United States
Environmental Protection
Agency
Atmospheric Sciences Research
Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S3-84-106 Dec. 1984
&EPA Project Summary
Green River Air Quality
Model Development:
Meteorological and Tracer
Data—July/August 1982
Field Study in Brush Valley,
Colorado
C. D. Whiteman, Richard N. Lee, Montie M. Orgill, and Bernard D. Zak
Special meteorological and atmos-
pheric tracer studies were conducted
during a three-week period in July and
August of 1982 in the Brush Creek
Valley of northwestern Colorado. The
experiments were conducted by the
U.S. Department of Energy's Pacific
Northwest Laboratory (PNL as part of
the U.S. Environmental Protection
Agency's (U.S. EPA) Green River
Ambient Model Assessment (GRAMA)
project. The objective of the field exper-
iments was to obtain data to evaluate a
model, called VALMET, which is being
developed at PNL under the GRAMA
project to predict dispersion of air pollu-
tants released from an elevated stack
located within a deep mountain valley in
the post-sunrise temperature inversion
breakup period. Three tracer experi-
ments were conducted in the valley
during a two-week period. In these
experiments, sulfur hexafluoride (SF6)
was released from a height of approxi-
mately 100 m, beginning before sunrise
and continuing until the nocturnal
down-valley winds reversed several
hours after sunrise. Dispersion of the
SF6 after release was evaluated by
measuring its concentrations in
ambient air samples taken from
sampling devices operated within the
valley. These samplers ware stationed
from the source to about 8 km down-
valley. An instrumented research
aircraft was also used to measure con-
centrations in and above the valley.
Tracer samples were collected by using
a network of radio-controlled bag
sampling stations, two manually
operated gas chromatographs, a con-
tinuous SF6 monitor, and a vertical SF6
profiler. In addition, basic meteorologi-
cal data were collected during the tracer
experiments. Frequent profiles of
vertical wind and temperature
structure were obtained with tethered
balloons operated at the release site and
at a site 7.7 km down the valley from
the release site. Experiments were
conducted in cooperation with the U.S.
Department of Energy's Atmospheric
Studies in Complex Terrain (ASCOT)
program. A great deal of supplementary
meteorological data is available from
the ASCOT program, including addi-
tional tethered balloon data, data from a
network of meteorological towers,
acoustic sounder data, and data from
laser anemometers.
This Project Summary was developed
by EPA's Atmospheric Sciences Re-
search Laboratory, Research Triangle
Park. NC. to announce key findings of
the research project that is fully docu-
mented in a separate report of the same
title (see Project Report ordering
information at back).
-------
Introduction
A Atmospheric tracer experiments were
conducted in the Brush Creek Valley in
the oil shale region of northwestern
Colorado during a three-week period in
July and August 1982. This report
presents the resulting data, which were
collected to evaluate the initial version of
an atmospheric transport and diffusion
model, called VALMET, developed for
individual valleys. The VALMET model is
being developed for the U.S. Environ-
mental Protection Agency (U.S. EPA) at
the U.S. Department of Energy's (U.S.
DOE's) Pacific Northwest Laboratory. The
U.S. EPA tracer experiments were con-
ducted as a supplement to a large mete-
orological field study that was designed
by the U.S. DOE's Atmospheric Studies in
Complex Terrain (ASCOT) program.
The Brush Creek Valley is a 25-km long
valley located about 50 to 70 km NNE of
Grand Junction in northwestern
Colorado. Brush Creek is a tributary to
Roan Creek, a major valley draining the
south side of Colorado's Roan Plateau,
located at the southern edge of the
Piceance Basin. The Brush Creek Valley
is a nearly linear, unobstructed valley,
draining from NW to SE. The valley is
deep (—650 m), narrow 3 km or less
between the upper sidewalls) and, other
than short box canyons on the east side,
has no major tributaries; average
sidewall slopes are 30 to 40 degrees. The
topography of Brush Creek is unusual in
that the valley floor has a rather steep
slope and the altitude of the ridgetops
changes little with up-valley distance.
The lowest 10 km of Brush Creek has a
slope of about 14 m/km. Upvalley from
the release site, the valley floor rises
more steeply, sidewalls become steeper,
and the valley attains a "v-shaped" cross
section.
The Brush Creek tracer experiments
were designed to provide the initial data
required to evaluate VALMET. The
approach taken was to collect
meteorological and tracer data to test the
full range of meteorological assumptions
and parametenzations used in modules
within the model. For example, the model
predicts that convective boundary layers,
will grow over heated surfaces after
sunrise, that upslope flows will develop
within these boundary layers, that
pollutants from the elevated nocturnal
plume will fumigate into the convective
boundary layers, and that they will be
transported out of the valley by the
upslope flows. Thus, within the restraints
of the resources available, it was
necessary to observe the development of
convective boundary layers over the
slopes, the upslope wind systems,
fumigation of pollutants, and transport of
pollutants up the slope. This required a
continued, elevated tracer release
within the valley during periods when a
strong nocturnal temperature inversion
had formed, and observation of the
subsequent transport and diffusion of the
tracer plumes as the valley temperature
inversion broke up the following sunrise.
Multiple experiments were run during
clear weather periods with a variety of
measurement systems to record the
changing meteorological and tracer
plume structure in the valley. The
experiments focused on the plume
breakup during the short post-sunrise
inversion breakup period Good spatial
time resolution of the observations was
necessary to record features of the
inversion breakup adequately. Manually
operated portable gas chromatographs
and a continuous tracer gas analyzer
were used to provide this time resolution
Good spatial resolution of the
instruments was necessary on a valley
cross section to view the expected
convective boundary layer and tracer
plume structure. To meet this need, a
network of surface-based bag samplers
was located throughout the valley,
including the valley sidewalls. Vertical
profiles were made through the elevated
plume using a vertical sulfur hexafluoride
(SF6) profiler balloon-borne sampling
device. A continuous tracer gas monitor
was operated from an aircraft to momtoi
tracer gas concentrations in the upper
valley atmosphere. Finally, tethered
balloon systems were used to make
observations of the changing
atmospheric structure within the valley.
This report describes the experimental
design and presents the meteorological
and tracer data collected in the U.S. EPA's
tracer experiments conducted in the
Brush Creek Valley of Colorado during
July and August, 1982. First, recom-
mendations for future work are
presented. Next is an initial evaluation of
the VALMET model. Then, the experi-
mental design is discussed, including
information on the topography of Brush
Creek Valley, the types and location of
instrument systems used, and the
weather conditions encountered during
the field experiments. A chapter is
provided on each of the data collection
and analysis systems, including the
tracer release system, the mobile
analysis laboratory, the bag sampling
system, the vertical SF6 sampling system,
the tethered ballon data collection
system, the portable gas chromatograph
system, the continuous tracer gas
analysis system, and the aircraft data
collection system.
Special features of the SF6 tracer data
set include:
• use of a vertical SF6 profiling system
to determine how the vertical
structure of the SF6 plume varied
with time
• extension of the bag-sampling net-
work to include tracer observations
high (150 m)on the valley sidewalls
• use of portable gas chromatographs
and SF6 monitors to observe rapid
variations in tracer concentrations
that occur during the post-sunrise
period when fumigations of the
elevated nocturnal plume occur on
the valley sidewalls
• use of a research aircraft to deter-
mine how pollutants are dispersed
into the upper reaches of the valley
following sunrise
Conclusions and
Recommendations
There were several advantages to
choosing the Brush Creek Valley for the
initial evaluation of VALMET. First , the
valley has a rather simple topography.
The narrow, 25-km-long valley has no
major changes in valley orientation along
its length. It has nearly equal sidewall
inclinations. The valley drains a plateau,
so that the ridges are at a constant
altitude regardless of location along the
valley axis. The valley has no major tribu-
taries. Second, the valley axis is oriented
from NW to SE so that the sidewalls will
be exposed to quite different isolation
during the post-sunrise temperature
inversion breakup period. The effect of
this unequal heating was a major
uncertainty in the model formation On
the basis of meteorological data collected
in wider Colorado valleys, and numerical
model results, the VALMET model was
developed under an assumption of
horizontal homogeneity of atmospheric
structure on a valley cross section. This
assumption could be readily tested in the
Brush Creek Valley, where the
narrowness of the valley and the NW-
SE orientation of the valley would clearly
maximize any horizontal gradients in
atmospheric structure between the
sidewalls. Third, the Brush Creek Valley
was heavily instrumented with
meteorological sensors by the ASCOT
-------
program. Access to their meteorological
data was a great benefit to the model
evaluation effort.
Along with the above advantages,
there was a major disadvantage to
conducting an initial evaluation of
VALMET with data from the Brush Creek
Valley. This disadvantage was related to
the short segment of the valley that was
accessible for tracer instrumentation.
VALMET is a two-dimensional model,
predicting concentrations on a cross
section oriented perpendicular to the
valley axis some distance down-valley
from a source. Restrictive assumptions
are present in VALMET regarding a
required homogeneity of the temperature
and wind structure in the along-valley
direction. The Brush Creek Valley,
however, is a short tributary valley that
flows into the Roan Valley a few
kilometers below the valley cross section
where most measurements are made.
Consequently, tracer plume carried down
the Brush Creek Valley during the night is
carried into Roan Creek. Reversal of the
down-valley winds (to up-valley) after
sunrise results in a large part of the tracer
plume being carried up the Roan Creek
Valley, rather than being carried back up
the Brush Creek Valley as assumed in the
model. Evaluation of VALMET was
complicated by this violation of a major
assumption in the model, which had been
designed for longer valleys.
The nocturnal plume was carried down
the valley, as expected. The nocturnal
plume, although released above the
valley center, was found to be displaced
towards one sidewall as it was
transported down the valley. The valley is
not strictly linear, but turns slightly with
down-valley distance. Because the plume
was displaced towards the "outside" of
the turn , it is conceivable that mertial
effects were responsible for the
displacement of the plume from the valley
centerline. The nocturnal plume was
carried down the valley in a rather strong
"jet" of down-valley winds, with the level
of maximum winds at about release
height. The nocturnal model, based on
the Gaussian formulation, is incapable of
treating vertical shears in transport
winds but it approximates transport and
diffusion along the valley direction fairly
well when winds at release height are
used for transport.
Assumptions in the daytime portion of
the model were verified with actual mete-
orological and tracer data. The post-
sunrise period was characterized by the
growth of convective boundary layers
over the sunlit valley surfaces. The tracer
plume fumigated the valley sidewalls as
convective boundary layers grew
upwards into the remnants of the
nocturnal temperature inversion
containing the elevated tracer plume.
Tracer was carried from the valley by
upslope flows, which developed within
the growing convective boundary layers.
Corresponding subsiding motions over
the valley center were noted in the
temperature profiles at several of the
tethered balloon sites, but the limited
vertical resolution of the tracer plume did
not allow this feature to be seen in the
tracer concentration analyses.
Due to the northwest-southeast
orientation of the deep, steep-walled
valley, very significant differences
occurred in the timing and rates of
convective boundary layer growth on the
opposing sidewalls following sunrise. As
a result of the unequal heating of the
different sidewalls, a cross-valley flow
developed, carrying the elevated plume
towards the warmer sidewall. Due to the
cross-valley advection, tracer concentra-
tions were higher on this sidewall than
predicted by the model. A future modifi-
cation of the VALMET model will be
required to handle this situation properly
in narrow valleys where post-sunrise
insolation on the opposing sidewalls is
quite different. The Brush Creek tracer
experiments were the first direct experi-
mental confirmation of the importance of
this physical effect on tracer plume
dispersion
The short length of the Brush Creek
Valley, as expected, affected the results
of the tracer experiments. The primary
effect, from initial analyses, seems to be
that the tracer concentrations in the
valley fell more rapidly than expected
after the post-sunrise wind reversal. This
is thought to be due to the nocturnal
plume being carried largely up Roan
Creek after the wind reversal rather than
reversing direction to come back up
Brush Creek
The experiments described in this
report should be considered as initial
experiments designed to provide a better
understanding of the basic physics of
valley meteorology. The VALMET model
appears to have promise in predicting air
pollution concentration in deep valleys.
Further work is recommended to
complete a full analysis of the data from
the 1982 experiment, and to evaluate and
improve the VALMET model with these
data.
-------
C. D. Whiteman, Richard N. Lee, andMontieM. Orgill are with Pacific Northwest
Laboratory, Rich/and, WA 99352; and Bernard D. Zak is with Sandia National
Laboratory, Albuquerque. NM 87185.
Alan H. Huber is the EPA Project Officer (see below).
The complete report, entitled "Green River Air Quality Model Development-
Meteorological and Tracer Data—July/August 1982 Field Study in Brush
Valley, Colorado," (Order No. PB85-125 490; Cost: $ 16.00, subject to change}
will be available only from;
National Technical Information Service
5285 Port Royal Road
Springfield, VA22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Atmospheric Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
US GOVERNMENT PRINTING OFFICE, 559-016/7866
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
Official Business
Penalty for Private Use $300
------- |