United States       Office of Air Quality       EPA-450/4-81-003
            Environmental Protection   Planning and Standards     July 1981
            Agency         Research Triangle Park NC 27711
            Air
c/ERA      Guideline for Use of
            Fluid Modeling to  Determine
            Good Engineering Practice
            Stack Height

-------
                                    EPA 450/4-81-003
        GUIDELINE  FOR USE  OF  FLUID MODELING
TO DETERMINE GOOD  ENGINEERING PRACTICE STACK HEIGHT
                     June  1981
   UNITED STATES ENVIRONMENTAL  PROTECTION AGENCY
        Office of Air, Noise, and  Radiation
   Office of Air Quality  Planning  and Standards
   Research Triangle Park,  North Carolina 27711
       U,3  trwiroryvv-rital Protection Agency
       Re,.}-.:.)  "•, Library
       230 3c\;-.;,  Dearborn Street
       Chicago, iiiincis  60604

-------
U,S. Environmental Protection Agency

-------
                          ACKNOWLEDGMENT

     This guideline was prepared by Alan Huber, formerly of the Monitoring
and Data Analysis Division, Office of Air Quality Planning and Standards,
Environmental Protection Agency, now assigned to the Agency's Meteorology
and Assessment Division.  Appreciation is extended to Dr. William Snyder
of the Meteorology and Assessment Division, Environmental Sciences
Research Laboratory, for his helpful comments and discussion concerning
fluid modeling.   His fluid modeling guideline document,  referenced
herein, serves well in establishing EPA standards to be followed in the
conduct of such studies.  Changes to the June 1979 draft were based
partially on comments and suggestions received during the period of
public comment and on EPA's assessment of its own demonstration studies
conducted by staff at the Agency's Fluid Modeling Facility.  Special
appreciation is extended to Robert E. Lawson, Jr., of the Meteorology
and Assessment Division, Environmental Sciences Research Laboratory,  for
taking the responsibility of conducting and reporting a demonstration
study.  Overall, the changes to the June 1979 draft should simplify the
study design and provide a framework for both consistent reports and
reviews.

-------
                     TABLE OF CONTENTS

                                                            Page

1.0  INTRODUCTION 	    1

2.0  BACKGROUND	    3

3.0  BASIC CONCEPTS 	    5

     3.1  Dynamic Similarity Criteria 	    7
     3.2  Boundary Layer Conditions 	   13
     3.3  Surface Roughness, Terrain,
            and Building Scaling	14
     3.4  Plume Rise	15
     3.5  Concentration Measurements	18

4.0  REQUIREMENTS FOR A FLUID MODEL DEMONSTRATION 	   21

     4.1  Preliminary Design	22

          4.1.1  Model Surface and Its Boundary Layer ...   22
          4.1.2  Plume Rise	29
          4.1.3  Atmospheric Dispersion Comparability ...   31

     4.2  Determination of GEP Stack Height	37

          4.2.1  Demonstration of Adverse Effects 	   38
          4.2.2  GEP Stack Height	41

5.0  REPORT CHECKLIST 	   43

6.0  REFERENCES	47
                              m

-------
1.0  INTRODUCTION

     This guideline contains specifications for the use of fluid model-
ing to determine Good Engineering Practice (GEP) stack heiyht.   The
guidance is intended for use by the U.  S. Environmental Protection
Agency (EPA), by State and local air pollution control agencies, and by
industries and their consultants in the design and final  review of a fluid
modeling study determination of GEP stack height.   The Agency issues
guidelines in association with regulations in order to make clear any
requirements for data and to present criteria the  Agency will  use in
evaluating the adequacy of that data.  The specifications in this
guideline are necessary to assure consistency among studies.  It is very
important for both those conducting the fluid modeling study and those
reviewing the results to share a common set of criteria for reference.
     The aim of fluid modeling is to produce an accurate representation
of the atmosphere using the flow of air or water in a test facility,
e.g., wind tunnel or water channel.  Certain similarity criteria must be
considered if fluid modeling studies are to accurately reproduce atmospheric
phenomena.  A separate guideline entitled, "Guideline for Fluid Modeling
of Atmospheric Diffusion," (Snyder, 1981), reviews the fundamental
principles and practical applications of fluid modeling.   The aim of
that guideline is to establish the capabilities and limitations of fluid
modeling, and to establish EPA standards for the conduct of fluid modeling
studies.  This guideline is based on Snyder's state-of-the-art review.

-------
2.0  BACKGROUND

     As required by Section 123 of the Clean Air Act Amendments  of
1977, the Administrator has proposed regulations (Sections  1  and 18
of 40 CFR Part 51) to assure that the control  of any air pollutant under
an applicable implementation plan shall  not be affected by  (1) stack  heights
that exceed good engineering practice or (2) any other dispersion technique.
Good engineering practice (GEP) is defined with respect to  stack heights  in
Section 123 of the Clean Air Act Ammendments of 1977 as "the  height
necessary to insure that emissions from the stack do not result  in excessive
concentrations of any air pollutant in the immediate vicinity of the  source
as a result of atmospheric downwash, eddies and wakes which may  be created by
the source itself, nearby structures or nearby terrain obstacles."

     The scientific literature, in general, indicates that  a  case-
specific review is integral to assuring the prevention of adverse aero-
dynamic effects in the immediate vicinity of a given source.  However,
the literature also identifies a general formulation that establishes a
minimum height necessary to prevent significant effects of  nearby
structures.  The GEP formulation is a reasonable working rule, defined
as:
          H  = H + 1.5L                                  (1)

where:    H  = GEP stack height
          H = Height of the structure or nearby structure
          L = Lesser dimension (height or width) of the structure or
          nearby structure.

-------
The basis of the formulation and a summary  of extensive  scientific
literature on the subject can be found  in the "Guideline for  Determination
of Good Engineering Practice Stack Height"  (EPA,  1980).

     Proposed regulations (40 CFR Part  51)  to implement  Section 123  of
the 1977 Clean Air Act Amendments allow the stack heights  near structures
as determined by Equation 1  to be used  in some cases  as  the maximum
creditable stack height which may be used in establishing  a source's
emission limitation for a State Implementation Plan.   The  GEP creditable
stack height, based on nearby terrain features, must  be  determined on a
case-by-case basis through the use of appropriate field  or fluid modeling
studies.  Field or fluid modeling studies may also be used by the source
operator to show that a stack height greater than determined  by Equation 1
is needed to prevent excessive pollutant concentrations.  This guideline
is appropriate when fluid modeling is used  to determine  GEP  stack height.
An excessive concentration,  for the purpose of determining GEP stack
height, is defined in the regulation as a maximum ground-level  concentration
monitored or modeled in the presence of nearby structures  or  terrain
obstacles that is 40 percent or more, in excess of the maximum ground-level
concentration, monitored or modeled for the same  orientation  and stack
parameters in the absence of downwash,  wake, or eddy  effects  produced by
nearby structures or terrain.

-------
3.0  BASIC CONCEPTS

     The basic concepts for designing a fluid modeling study are out-
lined in the following subsections.  The construction of a fluid model
requires that the flow in the test facility (e.g., wind tunnel) be
appropriately fixed and along with the surface roughness, terrain,
and/or buildings, scaled to accurately reproduce atmospheric phenomena.
The stack and plume from the source must also be similarly scaled if
dispersion patterns in the fluid model are to simulate those in the
field.  The fluid model encompasses the entire situation within the four
walls of the test facility that is designed to accurately simulate
atmospheric flow in the field.  Consideration of each of the concepts
outlined in the following subsections leads to requirements for data and
the reporting of that data as given later in Section 4.  Specific
references to requirements given in Section 4 are underlined.

     A detailed formulation and discussion of the fundamental  principles
for fluid modeling of atmospheric phenomena is presented by Snyder
(1981).  A summary of the important criteria is presented here.  Certain
similarity criteria must be considered if fluid modeling is to accurately
reproduce atmospheric phenomena.  The dynamics of the flow in  the fluid
model must accurately simulate those in the field.  The effects of
surface conditions in the field upstream of the modeled area must be
accounted for in the fluid model by developing appropriate boundary
layer conditions.  The necessary surface roughness, terrain, and buildings
are included in the construction of the fluid model.  The plume trajectory
in the fluid model must be similar to that in the field if air quality
impact is to be evaluated.
                                 5

-------
     The purpose of the fluid  modeling  study  for  determining GEP  stack
height is to demonstrate the stack height  needed  to  avoid  excessive
concentrations caused by the effects  from  nearby  structures or  terrain
obstacles as specified by stack height  regulations.   GEP stack  height
is appropriately determined for the situation under  atmospheric con-
ditions that result from those surface  influences of highest extent.

     Above some minimal reference wind  speed, e.g.,  3 m/s, the  flow
pattern near the structure or  terrain obstacle in the field is  independent
of wind speed, as is reasoned  in the  discussion presented  in Section 3.1
on Reynolds number independence.   The greatest effect on the plume
should occur when plume rise near the source  is lowest.  For most sources,
even those with a relatively high exit  velocity,  high wind speeds found
to occur occasionally at most  locations will  result  in  significantly
reduced plume rise and thus the greatest potential for  ground-level
concentrations in excess of those in  the absence of  structure or
terrain obstacle influences.

     The wind speed that will  result  in the determination  of greatest
GEP stack height is seen for all  foreseeable  situations to exceed 6 m/s.
The atmosphere is characterized by a  generally neutral  state of stability
when the surface wind speed at a height of 10m is greater than 6 m/s
(Turner, 1970).  Thus, the critical conditions of stability for de-
termining GEP stack height are expected to be associated with a neutral
(adiabatic) atmosphere.  Specific guidance for fluid modeling of
an adiabatic atmosphere is given herein.   Guidance for  modeling

-------
a nonadiabatic atmosphere is not provided; the need to model these situ-
ations requires case-by-case consideration.

     To define GEP stack height for a specific stack, measurements in
the wake of and in the absence of either the structure or terrain
obstacle are needed to assess the increase in maximum concentrations.
The concentration increase must be assessed to determine whether the in-
crease constitutes an excessive concentration.  Concentrations in the
wake of the structure or terrain obstacle are considered excessive if
the maximum ground-level concentrations are at least 40 percent greater
than the maximum in the absence of their influences.  Wind-tunnel modeling
is ideally suited for this type of determination since the model structure
or terrain feature being studied can be easily removed to assess its
effect.  More importantly, a properly designed wind-tunnel  study can
account for the aerodynamically induced influences affecting the dispersion
of the stack effluent.

3.1  Dynamic Similarity Criteria

     To rigorously model the dynamic behavior of atmospheric flow, five
nondimensional parameters must be matched between the model  and the
field.  These parameters, as discussed by Snyder (1981), are:
     1.  Froude number, Fr = UR/ / gL6TR/T0;
     2.  Ross by number, Ro = UR/L£2R;
     3.  Reynolds number, Re = URL/v;

-------
     4.  Peclet number, Pe =
     5.  Reynolds-Schmidt number, Re-Sc = UDL/a.
                                           K
where:
     UR = reference velocity;
     ftR = reference angular velocity;
      L = reference length;
     6TR = temperature deviation from adiabatic atmosphere;
     T  = temperature of adiabatic atmosphere;
     g = gravitational constant;
     v = kinematic viscosity (momentum diffusivity);
     K = thermal diffusivity;
     a = molecular diffusivity.

     The Froude number represents the ratio of inertial forces to
buoyancy forces on a local air parcel.  Calculation of the Froude
numbers should be based on the measured temperature profiles.  A large
value of the Froude number implies that buoyancy forces are smdll relative
to inertial forces.  Thus, atmospheric flows having large Froude numbers
are considered neutral (adiabatic).  A truly adiabatic atmosphere
has an infinite Froude number.  Isothermal flow must be maintained in
the test facility tp_ adequately model ar^ adiabatic atmospheric flo w.
This similarity criterion for fluid modeling of adiabatic atmospheric
flow can be easily met by insuring that air in the room containing the
facility and the fluid temperature in the test facility are equal.  This
is especially necessary when the flow speed through the test facility is
slow.  Steps 2a and 3a of Section 4.1.3 are required to satisfy this
similarity criterion.
                                 8

-------
     The Rossby number represents the ratio of the inertia! forces
to the Coriolis force on a local air parcel.  Coriolis force results
in the wind vector changing direction with increasing height above the
surface.  If the Rossby number is large, the Coriolis force is relatively
small and thus does not have a significant effect on a dispersing pollutant.
Snyder suggests that the Rossby number is sufficiently large at downwind
distances (L) less than about 5 km to ignore Coriolis force for modeling
dispersion in adiabatic atmospheric flow over flat terrain.  Nc in-
formation is available to assess the effect of Coriolis force in regions
of complex terrain.  This implies that fluid modeling should be limited
to areas within 5 km of the source, since Coriolis force cannot presently
be simulated.  Mathematical models of atmospheric dispersion in current
use do not account for Coriolis forces.  Therefore, while fluid
modeling should be_ limited tp^ areas within 5_ krn^ of_ the source, its use
for modeling larger areas may as^ for mathematical models be_ similarly
justified, when necessary.  The Rossby number criterion js^ not a^ critical
modeling parameter, although vt poses a^ limitation on_ the use of_ fluid
modeling.  There are no special data requirements to satisfy this similarity
criterion.

     The Reynolds number represents the ratio of the inertial  forces
to the frictional forces on a local air parcel.  When the modeling
medium is air, the reference velocity must be increased by the same
amount as the reference length is reduced in order to match the Reynolds
number.  In water, the reference velocity need only be increased by 1/15
the reduction in the reference length, since the kinematic viscosity of

-------
water is 1/15 that of air.   However,  using  water  as  the  medium requires
much more energy for equal  rates of flow.   This physical  limitation
generally results in water  tunnels having  to  be smaller  in  size and
operated with lower flow rates than wind tunnels.   The reference
length in atmospheric dispersion problems  must be modeled at  a reduced
scale of several orders of  magnitude, making  an equivalent  increase  in
the reference velocity impractical.   Thus  if  strict  adherence to the
Reynolds number criterion were required, no atmospheric  flows could  be
modeled.

     Various arguments to justify the use  of  smaller Reynolds numbers in
fluid modeling compared to  those in the atmosphere are found  in the
literature.  The best argument appears to  be  the  principle  of Reynolds
number independence.  This  principle is based upon the hypothesis that
in the absence of buoyancy  and Coriolis effects,  the pattern  of tur-
bulent flow is similar at all  sufficiently high Reynolds numbers.  If the
Reynolds number is large, frictions!  forces are relatively  small com-
pared to the inertia! forces.   For Reynolds number independence to hold,
the frictional forces must  remain relatively  small and have little
effect on the overall flow  as  the Reynolds number is decreased.  A large
amount of experimental evidence now exists to support this  principle.
For atmospheric flows, Reynolds number independence appears to apply
except in the very smallest scale of the turbulent flow  very  close to
the ground or other physical  boundary.  In effect, the reference length  L
(scale of flow structure examined) is small where the frictional forces
are important.  Flow very near structures  or  terrain features may not be
Reynolds number independent.
                               10

-------
     Practice indicates that sufficiently large Reynolds numbers are
attainable at least for modeling the flow over sharp-edged geometrical
structures or terrain features in ordinary meteorological wind tunnels.
However, more work must be done to determine if simulation of flow over
more streamlined surfaces can be sufficiently modeled.  Frictional
forces have very little effect if the general flow is detached from the
surface or other physical boundary.  Flow over streamlined surfaces is
less susceptible to detachment and thus is more sensitive to the value
of its Reynolds number.  Reynolds numbers in atmospheric flow generally
are sufficiently large for the independence principle to hold.  Fortu-
nately, flow similarity is generally observed at the lower values of
Reynolds number attainable in fluid modeling, provided the flow is
locally detached or if the area of study is sufficiently above the
surface.  This has led modelers in some situations to force flow de-
tachment by adding roughness to the model structures or terrain features.
The Reynolds number criterion is_ a_ critical  modeling parameter.  Step 4_
lH Section 4.2.1 requires a_ test for Reynolds number independence.

     The Peclet number is most easily discussed by writing it as the
product of the Reynolds Number and Prandtl Number:
          URL
     Pe = JJ-£= Re-Pr.

The Reynolds-Schmidt Number can similarily be written as the product of
the Reynolds Number and the Schmidt Number:
     Re-Sc = URL  v = Re-Sc.
              v   a
                                 11

-------
Both the Prandtl  number and the Schmidt number are properties  of the
fluid.  The Prandtl  number is the ratio of the momentum diffusivity to
the thermal diffusivity.  The Schmidt number is the ratio of the mo-
mentum diffusivity to molecular diffusivity.  These numbers  are somewhat
different between air and other fluids.  This would seem to  preclude
using any model medium other than air in simulations of atmospheric
flow.  A high Reynolds number is necessary in the wind tunnel  flow to
match the Peclet number or Reynolds-Schmidt number found in  the field.
Arguments similar to those constructed for Reynolds number independence
are used to justify the neglect of the Peclet number and Reynolds-
Schmidt number as modeling criteria, provided the Reynolds number is
sufficiently large.   Both heat and mass are regarded as passive
quantities in connection with most environmental  atmospheric dispersion
problems.  Thus, if the Reynolds number of the main structure of the
flow is sufficiently large, advection and the larger scale turbulent
motions are totally responsible for the transport and dispersion of a
passive pollutant.  That is, molecular or thermal diffusion  acts mostly
to smooth out the very small-scale discontinuities of concentration or
temperature.  Molecular or_ thermal diffusion J_s_ assumed t£ contribute
negligibly tp_ the dispersion of_ the source plume within the  simulated
turbulent atmospheric boundary layer, provided the Reynolds  number js_
large enough with the Peclet number and Reynolds-Schmidt number being
themselves unimportant.
                                 12

-------
3.2  Boundary Layer Conditions

     The effects of upstream surface conditions on the velocity of the
wind result in a variation with height described generally by some
theoretical distribution such as a logarithmic or power law profile.
The profile is characterized by the depth of the boundary layer and a
representative surface roughness length.  Turbulence intensity of the
wind naturally decreases with height above the surface roughness.  The
profiles of mean velocity and turbulence intensity are very significant
characteristics that should be very closely matched in the model.

     Measurement of vertical profiles of Reynolds stress throughout the
region of interest is especially useful  in characterizing the surface
friction velocity u*, which is a parameter used in representing the
velocity near the surface by a logarithmic profile.  Measured profiles
in the field will not likely be available at most sites where 6EP stack
height is to be determined.  In Section 4, general modeling criteria  are
3b. 3c_ jm Section 4.1.3 requires mean velocity, turbulence intensity, and
Reynolds stress profiles tp^ be_ measured at_ several positions^ throughout
the model.

     Consideration of additional flow characteristics of the atmospheric
boundary layer would be desirable.  However, specific guidance is not
possible.  The purpose of specifying necessary modeling criteria for  the
boundary layer is to first insure that dispersion throughout the modeled
flow correctly provides dispersion patterns comparable to those given by
                                 13

-------
recommended air quality modeling techniques as described in the Guideline
on Air Quality Models (EPA, 1978).   Concern for the similarity of
additional flow characteristics is  not necessary if the model  boundary
layer dispersive characteristics are documented through measurements in
the test facility to fall between estimated values for Pasquill-Gifford
stability category C and D as prescribed by Turner (1970).   Step 4 in
Section 4.1.3 requires documentation of_ dispersion from the source.

     A documentation and test of comparability of the fluid model
boundary layer conditions to flow over flat terrain in the  absence of
any buildings can be done simply.  Problems, however, arise in doing the
same for flow over complex terrain  or over urban areas where local
differences near the surface result from different surface  features.
Because of these differences one cannot establish that the  fluid model
boundary is nondeveloping or that the dispersive characteristics can be
represented by a general categorization.  In order to evaluate the fluid
model boundary layer conditions, it is necessary to first document the
flow in absence of the complex terrain or urban structures  as prescribed
above.  Differences found for flow over the complex terrain or urban
area can then be related to the increased surface roughness.  Section
4.2.1 requires measurements sufficient tp_ document such differences.

3.3  Surface Roughness, Terrain, and Building Scaling
     Minute geometric details of terrain or structures do not signif-
icantly affect atmospheric flow.  Thus, such detail need not be
considered in a fluid modeling simulation.   Objects about the same size
                                 14

-------
as the characteristic surface roughness length need not be produced in
geometrical form but an equivalent roughness must be established.  For
example, gravel can be added to the surface of the terrain or buildings
to establish an equivalent roughness.

     Major terrain features and building structures must be scaled
without geometric distortion.  The amount of reduction in scale is
limited by the requirement for the flow to be Reynolds number independent.
Discussion and some guidance on selecting the proper model surface
roughness, terrain height and building height are given by Snyder (1981).
The fluid modeler's decisions must be based on a number of interacting
concerns, including the size of the area to be modeled, the necessary
boundary layer depth, the desired diffusion characteristics,  and Reynolds
number independence.  Often, the fluid modeler's experience is the best
guide.  Data requirements and criteria which EPA will  use to  evaluate
the resulting study are given in Section 4.  It is recognized that each
fluid modeler may have somewhat different approaches to selecting these
design parameters.  This is acceptable if the results meet the report
requirements.  Section 4.1.1 outlines the requirements for modeling the
surface and its boundary layer.

3.4 Plume Rise

     The atmospheric conditions that are considered here in determining
GEP stack height are characterized as neutral  stability (adiabatic) with
high mean wind speed.  Under such conditions,  plume rise near the source
where its rise is dominated by momentum flux will be small while its
rise farther downwind may be largely due to buoyancy flux.  A thorough
review of the issues relating to fluid modeling of plume rise is presented
by Snyder (Section 3.1, 1981).
                                  15

-------
     In general, plume rise near the  source in an  adiabatic  atmosphere
has proven to be well-described for most conventional  sources  by  the
Briggs (1975) formulation,
                                                          2
3
BT
                        4H
H
+
L^'
B
Hs
3
M



gD2ws(pa-Ps)|
4p IPH
^Ma s
r 12
X
Hs


—     (2)
H      U'
           = iT
for gi = 1/3 + U/W^» e2 = 0-6 where Lm is a momentum length scale and LD
                  5                  HI                                 O
is a buoyancy length scale, defined as
                                                            (3)
Lm _ T
H ~ "2.
5
'psWs
U^
a
11/2
1 D
H
s
L . (w "\ 3

TT= 4
s
1
Fr^"
s.
D "s
H U
s I J
. . . fu ^3
1
~ 4

1
fr'i
a
D s
He lu 1
SJ
                                                            (4)
where:
     p  = stack effluent density
     pa = ambient air density
      a
     W  = stack effluent exit speed
     D = stack exit diameter
     HS = stack height
     U = mean wind speed at height of stack
     x = streanwise coordinate
     q = acceleration due to qravity
                                16

-------
     Frs = Ws/[gD(pa-ps)/ps]1/2
     Fra = Ws/[gD(pa-ps)pa]1/2

     The first term in Equation 2 represents the contribution due to
source momentum and the second term represents the contribution from
source buoyancy.  Close to the stack, the initial  momentum term will  be
important, whereas, the buoyancy term in most cases will  ultimately
dominate.  Matching the parameters in the above plume rise equation
should insure comparable plume rise in the fluid model  so long as the
plume is not downwashed into the wake of the stack, buildings, or
elevated terrain.

     A GEP stack height determination must examine the effect of the
nearby structure or terrain obstacle on the plume from the source.   GEP
stack height is limited to the height necessary to avoid  excessive
concentrations as explained in Section 2.  Immediate plume rise near the
source may in some situations significantly effect the stack height
necessary to avoid the most adverse effects.  The model  study must be
designed to demonstrate plume rise near the source to be  comparable to
the estimate by Equation 2 in the absence of buildings or terrain.  This
can be satisfied by correctly modeling the momentum length scale as
presented in Section 4.1,2.  EPA has based its definition of "excessive
concentrations" on model results for sources not having significant
plume rise downwind from the stack and thus the demonstration will  be
consistent.  The model study design is considerably simplified by not
having to model the buoyancy length scale which generally controls  the
resulting plume rise further downwind of the source.  Also the results
of the fluid model study will be generally more reliable.

                               17

-------
     When stack downwash occurs,  special  consideration must be given to
the flow around the stack.   It is essential  to assure that the flow
within the boundary layer around  the stack is turbulent.   Common practice
in fluid modeling is to use a trip wire or fence or other surface
roughness to force the boundary layer flow to be turbulent.  This is not
necessary for modeling rectangular stacks since their sharp corners
force flow separation.  In  all cases, it is  necessary to  assure that the
stack effluent exhaust is fully turbulent.  Data requirements and
criteria that EPA will use  tp_ evaluate the representativeness of_ the
plume rise jji the fluid model are presented  ir\_ Section 4.1.2.

3.5  Concentration Measurements

     Concentrations measured in a fluid modeling study should be related
to those in the field through the nondimensional concentration, C = xUH2/Q
as presented by Snyder (Section 3.5, 1981),  where
     X = mass concentration of pollutant (gm/m3),
     U = reference wind speed (m/s),
     H = characteristic length (m), and
     Q = pollutant emission rate (gm/s).

     The sampling time for measurements taken as part of  the fluid
modeling study must be long enough to provide steady-state averages.
Fluid modeling is designed to correspond to  conditions in the field for
which the wind direction is steady.  It is essential that the maximum
ground-level concentrations be shown to represent the steady-state
average values since they are crucial in the demonstration of excessive
                               18

-------
concentration.  Data requirements and criteria which must be_ considered



ijT^ establishing this fact are presented ij^ Section 4^.





     In absence of effects of building and/or terrain, the pattern of



concentrations in the fluid model should be comparable to those estimated



by mathematical models recommended by EPA.  Steady-state average concentratioi



measured in the fluid model should thus correspond to one-hour average



concentrations in the field.  In those situations where persistence of



the wind direction can be assumed, fluid modeling can be used for estimating



average concentrations for longer periods of time.  Fluid modeling



studies can also be used for estimating average concentrations for



periods having variability in the wind direction by including a separate



examination of the flow for several  directions.   Concentrations can then



be estimated by considering the frequency of wind direction.
                               19

-------
20

-------
4.0  REQUIREMENTS FOR A FLUID MODEL DEMONSTRATION

     Section 3 presented a summary of the basic design concepts used
in developing a fluid modeling study that is comparable to field
conditions.  Elements necessary for determining a GEP stack height
are specified in this section.  Specific guidance on requirements
for data and the reporting of that data are given.  Fluid modeling
studies can be adapted to meet detailed specifications since essential
characteristics can be controlled.  Atmospheric flow is extremely complex
in that profiles of its characteristics can vary with time and space.
In general, profiles in the field rarely are available at the sites
where a GEP stack height is to be determined. Therefore, it is only
necessary for the modeling study to be designed to meet the general
atmospheric conditions, given here as comparable to those used in air
quality models recommended by EPA.  Where detailed field information
describing the situation is available, the modeling study should be
designed to best assimilate it.  For these situations, the fluid model
should be shown to be comparable to the field.

     The requirements given here are based on the general guidance
by Snyder (1981).  Readers and users of this guideline should be
familiar with the presentation by Snyder.  Deviation from his general
guidance occurs in this guideline in a few instances where the objectives
were judged to be met without additional detail.

     Section 4.1 presents important criteria that should be considered
prior to the construction of the fluid model for the actual situation.
                                21

-------
The model scale should first be determined.   Then a test for atmospheric
dispersion comparability should be conducted as  specified.   Section 4.2
specifies the procedure and reporting requirements for determining GEP
stack height.  The step-by-step procedure leads  simply to a  satisfactory
study provided sufficient preliminary consideration is given to  the
criteria presented in Section 4.1.

     Specific reporting requirements presented here should be followed
as outlined.  All  supplementary data taken as part of the study  should
also be incorporated into the report.  A separate appendix describing
the fluid modeling facility and instrumentation  used in the  conduct of
the study should be attached.  Normal operating  conditions and associated
parameters should be described.  A daily log should be kept  during the
conduct of the study since the Agency may wish to conduct an audit and
review.  Since on-site visits and demonstration  of repeatability of some
measurements may be requested as part of quality assurance procedures,
a proposed plan of study must be sumbitted to the reviewing  Agency.

4.1  Preliminary Design
     4.1.1  Model  Surface and its Boundary Layer

     The size of all building structures and the general topography in
the vicinity of the source should be examined and the area to be deter-
mined.  A roughly cubical building or other major structure, or  a three-
dimensional hill upstream of the source should be included if its height
exceeds l/20th of the distance from the source.   An obstruction  whose
crosswind dimension is large compared to its height (width greater than
10 times its height) should be included if its height is greater than
                                 22

-------
l/30th of its distance upstream.  For tall obstructions (height greater
than width), the width replaces the height scale in the above determination
of the critical distances.  If possible, ridges even farther upstream
should be included.  In areas having undulating terrain, the hill  or ridge
height is defined as the elevation difference between its peak and local
trough.  A detailed topographic map and discussion concerning the
selection of the size of the modeled area should be presented in the
study report.

     Additional parameters and criteria should also be considered  in the
selection of the scale of the modeled area.  At this stage of experimental
design, the fluid modeler should select design parameters that can be shown
to satisfy necessary requirements.  The fluid modeler's experience is likely
the best guide in planning the study.

     1.  The buildings, other structures, and/or terrain should be im-
mersed in an appropriate boundary layer that can be characterized  as
representing atmospheric dispersion between that for Pasquill-Gifford
category C and D over flat terrain (Turner, 1970).  The depth of the
model boundary layer, 6, should be scaled to represent 600 m above the
general level of terrain in the field, independent of surface roughness
and wind speed.  The depth of the model boundary layer is not critical
so long as the boundary layer up to the level of the stack and its plume
are scaled appropriately.  The design wind speed will  not be considered
excessive so long as the speed is less than the speed that is exceeded
less than 2 percent of the time (i.e., 98th percentile wind speed).   This
should be based on frequency distributions from at least one year  of
wind records representative of the source location.   A frequency distribution
                                 23

-------
based on categories of specific design wind directions would only be
appropriate if on-site meteorology is used.  Wind speeds greater than the
98th percentile speed could be justified if air quality violations are a
problem at higher wind speeds.  In most cases,  the wind records  must be
extrapolated to estimate the wind speed at stack height.  The reviewing
agency should determine the availability of appropriate wind records
and their appropriate extrapolation to stack height.

     2.  The surface roughness length z0 and the friction velocity u*
should be derived from the mean velocity profile:

             _U_= 2.5 ln-4^— ,                      -     (5)
                               L
                *
in the range, 1.5li  < z < 1 . 5li  + 100 m, where z is the height,  h  is
                  r ~   -     r                                  r
the general height of the surface roughness elements,  and d is the
displacement height (neglected for z0<0.2 m, full  scale).  Simiu and
Scanlan (1978) suggest that reasonable values of d in  cities may be
estimated using the formula
                              d=H~-z0/k,                          (6)

where H is the general roof-top level and k is the von Karman constant
(0.4).  Values of the surface roughness length, z0, for various  types
of surfaces are presented in Table 1  as a guide for a  comparison.
Actual values over urban areas with tall buildings or  near elevated
terrain may be substantially larger.   Values of the friction velocity,
u*, are dependent on the value of z0.  Values for u* as suggested by
Counihan (1975) are presented in Figure 1 , as a guide  for a comparison.
                                 24

-------
Table 1.  Values of Surface Roughness Length (ZQ)  for Various  Types
          of Surface (from Simiu and Scanlan, 1978).
          Type of Surface                                   o
                                                           (cm)
 Sand                                                  0.01  -   0.1
 Sea Surface                                         0.003a  -  0.5b
 Snow Surface                                           0.1  -   0.6
 Mown Grass H).01m)                                   0.01  -     1
 Low Grass, Steppe                                        1  -     4
 Fallow Field                                             2  -     3
 High Grass                                               4  -    10
 Palmetto                                                10  -    30
 Pine Forest (Mean height of trees: 15m;
              one tree per 10m2; zd=12m)                 90  -   100
 Outskirts of Towns, Suburbs                             20  -    40
 Centers of Towns                                        35  -    45
 Centers of Large Cities                                 60  -    80
aWind speed at 10m above surface = 1.5m/sec.
 Wind speed at 10m above surface > 15m/sec.
                              25

-------
                        E
                         6
                                       o
•i- -Q    rd  >
 O -r-  S_
•i- -O  3
 S-  (O  O
M-
     O>  "O
 O) -C  Ol
 (J -t->  _C
 (C     IO
M-  E  to
 S- •!-  Q
 13
 (/)    o
     N    •
-O    -—~
 c ^:  un   •
 re +->  r-^ *-*
     CT) O^t CTl
  "  sz  i— r~.
 Q.  

^2E§
 S-     O  S-
   M-
 O -r-      X
 CL  2  s-  0)
        ,  C
 O   * fO -r-
                                   13  O ID  S-
                                  •r-  o  C  OJ
                                   S-  r-  =5  2
                                   fO  O)  O  O
                                  >  > JD  Q.
                                   O)
                                   cn
26

-------
The vertical  profile of -uw (i.e., the correlation between the  fluctuating
velocity in the streamwise and vertical  directions) may also be used
to estimate the value of u* as demonstrated in the example of Figure  2.
The value of u* is equal to the value of a -uw at the surface which should
be based on the entire profile.

     3.  The mean velocity profile through the entire depth of  the
boundary layer should be represented by a power law U/U^ = (z/6)p.  The
power law index p is dependent on the value of z .  Values for  p as
suggested by Counihan (1975) and Irwin (1979) are presented in  Figure 1,
as a guide for a comparison.

     4.  The surface of the model should be covered with roughness of
size e such that eu*/v _> 20, as suggested by Snyder (1981).  Similarily,
the step size in "stepped" terrain models should be of order e  and
surfaces of buildings or other structures should be covered with roughness
of size e.  Compromises may be appropriate for surfaces having  sharp
edges.  For sharp-edged surfaces, the flow is separated and likely not
significantly affected by roughness on its surface.  Protuberances in
the terrain, buildings, or other structures less than the size  e need
not be reproduced in detail in the model. Similarly, details in the flow
and dispersion pattern are not reproduced for scales less than  the size
£.

     5.  The flow over significant elevated terrain, buildings  or other
structures nearby the source is Reynolds number independent.  For design
purposes, a minimum Reynolds number U,,L/v greater than 11,000 is taken
here as sufficient without demonstration for sharp-edged obstacles.   The
                                27

-------
               1.0 -
               0.8 -
               16
               14
               0.2
                   V
                     £
                      w.
                                     1^=18171/5
a STATION 9.8m
* STATION 13
0 STATION 16
• STATION 19
• STATION 22
o STATION 24
                            0.5
    1.0
                                -uw /u 2
1.5
Figure 2.  Variation  of shear stress with  height measured at various
           downwind  positions in a wind  tunnel  boundary layer  (adiabatic
           flow).  Data from Zoric and Sandborn (1972), (as  in  Snyder 1981)
                                28

-------
reference velocity U,, is the mean velocity upstream at the height of the
obstacle and the reference length L is its lesser dimension (height or
width).  A test of Reynolds number independence should be conducted when
significant effects of flow over terrain or smooth-shaped obstacles are
being considered.

     6.  Blockage of flow (i.e., the ratio of cross-sectional  area of a
model to the cross-sectional area of the test section) is limited to a
percent for an ordinary wind tunnel and to 10 percent in a tunnel with
a properly adjusted ceiling.

     4.1.2  Plume Rise

     The Briggs (1975) formulation as presented in Section 3.4 is
adopted here to provide an estimate of the plume center!ine height for
the source in the field.  The GEP demonstration requires only  that the
stack effluent density, exit speed, and diameter be appropriately
scaled to model the momentum length scale as discussed in Section 3.4.
Vertical profiles of concentration through the plume center!ine can be
used to provide a measurement of the elevated center!ine height in the
fluid model, where reflection from the surface, and influences from
terrain, buildings or other structures are not significant. The measured
elevated plume centerline height should be comparable to estimates as
discussed in Section 3.4.  In all cases the plume height in the model
should be representative of plume rise in the field near the  stack.
However, the plume height farther downwind of the source will  not be
representative of plume rise in the field for highly buoyant sources.
                                 29

-------
     The exhaust from stacks is usually fully  turbulent in  the  field.
The effluent Reynolds number cannot be matched to  assure similarity.   It
is sufficient, however, for the fluid model  effluent  Reynolds  number  to
exceed a critical  value following  the arguments of Reynolds number
independence.  The fluid model  stack effluent  Reynolds  number,  W D/v,
should be guided (in order of decreasing "correctness") by  the  following:


     (a)  Fix the  effluent Reynolds number to  be as large as  possible,
          preferably greater than  15,000.

     (b)  If it is necessary to reduce the effluent Reynolds  number
          below 2,000, trip the flow to ensure a fully  turbulent
          exhaust.  A smoke visualized effluent should  be used  to
          demonstrate a fully turbulent exhaust.

     (c)  If it is desired to reduce the effluent  Reynolds  number
          below 300, it will be necessary to do some  experimentation  to
          determine under what conditions the  plume will  simulate the
          behavior of a plume in the field.


Plume rise must be fixed by matching each of the following  ratios:

               Ws  ps
               ___
               V Pa  Hs
which results in matching the momentum length scale.  To model  situation
with stack effluent downwash around the lip of the stack,  the  flow within
the boundary layer around the stack must be turbulent.
                                 30

-------
     4.1.3  Atmospheric Dispersion Comparability


     In the absence of buildings, other surface structures,  or  large

roughness and/or elevated terrain, dispersion in the fluid model must

show comparability to that described for the atmosphere by the  basic

Gaussian plume distribution (Turner, 1970).   Concentration measurements

for this test of comparability must be compared with values  representative

of field estimates given between estimates for Pasquill-Gifford category

C and D.  The procedure for demonstrating this comparability is outlined

below.  The purpose of this test is to provide an evaluation of the

model flow in absence of buildings, other surface structures or large

roughness, and/or elevated terrain.  This test will  insure that each

study shares some common ground and demonstrates comparability  to

recommended modeling techniques for atmospheric dispersion over flat

terrain (EPA, 1978).



Step 1:
     (a)  Select model scale and the model  flow velocity.   In
          choosing the scale, consideration should  be given to  all
          criteria as outlined in Sections  4.1.1  and 4.1.2  The flow
          velocity should be matched at the height  of the  proposed
          stack.

     (b)  Select the position where the model  stack will be placed.

     (c)  Select the method for providing a fully developed and
          appropriate boundary layer at and downwind of the stack.

     (d)  Report a detailed description of the fluid model.
                               31

-------
Step 2:
     (a)  Measure the mean temperature 9  (°K)  near  the model  surface  and
          at several  positions  within  the freestream  flow.  Additional
          profiles are necessary if operating  conditions  change.   These
          measurements are not  necessary  if model freestream  speeds
          exceed 3 m/s and the  facility is  temperature controlled  during
          the study.

     (b)  Take vertical  profiles of the mean velocity U  (m/s),  and
          the longitudinal turbulence  intensity(u^yu, vertical
          turbulence  intensity(wz )^/u, and -uw  (m2/s) near the position
          where the stack will  be placed, downwind  at the end of the
          planned study area, and midway  between  these two positions
          (3 profiles each).

     (c)  Take lateral profiles of the mean velocity  and  longitudinal
          turbulence  intensity  along the  model surface and two  elevated
          profiles bracketing the range of plume  heights  evaluated in
          the study near the  position  where the  stack will be placed  and
          near the end of the planned  study area  (4-6 profiles  each).
Step 3:
     (a)  Report and evaluate the temperature profiles.   The  profile of
          mean temperature should be uniform.   A deviation  from a  uniform
          profile would indicate that air within the  facility building
          is not well  mixed.

     (b)  Report and evaluate the velocity profiles.   Report  the vertical
          profiles of mean velocity on log-linear scaled  paper and
          estimate the values for the effective surface roughness  length
          z and the friction  velocity u*  at each position,  per
          Equation 5.   Estimate these values by determining the best
          fit to the data representing the lowest 100 m,  full  scale,
          above the height of the surface roughness elements.   Replot
          the profiles of mean velocity on linear scaled  paper
          and estimate the power law index p.   The model  values of z ,
          u*, and p should be consistent  with guidance presented in Table 1
          and Figure 1, representing atmospheric flow over  flat terrain
          with z <0.2 m and 6 = 600m.  Report the profiles  of turbulence
          intensity.  Figure 3 is presented for consideration to be used
          as a guide.   Values of turbulence intensity representative of
          conditions in Figure 3 fo z >0.2 m may indicate the model
          flow is too turbulent.  The best test lies  with the evaluation
          of concentration measurements as discussed  below.  The profiles
          of mean velocity and profiles of turbulence intensity should
          all be similar throughout the study area.   Significant differences
                               32

-------
                  0.05  0.10  0.15  0.20  0.25  0.30  0.35  0.40
                                VJ/L
Figure 3.  Variation of longitudinal  turbulence intensity  with height
           under adiabatic conditions (from Snyder, 1981).
                                33

-------
          in either the downwind  direction  and/or  the  lateral direction
          may indicate a deficiency  in  the  model design or  in the  facility
          operation.   No guidance can be  presently specified for deciding
          how much deviation  is unacceptable.  These profiles should  be
          used only to provide a  qualitative  assessment.

     (c)   Report and  evaluate the profiles  of -uvv  after dividing by  its
          estimated value at  the  surface.   The value at the surface  is
          equal  to the surface friction velocity squared u*2, which
          should be comparable to the estimate determined from  the
          velocity profile.   These profiles should be  used  only to
          provide a qualitative assessment.
Step 4:
     (a)   Position a model  stack,  so  that  the  top of the model  stack  is
          at a height representing 100 m above the ground.   It  is desirable
          at each demonstration  of comparability to use the  same height.
          A 100 m high stack was selected  because it is believed to be
          generally representative of the  height of stacks for  which  GEP
          demonstrations are conducted.  A 200 m high stack  should be used
          for situations where GEP stack height is above 170 m.  Design
          the model  stack so that its internal diameter is equal to
          0.05 times the stack height.  Fix the flow rate of a  nonbuoyant
          stack exhaust containing a  tracer so that the exhaust velocity
          is 1.5 times the mean  velocity at stack top.  This should allow
          concentration measurements  of the tracer to be taken  in absence
          of either plume rise or stack downwash.  Concentration measurements
          for this situation are required  to demonstrate comparability to
          atmospheric dispersion between that  estimated using Pasquill-
          Gifford dispersion parameters for category C and D as presented
          in Turner (1970).

     (b)   Take vertical  and  lateral profiles of concentration through the
          plume center!ine near  the quarter intervals between the source
          and the end of the planned  study area  (3 profiles  each).  At
          least one of these profiles should clearly show the elevated
          plume centerline height in  order to  provide an evaluation of
          the represenative plume rise near the source.  Otherwise, an
          additional profile is  necessary.   Take a ground-level longitudinal
          profile of concentration downwind along the surface ground-level
          centerline to the end  of the study area (1 profile).  Determination
          of the surface ground-level centerline should be supported
          by several ground-level  concentrations in the lateral.
                                 34

-------
(c)   Convert model  concentrations to equivalent field  values with
     the form xUs/Q (m~2).   Plot each vertical  and  lateral  profile
     of concentration measurement separately along  with  plume
     estimates for  both Pasquill-Gifford category C and  D  (Turner,
     1970).   Use the mean wind speed at the  height  of  the  top  of
     the stack U  as the reference wind speed.   The distribution
     of measured values should fall  between  the two estimated
     distributions.  Estimated plume distributions  incorporating
     estimated dispersion parameters may be  presented.   Distinct
     elevated center!ine concentrations should  fall  between the
     estimates.

     Plot the ground-level  longitudinal profile of  concentration
     measurement along with plume estimates  for both Pasquill-
     Gifford category C and D (Turner, 1970).   The  distribution of
     measured values should fall  between the two estimated distri-
     butions with an additional  allowance where these  two  distri-
     butions overlap as presented in Figure  4.   The dashed lines
     in Figure 4 allow a factor of two differences  in  the  over-
     lapping region.  The concentrations should fall  between the
     estimates at least downwind to the distance of maximum
     ground-level concentration.   It is critical that  the  ground-
     level  concentration measurement here and the above  vertical
     and lateral concentration measurement not  be representative of
     estimates for  more stable situations than  category  D.
     Representation by categories more unstable than C are not
     considered critical since a determination  of the  GEP  stack
     height will likely be less than that resulting under  more
     stable conditions since the more turbulent atmospheric flow
     should somewhat overshadow the local  building/terrain effect.
                            35

-------
            2x10 5
            5x10'6 —
          CSI
           E
           cf
           a

           X2x10'6
            5x10'7 —
            2x10'7
                0.5
                                 DISTANCE, km
Figure 4.  Ground-level  concentration with distance for a  100m  high
           plume, estimated  with  Pasquill-Gifford dispersion  parameters
           for stability category C and D (Turner, 1970).
                                  36

-------
4.2  Determination of GEP Stack Height





     Requirements and procedures for evaluating the model  boundary layer



characteristics and dispersion for the actual  situation are specified



below.  The results of the previous section establish fluid model  comparability



to atmospheric dispersion as estimated by recommended mathematical



models in the absence of buildings, other structures or large roughness,



and/or elevated terrain.  Differences between  the model boundary layer



characteristics and dispersion for the standard situation  established in



the previous section and the actual situation analyzed below should be



related to real expected differences in the field due to the effects of



buildings, other structures or large roughness, and/or elevated terrain.
                               37

-------
     4.2.1  Demonstration of Adverse Effects

     The procedure for demonstration and  documentation  of  the  adverse

effects of buildings and/or elevated terrain  nearby  the source is  outlined

below.

Step 1:  Place the model  topography  into  the  test  facility.


     (a)  For modeling the situation of a few isolated  buildings  in
          flat terrain, this is  a  simple  matter.   The buildings are
          simply immersed in the boundary layer designed for satisfying
          the requirements of Section 4.1.3.

     (b)  Additional complexity  arises for a  model of flow over a
          general  urban area and/or  elevated  terrain.   In  addition
          to constructing the model  of buildings and/or elevated
          terrain, surface roughness used in  covering the  model  and
          general  roughness elements upwind of the model to provide
          appropriate boundary layer characteristics for the situation
          may be different from  those used in the  atmospheric  dis-
          persion comparability  test (Section 4.1.3.).

     (c)  Report a detailed description of the fluid model.

Step 2:

     (a)  Take profiles of mean  velocity  and  longitudinal  turbulence
          intensity as in Step 2 (b) Section  4.1.3   (3  profiles each).

     (b)  Take profiles of mean  velocity  and  longitudinal  turbulence
          intensity as in Step 2 (c) Section  4.1.3  (4-6 profiles  each).

     (note)  There is no need to repeat measurements that  would have
     values identical to those satisfying requirements  in  Section  4.1.3.
     (i.e., measurement in the same  boundary  layer beyond  the  influence
     of the building and/or terrain).
Step 3:
     (a)  Follow Step 3 (b) Section 4.1.3  to  determine  the model
          values of z , u*, and p.   They should  be  consistent with
          guidance presented in Table 1  and  Figure  1.   Specific
          guidance for flow over areas of  elevated  terrain  is not
          available since little data correlating the  flow  character-
          istics to the height and  separation distances of  terrain
                                 38

-------
          features has been taken.  In general, z  in areas with elevated
          terrain should be much larger than the largest values
          found in Table 1.  Significant differences may be found
          at each location due to differences in the local  surface
          roughness.  A discussion of such differences should be pre-
          sented in the study report.

     (b)  Follow Step 3 (c) Section 4.1.3.

Step 4:  Test for Reynolds number independence.

     (a)  For sharp-edged obstacles having a Reynolds number U,,L/v
          greater than 11,000, no demonstration test is required.

     (b)  Where the effects of flow over elevated terrain or smooth-
          shaped obstacles are being evaluated, a Reynolds number test
          is required.  A full evaluation of Reynolds number inde-
          pendence would require a demanding research project for each
          situation.  The simple test required here should, however,
          be sufficient enough to provide a critical evaluation.
          Position a small source emitting a tracer at the site of
          the GEP stack in question at a height equal to the building
          or elevated terrain whose effects are in question.  The
          source should be nonbuoyant and have no plume rise.  Take
          a longitudinal surface-level profile of concentration along
          the downwind direction.  Repeat the profile after at least
          doubling the freestream wind speed U^.  Take a vertical  profile
          of the mean velocity at the site of the stack for this new
          situation.  Plot and compare the two profiles of concentration
          xLI/Q (m~2) using the freestream wind speed as the reference
          speed.  Differences in concentration should not be greater than
          10 percent.  Reconsideration of the model design is necessary
          where greater differences are observed.

Step 5:  An evaluation of the plume from the stack in question must be
made.  In general, the fluid modeler should first examine the plume
through a visualization technique, i.e., photographs of smoke exhaust.
Then a decision can be made as to the height for which GEP credit can  be
justified.  Photographs and/or measured data taken as part of this
process must be included in this report.  Full documentation as outlined
below is required for the actual determination of GEP stack height.

     (a)  Take vertical and lateral profiles of concentration through
          the plume centerline at positions one-fourth and one-half of
          the distance between the source and the end of the study area.
          Also, do the same at the end of the study area and at the
          position of maximum ground-level concentrations.   The value  of
                                  39

-------
     maximum ground-level  concentration must be unquestionably
     determined.   This  requires a longitudinal surface-level
     profile along the  plume centerline, supported by 2 to 4 partial
     lateral  profiles including one across  the position of maximum
     ground-level  concentration.

     In  some situations, it may be necessary to model at a scale
     such  that  the likely  maximum ground-level concentration falls
     downwind beyond the modeled area.  This is very undesirable,
     and should be avoided whenever possible.  Before such an
     approach is  planned,  the agency should first review a proposal,
     discussing its necessity.  The diffusive characteristics
     for a limited region  beyond the modeled area may be
     obtained by  extrapolating the measured values, only for
     situations of flow over generally homogeneous terrain and/or
     uniform urban environments.  For these situations, additional
     vertical and  lateral  profiles of concentration at the position
     three-fourths between the source and the end of the study area
     should be  made in  lieu of measurements at the maximum.  Extrap-
     olation should be  limited to distances equivalent to one-half
     the distance  between  the source and end of the modeled area.
     Ground-level  concentration profiles for several stack
     heights  having their  maximum value falling within the
     modeled area  should be made.  These profiles can be used
     to  extrapolate a maximum value for higher stack heights,
     and support  the value obtained by extrapolating the measured
     vertical and  lateral  profiles of concentration.

(b)   Convert model concentrations to equivalent field values with
     form  xUs/Q(m"2).   Take the mean wind speed at the height of
     the top of the stack  U  as the reference wind speed.  Report
     each  vertical  profile of concentration measurement.  In absence
     of  reflection from the surface, the plume centerline can be
     estimated  as  the vertical position of  the maximum concen-
     tration.   At  least one value must be compared to the esti-
     mated plume  rise as discussed in Section 4.1.2.  Additional
     vertical profiles  must be measured if  the above required
     profiles do  not provide sufficient information.  Report
     each  vertical and  lateral profile of concentration measure-
     ments separately along with estimated  plume distributions
     incorporating estimated dispersion parameters.

     In  areas of  elevated  terrain downwind  from the source, such
     estimates  may be difficult and perhaps meaningless.  In
     these instances a  discussion relating  the plume behavior to
     anticipated  effects of the terrain is  needed.  Plot the longi-
     tudinal  and  lateral profiles of concentration measurements.
                            40

-------
          The maximum must be unquestionably determined.   Two  repeated
          measurements at the positions of the maximum should  be  taken
          and reported as support that the concentration  does,  in fact,
          represent the steady-state average.  Differences in  the
          three concentrations should not be greater than 10 percent
          of their values.

          In some situations of flow over generally homogeneous terrain
          and/or uniform urban environments, the maximum  value may be
          obtained by extrapolation as discussed above.   For these
          instances, it is necessary that the appropriate dispersion
          parameters can be derived from the vertical  and lateral
          profiles of concentration measurements.  The maximum ground-
          level concentration can then be estimated by inserting
          the derived dispersion parameters into the Gaussian  plume
          formula (Turner, 1970).  Extrapolation of measured longitudinal
          ground-level profiles should support the estimated maximum
          ground-level concentration.  Approval for such  a study  plan
          must be granted by the reviewing agency prior to the study to
          allow agency experts to provide a critical assessment.

     4.2.2  GEP Stack Height

     The stack height for which full documentation has been provided in

the previous sub-section is GEP if the maximum surface-level concen-

tration is 40 percent or more in excess of the maximum in the  absence of

downwash, wake, or eddy effects produced by nearby structures  or  terrain.

The procedure for validating the proposed GEP stack height is  presented

below.
Step 1:
          Remove the building(s) or elevated terrain  in  question.   This
          is a simple matter in the case of buildings nearby  the  stack.
          The situation near elevated terrain is  complicated  since
          removal of the terrain feature in question  may result in  an
          unrealistic discontinuity in the topography.   A similar difficulty
          arises where a high plateau is upwind of the source.  In
          such instances it may be necessary to remove all  upwind
          terrain and replace its area with appropriate  surface roughness.
          The surface roughness elements must be  shown to result  in  an
                                 41

-------
Step 2:
          appropriate zo and u*.   This  requires -that a  vertical  profile
          of mean velocity,  longitudinal  turbulence intensity,  and
          shear stress be measured upwind of the  source at  the  position
          where the elevated terrain feature was  located.   Estimate  z0
          and u* guided by Step 3  (a,b)  Section 4.2.1.   The boundary
          layer should be appropriately  characterized as with the
          actual topography.
          Determine the maximum ground-level  concentration.   Document
          fully as required by Step 5,  Section  4.2.1.
Step 3:

          The proposed stack height is  creditable as GEP  if  the  maximum
          ground-level concentration determined in Step 5, Section  4.2.1
          is at least 40 percent in excess  of the maximum ground-level
          concentration determined in Step  2  above.  Discussion  relating
          the increased maximum ground-level  concentration measured
          in the presence of the building(s)  or terrain in question
          to anticipated effects due to downwash, wakes,  or  eddies
          should be presented in the report.
                                 42

-------
5.0  REPORT CHECKLIST

     The fluid modeling study report should include the five items that
are outlined below.  The report should completely document the design
and operation of the model study.  Three tests should be conducted.   The
data collected for these tests must allow for conclusions to be drawn
concerning the atmospheric conditions simulated by the fluid model,  and
the cause of increased maximum ground-level concentrations for a stack
in the presence of nearby building(s) and/or terrain.  The height of the
stack examined in the study is creditable as GEP if the maximum ground-
level concentration is at least 40 percent greater in the presence of
the nearby building(s) and/or terrain than that measured in their absence.
The report should include at least the following:

          1. A detailed topographic map and discussion concerning the
selection of the size of the modeled area and meteorological  parameters.
          2.  Discussion of model design and similarity criteria.
          3.  An evaluation of the test facility in the absence of building(s),
other surface structures, or large roughness and/or elevated terrain
including:

         (a)  a detailed description of the fluid model including features
of the scale model, surface roughness, velocity profile, and method  used
to provide the fully developed boundary layer,

         (b)  one representative vertical profile of mean temperature if
model free-stream speed is less than 3 m/s,
                                 43

-------
          (c)  three vertical  profiles  of mean velocity,

          (d)  three vertical  profiles  of longitudinal and vertical
turbulence intensity,

          (e)  three vertical  profiles  of uw,

          (f)  four lateral  profiles  of mean velocity,

          (g)  four to six lateral  profiles of longitudinal turbulence
intensity,
          (h)  effective surface  roughness length  z  , friction velocity
u*, and velocity power law index  p, determined by  evaluating the mean
velocity profiles and the shear stress  profile,

          (i)  three vertical  and lateral profiles of concentration
through the elevated centerline of the  plume,

          (j)  one ground-level longitudinal profile of concentration
downwind along the plume centerline,

          (k)  evaluation of comparability of measured concentrations to
plume estimates for Pasquill-Gifford  category C  and  D (Turner, 1970),

          (1)  evaluation of the  measured elevated centerline of the plume.

4.  Documentation for the GEP stack height test  in the presence of
building(s), other surface structures,  or large  roughness and/or elevated
terrain should include:
                                 44

-------
          (a)  detailed description of the fluid model  including  features
of the scale model, surface roughness, velocity profile,  and  the  method
used to provide the fully developed boundary layer,

          (b)  three vertical profiles of mean velocity,

          (c)  three vertical profiles of longitudinal  and  vertical
turbulence intensity,
          (d)  three vertical profile of uw ,
          (e)  effective surface roughness length ZQ, friction velocity
u*, and velocity power law index p, determined by evaluating  the  mean
velocity profiles and the uw profi1es,
          (f)  possible test for Reynolds number independence,
          (g)  four vertical  and lateral  profiles of concentraion
through the elevated center!ine of the  plume including profiles at  the
position of maximum ground-level concentration,
          (h)  one ground-level longitudinal  profile of concentration
downwind along the plume centerline at  ground-level,
          (i)  two to four lateral ground-level  profiles including  one at
the position of maximum ground-level  concentration,
          (j)  discussion supporting  the  unquestionable determination of
the maximum ground-level concentration,
          (k)  evaluation of  the measured elevated centerline of  the plume.
                                 45

-------
     5.  Documentation for the GEP stack  height  test  in the absence of
building(s) or elevated terrain considered  in justifying the stack
height should include:
          (a)  the same as steps (g),  (h),  (i),  (j),  and (k) above,
          (b)  discussion relating the  increased maximum ground-level
concentration measured in the  presence  of the building(s) or elevated
terrain in question to anticipated effects  due to downwash, wakes, or
eddies,
     6.  A separate section or appendix describing the fluid modeling
facility, instrumentation used in the conduct of the  study, and their
normal operating conditions and associated  parameters.
                                 46

-------
6.0  REFERENCES

Briggs, G. A., 1975:  Plume Rise Predictions.  In Lectures on Air Pollution
and Environmental Impact Analysis, American Meteorological Society,
Boston, Massachusetts.

Counihan, J., 1975:  Adiabatic Atmospheric Boundary Layers:  A Review
and Analysis of Data from the Period 1880-1972.  Atmos.  Envir., 9_, (10),
pp. 871-905.

Environmental Protection Agency, 1978:  Guideline on Air Quality Models.
EPA-450/2-78-027, Research Triangle Park, North Carolina.

Environmental Protection Agency, 1980:  Guideline for Determination of
Good Engineering Practice Stack Height (Technical Support Document
for Stack Height Regulations), EPA-450/4-80-023,  Research Triangle Park,
North Carolina.

Irwin, J. S., 1979:  A Theoretical Variation of the Wind Profile Power-
Law Exponent as a Function of Surface Roughness and Stability.  Atmos.
Envir., 13., (1), pp. 191-194.

Simiu, E. and R. H. Scanlan, 1978:  Mind Effects  on Structures. John Wiley
and Sons, New York, New York, 458 p.

Snyder, William H., 1981:  Guideline for Fluid Modeling  of Atmospheric
Diffusion.  EPA-600/8-81-009, Research Triangle Park, North Carolina.

Turner, D. B., 1970:  Workbook of Atmospheric Dispersion Estimates.
Publication No. AP-26, Office of Air Programs, Environmental  Protection
Agency, Research Triangle Park, North Carolina.

Zoric, D. L. and V. A. Sandborn, 1972:  Similarity of Large Reynolds
Number Boundary Layers.  Boundary Layer Meteorol., 2_, pp.  326-33.

-------
TECHNICAL REPORT DATA
(P!eare read in>tru<.tion; on tlu reverse before comf'irting)
1. REPORT NO. 1 2.
EPA-450/4-81-003 |
4. TITLE AND SUBTITLE
Guideline for Use of Fluid Modeling to Determine Good
Engineering Practice Stack Height
7. AUTHOR(S)
9. PERFORMING ORGANIZATION NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
12. SPONSORING AGENCY NAME AND ADDRESS
3, H£CfP'E:MT'S ACCESSION NO.
I
5 REPORT DATE
July 1981
6 PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Guidelines for developing and reviewing fluid modeling studies for
determining good engineering practice stack height. Includes review of appropriate
fluid modeling theory and a specific report requirement checklist.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS |b. I DENTI F IERS/OPEN ENDED TERMS C. COSATI Field/Group
Air Pollution j Air Pollu
Good Engineering Practice Stack Height
Physical Modeling
Fluid Modeling
I
18. DISTRIBUTION STATEMENT 19 SCCU^.fv
tion Control 13-B
CLASS , r>r^ f~'fc->rtj 21 NC. OF PAGES
Unlimited 20 SFCUR.TY';LA:?3 Tmspajie' '22 PRICE
unclassified
^
EPA Form 2220-1 (Rev. 4-77)    PREVIOUS EDITION i s OBSOLETE

-------