-------
&ya —
t/( 1 + 0.5t/Tiy)l/2 arid similarly for axa.
For the CBL, we adopt Tix = Tiy = Q.7zi/w.
(Weil, 1988), where z% is the CBL depth and w. is
the convective velocity scale, and evaluate the rms
turbulence velocities, au and av, from the 
-------
0.5
0 2
N
e
-a
0.1
Ft. Lab Model
0.044 • 	
0.05 ir
0.02 Lj-l-lxJ	¦-
0.05 0.1 0.2
5
2
0.5 1
t/t.
Fig. 2. Dimensionless lateral dispersion versus
t/t..
small by comparison with field observations in
CBLs having a nonzero mean wind (crv/w. ~ 0.6;
Weil, 1988). However, the above value is more
compatible with the av/w. in zero-wind or free
convection conditions as determined from large-
eddy simulations (Schmidt and Schumann, 1989;
av/wm ~ 0.4) and earlier tank experiments (Willis
and Deardorff, 1974;  2), the data
from all heights collapse to essentially the same
curve as the puff tends to a vertically well-mixed
distribution.
By comparison with Fig. 3a, there are two
obvious differences in the concentration history for
the high buoyancy case (Fig. 3b). First, the C at
z/zi = 1 is about an order of magnitude greater
1000
PFL Cone Time Series
¦Ft, = 0.044
{X&, Z/Zj)
(0,0.1)
(0,0.25)
O (0.0.5)
O (0,1)
100 :
Model
{zjzi = 0.1)

0.1
1000
PFH Cone Time Series
Ft. = 0.70
(xft. Zfei)
(0,0.1)
(0,0.25)
-£> (0,0.5)
"O (0,1)
100
Model
(z/zi = 0.1)
CO
0.1
u -d W
0.01
3
2
4
0
1
t/t,
Fig. 3. Dimensionless concentration at four
heights as a function of t/t..
than the values within the mixed layer {z/zt < 1).
This is due to the significant puff lofting or the
maintenance of a z near Z{ (Fig. 1). Second, the
concentrations within the mixed layer are about an
order of magnitude smaller than those at the same
-------
heights for the low buoyancy case (Fig. 3a). Again,
the lower concentrations result from the significant
puff lofting for case PFH.
A preliminary comparison of the modeled
concentration history at z/zi = 0.1 with the data
is shown in Figs. 3a and 3b. For case PFL,
the model curve captures the correct overall trend.
For 0.4 < tjU < 2, the model overestimation is
probably real even though the data are scattered
and the measured acjC is typically 2 - 3 in this
time interval. The lower observed C is probably due
to the I overshoot and reduced a2, which are not
adequately replicated by the model. For the high
buoyancy case (Fig. 3b), the model also captures
the overall data trend for z/zj = 0.1 and the
correct order of magnitude of C for 0.5 
-------
TECHNICAL REPORT DATA
1. REPORT NO.
EPA/600/A-97/091
4. TITLE AND SUBTITLE
Models and Laboratory Experiments of Buoyant Puff Dispersion in a Convcctive
Boundary Layer
3.RJ
5.REPORT DATE
6.PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
'Weil, J.C., 3W.H. Snyder, 3R.E. Lawson, Jr., "R.S. Thompson, and 5M.S.
Shipman	
8.PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
'CIRES, University of Colorado
Boulder, CO
2Mechanical Engineering Depaartment, University of Surrey
Guilford, Surrey, England
3Same as Block 12
4 AMD/NERL/USEAP
RTP, NC 72211
sGeophex, Ltd.
Raleigh, NC
10 .PROGRAM ELEMENT NO.
i 1. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
National Exposure Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
13.TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA/600/9
1S. SUPPLEMENTARY NOTES
16. ABSTRACT
Buoyant puffs or thermals are generated by the sudden release of heat in the atmosphere, e.g., from an explosion. Much is known
about the rise and spread of thermals in a nonturbulent environment from both models and laboratory experiments (e.g., Turner,
1979) and recent experiments have added to our knowledge of thermals in a neutral environment capped by a stable layer
(Thompson et al., 1998). However, there have been relatively few studies of buoyant puff behavior in a turbulent environment
such as the convective boundary layer (CBL). The latter is pertinent to a number of atmospheric problems. In this paper, we
present: 1) a simple model of buoyant puff dispersion in the CBL, 2) results from experiments on puff dispersion in a laboratory
convection tank, and 3) a brief comparison of the two. The experiments-the first on buoyant puffs in a convection tank-were
conducted at the U.S. EPA Fluid Modeling Facility.	
17.	KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b.IDENTIFIERS/ OPEN ENDED TERMS
c.COSATI



18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (This Report)
21.NO. OF PAGES
20. SECURITY CLASS (This Pagej
22. PRICE
-------