See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/260186166

Testing of an urban Lagrangian dispersion model using Gaussian and non-

Gaussian solutions

Conference Paper · January 2004

CITATIONS READS

2 65

4 authors, including:

Eric R. Pardyjak Michael J Brown

University of Utah Los Alamos National Laboratory
215 PUBLICATIONS   2,221 CITATIONS    167 PUBLICATIONS   1,965 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Distributed Power Generation and Urban Energy Systems View project

The MATERHORN Project View project

All content following this page was uploaded by Michael J Brown on 15 February 2014.

The user has requested enhancement of the downloaded file.

NOTE: Use these buttons to print or save the form. DO NOT use the browser tool bar. SAVE PRINT

LA-UR-
Approved for public release;
distribution is unlimited.

Title:
TESTING OF AN URBAN LAGRANGIAN DISPERSION
MODEL USING GAUSSIAN AND NON-GAUSSIAN
SOLUTIONS

Author(s):
Balwinder Singh, University of Utah
Michael D. Williams, Los Alamos National Laboratory
Eric R. Pardyjak, University of Utah
Michael J. Brown, Los Alamos National Laboratory

Submitted to:
Symposium on Planning, Nowcasting, and Forecasting
in the Urban Zone, 2004 AMS Annual Meeting
Seattle WA, January 2004

SAVE PRINT CLEAR FORM

Los Alamos National Laboratory, an affirmative action/equal opportunity employer, is operated by the University of California for the U.S.
Department of Energy under contract W-7405-ENG-36. By acceptance of this article, the publisher recognizes that the U.S. Government
retains a nonexclusive, royalty-free license to publish or reproduce the published form of this contribution, or to allow others to do so, for U.S.
Government purposes. Los Alamos National Laboratory requests that the publisher identify this article as work performed under the
auspices of the U.S. Department of Energy. Los Alamos National Laboratory strongly supports academic freedom and a researcher’s right to
publish; as an institution, however, the Laboratory does not endorse the viewpoint of a publication or guarantee its technical correctness.
Form 836 (8/00)
J 6.6 TESTING OF AN URBAN LAGRANGIAN DISPERSION MODEL
`
USING GAUSSIAN AND NON-GAUSSIAN SOLUTIONS
1 2 1 2
Balwinder Singh , Michael D. Williams , Eric R. Pardyjak and Michael J. Brown
1 2
University of Utah, Los Alamos National Laboratory

1. INTRODUCTION 2. METHODOLOGY

The QUIC-Plume (Quick Urban & Industrial 2.1 Gaussian Plume model for uniform velocity
Complex) fast response transport and dispersion code profile flow
(Williams and Brown, 2002) uses the Langevin Under certain idealized conditions (steady state,
equations (Rodean, 1996) to model the turbulent horizontal homogeneity, constant wind speed, and
dispersion of a passive scalar and estimate constant eddy diffusivity), the conservation equation
concentrations in discrete volumes. QUIC-Plume has for a passive scalar emitted from a point source yields
been designed to be used with QUIC-URB (Pardyjak a Gaussian solution (e.g., Seinfeld and Pandis, 1998).
and Brown, 2001), a 3D fast response model that The vertical and crosswind dispersion parameters for
computes a mass consistent wind field around neutral conditions were estimated according to
buildings using empirical parameterizations. Draxler’s (1976) simplification of Pasquill’s (1971)
In the work presented here, baseline calculations suggested definition, where dispersion in the
have been run to ensure QUIC-Plume’s performance crosswind and vertical directions is directly proportional
under idealized conditions. QUIC-Plume was to the product of standard deviation of the respective
compared to a Gaussian model (uniform velocity wind velocities, elapsed time and a universal function
profile, elevated release, see Fig. 1) and a non- dependent upon atmospheric boundary layer
Gaussian model (power-law velocity profile, elevated parameters (Draxler, 1976). The friction velocity, u*,
release, see Fig. 2). In addition, a nested grid needed in the QUIC-Plume model was estimated by
capability was implemented into QUIC-Plume in order comparing the Draxler horizontal and vertical
to speed up the model and still obtain good results. dispersion parameters with the standard Pasquill and
Gifford dispersion parameters for neutral atmospheric
conditions (e.g., Seinfeld and Pandis, 1998).
U
2.2 Non-Gaussian Plume model for power-law
velocity profile flow
Berlyand (1975) and Huang (1979) have shown
that the passive scalar conservation equation can be
H solved for power-law velocity and vertical eddy
Z diffusivity profiles. The non-Gaussian solution has
been shown to match concentrations distribution well
X Uniform Flow in boundary layer shear flow (Brown et al., 1993). The
crosswind spread parameters for the non-Gaussian
Figure 1:Schematic of the uniform flow model were evaluated according to Pasquill (1971).
release case. The friction velocity was calculated within the QUIC-
U Plume model using the gradient of the boundary layer
power-law velocity profile.

2.3 QUIC-Plume set up

Within the QUIC-Plume code, a passive scalar
Z H was released continuously from an elevated point
source located in the domain. For the uniform flow test
case, a domain of size 100 m in the streamwise
Boundary Layer Flow
X direction, 100 m in the crosswind direction and 140 m
Figure 2: Schematic of the power law in the vertical direction was used with a release height
boundary layer flow release case. of 70m. The plume was released at a height such that
the plume would not impact the ground within the
computational domain. This was done to avoid near-
ground reflections. For the power-law velocity profile
* Corresponding author address: Balwinder Singh, case, a domain of size 100 m in the streamwise
University of Utah, Department of Mechanical direction, 100 m in the crosswind direction and 20 m in
Engineering, Salt Lake City, UT 84112, the vertical direction was used with a release height of
[email protected] 11m.
 For the shear flow case, QUIC-Plume was run
with a power law velocity profile with an exponent of
0.15. The friction velocity was computed internally
based on the gradient of the velocity profile. For the
uniform flow test case, QUIC-Plume requires
specification of the friction velocity, i.e., it cannot be
specified from velocity gradients. As mentioned above,
u* was calculated by equating the Draxler plume
spread equation to the Pasquill-Gifford equation and
solving for the friction velocity.
In QUIC-Plume, the concentrations are computed
by counting particles in a 3D array of sampling boxes.
Grid refinement tests showed that in order to
sufficiently match near source results, fine grid .
resolutions were necessary. To efficiently capture both
the near-source and far-field concentration variations, Figure 3: Near centerline plot of
the QUIC-Plume code was modified to incorporate concentration vs. streamwise direction for
embedded grids near the source to allow for localized an elevated release in a uniform flow.
high-resolution calculations.
The QUIC-Plume simulations were performed with
an outer grid resolution of approximately 10 m X 10m
X 10m in streamwise, crosswind and vertical
directions, respectively, and a fine embedded grid near
the source of approximately 10/7m X 10m X 10m in
streamwise, crosswind and vertical directions,
respectively. 50,000 particles were released for these
simulations. The concentration calculations were then
made for the specified sampling boxes and a
concentration value was assigned to the center
location of the given box.

3. RESULTS

3.1 QUIC-Plume Evaluation using the Gaussian

Figure 4: Near source plot of
Model (uniform velocity profile)
concentration vs. crosswind direction for
To compare the Gaussian and QUIC-Plume model
an elevated release in a uniform flow.
runs, the concentration was normalized as follows:
CU h h 2
C* =
Q
where
C * = Normalized concentration
C = Concentration (gm-3)
h = Height of point of release (m)
U h = Velocity at the of point of
-1
release (ms ).
Q = Source strength (gs-1)
Figure 3 shows the near centerline (4m offset)
concentration variation in the streamwise direction.
QUIC-Plume (QP) predicts the concentration along the
centerline fairly well. It is clear that the improved
resolution from the embedded grid gives a much better Figure 5: Far source plot of concentration
estimate of concentrations near the source. Close to vs. crosswind direction for an elevated
the source, the coarse grid misses both the amplitude release in a uniform flow.
of the peak and the location.
 Figure 6: Near source plot of vertical direction vs.
concentration for an elevated release in a
uniform flow.

Figure 8: Near centerline plot of concentration vs.

streamwise distance for a power law velocity profile.

Figure 7: Far source plot of vertical direction vs.

concentration for an elevated release in a uniform
flow.

Figures 4 and 6 show the QUIC-Plume model

Figure 9: Near source plot of concentration vs.
matches the Gaussian solution well in the lateral and
crosswind distance for a power law velocity profile.
vertical directions too when using the high resolution
grid. The blue curves show that the spatial resolution
of the coarse grid is not adequate near the source. As
can be seen from Figs. 5 and 7, QUIC-Plume also The QUIC-Plume model also accurately predicts
predicts the lateral and vertical profiles well as we the lateral and vertical concentration profiles near the
move farther away from the source. source as seen in Figs. 9 and 11, respectively. It is
also clear that close to the source, fine resolution is
3.2 QUIC-Plume Evaluation using the Non- needed to obtain accurate answers. As can be seen
Gaussian Model (power-law velocity profile): from Figs. 10 and 12, QUIC-Plume also predicts the
Figure 8 compares the near centerline (approx 3m lateral and vertical profiles well as we move further
offset) concentration computed by the non-Gaussian away from the source.
model and QUIC-Plume as a function of downwind
distance. As was the case for the uniform flow, QUIC-
Plume predicts the location and magnitude of the peak
fairly well using the high resolution grid. Near the
source, the coarse grid over estimates the
concentration.
 4. SUMMARY

The purpose of this work was to evaluate the

QUIC-Plume Lagrangian dispersion model under very
idealized conditions. This was accomplished by
comparing QUIC-Plume results with the Gaussian and
Non-Gaussian plume models using uniform and
power-law wind profiles, respectively. Comparisons of
longitudinal, lateral, and vertical profiles of
concentration show that the QUIC-Plume model
matches both the Gaussian and non-Gaussian model
results well. In order to speed up calculations, a
nested concentration grid capability was incorporated
into QUIC-Plume. This fine grid allows for better
concentration estimates near the source, while the
coarse grid gives acceptable concentration estimates
far from the source. In future, we will look at near-
Figure 10: Far source plot of concentration vs. surface releases in order to better evaluate the
crosswind distance for a power law velocity profile. reflection scheme in the code.

5. REFERENCES

Berlyand, M. E., (1975) “Contemporary Problems of

Atmospheric Diffusion and Air Pollution”, Hydromet
Press, 448 pp.

Brown, M.J., P. Arya, and W.H. Snyder, (1993)

"Vertical Dispersion from Surface and Elevated
Releases: An Investigation of a Non-Gaussian Plume
Model", J. Appl. Meteor., 32,490-505.

Brown, M.J., P. Arya, and W.H Snyder, (1995) "Plume

Descriptors Derived from a Non-Gaussian
Concentration Model”, Atmos. Env. 31, pp. 183-189.

Draxler, R. R., (1976) “Determination of atmospheric

diffusion parameters”, Atmos. Env., 10, 99-105.

Figure 11: Near source plot of vertical direction vs. Huang, C. H., (1979) ” A theory of dispersion in
concentration for a power law velocity profile. turbulent shear flow”, Atmos. Environ., 13, 453-463.

Pardyjak, E.R. and M.J. Brown, (2001) “Evaluation of a

Fast-Response Urban Wind Model-Comparison to
Single-Building Wind Tunnel Data”, Los Alamos
National Laboratory report LA-UR-01-4028.

Pasquill, F., (1974) ” Atmospheric Diffusion”, 2nd ed.,

John Wiley and Sons, NY, NY, 429 p

Rodean,Howard C., (1996) ”Stochastic

Lagrangian Models of Turbulence Diffusion” ,The
American Meteorological Society, Boston, MA, pp. 82.

Seinfeld, J.H. and S.N. Pandis, (1998) “Atmospheric

Chemistry and Physics”, John Wiley and Sons, Inc.,
New York, pp.1326.

Williams, M.D. and M.J. Brown, (2002) “Description of

the QUIC-PLUME model”, Los Alamos National
Laboratory report LA-UR-02-1246.
Figure 12: Far source plot of vertical direction vs.
concentration for a power law velocity profile.

View publication stats