REVIEW OF U. S. ENVIRONMENTAL PROTECTION AGENCY
ATMOSPHERIC SCIENCES RESEARCH LABORATORY
COMPLEX TERRAIN MODEL DEVELOPMENT PROJECT
Lawrence E. Niemeyer, Chairman
Sumner Barr
Donald Shearer
Brian Lamb
Jeffrey Weil
November 19-21, 1985
Prepared by
Research and Evaluation Associates, Inc
1030 15th Street, N.W., Suite 750
Washington, D.C. 20005
(202) 842-2200
727 Eastowne Drive, Suite 200A
Chapel Hill, N.C. 27514
(919)493-1661
-------
TABLE OF CONTENTS
Chapter Page
INTRODUCTION 1
COMPLEX TERRAIN MODEL DEVELOPMENT (CTMD) PROGRAM 2
Theoretical Considerations 3
Laboratory Flow Simulations . 6
Model Design and Development 8
Field Experiments 14
Model Testing 18
WAKE EFFECTS ON PLUME DISPERSION 20
ARCTIC NORTH SLOPE DISPERSION STUDY 22
CONCLUDING REMARKS AND RECOMMENDATIONS 24
Complex Terrain Model Development Program 24
Wake Effects Program 26
Arctic North Slope Dispersion Study 26
Recommendations 27
REFERENCES 30
APPENDIX A - Agenda 32
APPENDIX B - Peer Review Panel 35
APPENDIX C - Process Evaluation Report 37
APPENDIX D - ASRL Responses to Reviews' Comments 40
APPENDIX E - Review of the Panel Report and 46
Responses of the Laboratory Director
-------
INTRODUCTION
A peer review was held at Research Triangle Park, North Carolina
on November 19-21, 1985 to evaluate the EPA Complex Terrain Program
including three projects:
Complex Terrain Model Development (CTMD) Program,
Wake Effects on Plume Dispersion, and
Arctic North Slope Dispersion Study.
In order to effectively review the work, a number of reports were
provided in advance for preview and a series of presentations were
made during 1 1/2 days by several of the investigators. The preview
reports included:
EPA Complex Terrain Model Development,
, First Milestone Report - 1981, Lavery et al., 1982, EPA
- 600/3-82-036, April 1982;
, Second Milestone Report - 1982, Strimaitis et al., 1983,
" EPA - 600/3-83-015, April 1983;
, Third Milestone Report - 1983, Lavery et al., 1983, EPA
- 600/3-83-101, November 1983;
, Fourth Milestone Report - 1984, Strimaitis et al., 1985,
EPA - 600/3-84-110, February 1985.
EPA Complex Terrain Model Development, Description of a Computer
Data Base from Small Hill Impaction Study No. 1, Cinder Cone
Butte, Idaho, L. E. Truppi and G.C. Holzworth, 1985.
Scientific Assessment Document on Status of Complex Terrain
Dispersion Models for EPA Regulatory Applications, Schiermeier,
F. A. 1984, EPA - 600/3-84-103, November 1984.
Evaluation of Method for Estimating Pollution Concentrations
Downwind of Influencing Building, Huber, A. H., Atmospheric
Environment, 18, 2313-2338, 1984.
-------
Presentations were given by F. Schiermeier, A. Huber, P.
Finkelstein, G. Holzworth, B. Greene, L. Truppi, W. Snyder, T.
Lavery, and D. Strimaitis.
COMPLEX TERRAIN MODEL DEVELOPMENT (CTMD) PROGRAM
The CTMD program has been directed toward developing a model for
predicting concentrations of plumes impinging on elevated terrain
during stable atmospheric conditions. The selected scenarios are
known to produce high concentrations on elevated terrain and are also
the ones for which existing regulatory models VALLEY, COMPLEX I
and II have been highly criticized; the latter models have also
been shown to perform poorly under stable conditions (White, 1985).
Thus, a program designed to deliver a theoretically improved and
better performing model to estimate impingement concentrations is
clearly a worthwhile objective.
With the above objective in mind, the model development program
began with a theoretically-based, conceptual picture of the ambient
flow field. The conceptual picture was verified by laboratory
simulations, which provided much additional detail on the nature of
the flow. In particular, they illustrated the dependence of the flow
field on stratification and hill geometry. These simulations were
conducted about impinging plumes and their diffusion at isolated
three-dimensional hills. Based upon a theoretical foundation,
information from the simulations, and an analytical approach to the
diffusion problem (Hunt et al., 1979), a field program was designed
to verify the conceptual picture and the laboratory simulations for
-------
full-scale terrain features. The field program was designed to
provide quantitative data for testing the flow model and for
developing/verifying a complex terrain diffusion model. We commend
EPA for their foresight in conducting a planning workshop in 1979 to
aid in the establishment of the CTMD program. In addition, we
believe that a good balance has been maintained in the program
between theory, mathematical modeling, laboratory simulation, and
field experiments.
In the following section, we will discuss the key aspects of
this program: theory, laboratory simulations, model design and
development, field experiments, and model testing.
Theoretical Considerations
There are two major theoretical components that form the basis
of the CTMD program: the flow field about a hill, especially under
strongly stable stratification, and the diffusion of a plume in that
flow. The first theoretical prediction of the flow was by Drazin
(1961) who characterized strongly stable stratification by the
condition F < < 1, where F is the Froude number defined as:
F * -
Nh '
Here, u is the mean wind speed upwind of the hill, N is the
Brunt-Vaisala frequency, and h is the hill height. Drazin found that
the primary flow was irrotational (i.e., potential) in horizontal
planes about the hill, except for a small vertical region of order
u/N near the hill top.
-------
In a systematic set of experiments, Hunt, Snyder and Lawson,
(1978) confirmed this theory and found that one could essentially
divide the flow into a lower region, which was horizontally layered,
and an upper region in which fluid passed both over and around the
hill. The surface dividing these two regions has been termed the
"dividing streamline height" (Hc), and was found to be given
approximately by Hc = h (1-F). Hc is a key variable describing
the flow field and, consequently, an essential parameter of the
developed Complex Terrain Dispersion Model (CTDM).
The theoretical basis of the dispersion component of CTDM is the
diffusion equation analysis introduced by Hunt and Mulhearn (1973),
and later extended by Hunt et al. (1979). The latter analysis
considered diffusion in a deformed flow about an axisymmetric hill,
where the flow was described by potential theory. In the analysis,
diffusion occurred across streamlines which may be converging or
diverging. Two stratification limits were investigated, and
analytical results for the concentration field were provided for
each: 1) F < < 1 with diffusion taking place in the horizontally
flowing layer (below Hc; two-dimensional potential theory); and 2)
F > > 1 with diffusion taking place in the fully three-dimensional
flow about the hill (three-dimensional potential theory).
Under the CTMD program, the above analysis has been generalized
to potential flow about ellipsoidal obstacles. Thus, the model,
CTDM, can accommodate a variety of hill aspect ratios - width to
height.
-------
In summary, we believe a rational and fundamental theoretical
foundation exists for CTDM and has provided the proper framework for
developing this model.
Insofar as theoretical extensions are concerned, we believe a
logical one is an adequate flow field description for moderate to
weak stratification, F >^ 1. This is necessary not only when F _> 1
but also is required to properly describe the flow above Hc when
the "cut-off" hill approximation is used. At present, potential flow
is used in both of the above circumstances, and it underestimates the
streamline deformation that occurs; i.e., the stable stratification
enhances the deformation. Use of potential flow could possibly
explain the CTDM underprediction of concentrations during neutral and
weakly stable conditions at Cinder Cone Butte (Strimaitis, et al.,
1984; Table 8 page 76).
Linearized theories of stratified flow about hills have been
developed, and one by Smith (1980) was examined in the Fifth
Milestone Report (1985). Smith's theory is a linearization with
respect to hill slope (i.e., the slope must be small). D. Cristofaro
et al., (1985) found that Smith's model did reasonably well in
predicting vertical streamline displacement over a hill at F ~ 2, but
underestimated the lateral streamline displacement. One possible
cause for this is the separated flow downstream of the hill; this
should be examined. We believe Drazin's (1961) linearization with
respect to stratification should also be applied to this Froude
number regime because it can accommodate any hill slope.
-------
Laboratory Flow Simulations
The laboratory simulations conducted at the EPA Fluid Modeling
Facility represent innovative and insightful contributions to the
stable plume impingement program. This work is one leg of the triad
of fundamentals that support the model development, along with theory
and field experiments. Stratified flow experiments in the Fluid
Modeling Facility stratified towing tank served to demonstrate the
validity of the dividing streamline concept described by Sheppard
(1956) and Drazin (1961), and to form the basis for the model and
field effort.
The laboratory work also identified some limitations of the
theory and prompted expansion of the theoretical concepts. For
example, the existence of an upwind recirculation zone was first
identified in the laboratory. Also, the dependence of the dividing
streamline formulation on the angle of attack between the mean wind
and a ridge of moderate length was reported by Snyder et al. (1982).
The work guided the design of the field experiments as a stepwise
progression in scale from the smaller, more readily simulated field
prototypes such as Cinder Cone Butte to larger, more complex domains
such as the Tracy power plant site.
In addition to their role in the design phase, the laboratory
experiments have entered into the post-analysis of field data.
Cinder Cone Butte was simulated in the towing tank, and demonstrated
a strong sensitivity of the path of the material below the critical
dividing streamline that went around the hill to small changes in the
oncoming wind direction. In view of this, hourly mean concentrations
6
-------
of tracer were simulated by six tows with the hill rotated to
correspond with a particular 10-minute average wind direction. This
technique yielded good agreement with the field results. The
documentation of the lee side separation point as a function of
stability offers guidance to extending the validity of the CTDM
beyond the ridge in some cases.
In addition to specific results, the laboratory flow visualiza-
tion capabilities and close control of flow quantities for parametric
studies promotes the development of insight into the mechanisms of
terrain flows. This has been well applied in the main thrust of the
CTMD program to date and should continue to be a focus of the
laboratory work.
We suggest that future laboratory simulation activities include
investigating internal wave effects for lee side separation,
providing guidance on partitioning wave and turbulence energies in
field data, and studying vertical motion structure over more complex
geometries such as valleys. We also suggest the utility of
laboratory experiments in the domain of weak to moderate
stratification (Froude numbers of 1 to 2) for different obstacle
geometries. A general recommendation is to continue to use the
laboratory facility to document as many flow properties as are
appropriate to simulate and, through this step, help set priorities
for future program activities. There have been several points of
interaction between the laboratory simulation and the other program
elements and we believe this interaction is very important. In the
interest of continuing this relationship, more cooperative laboratory
7
-------
experiments should be designed and conducted jointly with fluid
modeling facility and other program personnel.
Model Design and Development
From the initial stages of this program, the design of the model
was pursued with the dual objective of realistically simulating plume
transport around or over a hill while realizing a model that was
computationally efficient for regulatory purposes. In the view of
the panel, this goal has not been compromised. The physical basis
for the initial modeling concepts has been confirmed through the
fluid modeling and field study programs.
The design of the current model incorporates the dividing
streamline concept through the LIFT and WRAP models in a
straightforward and reasonable approach. Terrain obstacles are
treated as best-fit ellipsoids and plume streamlines are obtained
from potential flow theory for these ellipsoids. While actual
terrain shapes are not modeled exactly, the fitted ellipsoids are an
approximation to the real world and plume streamlines in the model
seem to duplicate the results from lab and field tests.
Diffusion rates of plumes transported along the streamlines are
initialized from measured turbulence intensities in a manner that is
consistent with current theory for near-field diffusion. It appears
from the Tracy full-scale study, however, that vertical diffusion
rates, based upon unfiltered turbulent intensities, are
overestimated. Model concentrations, therefore, underestimate
observed concentrations. This may be related to the contributions of
wave motions in the approach flow. Further work is needed to examine
8
-------
the wave effects and a method for treating this problem must be
developed before the model can be applied in real situations.
The vertical dispersion parameter, az, at large times (t) is
assumed to have a t^/2 dependence consistent with statistical
theory and diffusive behavior. However, there is a competing
theoretical formulation (Pearson et a!., 1983) that predicts a
constant QZ at large t. The latter formulation appears to be
equally consistent with the lidar observations at large times and
therefore, should merit further consideration.
The Lagrangian time scale, x|_yť for the lateral turbulence
component is assumed to be infinitely large based on the analysis of
some data from the CTMD program, at relatively short range, (x < 5
km) and from some other experiments; however, there are field data
from some other sites that show a Oy tl/2 at large times in
stable conditions. The ay vs time dependence becomes more of an
issue at long range (x > 5 km), i.e, for hills further distant from
the source than investigated in either of the small hill studies or
the Tracy experiment. The panel feels that the parameterization of
the Lagrangian time scale merits much further consideration and is a
variable that may benefit from local measurements.
The flow field modeling presently addresses a single isolated
hill. However, in real complex terrain, there will always be other
obstructions and some may be "nearby". Other hills become especially
important in strongly stable conditions and for flow below Hc where
two-dimensional potential flow is used. The latter flow field is
especially sensitive to the assumed lateral boundary conditions. If
9
-------
other obstacles are present, they can lead to slight changes in wind
direction and speed, which can produce dramatic concentration changes
on an "isolated" hill. Perhaps the effects of other hills could at
least be explored in sensitivity analyses, during the remainder of
the CTMD program.
Turbulent diffusion of a plume in the distorted flow over or
around an obstacle is simulated through the development of
terrain-averaged correction factors applied to the flat terrain wind
speed and diffusion parameters. These factors account for obstacle
shape, atmospheric stability, and the plume streamline receptor
geometry. The use of average factors in place of the line integrals
present in the original plume theory was recognized by the panel as a
necessary simplification for computational efficiency. However, it
requires further theoretical justification and demonstration as to
its equivalence with the line integral approach. It is recommended
that the sensitivity of the model output be documented with respect
to the use of the average correction factors.
It is a positive feature that the model explicitly incorporates
wind direction variability. As the source - receptor distance
increases, wind direction changes become the dominant factor in
determining the extent to which a narrow plume impinges upon a
particular terrain obstacle. If it is assumed that impingement
physics are similar for obstacles near or far from the source, then
the model can be extended to larger scales through simple
incorporation of the correct wind direction frequency distribution.
However, greater uncertainty in predicted concentrations and "misses"
10
-------
can be expected for hills at larger distances because of the greater
sensitivity to wind direction.
The model design also explicitly accounts for portions of a
plume transported above and below Hc in a mass consistent manner.
While the approach is physically plausible, the results of plumes
released very near Hc should be carefully analyzed to confirm this
approach. In particular, the thickness of the transition region near
Hc has not yet been documented.
The number of adjustable model parameters appears to be
relatively small, which is a positive feature of the model design.
Values of these parameters or methods to determine the parameters
require further work. This is probably best accomplished through a
combination of model performance vs. field data tests and model
sensitivity analyses.
The model employs the distance to separation of streamlines
(xsep) in tne 1ee °f tne n1^ in tne calculation of the terrain
correction factors. The specification of xsep is based upon
results from fluid modeling studies. In contrast to the Hc
concept, the utility of the rules governing xsep have not been
fully confirmed in the atmosphere. Following a sensitivity analysis
of xsep, confirmation is still important and should be obtained
through careful analysis of the available field data.
Based upon the presentations, further use of the digitized
trajectories from the fluid modeling results for different stability
conditions should be fully exploited in the final development of the
model. As the model exists, flow over the hill is traced as neutral
11
-------
flow and no compression of streamlines, due to stratification, is
provided. The fluid modeling results should be used to improve the
model in this respect.
A major weakness in the model appears to be in the treatment of
low-level plumes in blocked flow, upwind of a ridge, i.e, Hogback
Ridge. Unfortunately, guidance from fluid modeling is not available
for this case so improvements must be derived from the field data.
However, given the high concentrations associated with this
situation, the panel strongly believes further efforts should
emphasize the blocked flow case.
Early indications from new users of CTDM suggest the
presentation of the modeling concepts could be improved and the model
could become more user friendly through the addition of a plume
trajectory processor. This feature will be useful for trouble-
shooting in model comparisons with field data or with other, possibly
more sophisticated models. Other diagnostic tools which provide
vertical and horizontal concentration distributions at various
downwind distances would also be very useful.
With respect to model applications, the panel has the following
concerns:
The present version of CTDM contains considerable
mathematical and numerical detail, especially when compared
to other applied dispersion models. We feel it was proper
to develop the model in this way, but it may now be
necessary to simplify CTDM somewhat. There are three
reasons for this: 1) to highlight the essential physics
from the less important detail; 2) to make applications by
the user community easier; and 3) to reduce the
computational time. Furthermore, the present model
algorithms (e.g., the COMPLEX I and II plume path
descriptions), is not so overwhelmingly better as to
justify the present numerical detail. A sensitivity
12
-------
analysis should be conducted to aid in the simplification
of the model.
Some guidance will be necessary to aid the user in specify-
ing the input data, especially in determining the
parameters (ellipsoid axes - a and b) describing the hill
nf mnrorn
of concern
Guidance will also be necessary in determining what input
data to use especially to calculate u, the temperature
gradient, Hc, and the turbulence parameters -- in the
absence of measurements at plume elevation or near the
terrain elevation.
Some guidance will be necessary to assure that the model is
not misapplied by use of nonrepresentative wind
observations in a domain of mesoscale variability common in
areas of terrain complexity. For example, a serious
challenge made to the Valley model in the late 70's
centered not on the model itself but on the application
wherein it was driven by wind observations from a tower
mounted on sloping terrain on one side of a broad valley.
The plume was assumed to traverse the valley under the
stability conditions associated with drainage flow in the
valley and impact the high terrain directly across the
valley from the source. This was an unrealistic scenario
since subsequent data on the opposite side showed a
downslope there as well. The confluent slope winds in fact
turned and flowed down the valley axis. The fine
additional work incorporated in CTDM to address the physics
of impaction can be invalidated in the application step if
the user does not apply reasonable principles of
terrain-influenced mesoscale wind fields.
Future model development activities could benefit from the
exploration of alternatives to the Gaussian plume. Particle-based
dispersion modules offer a flexibility in problems of geometric
complexity that can stymie plume models. The panel suggests that,
while it is probably beyond the scope of the present program to
develop such techniques, some attention be given to alternative
dispersion modules. This can be done economically through
coordination with scientists who are working in this area. Questions
13
-------
of computational requirements and other limitations may preclude the
incorporation of sophisticated particle dispersion modules, but they
may offer valuable complementary information.
As this model development reaches completion, the panel believes
it becomes even more important to interact with other groups modeling
complex terrain. Specifically, it appears that a separate task could
be created with the objective of direct interaction of the CTMD
program with other similar programs.
Participation by the model user community in the development of
the CTDM is an important facet of the program. It is recognized that
there has been close coordination with the user community, including
the private sector as well as other programs within EPA, e.g., OAOPS
and the EPA Regional Offices. In addition, we understand that as the
model is nearing completion it will be tested by the potential users
in the agency and by the Electric Power Research Institute. We
applaud this kind of coordination, for it is only through use that
the final model product can be tailored to meet the actual
requirements of the user community.
We also recommend further testing of the model to include
careful comparison with other possibly more sophisticated models,
using the different models' data archives. This could best be
implemented through a modeling workshop with invited participants
required to exercise their models on specific data sets.
Field Experiments
The field experimental portion of the CTDM program stands out as
a very good example of coordination between the needs for evaluation
14
-------
and measurement design. The design of the experiments is based on
both theoretical considerations and the findings made throuqh
laboratory simulations. Having well-specified the measurement
requirements needed to generate data for model development, the field
experimental plan followed with well defined and narrow specific
objectives, thus yielding a well focused experimental plan.
The field experimental design was finalized after the contractor
conducted preliminary or abbreviated field measurements at each of
the three field sites. This too provided a mechanism to refine the
experimental design based on specific field experience.
The actual design, which incorporated employment of two
atmospheric tracers as well as visual smoke tracer and the newer
remote sensing devices, optimized the amount of data that could be
generated during the field phase of work. Additionally, this use of
remote sensors has led to valuable insight and detail about the
spatial and temporal variability of the physical processes of plume
behavior. Ultimately, such insight serves as a visual standard
against which one can compare the model.
Quality assurance included audits of the raw measurements and
tracer assay as well as audits of the meteorological measurements.
Post experimental review and cross checks among the data were
conducted in an extensive and rigorous manner. All of this should
result in high quality data and increase confidence in the resulting
developments.
The distance extent of the Cinder Cone Butte experiments was to
about 2 km from the tracer source. The extent of the Hogback Ridge
15
-------
experiment was to 1 or 2 km, while for the Tracy experiments it was
to a maximum of about 5 km. While these distances are of primary
concern, the regulatory application will surely extend to greater
distances. Therefore, the most logical next step would be to
consider greater distances and to conceive ways that limited field
experiments or laboratory simulations can be employed to represent
plume behavior at longer distances.
The field experiments were designed with primary emphasis on
documenting plume impingement on the front face of the terrain
feature. While this initial impingement is of high concern, other
work has illustrated that lee-side concentrations can exceed
impingement concentrations observed on the windward side at least
during some conditions (Smith, T. B., Diffusion Study in Complex
Mountainous Terrain, 1965). Regulatory concern will surely extend to
include lee-side impingement concentrations and therefore further
work in that area should be a high priority.
The field experiments conducted at the Hogback Ridge and, to a
lesser degree, those conducted at Tracy, illustrated that emissions
made below the critical streamline Hc can and do pool in the
upstream area of the terrain feature. Further investigation of the
upstream pooling should be conducted because, when such pooling
occurs, the plume concentration in these regimes can reach
comparatively high values. Regulatory concern surely will include
such phenomena.
At the current stage of the program, initial data analyses have
been completed. It is fully realized that further analyses will be
16
-------
conducted. Such follow-on analyses should include a comparison of
the ground level concentration "footprint" from the SF5 with that
of the freon tracer. Since the emission points for the two tracers
were located with a small distance separation, this kind of
comparison should give some detail about the expected spatial
variability and in turn could be employed as information about the
inherent variability one should expect within the field data.
In the Cinder Cone Butte experiments and those conducted at the
Hogback Ridge, non-buoyant plumes were studied, while at Tracy
limited plume buoyancy was introduced as the plant was operated at
reduced loads for the experimental period (typically 10 to 25%
capacity of the 120 MW unit). Other work has shown that plume rise
from buoyant plumes is reasonably well predicted by current
techniques during stable conditions. However, these techniques do
not adequately represent buoyant plume rise during either neutral or
unstable conditions and thus this limitation became a major source of
error in model predictions (EPRI PMV & D Diagnostic Model Valida-
tion). As the CTMD program progresses to include concentration
predictions during neutral and unstable conditions, additional effort
will be needed to develop better plume rise algorithms for those
conditions and a broader range of plume buoyancy will need to be
represented than has been done to date.
The panel looks forward to the results of the planned effort in
complex terrain dispersion climatology. As outlined in a briefing
during the review, it will highlight an empirical analysis of the
ground level concentration patterns and their dependence on release
17
-------
configuration and meteorological conditions. This type of analysis
will be a valuable adjunct to the overall interpretation of the field
experiments.
The data sets from each of the field experiments are being
archived on magnetic tape for distribution to independent requesters.
A procedure has been established for this dissemination and a
document describing the tapes is available to accompany the data.
This is a valuable step in enhancing the ultimate use of the data
since well-designed field experiments often benefit from innovative
analyses by independent users years after the projects are
completed.
Model Testing
Model testing is part of the objective of CTMD and there is a
vigorous task underway toward this end. The program uses the
procedures recommended by an EPA-AMS workshop as well as some
procedures developed by program participants. Model testing is a
very active area of research at the present time with new heuristic
and statistical perspectives being developed continually. Model
testing programs (EPRI) and workshops (DOE/Savannah River Lab) should
be acknowledged and some of their recommendations adopted. Model
skill is the acid test that will determine the ultimate enthusiasm
for compliance so the testing methods should be objective, rational,
simple to understand, and should point the way to model improvement
(e.g., sensitivity of skill score to various inputs and model
properties).
18
-------
There appear to be certain sensitive parameters for which a
small error in input quantities yields large departures from
observation. One example is wind direction in the typical field
experiment with a moderate to low density of point samplers. Often,
because plumes are narrow at short travel distances, a small error in
trajectory can account for a "hit" or "miss" at a point sampler,
making it often impossible to distinguish between a good prediction
that is mislocated by a few meters or a totally incorrect model. The
problem is compounded in stable flow about obstacles by the observed
sensitivity of plume path to very small changes in angle of attack of
the ambient wind.
The panel urges the thoughtful design of sensitivity analyses of
the CTDM in order to wisely use the limited time available on the
project. A careful selection could save a lot of work and time over
an exhaustive testing of the full matrix of model parameters.
Model tests should be guided by insights of time and space
continuity of the tracer plume and the mechanics of stable plume
aerodynamics (e.g., why are the bad points bad?). In performing
these analyses, project personnel should remain open to the possi-
bilities of as yet un-modeled phenomena (e.g., density currents, non-
homogeneous turbulence, non-homogeneous surface energy budgets,
etc.).
Every effort should be made to seek out independent tracer
experiment data sets to use in testing CTDM. The planned (February
1986) workshop is a first step in this direction.
19
-------
Another aspect of model testing, which will be indirectly
addressed through the workshop, is quality assurance of the actual
model code. The probability for either logic errors or typing errors
increases as the length of the model programs increases. A OA effort
is required to minimize errors in the model.
At some point in the evaluation of the model, the model
performance should be discussed in the context of the overall
uncertainty of the field data.
WAKE EFFECTS ON PLUME DISPERSION
In addition to the primary work on Complex Terrain Model
Development, two other projects funded under the Complex Terrain
program were reviewed. The first of these, Wake Effects on Plume
Dispersion, was shown to be of significant interest to the needs of
the Environmental Protection Agency in providing guidance in matters
pertaining to Good Engineering Practice in the placement, design,
height, and use of chimneys near buildings. Moreover, it was
demonstrated that the plume dispersion, in the wake of surface
obstacles, also has a practical objective beyond the scientific
desire to improve the equations used to estimate dispersion. It
would be difficult to tabulate the many problems found by responsible
authorities in urban areas that require the information being
developed under the Wake Effects Project. However, caution is
recommended. While there are a great number of problems requiring
solution, there is not an overwhelming requirement to study the
problem at length. Engineering or best estimates are sufficient to
meet the requirements of decision makers. This is not to say that
20
-------
all facets have been considered. For example, the influence of
multiple buildings or multiple building complexes on surface
dispersion appears to be an area that has received little attention.
Judging from the literature citations in the recent publications,
there is a reasonable knowledge of previous work in this field.
However, recent work on the modification of the ISC model by ERT may
have been overlooked and needs to be reviewed for pertinent
information.
Future work plans appeared to emphasize tests of the effects of
building scales vs. boundary layer scales. However, the panel
believes this type of work has received considerable attention in
past studies. We suggest that there are other areas which deserve
greater emphasis. Specifically, most building diffusion problems
involve buoyant plumes and/or momentum jets. The recent
modifications to ISC by ERT have addressed the interaction of plume
rise and building downwash in a simple but realistic fashion. This
type of modeling should be tested in the wind tunnel.
A second area in need of experimental and theoretical work is
the effect of building orientation upon downwind diffusion. The
desired end product is a diffusion algorithm which incorporates the
incident flow angle. A third area of concern is the extension of the
existing algorithm to stable conditions. Fluid modeling experiments,
possibly using the water tow tank and coupled with field
measurements, may address this question.
The video imaging technique under development in this task
appears to have the potential to become a powerful tool and it
21
-------
deserves further attention. It should continue to be applied to the
areas mentioned above. As part of these applications, the technique
can be used to examine the nature of concentration fluctuations near
buildings. In cases of the release of toxic materials from
buildings, peak concentrations or the probability that threshold
concentrations are exceeded are the factors of most concern.
In summary, the work under the project titled, Wake Effects and
Plume Dispersion, is of interest to the scientific community, has
applications to the needs of EPA, and is of good scientific
quality.
ARCTIC NORTH SLOPE DISPERSION STUDY*
The second project reviewed under the Complex Terrain Program
was the Arctic North Slope Dispersion Study. The study, which was
initiated in August of 1985, is expected to provide valuable
information on dispersion pertinent to the important oil production
area along the North Slope of Alaska. The primary goal of the
research study is to develop accurate and documented methods for
modeling pollutant transport and diffusion from local sources on the
North Slope of Alaska. The work is of prime interest because the
climatological conditions are considerably different from the
climatological conditions under which the dispersion equations
currently in use were developed.
*In order to avoid conflict of interest, Dr. Lamb, who is the
principal investigator of the Arctic North Slope Dispersion Project,
did not participate in the review of this project.
22
-------
The different conditions include fundamental differences in
boundary layer wind, temperature, and turbulence structure from low
latitude sites due to different surface conditions (e.g., roughness,
sensible and latent heat balances), and due to different combinations
of wind and stability resulting from altered driving forces (e.g.,
more strong wind, stable lapse rate conditions). Therefore, unless
the study is performed, calculations made with existing equations and
dispersion coefficients may lead to significantly inaccurate
estimates of the impact of sources of air pollution in this area
important to the development of national oil reserves. The study
appears to be well designed and careful attention has been given to
the development of a reasonable project that can be performed within
the time and budget allocations specified.
23
-------
CONCLUDING REMARKS AND RECOMMENDATIONS
The dispersion model development for predicting air pollution
concentrations associated with plume impingement on elevated terrain
is progressing very well under EPA sponsorship. The model being
developed employs applications of real atmospheric processes and thus
is dynamically pleasing. The employment of data from extensive field
measurement efforts and from laboratory fluid modeling simulations
appears to be a wise choice. The latter seems to provide much added
insight regarding the physical processes of plume behavior in complex
terrain features, and thus provides a valuable means of understanding
the processes that must be represented by the mathematical model.
Furthermore, the way the study has progressed from a theoretical
basis, through design and testing in the FMF, through preliminary
field tests, to complete field tests, to model development and model
testing with feedback at each step along the way is an excellent
example of a sound scientific approach to a problem, an approach
which often is not possible because of budget or time constraints.
EPA is to be commended for the excellent manner in which the program
has been organized and managed.
Complex Terrain Model Development Program
The overall program is an excellent combination of theory, fluid
modeling and field studies. It reflects well upon the efforts of the
managerial and technical personnel.
The theory is a reasonable approximation of the dominant
physical factors. Model design incorporates the theoretical
24
-------
concepts into a useful code for regulatory purposes. It appears that
needed simplifications in the model do not compromise the theoretical
foundations.
Theory and existing model algorithms have been largely confirmed
by fluid modeling. Fluid modeling was used extremely effectively to
design the experiments, provide guidance to the modelers, and to
understand field study results.
The field studies were well-designed and executed and the
Quality Assurance efforts were rigorous. It should be noted that
field studies and fluid modeling tests have not addressed transport
beyond 5 km or large buoyant plumes. This is reflected in
formulation of the diffusion terms for longer travel times. The
model development requires additional attention to some specific
portions:
1) T-factor averages,
2) values of adjustable constants (such as
Lagrangian time scales),
3) flow blockage for 2-dimensional ridges,
4) flow above Hc and moderately stratified
flow in general,
5) confirmation of the treatment of the Hc
transition region, and
6) effects of upwind boundary conditions.
Similarly, with regard to model applications, the following
points should be made.
1) The model may need to be further simplified
for regulatory use.
25
-------
2) Guidance will be required for the user to
specify input data correctly.
3) Diagnostic tools such as plume trajectories
should be provided.
The model testing follows recommended procedures but more recent
developments in this area could be incorporated.
The sensitivity analyses should be carefully developed to test
variations in model parameters and modules.
A workshop is recommended for model intercomparisons on common
data sets.
Wake Effects Program
This program has provided valuable input for treating the wide
variety of sources which involve building downwash and enhanced
diffusion. Future work should emphasize buoyant plumes, building
orientation effects, stable plume cases, and, in a general sense,
multiple building effects. Further development and application of
the video processing technique are recommended.
Arctic North Slope Dispersion Study
The Arctic North Slope Dispersion Study, which began in August
1985, is expected to provide valuable information on dispersion
pertinent to the developing oil production area along the North Slope
of Alaska. The study plan is reasonable and well designed. Good
results are anticipated with application to important regions in
Alaska as well as other important arctic regions of the world.
26
-------
Recommendations
1. Lee recirculation and/or lee transport close to the surface can
cause high pollutant concentrations for some source terrain
configurations. The nature of leeward flow patterns should be
investigated as a primary part of the second phase of the complex
terrain program. Results from previous fluid modeling and field
tests should first be analyzed in greater depth with regard to lee
recirculation.
Emphasis upon the leeward impact of sources located both upwind
and downwind of an obstacle should be included in this work. We
anticipate that the CTMD progression from simple to full-scale
situations involving theoretical modeling, fluid modeling, and field
studies will be well-suited to the problem of leeward recirculation.
2. The CTMD project has focused to date on small scales in a
systematic plan of building on a documented knowledge base. One
logical extension of the present work is to extend the scale of
interest to include settings of frequent practical interest,
typically up to about 30 km. Under stable, horizontally layered
conditions, stack plumes may travel over these distances with limited
dilution before impacting elevated topography. Many of the concepts
developed for the CTDM will continue to be valid; but, the full scale
demonstration should probably include this range. New field
measurements need not be intensive campaigns as were carried out at
the first three sites, but may center on sampling ground level
concentration patterns from existing stacks as sources of opportunity
using, where possible, existing monitoring equipment. Other
27
-------
approaches such as cooperative field efforts with other programs may
also be a viable way of collecting specific field data.
3. The program should incorporate aspects of circulation and plume
behavior associated with topographic forms other than isolated hills.
Valleys, for example, are common settings for both industries and
population across the United States. The circulations within valleys
and basins depend on a large number of physical phenomena and certain
combinations of meteorological mechanisms can produce very high
ground-level concentrations of pollutants. The EPA need not build an
independent program for each important terrain form, but should rely
on continued collaboration with other research programs such as the
DOE/ASCOT program to meet its objectives. Some mechanisms that would
influence the concentration field and be of interest to EPA include:
a) valley ventilation, especially as it affects plume
dispersal when gradient level winds cause mechanical
mixing from the ridgetops downward, and the dependence
of this phenomena on stratification;
b) impingement of plumes, especially from large industrial
sources within valleys, on adjacent ridgetops during
conditions when gradient level winds cross the valley
axis (There is evidence that the maximum concentrations
for large sources may indeed occur during stable
conditions with cross-valley winds when the plume
is confined to the shear layer between the valley
and the above valley flows; e.g., the Westvaco Pulp
Mill in Luke, Maryland.);
c) the diurnal cycle of the valley circulation and
ventilation as it affects plume behavior during
conditions when the gradient level winds cross
the axis of the valley;
d) concentration patterns that result during locally
driven valley flow conditions; and
e) concentration patterns and buildup during multi-day
stagnation periods.
28
-------
4. Time steps smaller than 1-hour may be valuable to understanding
physical processes leading to the observed hourly concentrations.
However the panel feels that the basic 1-hour time step is an
integral part of the project objective and should be the main focus
of the modeling and data analysis. Special studies at higher time
resolution may be pursued to the extent that such resolution is
supported by the data.
29
-------
REFERENCES
Dicristofaro, D.C., Strimaitis, D.G., Greene, B.R., Yamartino, R.J.,
Venkatram, A., Godden, D.A., (1985) EPA Complex Terrain Model
Development Fifth Milestone Report-1985. EPA/600/3-85/069. U.S.
Environmental Protection Agency, Research Triangle Park, NC.
Drazin, P.G. (1961) On the steady flow of a fluid of variable density
past an obstacle. Tellus 13:239-251.
Hunt, J.C.R. and Mulhearn, P.J. (1973), Turbulent dispersion from
sources near two-dimensional obstacles. J. Fluid Mech.
.61:245-274.
Hunt, J.C.R., Puttock, J.S., and Snyder, W.H. (1979) Turbulent
diffusion from a point source in stratified and neutral flows
around a three-dimensional hill. Part I. Diffusion equation
analysis. Atmos. Environ. 13;1227-1239.
Hunt, J.C.R., Snyder, W.H., and Lawson, Jr., R.E. (1978) Flow
structure and turbulent diffusion around a three-dimensional
hill. Fluid modeling study on effects of stratification. Part
I. Flow structure. EPA 600/4-78-041. U.S. Environmental
Protection Agency Report, Research Triangle Park, NC.
Pearson, H.J., Puttock, J.S., and Hunt, J.C.R. (1983) A statistical
model of fluid-element motions and vertical diffusion in a
homogeneous stratified turbulent flow. J. Fluid Mech.,
129:219-249.
Sheppard, P.A. (1956) Airflow over mountains. Quart. J. R. Met.
Soc. 82:528-529.
Smith, T.B. Diffusion Study in Complex Mountainous Terrian (April,
1965), 2 Volumes. Contract DA-42-007-AMC-45 (R), AD 484087.
U.S. Army Chemical Corps, Dugway Proving Ground, Utah.
Smith, R.B. (1980) Linear theory of stratified hydrostatic flow past
an isolated mountain. Tellus 32:348-364.
Synder, W.H., Thompson, R.S., Eskridge, R.E., Lawson, R.E., Jr.,
Castro, I.P., Lee, J.T., Hunt, J.C.R., and Ogawa, Y. (1982) The
structure of strongly stratified flow over hills - dividing
streamline concept. Appendix A to EPA-600/3-83-015, U.S.
Environmental Protection Agency, Research Triangle Park, NC, pp.
320-375.
Strimaitis, D.G., Lavery, T.F., Venkatram, A. Dicristofaro, D.C.,
Greene, B.R., and Egan, B.A. (1984) EPA Complex Terrain Model
Development; Fourth Milestone Report-1984. EPA/600/3-84/110.
U.S. Environmental Protection Agency, Research Triangle Park, NC.
30
-------
White, F., Ed., (1985) Summary of complex terrain model evaluation.
Prepared under U.S. Environmental Protection Agency Cooperative
Agreement 810297 to the American Meterological Society.
EPA/600/S3-85/060. U.S. Environmental Protection Agency,
Research Triangle Park, NC.
31
-------
APPENDIX A
AGENDA - PEER REVIEW OF THE
COMPLEX TERRAIN MODEL DEVELOPMENT PROJECT
32
-------
AGENDA
Peer Review of the
Complex Terrain Dispersion Model Meeting
November 19-21,1985
TIME
TOPIC
SPEAKER
Tuesday, November 19, 1985
8:00 - 8:30 a.m.
8:30 - 9:00 a.m.
9:00 - 10:00 a.m.
10:00
10:15
10:30
11:30
1:00
2:00
2:15
2:30
2:45
10:15 a.m.
10:30 a.m.
11:30 a.m.
1:00 p.m.
2:00 p.m.
2:15 p.m.
2:30 p.m.
2:45 p.m.
3:00 p.m.
3:00 - 4:00 p.m.
4:00 - 5:00 p.m.
EPA - Environmental Research Center
Classroom 2
Research Triangle Park, NC
Closed Session
Peer Review Program Orientation
Opening Session
Introduction and Discussion of
Complex Terrain Program Components
Plume Dispersion in the Wake of
Surface Obstacles
Break
Dispersion Modeling in the Arctic
Complex Terrain Field Studies
Lunch
Data Collection and Quality Assurance
Complex Terrain Data Dissemination
Complex Terrain Dispersion Climatology
Break
Introduction to Complex Terrain Model
Development Activities
Fluid Modeling Aspects of Complex
Terrain Model Development
Optional Tour of Fluid Modeling Faculty
R. Patterson
C. Coley
F. Schiermeier
A. Huber
P. Finkelstein
G. Holzworth
B. Greene
L. Truppi
G. Holzworth
P. Finkelstein
W. Snyder
W. Snyder
-------
-2-
TIME
TOPIC
SPEAKER
Wednesday, November 20, 1985
8:00 - 8:15 a.m.
8:15 - 9:15 a.m.
9:15 - 10:00 a.m.
10:00 - 10:15 a.m.
10:15 - 10:45 a.m.
10:45 - 11:15 a.m.
11:15 - 11:45 a.m.
11:45 - 12:00 a.m.
12:00 - 1:30 p.m.
1:30 p.m.
Thursday, November 21. 1985
8:30 - 10:00 a.m.
10:00 - 10:15 a.m.
10:15 a.m.
Reviewers' Meeting
Complex Terrain Data Analysis
Complex Terrain Model Theory
Break
Continuation of Model Theory
Complex Terrain Model Evaluation
Future Directions for ASRL Complex
Terrain Model Development
Summary Statements
Lunch
Report Preparation Executive
Session
Reviewer Debriefing With
Dr. Alfred H. Ellison, Director, ASRL
Break
Report Preparation Executive
Session
Closed Session
T. Lavery
D. Strimaitis
D. Strimaitis
D. Strimaitis
P. Finkelstein
F. Schiermeier
Closed Session
Closed Session
Closed Session
-------
APPENDIX B
EPA-ASRL PEER REVIEW PANEL
35
-------
REVIEW PANEL
COMPLEX TERRAIN DISPERSION MODEL
November 19-21, 1985
Name: Mr. Lawrence E. Niemeyer, Chairman
Consulting Meterologist
Work Address: Route 1, Box 367
Boone, North Carolina 28607
(704) 264-9140
Name: Dr. Sumner Barr
Work Address: Mail Stop D 466
Los Alamos National Laboratory
Los Alamos, New Mexico 87545
(505) 667-2636
Name: Mr. Donald Shearer
Work Address: TRC Environmental Consultants, Inc. - Denver
7002 S. Revere Parkway, Suite 60
Englewood, Colorado 80112
(303) 792-5555
Name: Dr. Brian Lamb
Work Address: Laboratory for Atmospheric Research
Washington State University
Pullman, Washington 99164
(509) 335-1526
Name: Dr. Jeffrey Weil
Work Address: Martin Marietta
145 South Rolling Road
Baltimore, Maryland 21227
(301) 247-0700 (ext. 359)
-------
APPENDIX C
PROCESS EVALUATION REPORT
37
-------
ATMOSPHERIC SCIENCES RESEARCH LABORATORY
PEER REVIEW
Process Evaluation Results for the
Complex Terrain Dispersion Model Program
The Atmospheric Sciences Research Laboratory (ASRL) of the U.S.
Environmental Protection Agency convened a panel of scientific
experts on November 19-21, 1985, to review the Complex Terrain
Dispersion Model Program. The panel consisted of five scientists.
These reviewers were asked to evaluate the process involved in
preparing and implementing this review. This report presents their
opinions of the process for this specific meeting.
The evaluation instrument was designed to assess the following
aspects of the process: 1) Preview Materials; 2) Process and
Logistical Information; and 3) the Review Meeting. A section was
also provided for reviewers to give their comments and
recommendations. The reviewers were instructed to respond to 15
items by circling numbers from 1 to 5 (with 1 representing poor;
2-fair; 3-good; 4-very good; and 5-excellent).
Table presents a summary of the reviewers' ratings for the 15
items. Several aspects of the review were rated as being excellent.
Most categories were rated either good or very good. Specific
comments and recommendations made by the reviewers are also
presented.
-------
Table 1
PROCESS EVALUATION
PREVIEW MATERIALS
Number of Raters
1. Written Quality
2. Technical Quality
3. Utility for Outside
Reviewer
4. Adequacy of Time
to Preview
PROCESS & LOGISTICAL
INFORMATION
5. Meeting Purpose
6. Scheduling: Agenda/
Format
7. Reviewer Responsibilities
8. Overall P.R. Process
9. Timeliness of Notice
10. Timeliness of Logistical
Information
Poor
REVIEW MEETING
11. Adequacy of Time for Dis-
cussion with EPA Staff
12. Adequacy of Time for
Executive Sessions
13. Quality & Utility of
Presentations
14. Quality & Utility of
Materials Disseminated
15. Support Services &
Activities
TOTAL NUMBER OF RATINGS
**__
0
Fair
1
1
2
Good
*
3
1
2
_***_
2
***-^
4
1
13
Very Good
4
4
3
2
1
1
4
2
4
2
4
1
1
1
2
36
Excellent
1
1
2
4
3
1
3
1
3
2
**
2
23
*"A lot of material to cover in a few weeks."
**"0ne presentation was poor; others ranged from good to
excellent."
***"Variable, from fair to excellent."
****"Again, variable."
-------
APPENDIX D
ASRL RESPONSES TO REVIEWS' COMMENTS
40
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
ATMOSPHERIC SCIENCES RESEARCH LABORATORY
RESEARCH TRIANGLE PARK
NORTH CAROLINA 27711
DATE:
March 11, 1986
SUBJECT: Response to Peer Reviewers' Comments on CTMD Program
FROM:
TO:
Reviewers' Comments
Francis A. Schiermeier, Program Manager
Complex Terrain Model Development Program (MD-80)
Ronald K. Patterson
ASRL Peer Review Coordinator (MD-59)
,;.-/
We suggest that future laboratory
simulation activities include in-
ternal wave effects for lee side
separation, providing guidance on
partitioning wave and turbulence
energies in field data, and study-
ing vertical motion structure over
more complex geometries such as
valleys, (et seq, P. 8)
Further work is needed to examine
the wave effects and a method for
treating this problem must be de-
veloped before the model can be
applied in real situations. (P. 9)
Perhaps the effects of other hills
could at least be explored in
sensitivity analyses, during the
remainder of the CTMD program.
(P. 10)
The use of average factors in place of
the line integrals present in the orig-
inal plume theory was recognized by the
panel as a necessary simplification for
computational efficiency. However, it
requires further theoretical justifica-
tion and demonstration as to its equiv-
alence with the line integral approach.
(P. ID
While the approach is physically plaus-
ible, the results of plumes released
very near Hc should be carefully ana-
lyzed to confirm this approach. In
particular, the thickness of the tran-
sition region near Hc has not yet been
documented. (P. 12)
ASRL Response
We are currently pursuing activities
in each of these areas, and continue to
encourage interactions with other program
personnel (outside FMF).
We agree. Work is underway on this
problem, both at ERT and inhouse.
Sensitivity analyses are going to be
performed (see recommendations), but it's
not clear that we have the data to separate
out the effect of other hills on the flow.
We agree. The sensitivity of the model
output will be investigated with respect
to the use of terrain-averaged correction
factors.
We agree. The thickness of the transition
region has been an area of ongoing investi-
gation.
-------
Further use of the digitized tra-
jectories from the fluid modeling
results for different stability
conditions should be fully exploit-
ed in the final development of the
model, (et seq, P. 12)
The present version of CTDM contains
considerable mathematical and numerical
detail, especially when compared to
other applied dispersion models.
A sensitivity analysis should be con-
ducted to aid in the simplification of
the model. (P. 13-14)
It appears that a separate task could
be created with the objective of direct
interaction of the CTMD program with
other similar programs. (P. 15)
We also recommend further testing of
the model include careful comparison
with other, possibly more sophisticated
models using the different models'
data archives. (P. 16)
The most logical next step would be
to consider greater distances and
to conceive ways that limited field
experiments or laboratory simulations
can be employed to represent plume
behavior at larger distances. (P. 17)
Such follow-on analyses should include
a comparison of the ground level con-
centration "footprint" from the
with that of the freon tracer.
(P. 18)
The probability for either logic errors
or typing errors increases as the length
of the model programs increases. A QA
effort is required to minimize errors in
the model. (P. 21)
We agree. We believe the treatment of
the flow over the hill as neutral is a
shortcoming of the model. We have en-
couraged use of the fluid modeling data
since they were collected.
See response to model sensitivity topic
on page 28.
During the entire CTMD program, we have
been interacting closely with the EPRI
Plume Model Validation and Development
(PMV&D) program and with the DOE Atmos-
pheric Studies in Complex Terrain (ASCOT)
program, even to the point of conducting
joint field programs with both.
This is precisely what was done in the
February 1986 workshop when invited
participants reported on exercising the
preliminary CTDM on their respective data
sets.
We agree and would appreciate any specific
suggestions as to which direction to proceed.
In this case, field experiments would likely
be more useful than laboratory simulations.
We agree. These analyses have not yet been
done because of problems with some of the
data and plume rise calculations, but they
will be, if possible. Also, note that
frequently the releases of the two tracers
were above and below Hc and will give very
different patterns.
We agree. QA of completed model code is a
standard procedure within ASRL.
-------
At some point in the evaluation
of the model, the model performance
should be discussed in the context
of the overall uncertainty of the
field data. (P. 22)
While there are a great number of
problems requiring solution, there
is not an overwhelming requirement
to study the problem at length.
Engineering or best estimates are
sufficient to meet the requiremeants
of decision makers. (P. 22)
Recent work on the modification
of the ISC model by ERT may have
been overlooked and needs to be
reviewed for pertinent information.
(P. 23)
Future work plans appeared to em-
phasize tests of the effects of build-
ing scales vs. boundary layer scales.
However, the panel believes the type
of work has received considerable
attention in past studies (P. 23)
Conclusions: P 27-28
The model development requires addi-
tional attention to some specific
portions:
1) T-factor averages.
2) Values of adjustable constants
(such as Lagrangian time scales),
3) Flow blockage for 2-dimensional
ridges.
4) Flow above Hc and moderately
stratified flow in general.
We agree.
goal.
A worthwhile and important
We believe it is very important to find
the proper balance between the depth of
the research investigation and the needs
of decision makers. We have consulted
closely with the regulatory side of the
Agency and we will plan our research so
that both of these needs are adequately
met.
We have provided reviews of ERT modifi-
cations to the ISC model, have attended
meetings between ERT and OAQPS, and will
continue to seek further pertinent infor-
mation on ERT modifications.
Some study of building scales vs. boundary
layer scales is now ongoing. The purpose
is to evaluate experimental arrangements
for our facility. These tests provide a
necessary evaluation before we launch into
future studies with fixed scales.
We agree that all these need additional
attention and are either planned or are
in the midst of the effort. Specifically
New formulations for "T" factors are being
developed using other than potential flow
solutions.
Adjustable constants are being chosen
based upon an optimized parameter
adjustment scheme. Use of the Lagrangian
time scale in very stable flow is being
reconsidered.
A new approach to flow blockage is being
developed inhouse.
Lateral displacement of the stream-
lines above and near Hc is a problem.
-------
5) Confirmation of the treatment
of the Hc transition region.
6) Effects of upwind boundary
conditions.
Similarly, with regard to model
applications, the following points
should be made:
1)
2)
3)
The node! may need to be fur-
ther simplified for regulatory
use.
Guidance will be required for
the user to specify input data
correctly.
Diagnostic tools such as plume
trajectories should be provided
The model testing follows recommended
procedures but more recent develop-
ments in this area could be incor-
porated.
The sensitivity analyses should be
carefully developed to test vari-
tions in model parameters and modules.
A workshop is recommended for model
intercomparisons on common data sets.
Future work should emphasize buoyant
plumes, building orientation effects,
stable plume cases; and in a general
sense, multiple building effects.
Further development and application
of the video processing technique is
recommended.
New approaches using the Hunt-
Mulhearn theory are being considered.
Upwind boundary conditions are very
difficult to define in complex terrain,
The data are being evaluated to see
if different approaches can be taken
to the problem.
We agree with these points as well.
Specifically -
The model is being simplified, and
a manual written to help the user
understand the underlying structure
of the model.
Both guidance and a front-end pre-
processor for input data are being
developed.
Better diagnostic tools will be
developed for use with the model.
We agree. The most up-to-date pro-
cedures will be used in model testing
and validation.
We agree. We are planning to use some
innovative sensitivity analyses, in-
cluding the FAST (Fourier Amplitude
Sensitivity Tests) to evaluate which
inputs are most important.
This would be a good idea, but we have
no money in the budget for such a
workshop. We will try to do so if
funds are available in future years.
We agree with these needs. We presently
have plans for research into orientation
effects and further development of the
video processing techniques. The study
of buoyant plumes, stable plume cases,
and general multiple building effects
each present special difficulties. We
are interested in these problems and will
pursue them as time and budget permit.
-------
Recommendations: (P. 29-31) We take no issue with the long-range
recommendations of the panel. Pri-
orities will have to reflect available
funds and the needs of the regulatory
side of EPA. We had planned to start
some of this work in FY-87, but recent
budget cuts will put this research off
at least one year.
Throughout the report were contained numerous recommendations regarding future
direction of efforts or expansion/redirection of present efforts. Most of these
recommendations were addressed in the preceding ASRL response. At the beginning
of the peer review, we asked that you provide such recommendations and appreciate
your diligence in providing them. We will make our best effort to incorporate
as many as possible in both our present complex terrain program and in similar
follow-on programs.
-------
APPENDIX E
REVIEW OF THE PANEL REPORT AND
RESPONSES OF THE LABORATORY DIRECTOR
46
-------
j UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
ATMOSPHERIC SCIENCES RESEARCH LABORATORY
,tcf RESEARCH TRIANGLE PARK
PRO NORTH CAROLINA 27711
DATE: March 25, 1986
SUBJECT: Peer Review Report on the Complex Terrain Modeling Program
FROM: Jack H. Shreffler^J*-X.ff
Deputy Director, AWL (MD-59)
TO: Ronald K. Patterson
TPRO, ASRL (MD-59)
I have read the draft peer-review report on the Complex Terrain
Model Development (CTMD) Program and the written response of the Program
Manager. The Reviewers seemed exceptionally well versed on the direction
and results of the research effort.
The Peer Reviewers were uniformily laudatory concerning the CTMD
Program. They recognized that the success of this research program was
due in no small part to effective long range planning and execution. The
Program had its origins in a planning workshop held in 1979 and attended
by a range of EPA and outside scientists. According to the developed
scientific plan, a series of integrated laboratory, field, and theoretical
studies have been carried out with almost unprecedented stability in
funding and program direction.
The Program Manager, Frank Schiermeier, has responded fully to the
comments and recommendations of the panel. All significant comments
of the Reviewers centered on directions for future work in this area
of research. Those comments along with the results of a planning work-
shop held February 1986 will be folded together to determine the future
direction of the CTMD Program.
Concerning the evaluation of the review process, Appendix C,
we should try to make the preview material as useful and concise as
possible so that the time that we allow seems adequate to the reviewers.
cc: F. Schiermeier
C. Hosier
------- |