xvEPA
United States
Environmental Protection
Agency
Air Pollution Training Institute
MD20
Environmental Research Center
Research Triangle Park, NC 27711
EPA 450/2-82-007
March, 1983
Air
APTI
Course Sl:410
Introduction to
Dispersion Modeling
Self-instructional
Guidebook
-------
United States
Environmental Protection
Agency
Air Pollution Training Institute
MD20
Environmental Research Center
Research Triangle Park, NC 27711
EPA 450/2-82-007
March, 1983
Air
APTI
Course Sl:410
Introduction to
Dispersion Modeling
Self-instructional
Guidebook
Developed by:
Donald R. Bullard
Peter Guldberg
Marilyn M. Peterson
Northrop Services, Inc. :
P.O. Box 12313
Research Triangle Park, NC 27709
Under Contract No.
68-02-3573
EPA Project Officer
R. E. Townsend
United States Environmental Protection Agency
Office of Air, Noise, and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, NC 27711
-------
Notice
This is not an official policy and standards document. The opinions and selections
are those of the authors and not necessarily those of the Environmental Protection
Agency. Every attempt has been made to represent the present state of the art as
well as subject areas still under evaluation. Any mention of products or organiza-
tions does not constitute endorsement by the United States Environmental Protec-
tion Agency.
11
-------
Unit 1
Introduction to
Course Materials
Course Description
Course Goal and Objectives
Requirements for Successful Completion of this Course
Materials
Using the Guidebook
Instructions for Completing the Final Examination
1-1
-------
Course Description
This training course is a 35 Vi -hour self-instructional course
using slide/tape presentations, text materials, and reading
assignments dealing with dispersion models for industrial point
sources. Models and their use in determining air pollution
impact areas, such as the urban area in Figure 5-1, and
ground-level concentrations will be examined in two case
studies. Course topics include the following:
• Introduction to the regulations requiring air quality
model use
• Introduction to air quality models for industrial point
sources
• General characteristics of air quality models for industrial
point sources
• Review of UNAMAP, Version 4 models*
• Input data required for specific models
• Interpreting the output data from specific models
• Case studies
Figure 1-1. Dispersion modeling
concerns.
Course Goal and Objectives
Goal
The purpose of this course is to familiarize you with the
general concepts and specific data requirements of air quality
models for industrial point sources for you to use to make
competent decisions about the impact of air pollution on air
quality.
Objectives
Upon completing this course, you should be able to:
1. cite the specific parts of the Federal regulations that
require the modeling of air pollution concentrations.
2. name and describe the original air quality modeling
technique used in formulating State Implementation
Plans (SIPs).
3. describe one typical atmospheric pollution problem that
can be solved using air quality modeling.
4. describe the basic Gaussian approach for an atmospheric
dispersion model for industrial point sources.
•Version 5 is scheduled for release in 1983 and will add several new models
to UNAMAP. All of the models discussed in this course will remain
unchanged with the release of Version 5.
1-3
-------
5. explain the rationale for using the Gaussian distribution
in atmospheric dispersion models for industrial point
sources.
6. list the atmospheric models for industrial point sources
that are available on UNAMAP, Version 4.
7. list the limitations of Gaussian-based atmospheric disper-
sion models for industrial point sources.
8. describe the method of obtaining model input data for
industrial point sources.
9. explain typical input data for an atmospheric dispersion
model for industrial point sources.
10. choose the specific section of a given model's output
data that computes ground-level concentrations.
Requirements for Successful Completion
of this Course
In order to receive 3.5 Continuing Education Units (CEUs) and
a certificate of course completion, you must:
1. take a mail-in final examination.
2. achieve a final exam grade of at least 70%.
Materials
Additional Required Reading
EPA 450/2-78-027, Guideline on Air Quality Models,
April 1978.
Audiovisual
Slide/tape presentations:
• SI:410-1 Introduction to Air Quality Regulations
• SI:410-2 Introduction to Air Quality Modeling
• SI:410-3 Air Quality Modeling Summary
Supplementary
EPA 450/4-77-001, Guidelines for Air Quality Maintenance
Planning and Analysis, Volume 10 (Revised), Procedures for
Evaluating Air Quality Impact of New Stationary Sources,
October 1977.
NASA SP-322, A Review of Methods for Predicting Air Pollu-
tion Dispersion, 1973.
EPRI EA-1131, Appendix D: Available Air Quality Models,
Electric Power Research Institute, Palo Alto, CA,
August 1979.
1-4
-------
DOE/TIC-11223, Handbook on Atmospheric Diffusion,
U.S. Department of Energy, 1982.
45 Federal Register 52676, "Requirements for Preparation,
Adoption, and Submittal of Implementation Plans;
Approval and Promulgation of Implementation Plans,"
August 7, 1980.
Using the Guidebook
This guidebook directs your progress through the slide/tape
presentations, text material, and reading assignments. It con-
tains seven units consisting of reading, supplementary, and
audiovisual materials. The first unit introduces the course. The
second unit, containing three lessons, has two slide/tape
presentations and a reading assignment that will give an over-
view of air .quality regulations and air quality models. The next
four units will be self-paced, presented as text with review
questions. The last unit briefly summarizes the major points of
the course in a slide/tape presentation.
Completing the Review Exercises
Complete the review exercise for each lesson upon completing
the reading assignments and slide/tape presentations for that
lesson. If you answered any review exercise incorrectly, review
the reading assignment and/or slide/tape script. Then proceed
to the next lesson in the guidebook.
To complete a review exercise, place a piece of paper across
the page covering the questions below the one you are answer-
ing. After answering the question, slide the paper down to
uncover the next question. The answer for the first question
will be given on the right side of the page separated by a line
from the second question (Figure 1-2). All answers to review
questions will appear below and to the right of their respective
questions. The answers will be numbered to match the
questions.
Using the Slide/Tapes
The audiocassettes and slide sets have been numbered con-
secutively. Table 1-1 lists tape number, slide series numbers,
and appropriate lesson number. The script for each presenta-
tion can be found in the unit and lesson number listed.
You do not need to follow the script provided in the appro-
priate lesson as you view each slide/tape presentation. The
script is provided for you to use to review the content.
The audiocassettes provided with the course materials will
most likely have an audible slide change tone. Begin the tape
Review Exercise
i.
2.
3.
^-.
Question loulo
nlli cllo yllonnlit
Questionoli oul
h iiliionyic o
Question > m lot
.jlp nil i cllo yllon
^>""' >w r
1. Answer
llllO
2. Answer
*ji~"iil|'
Figure 1-2. Review exercise format.
Table 1-1. Crow-listing of slide/tape
presentations with unit
and lesson numbers.
Slide/tape
audio cassette
number
1
2
3
Slide
numbers
1-1
through
1-35
2-1
through
2-34
3-1
through
3-18
Unit
number
2
2
7
Lesson
number
1
2
1-5
-------
with the first slide showing and then advance the slides
manually as you hear each tone.
If you have requested audiocassettes with inaudible tones
that automatically advance the slides, begin the tape with the
first slide showing. Should you need to use these tapes for
manual advance, consult the scripts for slide change points.
Lesson Content
• Reading assignments (if ones in addition to this guidebook
are required)
• Slide/tape presentation: slide numbers and cassette number
(if applicable)
• Lesson goal and objectives
• Reading guidance (if applicable)
• Text of lesson (except where readings from other documents
are specified) or script from slide/tape presentations
• Review exercise and review exercise answers
If supplementary reading material is available, it will be
recommended in the appropriate lesson, but it is not required
for course completion.
Instructions for Completing
the Final Examination
Contact the Air Pollution Training Institute if you have any
questions about the course or when you are ready to receive a
copy of the final examination.
After completing the final exam, return it and the answer
sheet to the Air Pollution Training Institute. The final exam
grade and course grade will be mailed to you.
Air Pollution Training Institute
Environmental Research Center
MD 20
Research Triangle Park, NC 27711
1-6
-------
Unit 2
Introduction to Air
Quality Regulations
and Air Quality
Modeling
Lesson 1 Introduction to Air Quality Regulations
Lesson 2 Introduction to Air Quality Modeling
Lesson 3 Introduction to Case Studies
2-1
-------
Lesson 1
Introduction to
Air Quality Regulations
Slide/Tape Presentation
First, view the slide/tape presentation—cassette no. 1 and
keyed slides 1-1 through 1-35, Introduction to Air Quality
Regulations—then complete the reading assignment.
Reading Assignment
EPA 450/2-78-027, Guideline on Air Quality Models,
pp. 1-12.
Section 165 of the Clean Air Act Amendments of 1977,
p. 2-5 of this guidebook.
Title 40, Part 51.24(1) of the Code of Federal Regulations,
p. 2-5 of this guidebook.
National Ambient Air Quality Standards, p. 2-6 of this
guidebook.
Prevention of Significant Deterioration Increments, p. 2-6 of
this guidebook.
Supplementary Reading
45 Federal Register 52676, "Requirements for Preparation,
Adoption, and Submittal of Implementation Plans; Approval
and Promulgation of Implementation Plans," August 7, 1980.
Lesson Goal and Objectives
Goal
To familiarize you with the regulations that require air quality
modeling and the manner in which models are required to be
used in Control Strategy Evaluation, New Source Review, and
Prevention of Significant Deterioration programs.
2-3
-------
Objectives
Upon completing this lesson, you should be able to:
1. cite the specific part number of the Clean Air Act
Amendments of 1977 that requires air quality modeling.
2. cite the specific section of the Code of Federal Regula-
tions that requires air quality modeling be used in
estimating ambient concentrations.
3. name the three regulatory programs using air quality
models as specified in the Guideline on Air Quality
Models.
4. describe the concept of the "highest, second-highest"
concentrations.
2-4
-------
Clean Air Act: Section 165. Preconstruction Requirements Requiring Use of Air Quality
Models.
Sec. 165. (a) No major emitting facility on which construction is commenced after the date of
enactment of this part may be constructed in any area to which this pan applies unless—
(1) a permit has been issued for such proposed facility in accordance with this part setting
forth emission limitations for such facility which conform to the requirements of this
part;
(3) The owner or operator of such facility demonstrates that emissions from construction or
operation of such facility will not cause, or contribute to, air pollution in excess of any
(A) maximum allowable increase or maximum allowable concentration for any pollu-
tant in any area to which this part applies more than one time per year, (B) national
ambient air quality standard in any air quality control region, or (C) any applicable
emission standard or standards of performance under this Act;
(3) The Administrator shall within six months after the date of enactment of this part prom-
ulgate regulations respecting the analysis required under this subsection which regulations—
(D) shall specify with reasonable particularity each air quality model or models to be
used under specified sets of conditions for purposes of this part.
Any model or models designated under such regulations may be adjusted upon a determination,
after notice and opportunity for public hearing, by the Administrator that such adjustment is
necessary to take into account unique terrain or meteorological characteristics of an area poten-
tially affected by emissions from a source applying for a permit required under this part.
Title 40, Part 51.24 of the Code of Federal Regulations
(1) Air quality models. (!) The plan (State
Implementation Plan) shall provide for pro-
cedures which specify that—
(i) All estimates of ambient concentrations
required under paragraph (1) shall be based
on the applicable air quality models, data
bases, and other requirements specified in
the Guideline on Air Quality Models (OAQPS
1.2-080, U.S. Environmental Protection
Agency, Office of Air Quality Planning and
Standards, Research Triangle Park, NC
27711, April 1978).
(ii) Where an air quality impact model
specified in the Guideline on Air Quality
Models is inappropriate, the model may be
modified or another model substituted.
(iii) A substitution or modification of a
model shall be subject to public comment
procedures developed in accordance with
paragraph (r) of this section.
(iv) Written approval of the Administrator
must be obtained for any modification or
substitution.
(v) Methods like those outlined in the
Workbook for the Comparison of Air Quality
Models (U.S. Environmental Protection
Agency, Office of Air Quality Planning and
Standards, Research Triangle Park, NC
27711, April 1977) should be used to deter-
mine the comparability of air quality models.
(2) The Guideline on Air Quality Models is
incorporated by reference. On April 27,
1978, the Office of the Federal Register
approved this document for incorporation by
reference. A copy of the guideline is on file in
the Federal Register library.
(3) The documents referenced in this
paragraph are available for public inspection
at EPA's Public Information Reference Unit,
Room 2922, 401 M Street SW., Washington,
B.C. 20460, and at the libraries of each of
the ten EPA Regional Offices. Copies are
available as supplies permit from the Library
Service Office (MD-35), U.S. Environmental
Protection Agency, Research Triangle Park,
NC 27711. Also, copies may be purchased
from the National Technical Information
Service, 5285 Port Royal Road, Springfield,
Va. 22161.
2-5
-------
Table 2-1. National Ambient Air Quality Standard* (NAAQS).
Pollutant
Sulfur dioxide (SOj)
Total suspended
participates (TSP)
Carbon monoxide (CO)
Ozone (Oj)
Nitrogen dioxide (NO*)
Lead (Pb)
Averaging
time
Annual arithmetic mean
24 hours
3 hours
Annual geometric mean
24 hours
8 hours
1 hour
1 hour
Annual arithmetic mean
3 months
Primary
standards
80 ng/m*
(0.03 ppm)
365 iig/m*
(0.14 ppm)
75 /ig/m3
260 /«g/ms
10 mg/m3
(9 ppm)
40 mg/m3
(35 ppm)
240 0g/m3
(0.12 ppm)
100 ng/m'
(0.05 ppm)
1.5/^/m3
Secondary
standardi
1300 /jg/m3
(0.5 ppm)
60 pg/m3*
150 /ig/m3
Same as primary
Same as primary
Same as primary
Same as primary
Note: National standards other than those based on annual arithmetic means or annual
geometric means are not to be exceeded more than once per year. All standards are deter-
ministic except for ozone, which is based on a statistical definition.
National primary standards: the levels of air quality necessary, with an adequate margin
of safety, to protect the public health.
National secondary standards: the levels of ah- quality necessary to protect the public
welfare from any known or anticipated adverse effects of a pollutant.
•Guideline to be used assessing implementation plans.
Table 2-2. Prevention of Significant Deterioration (PSD) increments.
Pollutant
Paniculate matter:
Annual geometric mean
24-hour maximum
Sulfur dioxide:
Annual arithmetic mean
24-hour maximum
3-hour maximum
Maximum
I
5
10
2
5
25
allowable increase
Claw
II
19
37
20
91
512
OVm5)
III
37
75
40
182
700
Note: Increments other than those based on annual means are not to
be exceeded more than once per year. The full increment is
not to be used if it would result in a violation of a NAAQS.
Increment consumption is limited to half the maximum
allowable at State borders.
2-6
-------
Introduction to
Air Quality Regulations
Slide no. Script
1. Focusing slide—no narrative
2. The 1977 Clean Air Act Amendments require air quality
modeling to help improve our air quality and keep pollu-
tion concentrations below certain levels.
3. An air quality model is used to determine the effect of air
pollution on ambient air—that is, on the air that is around
us.
4. The regulations that were issued call for modeling in three
programs: Prevention of Significant Deterioration, New
Source Review, and Control Strategy Evaluations.
5. The first program, Prevention of Significant Deteriora-
tion, or PSD, was designed to prevent air quality from
deteriorating in areas where it is already better than
required by the National Ambient Air Quality Standards.
Modeling is used to verify that the air quality does not
exceed these standards.
6. Let's take an example. In recreational areas like national
parks, wilderness areas, and other protected areas, the
ambient air is to remain relatively free from industrial and
other pollution sources.
7. If sources are in the area, the air quality may change by
certain amounts over time, and this change is specified in
the regulations.
8. These concentration increases are called PSD increments.
They define the amount that pollutant concentrations can
increase from a set baseline for all future time. The PSD
increment system is divided into three land-use classes,
based upon the amount of air quality degradation to be
allowed. Class I lands, which include most wilderness areas
and national parks, are protected the most.
Selected visuals*
FOCUS
Introduction to
Air Quality Regulations
Prevention of Significant Deterioration
New Source Review
Control Strategy Evaluations
Prevention of
Significant Deterioration
* Illustrations included here, no live shots included.
2-7
-------
Slide no.
Script
Selected visuals
9. For instance, for short-term periods, like 3-hour or 24-hour
periods, maximum concentration increments may be
exceeded only once a year.
10. In other words, since there are 365 days per year, the
maximum 24-hour increment could be exceeded only one
day in that 365-day period.
11. The 1977 Clean Air Act Amendments do allow a pollution
source to apply for a variance through which increments
may be exceeded more than once.
12. New Source Review is the second program for which
modeling is used. New Source Review is tied to PSD and to
other programs. It concerns the effects on air quality of
either building new pollution sources or making certain
modifications to existing sources.
IS. When a new source of pollution is to be built, or an
existing source is to be changed, modeling can be used to
help predict that source's effect on air quality. However,
modeling is not always used. A new or modified source
must meet certain criteria before modeling is required.
AflCU
New Source Review
14. In general, the impact on air quality of a new source only
has to be modeled if it is a major source. The
determination of whether a source is major or not is based
on its potential emissions. Potential emissions are the
emissions at maximum design capacity after the
application of pollution control technology and operating
restrictions.
2-8
-------
Slide no.
Script
Selected visuals
15. A major source has the potential to emit 100 tons or more
per year of any pollutant regulated by the Clean Air Act
for certain designated source categories, and 250 tons or
more per year for all other sources. The sources in the
100-ton category include large fossil fuel-fired power
plants, kraft pulp mills, smelters, steel mills, and oil
refineries.
16. The third of our three programs requiring modeling is
called Control Strategy Evaluation.
17. Individual States are required to have State
Implementation Plans, or SIPs, for air pollution control.
These implementation plans describe the methods by
which each State intends to control air pollution within its
borders.
Q \
|oo •
O \
IOO' fi-fll
18. These plans are designed to ensure that each State meets
the National Ambient Air Quality Standards for each
pollutant believed to adversely affect public health or
welfare. These are known as criteria pollutants.
19. In order to evaluate whether or not the plans are effective,
the air quality must be modeled by an air quality model
that has been accepted for regulatory evaluations by the
United States Environmental Protection Agency.
NAAQS
o
|oo •
O O
O
n> i
2-9
-------
Slide no.
Script
Selected visuals
20. Short-term (24 hours or less) pollutant concentration
estimates are for four criteria pollutants: ozone, sulfur
dioxide, total suspended particulates, and carbon
monoxide.
21. These estimates of air quality are based on a concept
called the "highest, second-highest" concentration, which is
consistent with EPA's definition of when an air quality
standard is violated. That is, the short-term standard must
be exceeded two or more times in a year for a violation to
occur. The concept requires making air quality estimates
downwind of a pollution source at a number of different
points, called receptors.
O
22. The pollutant concentrations determined from modeling
are ranked from highest to lowest for each receptor site.
The highest concentration from each receptor's data is
discarded.
23. Then, the next observed single-highest concentration
determined from all of the receptor estimates is chosen as
the "highest, second-highest" concentration.
24. There are times when the "highest, second-highest"
method of selecting concentrations cannot be used: for
ozone, for an inadequate data base or unrepresentative
model, and for unidentified sources.
/ Exceptions to Highest, \
Second Highest Concentration
• Inadequate data base or
unrepresentative model
• unidentified »ources
2-10
-------
Slide no.
Script
Selected visuals
25. The first exception is ozone.
O O
Q
26. To determine expected violations for ozone, statistical
methods are used rather than the "highest, second-highest'
method.
27. Another exception is when the Regional Administrator
identifies an inadequate data base or an unrepresentative
air quality model. An inadequate data base occurs when
not enough data is available. An unrepresentative air
quality model is one that cannot adequately simulate a
particular physical situation.
/" \
t2i-
Ozone: }
Statistical
Method
Averages
Violations 1
Over Time/
/ O O <^j> O \
1 *^ XN « <^a^> 1
28. The last exception to using the "highest, second-highest"
concentration as an estimate is when maximum
concentrations are caused by sources that cannot be
identified. When air quality monitoring data from specific
sites indicate that existing concentrations are greater than
those predicted by the model, then a major source has not
been identified.
29. For example, during certain weather situations, high
pollution concentrations may be transported into an area
from an unknown source. When this occurs, the higher
measured concentration should be used instead of the
model results in specifying emission limits.
2-11
-------
Slide no.
Script
Selected visuals
30. Therefore, determining the pollutant concentrations on
which to base judgments about the air quality is not a
simple matter. Many techniques have been developed to
interpret information about clean air in wilderness areas,
cities, and around factories.
31. In summary, the 1977 Clean Air Act Amendments require
modeling to help keep the air clean. Regulations specify
that three programs use modeling: Prevention of
Significant Deterioration, New Source Review, and Control
Strategy Evaluation.
32. In the next lesson, we will introduce air quality models.
We will discuss some of the Gaussian plume point source
models available on the UNAMAP Series.
33. Credit: Crew
O O
o \
I r\ r\
Coming up: UNAMAP
Introduction to
Air Quality Regulations
Technical Contmt:
DonBoll.rd
Dnlgn: Marilyn PMenoa
Graphic* LnlteWhtM
Photography/Audio DivM Churchill
Nirradon: Rick Pilmr
34. Credit: NET/EPA Contract
Lecture development
and production by:
Northrop Services Inc.
under
EPA Contract No. 68-02-3573
35. Credit: NET
Training
2-12
-------
Review Exercise
1. Section
of the Clean Air Act Amendments of
1977 requires air quality modeling.
2. True or False? The specific section of the Clean Air Act
Amendments of 1977 that requires air quality modeling
lists the air quality models that must be used.
1. 165
3. The individual who has the authority given by the Clean
Air Act Amendments of 1977 to allow a modeler to adjust
a specific model in case of inadequacy is the
a. State health officer.
b. Regional Administrator.
c. Air Quality Modeler.
d. Governor of the State.
2. False
4. Title
Part
of the Code of Federal
3. b. Regional Administrator.
Regulations requires the use of models to estimate concen-
trations needed to carry out the State Implementation
Plan.
5. Name the three programs specified hi the Guideline on
Air Quality Models that require air quality modeling.
4. 40, 51.24
6. True or False? The Prevention of Significant Deterioration
means that no increase in air pollution concentrations will
be allowed.
5. • Prevention of Significant
Deterioration,
• New Source Review,
• Control Strategy Evaluation
The short-term PSD increments may be exceeded
a. every 24 hours.
b. once every week.
c. once every hour.
d. once a year.
6. False
For a new steel mill or oil refinery, New Source Review
will require modeling if potential emissions are greater
than
a. 100 pounds per day.
b. 1000 pounds per year.
c. 100 tons per year.
d. 10 tons per day.
7. d. once a year.
8. c. 100 tons per year.
2-13
-------
9. True or False? Potential emissions for a new source are
the amount of pollutant that would be released into the
air before the application of required control equipment.
10. Receptors for a given area have recorded the following
concentrations (in /ig/ms). Circle the one "highest, second-
highest" concentration for this specific area.
Receptor #1 Receptor #2 Receptor #3
387 297 311
276 389 324
401 392 356
9. False
11. True or False? There are no exceptions to using the
method of "highest, second-highest" concentrations as air
quality estimates.
10. 389
12. The PSD increment for 24-hour SOZ levels in a Class II
area is
a. 365 /tg/m*.
b. 5
c. 20
d. 91 /ig/m3.
11. False
12. d. 91 /tg/ms.
2-14
-------
Lesson 2
Introduction to
Air Quality Modeling
Slide/Tape Presentation
First, view the slide/tape presentation—cassette no. 2 and
keyed slides 2-1 through 2-34, Introduction to Air Quality
Modeling—then complete the reading assignment.
Reading Assignment
EPA 450/2-78-027, Guideline on Air Quality Models,
pp. 13-24.
Supplementary Reading
NASA SP-S22, A Review of Methods for Predicting Air
Pollution Dispersion, pp. 1-10.
EPRI EA-1131, Appendix D: Available Air Quality Models.
Electric Power Research Institute. Palo Alto, CA.
EPA 450/4-77-001, Guidelines for Air Quality Maintenance
Planning and Analysis, Volume 10 (Revised): Procedures for
Evaluating Air Quality Impact of New Stationary Sources.
Lesson Goal and Objectives
Goal
To familiarize you with the process called modeling—from
determining the need to model to obtaining results — using air
quality models.
Objectives
Upon completing this lesson, you should be able to:
1. recognize whether a factory should model under New
Source Review procedures, given stack emissions data.
2. name the two kinds of data required for air quality
modeling.
2-15
-------
3. name the three ways that estimates from an air quality
model may be used.
4. name and describe the two types of air quality model
analyses.
5. identify one screening model and one refined
model.
6. define background concentration.
Supplementary Reading Information
The publications given in the supplementary reading section
are generally beyond the scope of this course. However, it is
appropriate to summarize the information contained in the
readings.
A Review of Methods for Predicting Air Pollution Dispersion
gives some reasons for developing air quality models. It puts
subjects such as classification of models, source inventory dif-
ficulties, meteorological data, and plume rise and dispersion
techniques into proper perspective for the student. A Review
also points out the reason for using the Gaussian plume model
instead of more refined approaches such as the Navier-Stokes
formulation for atmospheric diffusion.
Appendix D: Available Air Quality Models aids the air
quality modeler by placing the models into categories that fit
into the specific needs of industry. A specific model can then
be selected to fit a specific modeling situation. Appendix D
also discusses the underlying theory and techniques of the cur-
rently available models. The section on local plume and puff
models explains why the Gaussian plume technique is so
popular in air quality modeling.
The Guideline, Volume 10 Revised, Procedures for
Evaluating Air Quality Impact of New Stationary Sources was
designed to be used for screening a new source when refined
air quality modeling may not be necessary. Volume 10 takes
the user through plume rise, mixing height, and ground-level
concentrations; in effect, all factors necessary to estimate pollu-
tion concentrations, for comparison to the National Ambient
Air Quality Standards, are calculated.
2-16
-------
Introduction to
Air Quality Modeling
Slide no.
Script
1. Title slide—no narrative
2. In the previous lesson, we introduced the 1977 Clean Air
Act Amendments. The Amendments led to the issuance of
Federal regulations that require air quality modeling.
Because of these regulations, three air quality programs
evolved. The three programs we discussed were Prevention
of Significant Deterioration, New Source Review, and
Control Strategy Evaluation.
3. To introduce air quality modeling, let's look at the process
called modeling. This process involves taking source and
meteorological data and analyzing it.
4. Consider this example. An existing power plant, that is a
major source, is to be modified with the addition of a new
coal-fired unit and a new smoke stack. Under New Source
Review, this power plant must analyze source and
meteorological data if its potential emissions will increase
by a significant amount.
5. The modification to the power plant will result in an
additional release of 120 tons per year of sulfur dioxide.
This amount exceeds the significance threshold value of 40
tons per year for sulfur dioxide, so modeling and a PSD
analysis will be required. There are different significance
threshold values for each criteria pollutant. For paniculate
matter, an increase of only 25 tons per year is defined as
significant.
6. In the first step in the modeling processTspurce dataware
collected. These data would include the planTs geographic
location, stack data such as height (noted by h) and
diameter (noted by d), the effluent's temperature (T,) and
velocity (v,), and the pollutant emission rate (Q).
Selected visuals41
Introduction to
Air Quality Modeling
• Prevention of Significant Deterioration
• New Source Review
• Control Strategy Evaluation*
Air Quality Modeling
110toiu/)eu .
Addltlam*! SOz '
Significant: 40 toiu/year SOz
Source Data
H Q
h
r*i
T.
i.
*-*•'. -fJg^ -™Tr^
'•*• t 't 'i <
••t'w-^'^--v
^V^n'
* Illustrations included here, no live shots included.
2-17
-------
Slide no.
Script
Selected visuals
7. Meteorological data from the surrounding area are also
collected. These data would include wind speed (u) and
direction, air temperature (T0), atmospheric stability class,
mixing height (L), and the height and location of any
obstacles around the plant site. The meteorological data
that are gathered must be measured at a representative
location of the plant's surroundings.
8. A representative location chosen for sampling of
meteorological or other data is called a monitoring site. A
location where model predictions of pollutant
concentrations are made is called a receptor site.
After the data have been collected, the second step is to
choose a model for the specific situation at the plant. A
model is simply a set of mathematical equations. It relates
the source emissions to pollutant concentrations in the
ambient air.
Meteorological Data
U p
t_
n. T.. L
Stability Claw
Wind Direction
Models?
PTPLU
PTXXX
RAM
VALLEY
10. The parameters that make up individual models can be
programmed into large computers, minicomputers, and
pocket calculators. The source data are entered into the
computer or calculator, and the model is run. The run
produces output that gives a picture of what happens to
pollution as it leaves the stack and is transported and
dispersed by the atmosphere.
Input. . . h, v,, T,, d, Q, T.,
Output . . . X, H. Oy, Oz
11. Now that you know what's involved in the modeling
process — that is, identifying the need for modeling,
collecting the data, choosing the model, and running the
model —let's look at air quality models in greater detail.
Modeling Process Summary
n*«d • collect d.U
12. As we said earlier, an air quality model is simply a set of
mathematical equations. The equations try to explain the
atmospheric interactions taking place as the pollution is
released and as it travels to a receptor. The equation
shown on this slide is for a Gaussian plume model.
v .
A " 'O, 0. I
2-18
-------
Slide no.
Script
Selected visuals
13. The model then provides a way of predicting how the
pollution from new or existing sources will affect the areas
downwind.
14. These predictions may be used three ways: in developing
air pollution control plans, in assessing environmental
impact, and in projecting future air quality trends.
15. Let's look at these in greater detail. The first use is in
developing air pollution control plans. For example, high
pollution concentrations are measured in an area
downwind from a source. Air quality modeling may be
used to identify the specific source that contributed to the
excessive concentrations. Once identified, air quality
engineers can take action to solve the problem.
Air Quality Model Predictions
1. developing air pollution control plan*
2. aMCMing environmental impact
3. projecting future air quality trend*
Developing Air Pollution
Control Plans
16. Second, air quality models can be used to assess
environmental impacts. For example, a new factory will be
constructed near an urban area. Modeling must be used by
industry consultants to predict how emissions from the
factory will affect ambient air quality. Permission to build
will be given only if air quality will be maintained after the
factory is in operation.
Assessing Environmental Impact
17. Third and last, air quality models can be used to project
future air quality trends. For example, a regional planning
agency has several options for industrial expansion in a
rural county. The impact of each option can be assessed
with an air quality model. The model results can be used
with other information to rank each option. In this way,
environmental factors, like a new industry's effect on air
quality, can be weighed and considered in the planning
process.
18. As we have seen, modeling lets us logically connect air
pollution sources to ambient air quality concentrations,
noted by the Greek letter "chi."
Projecting Future
Air Quality Trends
1982
1992
2-19
-------
Slide no.
Script
Selected visuals
19. Applying a model to a source should be based on two
factors: specific use and data requirements. Ask, "How will
the specific model be used?" and, "What data are
necessary to run this model?"
Model Application
• Specific U*c
• Oat* Requirement*
20. First, a specific model may be used to screen the pollution
source. That is, a model may be run with limited
meteorological data and receptor sites. For example,
meteorological data may consist of a small number of
possible wind speed and atmospheric stability class
combinations.
21. The screening method allows a fast estimate of whether the
source may cause the National Ambient Air Quality
Standards or PSD increments to be exceeded.
Screening
22. The models used for the estimate are not expensive to run,
and the mathematical equations are simple to solve. The
PTXXX models we'll see later are considered screening
models.
Screening
23. The United States Environmental Protection Agency
recommends using a screening model first in the modeling
process. If the screening model indicates that the source
may cause the National Ambient Air Quality Standards or
the PSD increments to be exceeded, then a refined analysis
must be made.
24. The models used in making a refined analysis are more
expensive to run than are screening models. This is
because they process large volumes of meteorological data,
can consider hundreds of receptor sites, and can simulate
complex situations such as downwash or particulate matter
deposition. The EPA single source, or CRSTER, model
that we will see later is considered a refined model.
Refined Analysis
2-20
-------
Slide no.
Script
Selected visuals
25. Both the screening and refined model analyses will predict
air quality based on source emissions and the
meteorological conditions.
Air Quality Predictions
26. However, air quality at a specific location also depends on
how much pollution is in the air before the source adds its
own pollution. This quantity is called background and is
represented by "XB" • The model will predict a
concentration, noted as "XP"-
27. To get the total expected concentration, the background
concentration is added to the model's predicted
concentration. The total concentration is then compared to
the National Ambient Air Quality Standards to see if
violations will occur. The new source impact, without
background, is compared to the PSD increment
28. In this lesson, we have looked briefly at the process of
modeling. We have seen that a need must be identified,
data collected, a model chosen, and the results obtained.
Screening
Refined
i Pradfcttd
Conc*ntr«Uoo
X,
Modeling Process Summary
^i
i •>;•
29. We also discovered that air quality model predictions can
be used to develop air pollution control plans, to assess
environmental impacts, and to project future air quality
trends.
Air Quality Model Predictions
1. developing air pollution control plans
2. assessing environmental Impact
3. projecting future air quality trend*
30. We saw that air quality models can be used to quickly
screen air pollution sources. We also saw that air quality
models can be used to make a more detailed analysis of
sources and their surroundings, if required.
fiat.
2-21
-------
Slide no.
Script
Selected visuals
31. In the next lesson, we will introduce two case studies that
illustrate practical uses of air quality modeling.
Coming up : Two Case Studies
1
32. Credit: Crew
Introduction to Air Quality Modeling
Technical Content: Peter Guldbeig
Donald Bollard
Dertgn: MarUyn Peteiion
Graphic* Kathy Ward
Photography/Audio. David Churchill
Narration: Rick Palmer
33. Credit: NET/EPA Contract
Lecture development
and production by:
Northrop Services Inc.
under
EPA Contract No. 68-02-3573
34. Credit: NET
Northiop
Environmental
Training
2-22
-------
Review Exercise
1.
2.
3.
4.
5.
6.
7.
8.
True or False? An expansion of a major source must be
modeled if the expansion will increase potential emissions of
paniculate matter by 25 tons per year or more.
The two kinds of data that must be collected for inclusion
in a moriVI arc data and data.
A location where model predictions are made is called
a _ sit*.
Ways that a prediction from an air quality model may be
used are
a. in developing air pollution control plans.
b. in assessing impacts.
c. in projecting air quality trends.
d. all of the above
True or False? Screening an industrial site with an air
quality model allows a quick look at whether the site is
violating NAAQS or PSD increments.
True or False? The PTXXX models are considered refined
models.
True or False? The EPA single source (CRSTER) model
is considered a refined model.
Existing air quality in a specific area before a new factory
is hnilr is railed the _. concentration.
1. True
2. source,
meteorological
3. receptor
4. d. all of the above
5. True
6. False
7. True
8. background
2-23
-------
Lesson 3
Introduction to Case Studies
Lesson Goal and Objectives
Goal
To introduce two practical cases of modeling.
Objectives
At the end of this lesson, you should be able to:
1. recognize one reason why case studies can be useful.
2. recognize the reason that air quality modeling was
necessary in each of the two cases.
3. recognize differences in terrain features and meteorology
of the two areas.
Introduction
In Lessons 1 and 2 of this unit, you learned the reasons that
air quality modeling is required for Prevention of Significant
Deterioration (PSD), New Source Review, and Control Strategy
Evaluation. You were also introduced to air quality
models—what the process is and what air quality models, in a
general sense, are. The reading assignments have also pointed
out a painful truth: models come in all sizes and approaches.
Consequently, the available models that can be discussed must
necessarily be narrowed, since there are so many. The models
become very complex as they attempt to fully explain all of the
physical processes that influence pollution as it is transported
and dispersed in the atmosphere. In this course, you will read
about one type of air quality model, the simple Gaussian point
source model. This model has been in use for two decades and
continues to be useful. It was among the first models to be
developed. You will read about eight Gaussian point source
models. A course about modeling would not be complete,
however, without introducing and examining at least one case
study in some detail.
2-25
-------
Case Studies
Two case studies will be considered in Unit 6 of this course.
They serve as examples of the practical way air quality models
have been used by industry. Not all of the point source models
that will be discussed were considered for use in the two cases.
The case histories concern an oil refinery and an iron-casting
plant. Since the two industrial processes are different, the
model approaches will be different. The locations are also dif-
ferent: one is in the Southwest, the other in the Great Lakes
area. By studying these cases of modeling, you will gain some
insight into models and their use.
Oil Refinery
The first case to consider is an oil refinery located in northeastern
Oklahoma (Figure 2-1). The oil company that operates the
refinery wants to expand the present facility, which will expand
the processing capabilities. The plans require building a new
stack that will be 35 meters high and 1.56 meters in diameter.
The new stack will be located in the vicinity of the older stack,
which is 35 meters high and 1.56 meters in diameter. Like the
existing stack, the effluent will be SOS, so there is concern that
the new addition will cause the facility to exceed the Class II
PSD increments for SOZ. The existing rate is 3.28 grams per
second (114 tons per year), and since the emissions exceed 100
tons per year of SO2, this is already a major source. The new
stack will have an effluent rate of 1.5 grams per second (52
tons per year), which is a significant increase in emissions, and,
therefore, requires that this source be modeled for PSD. (The
Caw study—oil refinery.
Location
Stack proposed
Existing stack
Effluent
Existing rate
New stack rate
Terrain
High point
River
City
Class I PSD areas
NE Oklahoma
35 m high
1.56 m diam.
35 m high
1.56 m diam.
SO,
3.28 g/s
1.5 g/s
Uneven
10 m above stack base
West
1.61 km east of refinery
None within 50 km
Figure 2-1. Oil refinery.
2-26
-------
significance threshold for SO, is 40 tons per year or more.)
The terrain around the area is uneven with the highest point of
land rising 30.8 meters above the base of the stacks. A river
runs just west of the refinery. An urban area is located 1.61
kilometers east of the refinery. There are no Federal Class I
areas within 50 kilometers.
Iron-casting Plant
The second case to consider is an iron-casting plant (melting
furnace) located in northeastern Michigan (Figure 2-2). The
company that owns it, a large automobile manufacturer, melts
iron ingots in large furnaces before casting automobile engine
blocks. No new construction is planned. The company must
demonstrate that its effluent does not significantly contribute
to the high concentrations of total suspended paniculate
matter (TSP) observed within the urban area that surrounds it.
The area presently exceeds the NAAQS for TSP. The 14 stacks
at the iron-casting plant that emit particulate matter average
50 meters in height, but range from 24 to 70 meters. The
diameters range from 1.3 to 1.53 meters. The effluent rates of
the stacks range from 100 grams per second to 3966 grams per
second. The terrain around the area is essentially flat. A very
large river runs just northwest. There are no Federal Class I
areas within 50 kilometers.
Case study—iron-casting plant.
Location
Proposed
construction
Demonstration
Stacks
Height range
Average height
Average diameter
Effluent rate
1 stack
2 stacks
4 stacks
1 stack
4 stacks
1 stack
Terrain
River
Class I PSD areas
Northeastern
Michigan
None
Effluent not
significant
contributor to
measured
high TSP
concentrations
14
24 to 70 m
50m
1.3 to 1.53m
3966 g/s
3246 g/s
2419 g/s
S94g/s
158 g/s
100 g/s
Flat
NW
None within 50 km
Figure 2-2. Iron-casting plant.
2-27
-------
Summary
We introduce the two case studies at this point in the course to
encourage you to think about them as you learn about each of
the models. As each model is described, think about whether it
would be useful in either of the two situations just described.
In Unit 6, these two case studies will be considered in more
detail. They will be analyzed to illustrate each physical situa-
tion and the application of every phase of the modeling
process. The models, and why they were chosen, will also be
discussed, and, finally, an interpretation of model results (out-
put) will be given. You should compare your choices of models
with the model used in each case.
Review Exercise
1.
2.
3.
4.
5.
True or False? In the first case study— the oil refinery— the
air quality impact of the new stack must be modeled
because the proposed expansion will increase SOt emissions
by a significant amount from a major stationary source.
True or False? One reason case studies of modeling are
included in this course is because they help you gain insight
into models and their practical use.
True or False? In the second case study— the iron-casting
plant — the facility must be modeled for New Source Review.
The two case studies were located in the
a. Northwest, Great Salt Lake area.
b. Southeast, Great Lakes area.
c. Southwest, Great Lakes area.
d. Northeast, Great Salt Lake area.
True or False? Both areas had to be modeled because of
PSD requirements.
1. True
2. True
3. False
4. c. Southwest,
Great Lakes area.
5. False
2-28
-------
Unit3
Operational Point
Source Atmospheric
Dispersion Models
Lesson 1 User Considerations in Applying Air
Quality Models
Lesson 2 Characteristics of Model Classes
Lesson 3 Applications of Air Quality Models
3-1
-------
Lesson 1
User Considerations in Applying
Air Quality Models
Lesson Goal and Objectives
Goal
To introduce you to input data, modeling issues, and some cur-
rent mathematical models available.
Objectives
At the end of this lesson, you should be able to:
1. list six air pollution problem areas that might require air
quality modeling.
2. describe the general output of air quality models.
3. define dispersion model.
4. list the three types of input data to air quality models.
5. define empirical model.
6. define numerical model.
7. classify the Gaussian plume model.
Reading Guidance
The reading assignment introduces some types of models, issues
to be considered, and the advantages and disadvantages of
these models. Certain topical subjects, like numerical models,
will not be discussed further in this course.
Introduction
A dispersion model is a mathematical representation of the
transport and diffusion processes that occur in the atmosphere.
We have an incomplete understanding of the complex physical
and chemical processes involved hi the transport, dispersion,
transformation, and deposition of pollutants. Because of the
turbulent nature of the atmosphere, some limitations to the
predictive ability of even the best model will always remain.
Uncertainties in emissions and meteorological data also add to
model error. Nevertheless, to the extent that models reflect our
3-3
-------
best understanding of the relevant physical processes, they
represent a logical and environmentally equitable basis for
decision-making.
Models are used in a variety of environmental planning
activities. Some examples include:
• new source review,
• control strategy evaluation for SIPs,
• stack design studies,
• control technology evaluation,
• regulatory variances, and
• fuel conversion studies.
The models that are capable of addressing these issues vary
in complexity, required input data, and form of output data.
Input Data
The input data required by an air quality model can be
broadly classified as:
• source factors,
• site factors, and
• meteorological factors.
Source factors are related to the location and operating
characteristics of pollutant emission sources. They include the
time variability of emissions and their potential for chemical
reaction, deposition, and removal from the atmosphere. Site
factors represent the effects of terrain on dispersion and the
location of sensitive receptors relative to emission sources.
Meteorological factors include all of the parameters that define
transport and dispersion of pollutant mass, such as wind and
temperature fields, turbulence, and surface roughness.
The actual model used may consider all of these issues,
although for certain applications, the model or modeler may
only implicitly consider some of them. Complex models require
entries for all of these issues; the simpler models do not.
Model input
• Source factors
• Site factors
• Meteorological factors
Output Data
The output of air quality models consists of air pollutant con-
centrations for certain averaging times at specific spatial loca-
tions. The time and space detail of the output depends on the
characteristics of the chosen model and the model's applica-
tion. For example, the sequence of annual average concentra-
tions of SOj over an urban area is sufficient for determining
long-term trends in air quality, but a detailed time and space
distribution of SOZ is required for assessment of short-term
extremes in siting new coal-fired power plants in complex
terrain.
Model output
• Pollutant concentrations in
time and space
3-4
-------
Mathematical Models
Mathematical models currently used in the air pollution field
range from simple empirical models to very complex numerical
models. The empirical models are based on the analysis of air
quality data, source emission data, and meteorological data.
The numerical models are derived from the basic physical and
chemical principles relating to the processes of transport, diffu-
sion, transformation, and removal. Empirical and numerical
models are usually partitioned according to the model's ten-
dency to emphasize data or physiochemical principles.
However, the differences are not always distinct. For example,
empirical models incorporate varying degrees of physical
insight, such as accounting for the transport and the spatial
distribution of emissions in the source-receptor relationships.
Conversely, numerical models rely on empirically determined
parameters, such as transformation rates, removal rate con-
stants, and coagulation coefficients. Thus, a family of models,
or model hierarchy, exists, ranging from simple rollback
models to highly complex photochemical models.
Model hierarchy
• Simple rollback
• Screening
• Refined
• Complex photochemical
Semi-empirical Models
Semi-empirical is often used as an intermediate category of air
quality models. The Gaussian models, most widely used at the
present time, are semi-empirical. These models are derived
from scientific principles (e.g., conservation of mass), but rely on
empirically defined parameters (e.g., dispersion rates).
Empirical Models
Empirical models, which are closely tied to meteorological and
emission data bases, allow a full exploration of available
information in these bases. Relying on meteorological observa-
tions allows the complexities of the atmospheric system to be
represented, even though some complexities are not fully
understood. Also, empirical approaches allow a simultaneous
check on data quality through standard statistical tests. Finally,
empirical models can usually be formulated and operated at
low cost.
However, depending on meteorological and emission data
bases, disadvantages may occur for empirical modeling. Some
empirical models require high quality data, which often do not
exist. Additionally, empirical models and their parameters are
very closely tied to the specific conditions under which they
were created. As a result, when they are applied to other
meteorological situations in the same locale, the models may
lead to incorrect conclusions. Careful selection of variables and
thoughtful interpretations of observed relationships can
counteract some of the disadvantages.
3-5
-------
Numerical Models
Numerical models are formulated from basic scientific concepts
associated with physical and chemical processes occurring in
the atmosphere. This formulation affords confidence in their
application over various ranges of conditions and areas, as well
as in their predictive ability. However, these models possess
computational complexities, and require extensive data input
and specifications of numerous model parameters. The semi-
empirical models share advantages and disadvantages of both
the empirical and numerical models.
Review Exercise
1. List six air pollution problem areas that might require air
quality modeling.
2. The output of air quality models consists of .
1. • new source review
• control strategy evaluation
for SIPs
• stack design studies
• control technology
evaluation
• regulatory variances
• fuel conversion studies
3. Define a dispersion model.
2. air pollutant concentrations
in time and space
4. List the three types of input data to air quality models.
A dispersion model is a
mathematical representation
of the transport and diffusion
processes that occur in the
atmosphere.
5. Define empirical model.
4. • source factors
• site factors
• meteorological factors
5. Empirical models are models
based on analyzing three
kinds of data: air quality,
emission, and meteorological.
3-6
-------
6. Define numerical model.
7. True or False? The Gaussian plume model is a
semi-empirical formulation.
6. Numerical models are
derived from basic physical
and chemical principles that
relate to the processes of
transport, diffusion, transfor-
mation, and removal.
7. True
3-7
-------
Lesson 2
Characteristics of
Model Glasses
Supplementary Reading
DOE/TIC-11223, Handbook on Atmospheric Diffusion, U.S.
Department of Energy, 1982, chapters 1, 2, and 10.
Lesson Goal and Objectives
Goal
To familiarize you with the characteristics that are used to
refine air quality model classes.
Objectives
At the end of this lesson, you should be able to:
1. list eight model characteristics.
2. describe the reason that the distances between grid points
in air quality models are limited by the meteorological
scales of motion.
3. state the reason air quality models may be called time-
varying models.
4. list two reasons that Lagrangian air quality models are
more capable of describing atmospheric processes than
Eulerian air quality models.
5. list two reasons that emission data inputs to air quality
models may be incorrect in estimating pollution
concentrations.
6. list two reasons that meteorological data inputs to air
quality models may be incorrect in estimating pollution
concentrations.
7. list two reasons that an air quality model may not be
representative of the problems in estimating pollution
concentrations.
3-9
-------
Model Characteristics*
The air pathway processes that control the fate of pollutants
from source to receptor are transport, diffusion, transforma-
tion, and removal. Because of the complexity of these processes,
as well as the complications introduced by terrain and the
pollutants themselves, there exists a large and diverse family of
air quality models.
The members of the family of air quality models, in the
empirical, semi-empirical and numerical categories, possess a
variety of characteristics that can be used to further refine
model classification. These characteristics are a result of the
ambient meteorological and topographical conditions, the time
and space scales inherent in the model application, the
mathematical procedures used to solve the system of equations,
and the pollutants and reaction mechanisms required to answer
the particular air quality question.
Time and Space Scales
Air pollution decisions can be described in terms of four
geographical subdivisions: site specific (local), regional,
national, and global (Figure 3-1). These form a reasonable
classification scheme for horizontal spatial and time scales of
air quality models. At the lower end of the scale, site-specific,
or local, situations include considerations such as emissions,
source characteristics, initial plume rise, initial phase of mix-
ing, local terrain, and initial transport. At the higher end, the
site-specific category is concerned with interacting plumes from
sources separated by 10 to 20 km.
Regional-scale problems range from an urban area or large
industrial complex to a region where urban areas are point
sources in the air quality models. For example, the lower limit
of the scale may be represented by a nocturnal urban plume,
while the northeast quarter of the continental U.S. represents
the upper limit of the regional scales.
National scales vary from half of the continental U.S. to the
entire continental U.S. For example, models have been used to
estimate the SOZ concentrations from existing sources west of
100 °W longitude for a high-coal-use electric scenario projected
to the year 2000. Currently, the Department of Energy is con-
ducting a national coal assessment that will estimate SO* con-
centrations over the entire continental U.S.
Although global decisions may be concerned with global
problems and models, once a pollutant crosses international
boundaries, international decisions may be required. The
Model characteristics
Time and space scales
Steady state or time dependent
Frame of reference
Pollutants and reaction mechanisms
Treatment of turbulence
Multiple plumes
Treatment of topography
Treatment of uncertainty
Regional
National
Global
Figure 3-1. Time and space scales.
•Source: EPRI EA-1311, Section 2, pages 2-1 through 2-13.
3-10
-------
models used to help resolve these problems will be of a scale
geometrically smaller than global. However, two important
global problems whose impacts are independent of national
boundaries are the effects on weather and climate of substan-
tial increases of CO* and fine panicles in the atmosphere.
The determination of time scales from the model application
point-of-view depends on the effects of the pollutant, the
regulatory standards, and the variability of emissions and
meteorology. Odor and taste perception is nearly instan-
taneous; possible acute toxic effects on humans and animals
occur over periods of hours; and chronic effects occur over
seasons and years (Figure 3-2). Regulatory standards are
usually closely related to time scales of expected effects. Emis-
sion variability depends on power demand curves and possible
accidental releases, while variability in meteorology depends-on.
turbulence, passing thunderstorms and weather fronts, and sta-
tionary air masses.
The air quality model (AQM) will calculate pollution con-
centrations at preselected times and locations called grid points
(Figure 3-3). The user determines the times and grid locations
desired. The outcome of an AQM is highly dependent on the
availability of the meteorological input data. That is the
distance between sampling stations and the time period for
which the data is averaged. The model's ability to calculate at
optimum grid points, using the smallest possible distances at
time intervals, is called resolution. Consequently,
meteorological observing stations cannot detect weather distur-
bances that are smaller than one-half the distance from one
station to the next. The ability to "see" only certain sizes of
phenomena limits the model's forecasting ability.
For example, using Table 3-1, local atmospheric phenom-
ena, such as sea breezes, are approximately 2 kilometers at the
smallest, or Lmin (Figure 3-4). This means the smallest grid size
distance that can be used for calculations in a local model is
one-half of Lmi*. Local grid sizes are not closer than 1 kilo-
meter apart. The time average of the calculations is on the
order of an hour. A large number of calculations are needed to
estimate concentrations across atmospheric scales. There is a
limit on the number of grid points that can be economically
used to fit the scales. There are approximately 100 grid points
in each horizontal direction.
Models with length scales less than global or hemispheric
require time-dependent, lateral boundary conditions so that
features with Lmax/4 scales are properly resolved. This limita-
tion on the scales spanned by atmospheric models implies that
a user can expect broad coverage or detailed interaction, but
not both. Boundary conditions must always be specified, while
subgrid scale processes must always be given parameterized
values.
Figure 5-2. Pollution effects.
Figure S-3. Grid points for
regional scale.
1 km
Figure 3-4. Grid points for
local scale.
3-11
-------
Table 3-1. Atmospheric scales: model scope, characteristic length
and time scales of phenomena, and examples.
Atmospheric
scales
Global
Hemispheric
Continental
Regional
Local
Convective
Turbulent
Model
Grid
(km)
400
200
100
20
1
0.04
0.01
Lmin
(km)
800
400
200
40
2
0.08
0.02
1*0,
(km)
40,000
20,000
10,000
2,000
100
4.00
1.00
Length
L-X/4
(km)
10,000
3,000
1,000
200
10
1
0.2
Time
T-P/4
S m
10 d
3d
Shr
1 hr
15 min
1 min
Phenomena
season
climate zone
spell
storm track
air mass
anticyclone
cyclone
front
squall
sea-breeze
heat island
shower
tornado
plume
eddy
gust
Source: Adapted from Atmospheric Modeling Relative to Fuel Use Strategy, presented
at BNL Conf. on Energy Related Modeling and Data Base Management, May 12-14,
1975, p. 8.
The current state of affairs in understanding the atmos-
pheric phenomena in Table 3-1 is as follows:
• There is much to learn about climate change due to solar
radiation and the distribution of continents and moun-
tains. But the greatest challenge is to understand the
causes for very small changes in climate that have major
impacts on society in areas with marginal climates. The
forcing functions for these small changes are radiation
fluxes and turbulent fluxes of heat, moisture and momen-
tum from the surface, along with the radiative influences
of clouds. Little is known about the effects of these
phenomena on local and regional climate.
• Anticyclones and cyclones, outside of the tropics, are the
easiest important phenomena to understand, treat
theoretically, and predict (Figure 3-5).
• Regional, local, and convective scales include very com-
plex processes that depend on underlying topography,
latent heat releases, and nonlinear interactions and feed-
back mechanisms between scales that are larger and
smaller than the phenomenon in question. These areas
suffer from a lack of data from routine measurements,
special field projects, and numerical simulation.
Figure 3-5. Anticyclones and
cyclones.
3-12
-------
• Considerable advances have occurred in modeling tur-
bulent boundary layer flows over homogeneous terrain and
turbulent diffusion over uniform surfaces, for distances out
to 10 to 20 km. However, over irregular terrain and urban
areas, the prediction of pollutant trajectories is still dif-
ficult (Figure 3-6).
This paucity of atmospheric data is reflected in the current
status of weather forecasting (Table 3-2). Statistical techniques
are used beyond prediction periods of five days; equations of
fluid dynamics for periods of 18 hours to several days; simple
translation or extrapolation of patterns is used for two to six
hours; and persistence of observed conditions is reliable for two
hours. In summary, the physical science of atmospheric predic-
tion is relatively advanced for time scales of 18 to 72 hours;
beyond this period, statistical science dominates, and for
shorter periods, instrumentation and data processing dominate.
Since AQM's cannot be better than their atmospheric
elements, the conclusions given above concerning the time and
space modeling scales in Table 3-1, the state-of-affairs in
atmospheric sciences, and the state-of-forecasting in Table 3-2,
are equally valid for AQM's. Additionally, the grid and
resolvable scales listed in Table 3-1 can be used for
characteristic lengths in the AQM's.
Smooth
Rough
Figure 5-6. Smooth to rough
transition.
Table 3-2. Techniques of weather forecasting.
Period
Beyond 3 months
1 to 3 months
6 to SO days
S to 5 days
18 to 72 hours
2 to 18 hours
0 to 2 hours
Name
(experimental)
seasonal
outlook
extended
intermediate
short range
nowcast
Domain of data
global
global
global
hemisphere
hemisphere
300 to
3000km
0 to 300 km
Method
climatological
statistics
statistics
statistics
statistics and
dynamics
dynamics
statistics
dynamics (6 to 18)
translation (2 to 6)
persistence
Quality
vague
unproved
vague and
little proved
good to poor . . .
erratic
good except for
precipitation
fair
as good as data
processing, com-
munications, and
display
Source: Adapted from Atmospheric Modeling Relative to Fuel Use Strategy, presented at BNL
Conf. on Energy Related Modeling and Data Base Management, May 12-14, 1975, p. 10.
3-13
-------
Steady-state or Time-dependent Models
Models are steady state or time varying, depending on whether
or not time is explicit in their formulation. If the system of
equations governing the phenomena being studied in the model
depends on time, the model is time varying. If the system
represents the average state of phenomena over a certain
period of time, the model is steady state.
Steady-state models are applicable when the time and space
scales are sufficiently small, or when the desired output is suffi-
ciently coarse that variability in the effects of pollutants, emis-
sions, and meteorology can be ignored or averaged out (Figure
3-7). For example, the steady-state Gaussian plume can be
used over site-specific scales if the winds and atmospheric ther-
mal structure are nearly uniform over the time period of
interest. Steady-state models can be used for certain policy^
and standard-setting decisions on the regional and global
scales. However, for technological assessments on the regional
and global scales, time-varying models must be used to account
for the variability in meteorology and emissions. Whenever
steady-state models can be used in place of time-varying
models, there is a saving in computer time and cost.
Wind
Figure 3-7. Steady-state meteorology
conditions.
Frame of Reference
Air quality models, except for some empirical ones, are related
to a coordinate system, or reference frame (Figure 3-8).
Reference frames may be fixed at the earth's surface, at the
source of the pollutant (for either fixed or moving sources), or
on a puff of pollutant as it moves downwind from the source.
Reference frames fixed at the earth's surface or on the source
are called Eulerian (because of their relation to the advecting
and diffusing pollutant), while frames fixed on a puff of pollu-
tant are called Lagrangian.
The advantage of Lagrangian models over Eulerian models,
and vice versa, depends on the class of models and the
availability of the proper input data. Turbulent diffusion of
pollutants is more easily formulated in the Lagrangian sense,
but most of the pollutant concentration data have been
obtained in the Eulerian sense. No adequate theoretical basis
exists for converting Eulerian diffusion data to Lagrangian
data. Hence, some of the advantages of Lagrangian models
given below may suffer from this lack of adequate input data.
Lagrangian models are more capable than Eulerian models
of accounting for source locations and emission rates and of
describing diffusion as the pollutants are carried by the wind.
On the other hand, Eulerian models are more capable of
accounting for topography, atmospheric thermal structure, and
reactive pollutants from many sources. Lagrangian trajectory
models are less costly to run than Eulerian models and are
i
Figure 3-8. Reference frame
coordinate system.
3-14
-------
adequate for regional long-term assessments, while three-
dimensional Eulerian models are best for the analysis of such
phenomena as a photochemical smog episode in the Los
Angeles basin.
Pollutants and Reaction Mechanisms
Air quality models describe the fate of airborne gases and par-
ticles. As these pollutants travel over their pathways, physical
and chemical reactions may occur. As shown in Figure 3-9, the
categories of mechanisms are nonreactive, reactive (photo-
chemical and nonphotochemical), gas-to-particle conversions,
gas/particle processes, and particle/particle processes. In addi-
tion, the gases and panicles may be radioactive, in which case
the models must contain some provisions for accounting for
radioactive decay and the production of subsequent radioactive
elements.
Nonreactive models have been constructed to determine the
fate of automobile emissions of CO and emissions of paniculate
matter from fossil-fuel power plants (Figure 3-10). Reactive
models have been developed to determine the formation of
sulfate deposits from SO* emissions from coal-fired power
plants. On hot, sunny days, in areas like the Los Angeles basin
and Houston, complex photochemical models predict the
formation and concentration of oxidants from hydrocarbon
and NOX emissions for both moving and stationary sources
(Figure 3-11).
Both the SOj/sulfate and photochemical models have gas-to-
panicle and gas/panicle components. The gas-to-panicle
components account for the production of panicles directly
from gases via gaseous reactions or via condensation. The
gas/panicle components in the models account for panicle
growth by condensation or by absorption of gases. Panicle/par-
ticle processes are accounted for in aerosol models. These are
similar to nonreactive gas models, except that particle com-
ponents are added to the equations. These components include
coagulation (collision followed by sticking together), breakup,
condensational growth, and diffusion.
Nonreactive
Reactive
Gaa-to-particle
Gas/panicle
Panicle/panicle
4«
Figure 5-9. Airborne gat and
particle conversions.
Figure $-10. Nonreactive emissions.
Figure 3-11. Reactive emissions.
3-15
-------
Treatment of Turbulence
Atmospheric turbulence, shown in Figure 3-12, is the
mechanism that dilutes and mixes the gaseous and paniculate
pollutants as they are transported by the mean wind. Tur-
bulence is one of the least known, but one of the most impor-
tant, phenomena in the atmosphere, and it is produced when
certain gradients in the wind, temperature and humidity fields
occur in the atmosphere. For these reasons the formulation of
turbulence in air quality models ranges from the simple (a
well-mixed or stirred volume) to the complex (accounting
for both local and historical influences of turbulence on
velocity fields).
Atmospheric turbulence in a model may be provided for by
a well-mixed volume, semi-empirical diffusion coefficients,
eddy diffusivity, Lagrangian statistics, or more complex-tur-
bulence models. The well-mixed volume approach, as in roll-
back and simple box models, basically ignores turbulence
except in a loosely implicit manner. Semi-empirical diffusion
coefficients are the main parameters in the current air quality
models: Gaussian plume and puff (Figure 3-13). These coeffi-
cients have been determined from field diffusion studies over
flat terrain and usually under neutral stability conditions. Most
working grid and multibox models use the eddy diffusivity
formulation, which is based on theoretical, physical, and
numerical studies of the planetary boundary layer.
To account for some of the physical inconsistencies hi the
eddy diffusivity formulation, more complex formulations have
been developed. These turbulence models, which contain many
parameters and new dependent variables, increase the number
of equations in air quality models. The models also increase
the number of parameters that need to be specified, introduce
new uncertainties, and increase the computer costs of running
air quality models.
In spite of these additional considerations, these complex
turbulence formulations are seen as necessary for numerical
stability and accuracy in grid models, as well as for analysis and
prediction of complex reaction mechanisms in power plant
plumes.
Another approach for introducing more realistic turbulence
into a diffusing system is to apply Lagrangian statistics. The
statistics of turbulent diffusion following puffs of pollutants are
mathematically more simple than those of the Eulerian
approach. Most field data, however, are obtained in the
Eulerian sense, and the relationship between Eulerian and
Lagrangian statistics is unknown for atmospheric turbulence in
the planetary boundary layer. Therefore, the only real advance
in this area has been for numerically simulated turbulent
fields, not for diffusion fields in the real atmosphere.
Figure 3-12. Atmospheric
turbulence.
Figure 3-13. Gaussian plume
and puff.
3-16
-------
Multiple Plumes
The assumption that nonreactive pollutant plumes can be
added together is illustrated hi Figure S-14. Suppose an air
quality model is used to calculate concentrations of an i-th
pollutant (c() in the plumes from Plant A and Plant B. Con-
centrations of the i-th pollutant from the combined plants are
calculated by simply adding the contributions from Plant A, c,
(x,t;A), and Plant B, c/ (x,t;B), together for common spatial
(x) and tune (t) positions. The validity of this assumption
depends on physical and chemical noninteraction between the
i-th pollutant in the two plumes. The principal advantage of
the property is the saving in computer time and storage for
regional and national assessments where tens to hundreds of
sources must be considered.
Treatment of Topography
Surface conditions and topographic features generate fields of
turbulence, modify vertical and horizontal winds, and change
the temperature and humidity distributions in the boundary
layer. All of these changes modify the transport and diffusion
of pollutants. An important characteristic of air quality models
is the manner in which surface conditions and topography are
treated.
Topography is characterized in air quality models as homo-
geneous flat terrain, nonhomogeneous flat terrain, simple
terrain, and complex terrain. Examples of homogeneous flat
terrain are shown in Figure 3-15; the grasslands of western
Kansas, the corn fields of Ohio and Iowa, and the pine forests
of the Southeast. The greatest number of experimental diffu-
sion studies have taken place over homogeneous terrain; there-
fore, air quality models based on diffusion coefficients and
eddy diffusivities are most applicable for this topographic
category.
Figure 3-14. Plume additivity.
Figure 3-15. Terrain features.
3-17
-------
Nonhomogeneous flat terrain includes water-land transitions,
transitions from grasslands to forests, and transitions from
irrigated farmland to desert areas.
Simple terrain includes street canyons in urban areas, simple
deep valleys, sharp-edged cliffs, and simple hills. Figure 3-16
depicts two views of the flow of pollutants around a simple hill
under stable atmospheric conditions.
Figure 3-17 indicates the effects of lateral drainage winds,
known as katabatic winds, on pollutants. Even though the
pollutants may initially escape from the valley, the nighttime
drainage situation has a tendency to recirculate the pollutants
so that they finally accumulate in the valley.
Figure 3-18 illustrates the terrain downwash effect of siting a
power plant in the lee of a sharp-edged cliff. Terrain
downwash may produce high concentrations of pollutants near
the source. S and S1 represent stagnation points in the flow
field, while the dashed line represents the surface of separation
between streamline flow and the revolving rotor.
Figure 3-18. Terrain downwash.
Complex terrain consists of mountain ranges and deep
valleys such as those found in the Rocky Mountains and the
Northern Appalachians. In these areas the wind and tem-
perature fields are very complex, as are the distribution and
intensity of atmospheric turbulence. At latitudes closer to the
equator, photochemical reactions should be more pronounced
than at latitudes toward the poles because the intensity of
sunlight is greater. Because of these complexities and the
sparseness of monitoring stations, experimental data on flow,
diffusion, and photochemical reactions in complex terrain is
fragmentary and incomplete. Better data and models in com-
plex terrain are needed because of the availability of energy in
Figure 3-16. Transport of a
plume around the
side of a hill under
stable atmospheric
conditions.
Figure 3-17. Pollutant
accumulation
in valley.
3-18
-------
the mountainous West, which increases the sitings of power
plants, and the requirement for preventing significant
deterioration of pristine areas.
Treatment of Uncertainty
Many sources of errors and/or uncertainties occur in pollutant
concentrations obtained from air quality models. The emissions
data that are inputs to these models may be hi error because of
incorrect source strengths and locations, unaccounted for time
variability hi emission rates, uncertainties in stack parameters,
and incorrect calculations of plume rise after emissions.
Meteorological data that are used for model inputs and for
evaluating model parameters may be in error because of incor-
rect calculations of plume rise after emissions. Meteorological
data that are used for model inputs and for evaluating model
parameters may be in error because of incorrect wind speed
and direction, poorly specified dispersion parameters, and
incorrect determination of the atmospheric thermal structure.
The air quality model itself may not be representative of the
problems in question because of incomplete knowledge of
chemical and physical interactions of gases and particles, incor-
rect formulations of removal processes, and poorly specified
boundary conditions.
The treatment of physical and chemical uncertainties has
resulted in the following classification scheme for models:
• Deterministic, if the formulation of the model is in terms
of specific constants, functions, and parameters.
• Stochastic, if the information concerning the physical
process is not entirely known but the underlying structure
is. That is, the detailed information is random, but
specific tools exist to solve the system.
• Adaptive, if the basic structure of the process is unknown.
However, in this system more is learned as one sets about
determining the solution and the basic structure begins to
evolve. Examples of this type of system in air quality
modeling are complex photochemical reaction schemes,
gas/particle and particle/particle systems.
Although most current air quality models are deterministic,
some models are truly stochastic and a few are adaptive. In
fact, all air quality models representing specific real-world
situations have features that are either explicity or implicitly
adaptive. The basic reasons for adaptive systems are uncer-
tainty in meteorological data; incomplete knowledge of
atmospheric turbulence; ignorance of chemical reactions and
reaction rates in the atmosphere; incomplete knowledge of the
chemical species from, and the emission characteristics of,
natural sources; and lack of data and knowledge of gas-to-
particle and particle/particle interactions.
Model clarification* for
uncertainty
Deterministic
Stochastic
Adaptive
3-19
-------
Current investigations are trying to remove, or at least
reduce, these areas of uncertainty. For example, there are cur-
rently three large field experiments in the northeastern quarter
of the United States dealing with the conversion of SOt to
sulfate: EPA's Sulfate Transport and Transformation in the
Environment (STATE), DOE's Multistate Atmospheric Power
Production Pollution Study (MAPSS), and EPRI's Sulfate
Regional Experiment (SURE).
Field experiment!
EPA's STATE
DOE's MAPSS
EPRI's SURE
Review Exercise
1. List the eight model characteristics discussed in this lesson.
2. True or False? A limitation on length scales is used in air
quality models because the user can either get fine detail
or broad coverage of pollution estimates but not both.
1. • time and space scales
• steady state or time
dependent
• frame of reference
• pollutants and reaction
mechanisms
• treatment of turbulence
• multiple plumes
• treatment of topography
• treatment of uncertainty
3. True or False? A model is called time varying because its
equations depend on time.
2. True
4. The Lagrangian models are more capable of describing
reality than the Eulerian models because
a. the Lagrangian uses anemometer information from
the NWS.
b. Lagrangian mathematics are simpler to solve than
Eulerian mathematics.
c. Lagrangian models can acccount for source location and
emission rates.
d. Eulerian models require meteorological information
from balloons for accuracy.
3. True
5. Emission data may be incorrect because of
a. uncertainties in stack parameters.
b. incorrect calculations of emission plume rise.
c. both a and b
4. c. Lagrangian models can
account for source location
and emission rates.
5. c. both a and b
3-20
-------
6. Meteorological data can be incorrect because
a. windspeed and direction are taken with an anemometer
only.
b. windspeed and direction are taken with a tethered
balloon only.
c. dispersion parameters may be poorly specified.
d. the stability of the atmosphere is well known and need
not be measured.
7. An air quality model used may not be representative of a
modeling situation because
a. modeling is accurate only in California.
b. modeling is not required under any circumstances.
c. removal processes are never required in a model.
d. boundary conditions are poorly specified.
6. c. dispersion parameters may
be poorly specified.
7. d. boundary conditions are
poorly specified.
3-21
-------
Lesson 3
Applications of
Air Quality Models
Lesson Goal and Objectives
Goal
To familiarize you with the way air quality models are applied
in making decisions.
Objectives
At the end of this lesson, you should be able to:
1. name the nine model subdivisions that aid in deciding if
professional modeling consultants are required.
2. match air pollution applications and geographical deci-
sion scales to their representative sizes.
3. decide from a site-specific fuel choice problem which
model subclass to use in determining air quality.
Introduction
Air quality models treat air pollutants as they travel between
the source and the receptor. The models accept the emission
characterisics of pollutants as input and produce estimates of
ambient air concentrations and material deposited on surfaces
as output. This output is then used to analyze the impacts and
effects of pollutants on receptors, weather, and climate. Model
applications may be summarized as follows.
Applications and Decisions
Air pollution applications and decisions can be geographically
divided into site-specific areas, with horizontal spatial scales from
1 to 20 km; regional, with scales from 20 to 1000 km; national,
with 1000 km to continental United States; and global, with
hemispherical to global. Important decisions on the local, or
site-specific, scale are concerned with the choice of fuel during
air pollution episodes or on a continuing basis, the type of
abatement technology that should be employed -in a power
plant, and the choice of when and where to monitor air
Air pollution applications
and decisions.
Area
Site specific
Regional
National
Global
Horizontal spatial
scale
1 to 20 km
20 to 1000 km
1000 km to
continental U.S.
Hemispherical to
global
3-23
-------
quality. In addition, new power plants must be analyzed in
regard to their production of incremental changes in air
quality over pristine areas.
Decisions on the regional scale involve plant siting of large
power units, the assessment of the effects on air quality as a
result of given or new technologies, when and where to monitor
air quality, land use planning, and setting of emission stan-
dards to satisfy ambient air quality regulations. National deci-
sions should address technological assessment, choice of
research and development programs, and the national energy
policy. Finally, global decisions should treat assessment
problems, such as the increase of CO2 and fine particles in the
air, or address the international energy policy.
Subclasses
In arriving at these decisions, air quality models are often used.
The nine model subclasses are as follows: rollback, statistical,
local Gaussian plume and puff, regional trajectory, box and
multibox, grid, particle, global, and physical. Table 3-1 sum-
marizes the characteristics of these nine model subclasses.
These characteristics help decision makers understand the type
and complexity of the model subclass and aid in determining
whether the modeling work should be done in-house or by a
consulting firm.
Decision Table
Table 3-2 indicates what class of models should be used for a
given decision. In addition to the model class, the table
indicates the important time scales that should be used, and
whether or not the models should contain transport and diffu-
sion, transformation, and removal components. The use of
these tables will aid decision makers in realistically treating the
air pathway segment of the air pollution decision process.
S-24
-------
Table 3-1. Summary of model subclanes.
Model
subclass
RoUback
Statistical
Gaussian
plume
and puff
Regional
trajectory
Box and
multibox
Grid
Particle
Global
Physical
Model
da-
Empirical
Empirical
Semi-
empirical
Semi-
empirical
Semi-
empirical
and
meteoro-
chemical
Numerical
Numerical
Numerical
Empirical
Geographical
subdivisions
Local
Regional
National
Local
Regional
Local
Regional
National
Local
Regional
Local
Regional
Local
Regional
Global
Local
Steady state
or
time dependent
Steady state
Steady state
Tune dependent
Steady state
Time dependent
Tune dependent
Steady state
Time dependent
Steady state
Time dependent
Time dependent
Time dependent
Time dependent
Frame
of
reference
Eulerian
Eulerian
Eulerian
Lagrangian
Lagrangian
mixed
Lagrangian
and
Eulerian
Eulerian
Lagrangian
Eulerian
Mixed
Lagrangian
and
Eulerian
Eulerian
Mixed
Eulerian
and
Lagrangian
Type
of
polluumta
Gases and
panicles
Gases and
panicles
Gases and
particles
Gases and
panicles
Gases and
panicles
Gases and
panicles
Gases and
particles
Gases and
panicles
Gases and
panicles
Reaction
nmrhaniMm
Nonreactive
Nonreactive
Reactive
Gas-to-
particle
Nonreactive
Reactive
Nonreactive
Reactive
Nonreactive
Reactive
Gas-to-
panicle
Nonreactive
Reactive
Gas-to-
particle
Nonreactive
Reactive
Gas-to-
particle
Nonreactive
Reactive
Nonreactive
Treatment
of
turbulence
Well-muted
Well-mixed
Diffusion
Coefficients
Diffusion
Coefficients
Eddy
Diffusivities
Well-mixed
Eddy
Diffusivities
Eddy
Diffusivities
Complex
Formulation
Eddy
Diffusi vines
Eddy
Diffusivities
Not
applicable
Plume
additivity
Not applicable
Not applicable
Yes and no
Yes
Yes and no
Yes and no
Yes and no
Yes
Not applicable
Treatment
of
topography
Homogeneous
to simple
terrain
Homogeneous
to simple
terrain
Homogeneous
to complex
terrain
Non-
homogeneous
to complex
terrain
Homogeneous
to simple
terrain
Homogeneous
to complex
terrain
Homogeneous
to complex
terrain
Non-
homogeneous
to complex
terrain
Homogeneous
to complex
terrain
Treatment
of model
uncertainty
Deterministic
Stochastic
Adaptive
Deterministic
Deterministic
Stochastic
Deterministic
Deterministic
Deterministic
Stochastic
Deterministic
Deterministic
3-25
-------
Table 3-2. Decisions and model applications.
Decision
Site specific
(1 to 20 km spatial scale)
Fuel choice
Abatement technology
Incremental changes
Monitoring
Regional
(20 to 1000 km
spatial scale)
Plant siting
Technological assessment
Monitoring
Land use
Standard setting
National
(1000 km to continental
U.S. spatial scale)
Policy
Research and
development
Technological assessment
Global
(hemispherical to global
spatial scales)
Policy
Assessment
Model
lubclass
Rollback
Statistical
Gaussian plume
and puff
Rollback
Statistical
Gaussian plume
and puff
Gaussian plume
and puff
Grid
Gaussian plume
and puff
Regional
trajectory
Grid
Multibox
Regional
trajectory
Multibox
Regional
trajectory
Regional
trajectory
Regional
trajectory
Rollback
Rollback
Regional
trajectory
Regional
trajectory
Regional
trajectory
Simple global
model
Complex global
model
Simple global
model
Averaging time
or
temporal scales
Weekly to annual
Hourly to daily
2 min to 2 hr
Weekly to annual
Hourly to daily
2 min to 2 hr
2 min to 2 hr
2 min to 2 hr
2 min to 2 hr
2 hr to 4 d
2 hr to 4 d
2hrto4d
Monthly to annual
2 hr to 4 d
2 hr to 4 d
2 hr to 4 d
2 hr to 4 d
Weekly to annual
Annual
Monthly to annual
10 hr to 6 d
Monthly to annual
Monthly to annual
2 d to 2 wk
Monthly to decades
Need for transport
and
diffusion components
No
No
Yes
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
Yes
Yes
I
Yes
Yes
Yes
Yes
Need for
transformation or
wet and dry removal
components
No
No
Yes and no
No
No
Yes and no
Yes and no
Yes and no
Yes and no
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
Yes
Yes
Yes
Yes
Yes
Yes
3-26
-------
Review Exercise
1. Air quality models are organized into subclasses. One model
subclass is rollback. Name three others.
2. Given the following lists of geographical scales and repre-
sentative sites, match them appropriately.
Geographical scales
A. site specific
B. regional
C. national
D. global
Representative sizes
a. hemispherical
b. 1 to 20 kilometers
c. 1000 km to continental U.S.
d. 20 to 1000 kilometers
1. • statistical
• Gaussian plume and puff
• regional trajectory
• box and multibox
• grid, particle, global,
physical
3. Given the table that summarizes model applications, give
the model subclass that is recommended for use for a
weekly to annual, site-specific, fuel-choice problem at a
power plant (see Table 3-2).
2. A., b. 1 to 20 kilometers
B., d. 20 to 1000 kilometers
C., c. 1000 km to
continental U.S.
D., a. hemispherical
3. rollback
3-27
-------
Unit 4
Gaussian Point
Source Atmospheric
Dispersion Models
Lesson 1 Introduction to UNAMAP, Version 4
Lesson 2 PTXXX Models: PTMAX, PTDIS, and PTMTP
Lesson 3 PTPLU
Lesson 4 CRSTER
Lesson 5 RAM
Lesson 6 MPTER
Lesson 7 VALLEY
4-1
-------
Lesson 1
Introduction to UNAMAP,
Version 4
Reading Guidance
The information of availability letter that you will read as part
of this lesson concerns all 21 UNAMAP models found in Ver-
sion 4. The only models mentioned in the letter you should
read carefully are the PTMAX, PTDIS, PTMTP, PTPLU,
CRSTER, RAM, MPTER, and VALLEY. This letter shows
you how the UNAMAP verions are announced.
Supplementary Reading
DOE/TIC-11223, Handbook on Atmospheric Diffusion, U.S.
Department of Energy, 1982, chapter 4.
Lesson Goal and Objectives
Goal
To familiarize you with the purpose and general characteristics
of the UNAMAP, Version 4* computer package.
Objectives
At the end of this lesson, you should be able to:
1. write the meaning of the acronym UNAMAP.
2. choose a statement that describes the purpose of
UNAMAP.
•Version 5 is scheduled for release in 1983 and will add several new models
to UNAMAP. All of the models discussed in this course will remain
unchanged with the release of Version 5.
4-3
-------
Introduction
In Unit 3, you read about the general characteristics of air
quality models—time and space scales, frame of reference,
pollutants, relationship of model to time, and treatment of tur-
bulences and topography. You discovered that models are not
easy to classify. We also discussed using models to make
decisions.
Review
Time and space scales
Frame of reference
Pollutants
Relationship of model
to time
Treatment of turbu-
lence and topography
Background
The package of computerized dispersion models that U.S. EPA
provides is called UNAMAP. UNAMAP is an acronym for
User's Network for Applied Modeling of Air Pollution.
Although not all UNAMAP models are approved, U.S. EPA-
approved models have been available on computer tape since
1973. UNAMAP is provided as a public service; it is not a
Federally mandated requirement. It also serves to ensure that
the same models are available to all users. The Meteorology
and Assessment Division (MD) of the Environmental Science
Research Laboratory (ESRL) at Research Triangle Park, North
Carolina, is the agency that decides which models will be on
the package.
UNAMAP
User's Network for
Applied Modeling of
Air Pollution
Selection for UNAMAP
UNAMAP contains EPA guideline models and other state-of-
the-art Gaussian dispersion techniques. Not all of the models
mentioned in the Guideline on Air Quality Models, EPA
450/2-78-027 (Figure 4-1), are included in the UNAMAP
package. Those that are excluded are generally of limited use
in regulatory activities, are exceedingly complex and expensive
to run (e.g., numerical models), or have undergone only
limited testing.
OAQPS Guideline
Series
Guideline on Air
Quality Models
Figure 4-1. Guideline.
4-4
-------
Availability
The UNAMAP, Version 4 is available without cost to certain
qualified users. For example, Federal, State, and local air
pollution modelers may use the package free. The Regional
Meteorologist in the user's area can grant permission for
obtaining the package or using the computer located in
Research Triangle Park, NC. The U.S. EPA computer is a
Sperry-Rand UNIVAC Series 1100 system model.
Other users, such as private industry, may obtain copies of
the UNAMAP tape from the National Technical Information
Service (NTIS) located in Springfield, Virginia. These users
must pay a fee of $840 for the complete model series. The
users will receive model software that is designed to run on a
UNIVAC computer. For other computer systems, changes to
the software, sometimes extensive, must be made. An IBM ver-
sion of the UNAMAP tape is available from HMM Associates
Inc. of Waltham, MA.
The UNAMAP series is updated periodically. As newer
models are created that have wider use and appeal, and other
models gain popular usage, they will be added to the package.
National Technical
Information Service
(NTIS)
Springfield, Virginia
UNAMAP $840.00
4-5
-------
\
I
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
RESEARCH TRIANGLE PARK
NORTH CAROLINA 27711
April 17, 1981
Information on availability of UNAMAP (Version A):
The Environmental Operations Branch makes available air quality dispersion
models through its system UNAMAP (User's Network for Applied Modeling of
Air Pollution). UNAMAP is in two forms: 1) EPA users access UNAMAP on
EPA's UNIVAC 1100 computer at Research Triangle Park, N.C. 2) Users
outside EPA can purchase a magnetic tape containing the FORTRAN source
codes and test data from NTIS so that the specific models of interest can
be installed and executed on the user's computer.
UNAMAP (Version 4) is the latest UNAMAP update and became available March
12, 1981. It contains FORTRAN source code for 21 air quality simulation
models. The contents of the tape as well as brief abstracts and references
are given in the attached description of UNAMAP. The tape furnished in
ASCII, 9 track, 1600 bits per inch, odd parity. The following information
is furnished for ordering the magnetic tape.
Tape name
Accession number
Price
Available from:
UNAMAP (Version 4)
PB 81 164 600
$ 840 for North American Purchasers
$1360 for all others
Computer Products (703) 487-4763
National Technical Information Service
U.S. Department of Commerce
Springfield, VA 22161
Users Guides are furnished with each tape as well as a copy of the print-out
resulting from execution of the models using the test data included on the
tape.
Sincerely yours,
Mailing address:
Chief, EOB
Mail Drop 80, EPA
RESCH TRI PK, NC 27711
D. Bruce Turner/NOAA (919) 541-4564
Chief, Environmental Operations Branch
Meteorology and Assessment Division
Environmental Sciences Research Laboratory
4-7
-------
UNAMAP (VERSION
WHAT IS UNAMAP?* *****
********
UNAMAP IS AN ACRONYM FOR USER'S NETWORK FOR APPLIED MODELING OF AIR POLLUTION. THIS IS A COLLECTION
OF FORTRAN SOURCE CODES FOR AIR QUALITY SIMULATION MODELS (AQSM).
UNAMAP EXISTS IN TWO FORMS:
1) SOURCE CODES AND EXECUTABLES RESIDE IN EPA'S UNIVAC 1110 AT RESEARCH TRIANGLE PARK, NC. THESE
PROGRAMS CAN BE READILY ACCESSED BY EPA USERS.
2) A MAGNETIC TAPE CONTAINING FORTRAN SOURCE CODES AND TEST DATA IS AVAILABLE FROM THE NATIONAL
TECHNICAL INFORMATION SERVICE, U. S. DEPARTMENT OF COMMERCE, SPRINGFIELD, VA 22161.
THIS TAPE HAS 23 FILES. TAPE NAME: UNAMAP (VERSION 4), ACCESSION NUMBER: PB 81 164 600,
PRICE: $840.
************* 'BACKGROUND* *************
SINCE 1973, UNAMAP HAS SERVED AS A SOURCE FOR AQSM'S IN COMPUTER COMPATIBLE FORM. THESE MODELS INPUT
EMISSION AND METEOROLOGICAL DATA TO CALCULATE PROJECTED AIR POLLUTANT CONCENTRATIONS. UNAMAP IS BASICALLY
STATE-OF-THE-ART DISPERSION RESEARCH ALGORITHMS SUPPORTED BY EPA'S OFFICE OF RESEARCH AND DEVELOPMENT.
AS AN ADDITIONAL SERVICE TO THE REGULATORY GROUPS IN EPA AND TO THOSE TRYING TO CONFORM TO REGULATIONS,
UNAMAP CONTAINS "GUIDELINE MODELS." GUIDELINE MODELS ARE THOSE IDENTIFIED BY EPA'S OFFICE OF AIR QUALITY
PLANNING AND STANDARDS IN "OAQPS GUIDELINE SERIES, GUIDELINE ON AIR QUALITY MODELS, EPA-450/2-78-027,
OAQPS NO. 1.2-080, APRIL 1978. (GUIDELINES HAVE BEEN PROPOSED AND PRESENTED AT THREE PUBLIC MEETINGS IN
WASHINGTON, SEATTLE, AND CHICAGO IN OCTOBER 1980. BASED ON COMMENTS RECEIVED AT THESE MEETINGS AND SUBMITTED
IN WRITING TO THE DOCKET BY DECEMBER 1. 1980, REVISIONS WILL BE MADE. THE REVISED PROPOSED GUIDELINES
WILL BE PRESENTED AND DISCUSSED AT A PUBLIC MEETING.)
IN THE PAST SOME PERSONS HAVE TENDED TO EQUATE UNAMAP WITH GUIDELINE MODELS. OUR INTENTION IS THAT
UNAMAP WILL INCLUDE GUIDELINE MODELS, IN SO FAR AS IS POSSIBLE, BUT THAT UNAMAP PRIMARILY REPRESENTS
STATE-OF-THE-ART DISPERSION RESEARCH ALGORITHMS. (SOME OF THESE MODELS MAY EVENTUALLY BECOME CANDIDATES
FOR GUIDELINE STATUS.)
VERSION 3 OF UNAMAP WAS MADE AVAILABLE IN MARCH 1978. IT CONTAINED 11 AQSM'S. THREE CHANGES WERE
ISSUED TO PURCHASERS OF UNAMAP (VERSION 3):
CHANGE 1 23 AUGUST 1978
CHANGE 2 5 JULY 1979
CHANGE 3 16 AUGUST 1979
UNAMAP (VERSION 4) RELEASED DECEMBER 1980.
IT IS ANTICIPATED THAT UPON AVAILABILITY OF THE 1981 MODELING GUIDELINES, THAT THE SECTION ON GUIDELINE
MODELS WILL BE REVISED AND UNAMAP (VERSION 5) WILL BE ISSUED. THIS IS CURRENTLY SCHEDULED FOR DECEMBER
1981.
********** CHANGES TO UNAMAP **********
SINCE IT IS RARE FOR A PIECE OF COMPUTER CODE TO BE COMPLETELY ERROR FREE UNDER ALL POSSIBLE COMBINATIONS
OF ACCEPTABLE INPUT, ERRORS ARE DETECTED OR BROUGHT TO OUR ATTENTION. WE TRY TO BRING TO THE ATTENTION OF
USERS OF UNAMAP THE CORRECTIONS TO RESOLVE SUCH ERRORS.
DISCUSSION OF UNAMAP SUPPORT
IN THE PAST THE ENVIRONMENTAL OPERATIONS BRANCH HAS SUPPORTED ALL FORTRAN COOES ON UNAMAP FROM THE
STANDPOINT OF HAVING AT LEAST ONE INDIVIDUAL IN THE BRANCH FAMILIAR WITH THE LINES OF PROGRAM CODE AND
THEIR FUNCTION AND HAVE MADE CODING CHANGES AS REQUIRED TO CORRECT THE CODE AS PROBLEMS OCCUR. WE HAVE
ALSO TRIED TO PROVIDE LIAISON WITH USERS IN ATTEMPTING TO ADVISE ON THE MOST APPROPRIATE MODEL (IF ANY)
TO USE FOR GIVEN SITUATIONS IN SO FAR AS THOSE SITUATIONS CAN BE ADEQUATELY DESCRIBED IN A PHONE CONVERSATION
OF REASONABLE LENGTH.
THERE ARE THREE PROFESSIONALS ON OUR STAFF THAT IN ADDITION TO OTHER RESEARCH DUTIES ASSIST ON UNAMAP.
THESE PERSONS ARE BRUCE TURNER, WILLIAM PETERSEN, AND JOHN IRWIN. WE HAVE BEEN FORTUNATE TO HAVE A
PART-TIME STUDENT EMPLOYEE, THOMAS PIERCE, OVER THE LAST YEAR AND A HALF WHO HAS ALSO BEEN OF GREAT
ASSISTANCE ON UNAMAP.
THE OFFICE OF AIR QUALITY PLANNING AND STANDARDS HAS GIVEN ASSISTANCE IN SUPPORT OF UNAMAP ESPECIALLY
IN REVISING AND MAINTAINING CRSTER AND VALLEY. WE ARE GRATEFUL TO JEROME MERSCH AND EDWARD BURT FOR THAT
SUPPORT.
4-8
-------
-2-
UNAMAP (VERSION 4) - CONTENTS
UNAMAP
DESCRIPTION
SECTION 1. MODELS
COMPLEX!!
COMPLEXI
BLP
POSTBLP
TOTAL
LINES
OF
SOURCE
CODE
NUMBER
OF
SUB
UN I VAC
1100
CORE
PROGS. REQUIRED
UNIVAC
1100
COST
OF
TEST
NUMBER
OUTPUT
PAGES
OF
TEST
FOR EVALUATION.
2811
2794
2062
929
8
8
21
5
51K
50K
20K
17K
S 0.40
$ 0.40
$ 0.75
$ 0.43
11
11
9
14
SECTION 2. GUIDELINE MODELS (1978).
RAM
RAMMET
CRSTER
CRSMET
COM
CDMQC
APRAC
HIWAY
HIWAYI
VALLEY
TEM8
TCM2
SECTION 3. MODELS
PAL
PTPLU
MPTER
HIWAY2
HIWAY2I
ISCST
ISCSMET
ISCLT
SECTION 4. MODELS
PTMAX
PTDIS
PTMTP
PTMAXI
PTDISI
PTMTPI
4547
332
1728
422
1313
1988
2015
1242
1303
1006
3778
2004
PROPOSED SEP
3484
957
2479
1298
1299
2756
332
3503
OF HISTORICAL
460
625
661
461
644
673
14
0
4
0
5
11
19
6
6
2
12
8
80 FOR 81
18
5
7
4
4
9
0
15 EXP1
EXP2
INTEREST.
2
3
3
2
3
3
64K
28K
20K
48K
32K
UK-
14K
39K
41K
$ 1.14
$ 2.31
$ 5.54
$10.14
$ 1.11
$ 0.92
$ 0.24
S 0.36
S 0.55
20
23
10
23
3
2
12
9
11
GUIDELINES.
43K
12K
48K
8K
65K
67K
69K
9K
9K
10K
$ 1.13
$ 0.38
$ 0.32
S 1.11
S 8.30
S10.35
$ 8.51
$ 0.13
S 0.22
$ 0.16
6
2
3
2
36
19
17
2
6
4
SECTION 5. MISCELLANEOUS.
TPHI5
TPRN25
251
318
1
1
SET RANDOM NUMBERS
ONE YEAR SURFACE DATA
(FOR INPUT TO MET PREPROCESSOR)
ONE YEAR MIXING HEIGHT DATA
(FOR INPUT TO MET PREPROCESSOR)
PROGRAM TO TRANSLATE MET DATA
TO UNFORMATTED
ONE YEAR MET DATA OUTPUT
FROM MET PROCESSOR (FORMATTED)
LINES OF DATA
366
8784
368
8785
4-9
-------
-3-
WE HAVE ALSO HAD ASSISTANCE FROM THE DATA MANAGEMENT AND SYSTEMS ANALYSIS SECTION OF THE ATMOSPHERIC
MODELING AND ASSESSMENT BRANCH. SPECIFICALLY JOAN NOVAK, CHIEF OF THAT SECTION, ADRIAN BUSSE, AND ALFREIDA
RANKINS HAVE ASSISTED CONSIDERABLY.
WITH THE ADDITION OF A NUMBER OF MODELS ON VERSION 4 WITH NO CHANGES IN STAFF AND WITH EXTRAMURAL
RESOURCES DECREASING MARKEDLY (NOT INCLUDING THE EFFECTS OF INFLATION), WE REGRET TO SAY THAT WE WILL NOT
BE ABLE TO SUPPORT ALL MODELS UNIFORMLY. IN FACT, SOME OF THE MODELS ARE PLACED IN UNAMAP FOR YOUR
EXAMINATION WITH NO SUPPORT.
********** UNAMAP PROGRAM DESCRIPTIONS
NOTE!
SERIES 600 EPA PUBLICATIONS ARE AVAILABLE FROM:
ENVIRONMENTAL RESEARCH INFORMATION CENTER
OFFICE OF RESEARCH AND DEVELOPMENT
ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OH 45268
PHONE: COMM'L (513)684-7562 FTS 684-7562
SERIES 450 EPA PUBLICATIONS ARE AVAILABLE FROM:
LIBRARY
MAIL DROP 35, EPA
RESRCH TRI PK, NC 27711
PHONE: COMM'L (919)541-2777 FTS 629-2777
COMPLEX II
COMPLEX II IS A MULTIPLE' POINT SOURCE COPE WITH TERRAIN ADJUSTMENT. THE MODEL SPECIFICATIONS FOR
TESTING WERE SUGGESTED BY TEAM "B" ON COMPLEX TERRAIN AT THE REGIONAL WORKSHOP ON AIR QUALITY MODELING IN
CHICAGO IN FEBRUARY 1980. IT IS A SEQUENTIAL MODEL UTILIZING HOURLY METEOROLOGICAL INPUT AND ASSUMES THAT
HOURLY AVERAGED PLUMES HAVE NORMAL DISTRIBUTIONS IN BOTH THE HORIZONTAL AND VERTICAL.
THERE IS NO USERS GUIDE FOR COMPLEX II AND NO PUNS TO DEVELOP ANY AS OF DEC 80.
A*******************************************************
COMPLEX I
COMPLEX I IS A MULTIPLE POINT SOURCE CODE WITH TERRAIN ADJUSTMENT. THE MODEL SPECIFICATIONS FOR
TESTING WERE SUGGESTED BY TEAM "B" ON COMPLEX TERRAIN AT THE REGIONAL WORKSHOP ON AIR QUALITY MODELING IN
CHICAGO IN FEBRUARY 1980. IT IS A SEQUENTIAL MODEL UTILIZING HOURLY METEOROLOGICAL INPUT. IT ASSUMES A
NORMAL DISTRIBUTION IN THE VERTICAL AND A UNIFORM DISTRIBUTION ACROSS A 22.5 DEGREE SECTOR: THEREFORE IT
REPRESENTS A SEQUENTIAL MODELING BRIDGE. BETWEEN VALLEY AND COMPLEX II.
THERE IS NO USERS GUIDE FOR COMPLEX I AND NO PLANS TO DEVELOP ANY AS OF DEC 80.
**********************************************************
BLP
BLP (BUOYANT LINE AND POINT SOURCE DISPERSION MODEL) IS A GAUSSIAN PLUME DISPERSION MODEL DESIGNED TO
HANDLE UNIQUE MODELING PROBLEMS ASSOCIATED WITH ALUMINUM REDUCTION PLANTS, AND OTHER INDUSTRIAL SOURCES
WHERE PLUME RISE AND DOWNWASH EFFECTS FROM STATIONARY LINE SOURCES ARE IMPORTANT.
SCHULMAN, LLOYD L., AND JOSEPH S. SCIRE. "BUOYANT LINE AND POINT SOURCE (BLP) DISPERSION MODEL
USER'S GUIDE." DOCUMENT P-7304B. ENVIRONMENTAL RESEARCH AND TECHNOLOGY, INC., CONCORD, MA. (NTIS
ACCESSION NUMBER PB 81 164 642.)
SCHULMAN, LLOYD L., AND JOSEPH S. SCIRE. "DEVELOPMENT OF AN AIR QUALITY DISPERSION MODEL FOR ALUMINUM
REDUCTION PLANTS." DOCUMENT P-7304A. ENVIRONMENTAL RESEARCH AND TECHNOLOGY, INC., CONCORD, MA (NTIS
ACCESSION NUMBER PB 81 164 634.)
*******************************************************
RAM
GAUSSIAN-PLUME MULTIPLE-SOURCE AIR QUALITY ALGORITHM. THIS SHORT-TERM GAUSSIAN STEADY-STATE ALGORITHM
ESTIMATES CONCENTRATIONS OF STABLE POLLUTANTS FROM URBAN POINT AND AREA SOURCES. HOURLY METEOROLOGICAL
DATA ARE USED. HOURLY CONCENTRATIONS AND AVERAGES OVER A NUMBER OF HOURS CAN BE ESTIMATED. BRIGGS PLUME
RISE IS USED. PASQUILL-GIFFORD DISPERSION EQUATIONS WITH DISPERSION PARAMETERS THOUGHT TO BE VALID FOR
URBAN AREAS ARE USED. CONCENTRATIONS FROM AREA SOURCES ARE DETERMINED USING THE METHOD OF HANNA, THAT IS,
SOURCES DIRECTLY UPWIND ARE CONSIDERED REPRESENTATIVE OF AREA SOURCE EMISSIONS AFFECTING THE RECEPTOR.
SPECIAL FEATURES INCLUDE DETERMINATION OF RECEPTOR LOCATIONS DOWNWIND OF SIGNIFICANT SOURCES AND DETERMI-
NATION OF LOCATIONS OF UNIFORMLY SPACED RECEPTORS TO ENSURE GOOD AREA COVERAGE WITH A MINIMUM NUMBER OF
RECEPTORS.
4-10
-------
-4-
TURNER, D. BRUCE, AND NOVAK, JOAN HRENKO, 1978: USER'S GUIDE FOR RAM, VOL. I. ALGORITHM DESCRIPTION
AND USE. EPA-600/8-78-016A (NTIS ACCESSION NUMBER PB 294 791), VOL. II. DATA PREPARATION AND LISTINGS.
EPA-600/8-78-016B (NTIS ACCESSION NUMBER PB 294 792.) U.S. ENVIRONMENTAL PROTECTION AGENCY, RESEARCH
TRIANGLE PARK, NC. (NOVEMBER 1978).
NOTE- RAM HAS BEEN REVISED IN 1980. BE SURE TO EXAMINE INFORMATION IN THE SOURCE CODE TO PREPARE RUNSTREAMS ETC.
THE REVISION IS FOR ADDED USER CONVENIENCE AND OPTIONS. ALL CALCULATIONS PRODUCED WITH THE ORIGINAL RAM CAN BE
REPRODUCED WITH THE CURRENT RAM WITH NO CHANGE IN NUMERICAL RESULTS.
*******************************************************
CRSTER
THIS ALGORITHM ESTIMATES GROUND-LEVEL CONCENTRATIONS RESULTING FROM UP TO 19 COLOCATED ELEVATED STACK
EMISSIONS FOR AN ENTIRE YEAR AND PRINTS OUT THE HIGHEST AND SECOND-HIGHEST 1-HR, 3-HR, AND 24-HR CONCENTRATIONS
AS WELL AS THE ANNUAL MEAN CONCENTRATIONS AT A SET OF 180 RECEPTORS (5 DISTANCES BY 36 AZIMUTHS). THE
ALGORITHM IS BASED ON A MODIFIED FORM OF THE STEADY-STATE GAUSSIAN PLUME EQUATION WHICH USES EMPIRICAL
DISPERSION COEFFICIENTS AND INCLUDES ADJUSTMENTS FOR PLUME RISE AND LIMITED MIXING. TERRAIN ADJUSTMENTS
ARE MADE AS LONG AS THE SURROUNDING TERRAIN IS PHYSICALLY LOWER THAN THE LOWEST STACK HEIGHT INPUT.
POLLUTANT CONCENTRATIONS FOR EACH AVERAGING TIME ARE COMPUTED FOR DISCRETE, NON-OVERLAPPING TIME PERIODS
(NO RUNNING AVERAGES ARE COMPUTED) USING MEASURED HOURLY VALUES OF WIND SPEED AND DIRECTION, AND ESTIMATED
HOURLY VALUES OF ATMOSPHERIC STABILITY AND MIXING HEIGHT.
MONITORING AND DATA ANALYSIS DIVISION, 1977: USER'S MANUAL FOR SINGLE-SOURCE (CRSTER) MODEL. U. S.
ENVIRONMENTAL PROTECTION AGENCY. RESEARCH TRIANGLE PARK, NC. EPA-450/2-77-013. (NTIS ACCESSION NUMBER
PB 271-360).
COM
THE CLIMATOLOGICAL DISPERSION MODEL DETERMINES LONG TERM (SEASONAL OR ANNUAL) QUASI-STABLE POLLUTANT
CONCENTRATIONS AT ANY GROUND LEVEL RECEPTOR USING AVERAGE EMISSION RATES FROM POINT AND AREA SOURCES AND
A JOINT FREQUENCY DISTRIBUTION OF WIND DIRECTION, WIND SPEED, AND STABILITY FOR THE SAME PERIOD.
BUSSE, ADRIAN D., AND ZIMMERMAN, J.R., 1973: USER'S GUIDE FOR THE CLIMATOLOGICAL DISPERSION MODEL.
U.S. ENVIRONMENTAL PROTECTION AGENCY. RESEARCH TRIANGLE PARK, NC. ENVIRONMENTAL MONITORING SERIES,
EPA-R4-73-024, 131 P. (NTIS ACCESSION NUMBER PB 227-346).
*******************************************************
CDMQC
THIS ALGORITHM IS THE CLIMATOL06ICAL DISPERSION MODEL (COM) ALTERED TO PROVIDE IMPLEMENTATION: OF
CALIBRATION, OF INDIVIDUAL POINT AND AREA SOURCE CONTRIBUTION LISTS, AND OF AVERAGING TIME TRANSFORMATIONS.
THE BASIC ALGORITHMS TO CALCULATE POLLUTANT CONCENTRATIONS USED IN THE COM HAVE NOT BEEN MODIFIED, AND
RESULTS OBTAINED USING COM MAY BE REPRODUCED USING THE COMQC.
BRUBAKER, KENNETH L., BROWN, POLLY, AND CIRILLO, RICHARD R., 1977: ADDENDUM TO USER'S GUIDE FOR
CLIMATOLOGICAL DISPERSION MODEL. PREPARED BY ARGONNE NATIONAL LABORATORY FOR THE U.S. ENVIRONMENTAL
PROCTECTION AGENCY, RESEARCH TRIANGLE PARK, NC. EPA-450/3-77-015. (NTIS ACCESSION NUMBER PB 274-040).
*******************************************************
APRAC
STANFORD RESEARCH INSTITUTE'S URBAN CARBON MONOXIDE MODEL COMPUTES HOURLY AVERAGES FOR ANY URBAN
LOCATION. REQUIRES AN EXTENSIVE TRAFFIC INVENTORY FOR THE CITY OF INTEREST. REQUIREMENTS AND TECHNICAL
DETAILS ARE DOCUMENTED IN:
USER'S MANUAL FOR THE APRAC-1A URBAN DIFFUSION MODEL COMPUTER PROGRAM (NTIS ACCESSION NUMBER PB 213-091.)
ADDITIONAL INFORMATION IS AVAILABLE ON APRAC FROM:
A PRACTICAL, MULTIPURPOSE URBAN DIFFUSION MODEL FOR CARBON MONOXIDI (NTIS ACCESSION NUMBER PB 196-003).
FIELD STUDY FOR INITIAL EVALUATION OF AN URBAN DIFFUSION MODEL FOR CARBON MONOXIDE (NTIS ACCESSION NUMBER
PB 203-469).
EVALUATION OF THE APRAC-1A URBAN DIFFUSION MODEL FOR CARBON MONOXIDE (NTIS ACCESSION NUMBER PB 210-813.)
DAB8ERDT, WALTER F.; LUDWIG, F.L.; AND JOHNSON, WARREN B., JR., 1973: VALIDATION AND APPLICATIONS OF AN
URBAN DIFFUSION MODEL FOR VEHICULAR POLLUTANTS, ATMOS. ENVIRON., 7, 603-618.
JOHNSON, W.B.; LUDWIG, F.L.; DABBERDT, W.F.; AND ALLEN, R.J., 1973: AN URBAN DIFFUSION SIMULATION MODEL
FOR CARBON MONOXIDE. J. AIR POLL. CONTROL ASSOC. 23, 6, 490-498.
*******************************************************
4-11
-------
-5-
HIUAY
COMPUTES THE HOURLY CONCENTRATIONS OF NON-REACTIVE POLLUTANTS DOWNWIND OF ROADWAYS. IT IS APPLICABLE
FOR UNIFORM WIND CONDITIONS AND LEVEL TERRAIN. ALTHOUGH BEST SUITED FOR AT-GRADE HIGHWAYS, IT CAN ALSO BE
APPLIED TO DEPRESSED HIGHWAYS (CUT SECTIONS).
ZIMMERMAN, J.R.: AND THOMPSON, R.S., 1975: USER'S GUIDE FOR HIWAY: A HIGHWAY AIR POLLUTION MODEL.
U.S. ENVIRONMENTAL PROTECTION AGENCY, RESEARCH TRIANGLE PARK, NC. ENVIRONMENTAL MONITORING SERIES,
EPA-650/4-74-008, 59 P. (NTIS ACCESSION NUMBER PB 239-944).
*******************************************************
VALLEY
THIS ALGORITHM IS A STEADY-STATE, UNIVARIATE GAUSSIAN PLUME DISPERSION ALGORITHM DESIGNED FOR ESTIMATING
EITHER 24-HOUR OR ANNUAL CONCENTRATIONS RESULTING FROM EMISSIONS FROM UP TO 50 (TOTAL) POINT AND AREA SOURCES.
CALCULATIONS OF GROUND-LEVEL POLLUTANT CONCENTRATIONS ARE MADE FOR EACH FREQUENCY DESIGNATED IN IN ARRAY DEFINED
BY SIX STABILITIES, 16 WIND DIRECTIONS, AND SIX WIND SPEEDS FOR 112 PROGRAM-DESIGNED RECEPTOR SITES ON A RADIAL
GRID OF VARIABLE SCALE. EMPIRICAL DISPERSION COEFFICIENTS ARE USED AND INCLUDE ADJUSTMENTS FOR PLUME RISE
AND LIMITED MIXING. PLUME HEIGHT IS ADJUSTED ACCORDING TO TERRAIN ELEVATIONS AND STABILITY CLASSES.
BURT, EDWARD W., 1977: VALLEY MODEL USER'S GUIDE, U.S. ENVIRONMENTAL PROTECTION AGENCY, RESEARCH
TRIANGLE PARK, NC, EPA-450/2- 77-018. (NTIS ACCESSION NUMBER PB 274-054).
*******************************************************
TEH
TEM (TEXAS EPISODIC MODEL) IS A SHORT-TERM, STEADY-STATE GUASSIAN PLUME MODEL FOR DETERMINING SHORT-TERM
CONCENTRATIONS OF NONE-REACTIVE POLLUTANTS.
STAFF OF THE TEXAS AIR CONTROL BOARD. USER'S GUIDE TO THE TEXAS EPISODIC MODEL. TEXAS AIR CONTROL BOARD,
PERMITS SECTION, 6330 HIGHWAY 290 EAST, AUSTIN, TEXAS 78723. (NTIS ACCESSION NUMBER PB 80-227 572)
TCM
TCM (TEXAS CLIMATOLOGICAL MODEL) IS A CLIMATOLOGICAL STEADY-STATE GAUSSIAN PLUME MODEL FOR DETERMINING
LONG-TERM (SEASONAL OR ANNUAL ARITHMETIC) AVERAGE POLLUTANT CONCENTRATIONS OF NON-REACTIVE POLLUTANTS, - __
STAFF OF THE TEXAS AIR CONTROL BOARD, USER'S GUIDE TO THE TEXAS CLIMATOLOGICAL MODEL (TCM). TEXAS AIR
CONTROL BOARD, PERMITS SECTION, 6330 HIGHWAY 290 EAST, AUSTIN, TX 78723 (NTIS ACCESSION NUMBER PB 81 164 626.)
************************************************ **.* *****
PAL ""
POINT, AREA, LINE SOURCE ALGORITHM. THIS SHORT-TERM GAUSSIAN STEADY-STATE ALGORITHM ESTIMATES
CONCENTRATIONS OF STABLE POLLUTANTS FROM POINT, AREA, AND LINE SOURCES. COMPUTATIONS FROM AREA SOURCES
INCLUDE EFFECTS OF THE EDGE OF THE SOURCE. LINE SOURCE COMPUTATIONS CAN INCLUDE EFFECTS FROM A VARIABLE
EMISSION RATE ALONG THE SOURCE. THE ALGORITHM IS NOT INTENDED FOR APPLICATION TO ENTIRE URBAN AREAS
BUT FOR SMALLER SCALE ANALYSIS OF SUCH SOURCES AS SHOPPING CENTERS, AIRPORTS, AND SINGLE PLANTS. HOURLY
CONCENTRATIONS ARE ESTIMATED AND AVERAGE CONCENTRATIONS FROM 1 HOUR TO 24 HOURS CAN BE OBTAINED.
PETERSEN, WILLIAM B., 1978: USER'S GUIDE FOR PAL - A GAUSSIAN-PLUME ALGORITHM FOR POINT, AREA, AND
LINE SOURCES. U.S. ENVIRONMENTAL PROTECTION AGENCY, RESEARCH TRIANGLE PARK, N.C., ENVIRONMENTAL MONITORING
SERIES EPA-600/4-78-013 (NTIS ACCESSION NUMBER PB 281-306).
PTPLU
PTPLU IS A POINT SOURCE DISPERSION GAUSSIAN SCREENING MODEL FOR ESTIMATING MAXIMUM SURFACE CONCENTRATIONS
FOR 1-HOUR PERIODS. PTPLU IS BASED UPON BRIGGS PLUME RISE METHODS AND PASQUILL-GIFFORO DISPERSION
COEFFICIENTS AS OUTLINED IN THE WORKBOOK OF ATMOSPHERIC DISPERSION ESTIMATES. PTPLU IS AN ADAPTATION AND
IMPROVEMENT OF PTMAX WHICH ALLOWS FOR WIND PROFILE EXPONENTS AND OTHER OPTIONAL CALCULATIONS SUCH AS
BUOYANCY INDUCED DISPERSION, STACK DOWNWASH, AND GRADUAL PLUME RISE. PTPLU PRODUCES AN ANALYSIS OF CONCENTRA-
TION AS A FUNCTION OF WIND SPEED AND STABILITY CLASS FOR BOTH WIND SPEEDS CONSTANT WITH HEIGHT AND WIND
SPEEDS INCREASING WITH HEIGHT. USE OF THE EXTRAPOLATED WIND SPEEDS AND THE OPTIONS ALLOWS THE MODEL USER A
MORE ACCURATE SELECTION OF DISTANCES TO MAXIMUM CONCENTRATION.
THERE IS NO USER'S GUIDE AVAILABLE FOR PTPLU. THE USER IS REFERRED TO THE SOURCE CODE FOR INPUT
FORMATS, ETC.
• A******************************************************
4-12
-------
-6-
MPTER
MPTER IS A MULTIPLE POINT-SOURCE GAUSSIAN MODEL WITH OPTIONAL TERRAIN ADJUSTMENTS. MPTER ESTIMATES
CONCENTRATIONS ON AN HOUR-BY-HOUR BASIS FOR RELATIVELY INERT POLLUTANTS (I.E., S02 AND TSP). MPTER USES
PASOUILL-GIFFORD DISPERSION PARAMETERS AND BRIGGS PLUME RISE METHODS TO CALCULATE THE SPREADING AND THE
RISE OF PLUMES. THE MODEL IS MOST APPLICABLE FOR SOURCE-RECEPTOR DISTANCES LESS THAN 10 KILOMETERS AND
FOR LOCATIONS WITH LEVEL OR GENTLY ROLLING TERRAIN. TERRAIN ADJUSTMENTS ARE RESTRICTED TO RECEPTORS
WHOSE ELEVATION IS NO HIGHER THAN THE LOWEST STACK TOP. IN ADDITION TO TERRAIN ADJUSTMENTS, OPTIONS ARE
ALSO AVAILABLE FOR WIND PROFILE EXPONENTS, BUOYANCY INDUCED DISPERSION, GRADUAL PLUME RISE, STACK
DOWNWASH, AND PLUME HALF-LIFE.
PIERCE, T. E. AND TURNER, D. B., 1980: USER'S GUIDE FOR MPTER: A MULTIPLE POINT GAUSSIAN DISPERSION
ALGORITHM WITH OPTIONAL TERRAIN ADJUSTMENT. EPA-600/8-80-016, U.S. ENVIRONMENTAL PROTECTION AGENCY,
RESEARCH TRIANGLE PARK, NC. 239 PP.
*******************************************************
HIWAY2
HIWAY2 IS A BATCH AND INTERACTIVE PROGRAM WHICH COMPUTES THE HOURLY CONCENTRATIONS OF NON-REACTIVE
POLLUTANTS DOWNWIND OF ROADWAYS. IT IS APPLICABLE FOR UNIFORM WIND CONDITIONS AND LEVEL TERRAIN. ALTHOUGH
BEST SUITED FOR AT-GRADE HIGHWAYS. IT CAN ALSO BE APPLIED TO DEPRESSED HIGHWAYS (CUT SECTIONS). HIWAY2
IS INTENDED AS AN UPDATE TO THE HIWAY MODEL.
PETERSEN, W. B., 1980. USER'S GUIDE FOR HIWAY2: A HIGHWAY AIR POLLUTION MODEL. U.S. ENVIRONMENTAL
PROTECTION AGENCY, RESEARCH TRIANGLE PARK, NC., EPA-600/8-80-018,70 P.
RAO, S. T. AND M. T. KEENAN, 1980: SUGGESTIONS FOR IMPROVEMENT OF THE EPA HIWAY MODEL. JAPCA, 30, 6,
pp 247-256.
*******************************************************
ISCST
THE INDUSTRIAL SOURCE COMPLEX SHORT TERM MODEL IS A STEADY-STATE GAUSSIAN PLUME MODEL WHICH CAN BE
USED TO ASSESS POLLUTANT CONCENTRATIONS FROM A WIDE VARIETY OF SOURCES ASSOCIATED WITH AN INDUSTRIAL
SOURCE COMPLEX. THIS MODEL CAN ACCOUNT FOR SETTLING AND DRY DEPOSITION OF PARTICULATES, DOWNWASH, AREA,
LINE AND VOLUME SOURCES, PLUME RISE AS A FUNCTION OF DOWNWIND DISTANCE, SEPARATION OF POINT SOURCES, AND
LIMITED TERRAIN ADJUSTMENT. AVERAGE CONCENTRATION OR TOTAL DEPOSITION MAY BE CALCULATED IN 1-, 2-, 3-,
4-, 6-, 8-, 12- AND/OR 24-HOUR TIME PERIODS. AN 'N' -DAY AVERAGE CONCENTRATION (OR TOTAL DEPOSITION) OR AN
AVERAGE CONCENTRATION (OR TOTAL DEPOSITION) OVER THE TOTAL NUMBER OF HOURS MAY ALSO BE COMPUTED.
BOWERS, J. F., J. R. BJORKLUNO AND C. S. CHENEY. "INDUSTRIAL SOURCE COMPLEX (ISC) DISPERSION MODEL
USER'S GUIDE, VOLUMES 1 AND 2." PUBLICATION NOS. EPA-450/4-79-030,031 (NTIS PB-80-133 044, 133 051), OFFICE
OF AIR QUALITY PLANNING AND STANDARDS. U. S. ENVIRONMENTAL PROTECTION AGENCY, RESEARCH TRIANGLE PARK,
NORTH CAROLINA 27711, DECEMBER 1979.
*******************************************************
ISCLT
THE INDUSTRIAL SOURCE COMPLEX LONG TERM MODEL IS A STEADY-STATE GAUSSIAN PLUME MODEL WHICH CAN BE USED
TO ASSESS POLLUTANT CONCENTRATIONS FROM A WIDE VARIETY OF SOURCES ASSOCIATED WITH AN INDUSTRIAL SOURCE
COMPLEX. THIS MODEL CAN ACCOUNT FOR SETTLING AND DRY DEPOSITION OF PARTICULATES, DOWNWASH. AREA, LINE AND
VOLUME SOURCES, PLUME RISE AS A FUNCTION OF DOWNWIND DISTANCE, SEPARATION OF POINT SOURCES, AND LIMITED
TERRAIN ADJUSTMENT.
ISCLT IS DESIGNED TO CALCULATE THE AVERAGE SEASONAL AND/OR ANNUAL GROUND LEVEL CONCENTRATION OR TOTAL
DEPOSITION FROM MULTIPLE CONTINUOUS POINT, VOLUME AND/OR AREA SOURCES. PROVISION IS MADE FOR SPECIAL
DISCRETE X, Y RECEPTOR POINTS THAT MAY CORRESPOND TO SAMPLER SITES, POINTS OF MAXIMA OR SPECIAL POINTS OF
INTEREST. SOURCES CAN BE POSITIONED ANYWHERE RELATIVE TO THE GRID SYSTEM.
SOWERS, J. F., J. R. BJORKLUND AND C. S. CHENEY. "INDUSTRIAL SOURCE COMPLEX (ISC) DISPERSION MODEL
USER'S GUIDE, VOLUMES 1 AND 2." PUBLICATION NOS. EPA-450/4-79-030,031 (NTIS PB-80-133 044, 133 051), OFFICE
OF AIR QUALITY PLANNING AND STANDARDS, U. S. ENVIRONMENTAL PROTECTION AGENCY, RESEARCH TRIANGLE PARK,
NORTH CAROLINA 27711, DECEMBER 1979.
*********************************
PTMAX
PERFORMS AN ANALYSIS OF THE MAXIMUM SHORT-TERM CONCENTRATIONS FROM A SINGLE POINT SOURCE AS A FUNCTION
OF STABILITY AND WIND SPEED. THE FINAL PLUME HEIGHT IS USED FOR EACH COMPUTATION. USES BRIGGS PLUME
RISE METHODS AND PASQUILL-GIFFORD DISPERSION METHODS AS GIVEN IN EPA'S AP-26, "WORKBOOK OF ATMOSPHERIC
DISPERSION ESTIMATES," TO ESTIMATE HOURLY CONCENTRATIONS FOR STABLE POLLUTANTS.
TURNER. D.B.: AND BUSSE, A.D., 1973: USER'S GUIDE TO THE INTERACTIVE VERSIONS OF THREE POINT SOURCE
DISPERSION PROGRAMS: PTMAX, PTDIS, AND PTMTP. PRELIMINARY DRAFT, METEOROLOGY LABORATORY, U.S. ENVIRONMENTAL
PROTECTION AGENCY, RESEARCH TRIANGLE PARK, NC. 27711
********************************************************
4-13
-------
-7-
PTDIS
ESTIMATES SHORT-TERM CONCENTRATIONS DIRECTLY DOWNWIND OF A POIWT SOURCE AT DISTANCES SPECIFIED BY
THE USER. THE EFFECT OF LIMITING VERTICAL DISPERSION BY A MIXING HEIGHT CAN BE INCLUDED AND GRADUAL
PLUME RISE TO THE POINT OF FINAL RISE IS ALSO CONSIDERED. AN OPTION ALLOWS THE CALCULATION OF ISOPLETH
HALF-WIDTHS FOR SPECIFIC CONCENTRATIONS AT EACH DOWNWIND DISTANCE. USES BRIGGS PLUME RISE METHODS
AND PASQUILL-GIFFORD DISPERSION METHODS AS GIVEN IN EPA'S AP-26, "WORKBOOK OF ATMOSPHERIC DISPERSION
ESTIMATES," TO ESTIMATE.HOURLY CONCENTRATIONS FOR STABLE POLLUTANTS.
TURNER, O.B.ANO BUSSE, A.D., 1973: USER'S GUIDE TO THE INTERACTIVE VERSIONS OF THREE POINT SOURCE
DISPERSION PROGRAMS: PTMAX, PTDIS, AND PTMTP. PRELIMINARY DRAFT, METEOROLOGY LABORATORY, U.S. ENVIRON-
MENTAL PROTECTION AGENCY, RESEARCH TRIANGLE PARK, NC 27711.
********************************************************
PTMTP
ESTIMATES FOR A NUMBER OF ARBITRARILY LOCATED RECEPTOR POINTS AT OR ABOVE GROUND-LEVEL, THE CONCENTRATION
FROM A NUMBER OF POINT SOURCES. PLUME RISE IS DETERMINED FOR EACH SOURCE. DOWNWIND AND CROSSWINO DISTANCES
ARE DETERMINED FOR EACH SOURCE-RECEPTOR PAIR. CONCENTRATIONS AT A RECEPTOR FROM VARIOUS SOURCES ARE
ASSUMED ADDITIVE. HOURLY METEOROLOGICAL DATA ARE USED: BOTH HOURLY CONCENTRATIONS AND AVERAGES OVER ANY
AVERAGING TIME FROM ONE TO 24 HOURS CAN BE OBTAINED. USES BRIGGS PLUME RISE METHODS AND PASQUILL-GIFFORD
DISPERSION METHODS AS GIVEN IN EPA'S AP-26, "WORKBOOK OF ATMOSPHERIC DISPERSION ESTIMATES," TO ESTIMATE
HOURLY CONCENTRATIONS FOR STABLE POLLUTANTS.
TURNER. D.B.: AND BUSSE, A.D., 1973: USER'S GUIDE TO THE INTERACTIVE VERSIONS OF THREE POINT SOURCE
DISPERSION PROGRAMS: PTMAX, PTDIS, AND PTMTP. PRELIMINARY DRAFT, METEOROLOGY LABORATORY, U.S. ENVIRONMENTAL
PROTECTION AGENCY, RESEARCH TRIANGLE PARK, NC 27711
TPHI5
TPHIS (TURNER AND PIERCE'S HIGH-FIVE PROGRAM) IS A PERIPHERAL PROGRAM WHICH READS DATA OFF AN HOURLY
CONC FILE (OUTPUT FROM RAM, CRSTER, OR MPTER) AND TABULATES END TO END AVERAGE CONCENTRATIONS FOR VARIOUS
AVG TIMES (UP TO 5) FOR THE NUMBER OF STATIONS CONTAINED ON THE TAPE/DISK FILE ASSIGNED TO THE RUNSTREAM.
THERE IS CURRENTLY NO USER'S GUIDE FOR TPHIS. USERS ARE REFERRED TO A LISTING OF THE SOURCE PROGRAM
FOR INFORMATION.
********************************************************
TPRN2S
TPRN25 IS A PERIPHERAL PROGRAM DESIGNED TO READ CONCS.OFF A DISK/TAPE FILE AND DETERMINE RUNNING
AVERAGES FOR FOUR OR FIVE AVG.TIMES. THE USER DESIGNATES THE FIFTH AVERAGING TIME AND THE STATION
NUMBERS (UP TO 50). THE OUTPUT THEN CONSISTS OF TABLES OF THE 25 HIGHEST CONCENTRATIONS FOR EACH AV6
TIME AND RECEPTOR.
THERE IS CURRENTLY NO USER'S GUIDE FOR TPRN25. USERS ARE REFERRED TO A LISTING OF THE SOURCE PROGRAM
FOR INFORMATION.
********************************************************
4-14
-------
Review Exercise
1. What does the acronym UNAMAP stand for?
2. UNAMAP is available
a. to make a profit for NTIS.
b. to require the air quality modeler to use U.S. EPA
models.
c. as a public service.
d. to make Mr. Turner famous.
e. as a repository for all U.S. EPA models.
1. User's Network for Applied
Modeling of Air Pollution
The agency responsible for deciding which models are
included in UNAMAP is the
a. Meteorology and Assessment Division of ESRL.
b. Source Receptor Analysis Branc'h of OAQPS.
c. Monitoring and Data Analysis Division of HERL.
2. c. as a public service.
4. The latest UNAMAP version number (as of 1982) is
3. a. Meteorology and Assess-
ment Division of ESRL.
What is one problem that may arise for persons desiring to
use UNAMAP?
a. If they do not use a UNI VAC computer, changes to the
model software may be necessary.
b. The user may have a machine with tape drives; this
package is supplied on disk only.
c. The user must have a computer with the basic language.
4. four (4)
5. a. If they do not use a
UNIVAC computer, changes
to the model software may be
necessary.
4-15
-------
Lesson 2
PTXXX Models:
PTMAX, PTDIS, and PTMTP
Lesson Goal and Objectives
Goal
To familiarize you with plume distribution, method of plume
rise, sigma y and sigma z, data entries, limitations, and use of
the PTXXX models, which include PTMAX, PTDIS, and
PTMTP.
Objectives
At the end of this lesson, you should be able to:
1. describe the plume distribution of the PTXXX models.
2. recognize the reason a time average of conditions is
necessary for a Gaussian distribution.
3. identify the reason Gaussian plume models may not give
accurate estimates of pollution.
4. define sigma y and sigma z.
5. name the two methods of entering data for the PTXXX
models.
6. state the reason the PTXXX models are used as screening
models.
Introduction
The first models to be discussed all belong to the series of
models known as the PTXXX models. Three model
acronyms-PTMAX, PTDIS, and PTMTP-all represent the
specific use intended for each model. For example, the
PTMAX model letters stand for Point Maximum, meaning the
model calculates the maximum ground-level pollutant concen-
trations for point sources. PTDIS is the Point Distance model
that determines the downwind profile of concentrations with
distance. PTMTP is the Point-Multi-Point model that can
calculate impacts for more than one point source on a field of
receptors. The PTXXX models were among the first opera-
tional air quality models to be used. They were derived from
PTXXX models
PTMAX
PTDIS
PTMTP
4-17
-------
the techniques in the Workbook for Atmospheric Dispersion
Estimates (WADE) by D. Bruce Turner (Figure 4-2). This
lesson will discuss the model's plume distribution input, limita-
tions, and use and output. While the PTXXX models evolved
from a common approach, each of the three variations was
intended to produce results in addition to the standard output,
maximum ground-level concentrations. This extra information
that the user has to specifically ask for is called an option.
Plume Characteristics
The models are called Gaussian because the pollutant mass
within the plume is assumed to follow a bell-shaped curve,
called the normal distribution (Figure 4-3). A normal, or
Gaussian, distribution is one in which the maximum concentra-
tions occur in the middle of the plume and taper exponentially
to almost zero at the edges, as in Figure 4-4. The edge of the
plume is defined by the point where the concentration drops to
10% of the centerline value. For example, if the maximum
concentration is 220 /tg/ms at centerline, then the edge would
occur where the concentration was 22 /*g/ms. The use of the
Gaussian distribution as a basis for plume descriptions is a
simplifying assumption.
Boundary Conditions
This one major assumption incorporates a number of other
supporting assumptions called boundary conditions. The first
supporting assumption is that the atmosphere and source are in
steady state. Being steady state means that the atmosphere and
source conditions are constant over a period of time. For the
PTXXX models, meteorology and emission conditions are
assumed to be invariant for a 1-hour period. Therefore, this is
not an instantaneous picture of conditions. Since, in reality,
both the atmosphere and source are variable over periods of
time, an average must be taken that uses many instantaneous
pictures.
The second supporting assumption is that no pollutant mass
is lost from the plume through chemical reaction or physical
deposition on a surface. This is called conservation of mass.
The third supporting assumption is that the plume does not
stretch in the downwind direction. This means that the pollu-
tant material through any slice, or cross section, of the plume
is the same as any other cross section of the plume; distance
from the source does not matter.
Workbook of
Atmospheric Dispersion
Figure 4-2. WADE.
Figure 4-3. Normal distribution.
220
22
— -I
/
1
1
1
1
- /
^.^
"C
\
\
\
\
\
\
Edge^-
CL
Figure 4-4. Plume distribution.
Assumptions
1.
2.
3.
4.
5.
Steady state
No removal
No downwind stretching
Stable pollutant
Average wind
4-18
-------
The last supporting assumption is that an average wind
speed and direction can be identified for the 1-hour period,
and that they are typical of the atmospheric layer that will
disperse the plume.
Accuracy of Model
Boundary conditions limit the model's ability to fully describe
the physical conditions of the source and atmosphere. This
means that models using the Gaussian distribution may not
estimate pollutant concentrations accurately (Figure 4-5). The
assumptions are the reasons that the model results are conser-
vative. That is, the estimates of downwind concentrations are
larger than may be observed at a real receptor. Using the
PTXXX models, a calculation for a new source will over-
estimate the source's effect on air quality. Three factors called
plume rise, sigma y, and sigma z must be input to. estimate
pollution concentration.
Plume Rise Method
The distance above the stack that the plume centerline will
climb before leveling off is called plume rise (Figure 4-6).
Plume rise, Ah, is calculated using formulas developed by
G. A. Briggs. The Briggs' formulas for stable or neutral
atmospheric conditions use the Pasquill-Gifford (P-G) stability
classifications A through D. (A through D are identified for the
computer as 1 through 4.) When the atmosphere is stable, P-G
classifications £ and F (computer identified as 5 and 6) are
used in the formulas. The plume rise, Ah, is added to the
physical height of the stack, h, resulting in the effective plume
rise, H (Example 4-1).
H is the calculated centerline of the plume, not the plume
edge. The centerline is where the maximum pollution concen-
tration occurs. Our first concern is the plume centerline's rela-
tionship to ground level. The relationship allows the calcula-
tion of an estimate of maximum ground-level concentrations
with distance from a source. Ground-level concentrations are
reduced as the plume rises. For example, let the conditions of
the atmosphere and the source remain constant and the
physical stack height increase from 10 to 20 meters, as in
Figure 4-7. The ground-level concentrations may decrease by a
calculated factor. Therefore, plume rise, Ah, is an important
calculation in the models.
Figure 4-5. Boundary condition*.
Figure 4-6. Plume rise method.
Example 4-1. Plume rise
calculation.
Figure 4-7. x difference due
stack height.
4-19
-------
Sigma y and Sigma z
The factors called sigma y and sigma z are the horizontal and
vertical dispersion parameters, and their distribution is called
binormal (Figure 4-8). In effect, they are the standard devia-
tions of the plume concentration distribution in horizontal and
vertical directions. The sigmas are measures of the Gaussian
distribution. The edge of the plume, at which the concentra-
tion is 10% of the centerline, is 2.15 standard deviations from
the center. The values for sigma y and sigma z, found in
graphs by Turner, give the rate of dispersion as a function of
stability class (A through F) and downwind distance. The
values of sigma y and sigma z estimate the width of a plume as
it travels downwind. Dispersion rates are also a very important
part of calculating pollution concentrations.
Computer Modes
The PTXXX models may be run in two modes: batch (Figure
4-9) and interactive (Figure 4-10). In the batch mode, all
inputs to the computer are on IBM cards or card images.
These cards are placed in a reader machine and the data is
transferred to the computer for processing. In the interactive
mode, the computer will ask for input data as they are needed.
The user must have a special typewriter keyboard, called a
remote terminal, in order to access the models in this mode.
The batch mode method is a faster and slightly less expensive
method than the interactive mode. However, the interactive
mode is usually more convenient. The PTXXX models' inputs
are very similar to each other.
2.5 a
Figure 4-8. Binomial distribution.
Figure 4-9. Batch mode
card reader.
Inputs
As shown in Table 4-1, all PTXXX models require inputs such
as source emission strength (Q), stack height (h,), stack gas
temperature (T,), stack diameter (d), stack gas velocity (v,),
ambient air temperature (Ta), and P-G stability category (1
through 6, A through F). The source strength, Q,, given in
grams per second, is the amount of pollution leaving the stack
per second. The stack height, h., given in meters, is the actual
height of the stack. The stack gas temperature, T,, given in
Kelvin, is the temperature at which the pollution leaves the
stack. The stack gas velocity, vj( given in meters per second, is
the speed at which the pollution leaves the stack. The stack
diameter, d, given in meters, is the width of the stack opening.
The ambient air temperature, T0, given in Kelvin, is the out-
side air temperature. The Pasquill-Gifford atmospheric stability
PTXXX
WIND DIRECTION?
NE
WIND SPEED?
Figure 4-10. Interactive mode
terminal.
4-20
-------
categories A through F, which are coded as 1 through 6, are
used to calculate values for sigma y and sigma z. The inputs
that are not the same for the variations of the PTXXX are
shown in Table 4-1 also.
Table 4-1. PTXXX inputs.
Input by u*er
Ambient air TB (K)
Wind speed u (m/s)
Source emission strength Q. (g/s)
Mixing height L (m)
Stack height h, (m)
Stack gas temperature T, (K)
Stability classes 1-6
Stack diameter d (m)
Stack velocity v, (m/s)
Receptor locations
(x,y coordinates)
Isopleth values (g/m*)
Averaging times (s)
PTMAX
yes
no
yes
no
yes
yes
yes
yes
yes
no
no
no
PTDIS
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
no
PTMTP*
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
no
yes
'Considers multiple sources
Limitations
The PTDIS and PTMTP models do require that the average
wind speed, u in m/s, be entered. However, the PTMAX has
predetermined wind speeds within the model for each stability.
These wind speeds are considered appropriate for each stability
category. The PTDIS and PTMTP models require a mixing
height, L, entry (Figure 4-11). The PTMAX model does not
consider any situation that involves a limitation to mixing. The
PTDIS and PTMTP models require an entry for the number of
receptor sites that are to be considered and the distance to
each from the source. The PTMAX is designed to calculate the
distance to the maximum concentration for each P-G stability «
category A through F and each of the predetermined wind
speeds.
Figure 4-11. Mixing height.
4-21
-------
The PTDIS model has an option that will help in drawing
concentration isopleths (lines of equal concentrations) on a
map, as shown in Figure 4-12. The model asks for the number
of isopleths desired in their strength. Also, the model needs to
know whether 16-point wind information or 36-point wind
information is used. The PTMAX and PTMTP models do not
consider isopleths.
y-aj
150
100
50
Source (
50
100
150
as (mj
1 1C
/
0 2(
\
/
10 3
X
/
>
30 4(
^
^.
*v
s*
=^
« 5
-•^
\
^t
f^~
.*•—
^-Isop
W &
-•--J
^»x
2 x 10'4 g
——
leths
K) 7(
— •«.
__!
10 8(
»•»
/m1
I
xlO-»
10 9C
g/m»
0 10
00
. x-azis (m)
Figure 4-12. Ground-level concentration uoplethi.
The PTMTP model will produce a time-averaged concentra-
tion for periods longer than 1 hour. The model merely adds
the hourly concentrations and divides by the number of con-
centrations involved in the addition. The PTMAX and PTDIS
models consider only hourly averages.
To the novice air quality modeler it may not be apparent
that the PTXXX models are simple models. They are used as
screening models. This means they will analyze a limited
number of sources, meteorological conditions, and receptors,
and will yield a conservative estimate of concentration. If more
detail about concentrations is needed, the PTXXX models
should not be used. As the other models are examined, the
simplicity of the PTXXX should become obvious.
Use and Output
Because of their simplicity, the PTXXX models in Figures 4-13
to 4-15 cost less than one dollar to produce. Most applications
are inexpensive.
The first example, Figure 4-13, is a run for the PTMAX
model. The input on the left side of the page produces the out-
put on the right. Under the output, column one provides all
six P-G stabilities; column two gives the predetermined wind
4-22
-------
speeds for each stability; column three gives maximum concen-
trations for each stability and maximum concentrations for
distances given in column four; column five gives the final
plume rise heights.
This output allows the user to identify the critical wind
speed for each stability. Critical wind speed means the wind
speed that allows the highest concentrations at a specific
stability to occur.
Figure 4-14 is an example of the PTDIS model run. Note the
similarity to the PTMAX interactive mode entries. Also note
the differences that have been mentioned. In section one, col-
umn one, are the preselected receptor distances; in column two
is final plume rise; in column three are the maximum concen-
trations; in columns four and five are the sigma y and sigma z
values at the receptor distances. Output section two gives the
information necessary for drawing concentration isopleths.
Each value of sigma y given in the half-width column is the
distance from the x-axis (mean wind direction) to the isopleth
value above. For example, the first isopleth value is 0.1 X 10"*.
The half-width is given at 300 meters as 57 meters. The par-
tially drawn isopleth was shown in Figure 4-12.
4-23
------- An error occurred while trying to OCR this image.
------- An error occurred while trying to OCR this image.
-------
Figure 4-15 is an example of a run for the PTMTP model.
Again note input similarities to PTMAX and PTDIS models.
Because the PTMTP handles multiple sources, the entries are
multiple: one for each stack. This example has four stacks.
Section one contains source entries. Section two has 14 receptor
distances in x (prec) and y (srec) coordinates. Section three
contains meteorological inputs. Section four is part of the
hourly concentrations at each receptor 1 through 6, 13, and 14.
Receptors 2 through 12 are not shown. Section five is the
3-hour average of the three hourly average concentrations given
in section four. The partial concentrations that contribute to
the total pollution are given in sections four and five.
ENTER ALPHANUMERIC TITLE (UP TO 64 CHARACTERS)
?
TEST OF PTMTP 10/11/79
ENTER NUMBER OF SOURCES TO BE CONSIDERED. MAX 25
4
ENTER SOURCE STRENGTH (G/SEC) FOR EACH STACK
4*287
ENTER PHYSICAL HEIGHT (M) OF EACH STACK
4*30
ENTER GAS TEMPERATURE (DEG K) OF EACH STACK
4*350
IS VOLUME FLOW KNOWN FOR EACH STACK? YES OR NO
1
NO
ENTER GAS vaOCITY (M/SEC) FOR EACH STACK
4*20
ENTER DIAMETER (M) OF EACH STACK
4*0.6
ENTER COORDINATES (KM) OF EACH STACK. ORDERED PAIRS
1..0., 1.05,0., 1.10,0., 1.15,0.
ENTER NU»ER OF RECEPTORS TO BE PROCESSED. MAX 30
14
ENTER COORDINATES (KM) OF EACH RECEPTOR. ORDERED PAIRS
.8,0., 1.02,0., 1.07,0., 1.12.0., 1.17,0.. 1.2,0.. 1.3,0.,
1.4,0., 1.5,0., 1.6,0., 1.7,0., 1.8,0., 1.9,0., 2.0,0.
ENTER HEIGHT (M) ABOVE GROUND FOR EACH RECEPTOR
14*0.
ENTER NUMBER OF HOURS TO BE AVERAGED. MAX 24
3
ENTER WIND DIRECTION (DEG) FOR EACH HOUR
265,270.275
ENTER WIND SPEED (M/SEC) FOR EACH HOUR
4,4,4
ENTER STABILITY CLASS FOR EACH HOUR
3*3
ENTER MIXING HEIGHT (M) FOR EACH HOUR
3*700
ENTER AMBIENT AIR TEMPERATURE (DEG K) FOR EACH HOUR
3*293
DO YOU WANT PARTIAL CONCENTRATIONS PRINTED? YES OR NO
t
YES
DO YOU WANT HOURLY CONCENTRATIONS PRINTED? YES OR NO
7
YES
Figure 4-15. PTMTP model run
(input).
4-26
------- An error occurred while trying to OCR this image.
-------
Review Exercise
The PTXXX models are Gaussian plume models. This
means that
a. the plume spreads vertically, but not horizontally.
b. the maximum concentrations occur in the middle of
the plume.
c. the models estimate time-averaged concentrations.
d. both b and c
2. Time averages of conditions, such as wind speed, are
necessary for a Gaussian distribution because
a. conditions are variable with time.
b. instantaneous wind readings are not possible.
c. instantaneous wind readings are possible but expensive.
d. steady-state conditions are not assumed.
1. d. both b and c
3. In Gaussian plume distributions, the edge of the plume is
defined as
a. the visible edge of the plume.
b. the point where concentration drops to 10% of the
centerline concentration.
c. being determined by each individual plume.
d. not being important to concentration estimates.
2. a. conditions are variable
with time.
One of the supporting assumptions for the Gaussian distri-
bution is steady-state conditions. Being steady state
means that
a. the atmosphere is sampled instantaneously.
b. the atmosphere is always the same no matter how long
the time period.
c. there aren't any variations in the atmosphere.
d. conditions are constant for a given period of time.
3. b. the point where
concentration drops to
10% of the centerline
concentration.
5. The PTXXX models are conservative in estimating down-
wind concentrations. This means that
a. concentration estimates are larger than actual
observations.
b. actual observations are always larger than the concen-
tration estimates.
c. concentration estimates are the same as actual
observations.
d. concentration estimates are impossible to make.
4. d. conditions are constant
for a given period of time.
5. a. concentration estimates
are larger than actual
observations.
4-28
-------
Plume rise calculations are important for estimating
ground-level concentrations because they
a. determine the distance of the centerline of the plume
above the ground.
b. calculate an instantaneous picture of centerline
concentrations.
c. calculate average concentrations at stack level.
d. both a and c
Sigma y and sigma z are also important for estimating
ground-level concentrations because they
a. determine the lateral and vertical spread of the
plume at specific distances downwind.
b. determine the concentrations at stack level.
c. are the horizontal and vertical dispersion parameters.
d. both a and c
6. a. determine the distance of
the centerline of the plume
above the ground.
8. The PTXXX models may be entered into the computer
by two methods. They are
a. interactive and modal.
b. interactive and batch.
c. modal and batch.
d. manual and modal.
7. d. both a and c
The PTXXX models are called screening models
because they
a. estimate ground-level concentrations accurately.
b. are refined, detailed models.
c. make conservative estimates of ground-level
concentrations.
d. both a and b
8. b. interactive and batch.
10. The PTXXX models calculate plume rise
using
a. Briggs' urban sigmas.
b. Briggs' plume rise formulas.
c. Moses and Carson plume rise formulas.
d. Turner's Workbook values.
9. c. make conservative
estimates of ground-
level concentrations.
11.
The three models of the PTXXX series are
PTMIN, PTMAX, PTMPP.
a.
b.
c.
PTPTP, PTMIN, PTDIS.
PTDIS, PTMAX, PTMTP.
10. b. Briggs' plume rise.
formulas.
11. c. PTDIS, PTMAX,
PTMTP.
4-29
-------
Lesson 3
PTPLU
Lesson Goal and Objectives
Goal
To familiarize you with the refinements of wind speed correc-
tions, stack-tip down wash, gradual plume rise, and buoyancy-
induced dispersion found in the PTPLU model.
Objectives
At the end of this lesson, you should be able to:
1. name the type of plume distribution used in formulating
the PTPLU model.
2. list the three technical options that differentiate the
PTPLU model from PTXXX models.
3. identify the model's useful range from the source.
4. describe the effect of each of the three technical options
on the calculations for ground-level concentrations.
5. choose the highest estimated concentration from an
example of PTPLU output.
Introduction
The fourth model discussed is PTPLU. It is an improved ver-
sion of the PTMAX model, discussed in the previous lesson.
Three technical options are included in PTPLU. These options
are also available in other UNAMAP, Version 4 models more
refined than PTPLU. More options indicate that the model is
a more detailed screening tool than the PTXXX models.
Before discussing the options in PTPLU, let's briefly review the
PTMAX model features.
The PTMAX model is a Gaussian (binormally) distributed
plume model for single sources. It is steady state and designed
for flat, rural areas. The dispersion coefficients are the
Pasquill-Gifford sigmas that are applicable only to flat, rural
situations. The concentrations are 1-hour averages. PTMAX
runs in either batch or interactive mode (Figures 4-16 and 4-17).
Figure 4-16. Batch mode.
PTPLU
I WIND DIRECTION
NE
, WIND SPEED?
Figure 4-17. Interactive mode.
4-31
-------
Input Parameters
The inputs required on the PTPLU are the same as PTMAX:
source strength, stack height, stack gas temperature, stack gas
velocity, stack diameter, and ambient temperature.
The first two physical inputs not required in PTMAX, but
required in the PTPLU, are anemometer height and mixing
height. The anemometer height is required to extrapolate the
winds taken at a lower level (usually 10 meters) to the height of
the stack opening (h,). The wind profile extrapolation uses a
power law formula and exponents (p) that are related to the
atmospheric stability category to estimate a representative wind
speed for plume transport and dispersion. The power law states
that wind speed (u) at stack height (z) is a function of wind
speed at a lower level (u.) (Equation 4-1).
(Eq. 4-1)
u = ue(z/10)'
The second input, mixing height (L), is used as shown in
Example 4-2 to realistically calculate ground-level concentra-
tions when the plume is restricted from dispersing vertically by
an elevated inversion (sometimes called a ltd). This condition
results in the plume eventually being well mixed at some
distance downwind of the source. This feature differs from
PTMAX in that the PTMAX model has no allowance for
plume mixing between the ground and an inversion.
Options
The three options of PTPLU are stack-tip down wash, gradual
plume rise, and buoyancy-induced dispersion (Figure 4-18).
These technical options adjust plume rise and dispersion rates.
Weather elements, such as wind speed, are not affected.
Caution: To ensure regulatory consistency, you should check
with the EPA Regional Meteorologist in your region before
using any of these options in a permit analysis.
Stack-tip Downwash
The first option, stack-tip down wash, is considered if the exit
velocity of the effluent is less than one and one-half times the
wind speed estimated at the stack opening. For example, if the
wind speed is 14 meters per second, the gas exit velocity at
the stack top must be less than 21 meters per second for the
effective stack height, H, to be adjusted downward toward the
ground. Stack-tip downwash increases ground-level concentra-
tions by lowering the plume's relationship to the ground.
Example 4-2. Mixing height
under elevated
inversion.
Buoyancy-induced
dispersion
Figure 4-18. PTPLU options.
4-32
-------
Gradual Plume Rise
The second option, gradual plume rise, allows the plume to
slowly continue rising with distance downwind from the source.
The plume will eventually cease rising when the temperature of
the plume is the same as the surrounding air. The point where
the plume becomes constant with height is called final rise.
The PTMAX model uses final rise for all calculations. By using
this option, PTPLU will calculate plume rise that is lower than
PTMAX. This lower plume rise results hi higher ground-level
concentrations. At final rise distance and beyond, both models
would give the same concentrations.
Buoyancy-induced Dispersion
The third option, buoyancy-induced dispersion, considers that
the plume spreads wider and higher than ambient turbulence
alone could initially spread it. This initial spreading takes place
during the tune the plume first comes out of the stack. It is
due to the larger amount of entrainment (mixing) of surround-
ing air into the plume.
Limitations
The limitations of PTPLU are similar to those of PTMAX.
Effects that cannot be simulated include building down wash,
pollutant removal or chemical reactions, multiple sources, and
fumigation. The PTPLU model remains a screening tool,
although it is more detailed than its predecessor, PTMAX.
Output
Figure 4-19 gives an example of the batch version of PTPLU.
Section one is the title. Section two gives input parameters.
Section three gives two calculated values: volumetric flow and
buoyancy flux used for plume rise. Section four contains out-
put for constant wind speed with height on the left and for
extrapolated wind speed with height on the right. For each
wind speed and stability combination, the maximum concen-
tration, the distance to maximum, and effective plume height
are given. Section five prints three caution messages. These
correspond to numbers in parentheses in Section four.
When the distance to the maximum concentration is greater
than 100 kilometers, the output will print consecutive 9's —for
example, 9.999 E+ 09 grams per second.
4-33
-------
Any effective plume height above 200 meters is considered
excessive and will be tagged —for example, a plume height of
824.9 (2). The 2 in parentheses keys a cautionary message that
care should be used when interpreting the computation. The
cost of running PTPLU is the same as that for the PTXXX
models, i.e., less than $1.00.
4-34
-------
i-rri.u (VUISION 11014)
AN IMPROVED POINT SOUItCK SCREENING MOIieL
rtXllllEf) UYi JOE CAT A LAND AND FRANK HALE
AEIIOOOMP, INC. - COSTA MESA, CA FOU THE
ENVIRONMCNTAL OPERATIONS OltANCH, EPA
»>INPUT PAHAMITEH3<«
"ITrLL"* PTPLU EXAMPLE HUN - INPUT UY T. PIERCE 12/11/10
•"OPTIONS*"
IF * I, USE OPTION
IF • 0, IGNORE OPTION
IOPTII) ' 0 (GHAD Pl.IMb RISE)
IOPTI1) • 1 (STACK UOWNWASII)
IOITU) * I (BUOY. INDUCED DISP.)
EKillT*" • 2.00 (M)
•••MKTEOHOLOaY*"
AMblENT AIR TEMPERATURE
MIXING HEIGHT
ANEMOMETER HEIGHT
WIND PROFILE EXPONENTS
171.01 (K)
1300.00 (M)
1.00 (M)
AiO.Ot, BtO.07,
DtO.lS, EiO.33,
CiO.ll
FlO. OS
•••SOUHCE***
IMISS ION RATE
STACK HEIGHT
EXIT TEMP.
EXIT VELOCITY
STACK DIAM.
VOI.UMLTHIC PLOW • ltl.1t (M"1/SEC>
>»CALCULATED PARAMETERS<«
UUOYANCY FLUX PARAMETER •
4ti.ii HTM/SEC*'))
uti.it (a/sec)
101.40 (M)
450.00 (K)
ll.«0 (M/SEC)
i.«« (M)
Section 2
Sections
PTI'1.0 EXAMPLE NUN - INPUT BY T. PIEKCE 11/13/10
STAUI L1TY
STAUILITY
2
1
STAUI 1.ITY
1
1
STAUILITY
4
4
4
STAUILITY
i
S
i
1
i
STAUI LITY
••••MINUS CONSTANT WITH HEIGHT*'**
WIND SPEED MAX CONG OlST OF MAX PLUME HT
(M/SEC) (0/CU M) (KM) (M)
.SI I.OOOOE'OO t.110 3190.5(1)
.11 0. OOOOEtOO 0.000 11)7.1(1)
.00 I.OOOOE'OO 0.000 1141. XI)
.it 3.0137E-04 i.ii4 1233. j(i)
.10 3.1S40E-04 1.551 074.9(1)
.SO 1.I01IE-04 1.114 111.1(2)
.00 3.I72IE-04 I.IS4 711.1(2)
••••MINUS CONSTANT WITH HEIGHT""
WIND SPEED MAX CONC OlST OP MAX PLUME HT
(M/SCC) IO/CU M) (KM) IM)
I. SO 0. OOOOEtOO .400 3200.5(2)
0.10 0. OOOOEtOO .000 2117. 1(2)
1.00 .OOOOEtOO .000 1740.7(1)
I.SO .SSilE-04 .7(4 1211.2(11
2.00 .12UE-04 .175 174.4(2)
1.50 .:OSOE-04 .170 019. t(2)
1.00 .4472E-44 .002 714.1(2)
4.00 .SS7IE-04 .411 SI7.4(2)
J.OO .(IOIE-44 .025 509.0(2)
"" INDS CONSTANT WITH HElarr**"
WINU SI'EliO MAX CONC OlST OF MAX PLUME HT
(M/SEC) (G/CU M) (KM) (M)
.00 .207IE-OS 14.15) 974.0(2)
.SO .I300E-OS II. Oil 111.1(2)
.00 .SII7E-OS .5)1 711.1(2)
.00 . D570H-C4 .7Sf $17.4(2)
.00 .H40E-04 .III SOI. 1(1)
7.00 .ISSIE-04 .411 421.4(2)
10.00 .I1I1E-04 .101 )5S.O(1)
12.00 ".092IE-04 .lii 921.1(2)
li.OO .0)7)E-04 .III 301.1(2)
"••WINIW inNSTANT WITH IHilGIIT*"*
WINU SPELI) MAX CONC DIST OP MAX PLUME HT
(M/SEC) «,YCU M) 00 .000
.01 .OOOOEtOO .000
.21 .1041E-04 .001
.00 .34I1E-04 .(11
.S3 .3100E-04 .114
.11 .4100E-04 .IS1
.71 .S400E-04 .SIS
.01 .II20E-04 .000
.32 .I20IE-04 .«I4
""STACK TOP WINDS (EXTRAPOLATED FROM
WIND SPEED MAX CONC DIST OF MAX
(M/SEC) (Q/CU M) (KM)
.11 .1IS3E-05 10.010
.SO .OI2IE-04 O.SI2
.10 .0110E-04 7.S13
.St .I340E-04 0.3S1
.01 .1S15E-04 S.SOI
.11 .I349E-04 4.040
13.01 .OS1IE-04 3.110
11.11 .00416-04 1.111
21.11 .3311E-03 1.431
"••STACK TOP WINDS ( EXTRAPOLATED PROM 7
WIND SPEED MAX CONC DIST OF MAX
(M/SEC) «I/CU Ml (KM)
0.13 .ooootin o.ooo
1.31 .OOOOetOO 190.910(3)
l.li .91t4gt|4 010.010(3)
2.40 .HUE-IS 12.111
3.11 .OIOIE-IS 13. SOI
4.11 .4I20E-OS 41.110
4.00 .1I13E-IS 11.101
l.ll .1II7E-46 14.401
1.21 .1II2E-I1 21.221
II. SI .4I4IE-IS 20. SO!
16.11 .4SI1E-OS 11.441
10.04 .1044E-OS 14.710
24.00 1. HUE-OS 11.201
13.01 2.0SIIC-IO 11.191
••••STACK TOP WINDS (EXTRAPOLATED FROM I
WIND SPEED MAX COHC DIST OF MAX
(M/SEC) (G/CU M) (KM)
1.47 1.1232E-OS S4.0IO
1.01 1.UI7E-OS SO. 490
9.10 1.11I7E-OS 47.111
11.01 1.S021E-OS 42.041
11.11 I.3I11E-OS 40.000
""STACK TOP WINDS (EXTRAPOLATED FROM 1
WIND SPEED MAX CONC DIST OF MAX
(M/SEC) (a/CU M) (KM)
12.04 .OIIOEtOO 090.001(3)
IS. 00 .OOOOEtOO 909.000(1)
11.91 .0090EtOO 009.990(1)
2S.20 .OOtOEtOO 099.999(3)
31.00 .OOOOEtOO 000.000(3)
1.1 METERS)"" Section 4
pLWp Iff CIW.ITOB 1
(M)
1451.1(1)
1731.0(1)
1425.1(1)
1017.1(1)
111.1(2)
100.2(1)
100. 5(1)
7.0 METERS)""
PLUME HT
(M)
2ISI. 2(1)
1112.0(1)
HIS. 1(2)
1011.1(2)
111.1(2)
100.2(1)
101.5(2)
501.4(1)
445.1(1)
1.1 METE1IS)""
FLUME HT
(M)
754.2(1)
041.1(1)
500.4(1)
477.1(1)
411.7(1)
351.1(1)
310.1(1)
200.1(1)
ISO. 4(1)
.0 Mirrens)""
PLUME HT
(Ml
1014.0(1)
1111.0(1)
1111.1(1)
114.0(1)
- III. 1(1)
514.0(1)
511.4(1)
434.3(2)
117.5(2)
311.0(1)
100.0(2)
213.2(2)
155.5(3)
111.0(2)
.1 METERS )••••
PLUME HT
(M)
111.1(1)
111.0(1)
100.4(1)
100.0(1)
211.1(1)
.1 METERS)""
PLUME HT
(M)
210.0(1)
171.7(3)
110.1(1)
157.0(2)
'50.1(3)
II) THE IIISTANCK m Tllli POINT OK MAX IL1M murt'NTn «Tinu IS sn nnBAT T»iT •nie a .MO el-inn ITV ic u/vi. , , ,/t,. „
TO
Pens IST uwa ENOUGH KOR THE PI.UME TO TRAVEL THIS FAB.
Section 5
(2) 11IE PLUME IS CAIXXJI^TED TO BE AT A HEIGHT WHERE CARE SHOULD OB USED IN INTERPRETING THE COMPUTATION.
(1) NO COMPUTATION WAS ATTEMPTED FOR THIS HEIGHT AS THE POINT OF MAXIMUM CONCENTRATION IS GREATER THAN 100 KILOMETERS
FIIOM THE SOURCE.
figure 4-19. Batch run of PTPLU model.
4-35
-------
Review Exercise
1. The PTPLU model is a
plume type model.
2. Which of these conditions are represented by the PTPLU
model? (More than one answer may apply.)
a. complex terrain
b. building downwash
c. wind speed profile with height
d. plume rise
e. multiple emission sources
1. Gaussian
3. The three technical options available in the model are
, . and
2. c. wind speed profile
with height
d. plume rise
4. True or False? The useful distance for calculating down-
wind concentrations with the PTPLU model is 5 kilometers.
3. • stack-tip downwash
• gradual plume rise
• buoyancy-induced dispersion
5. In the previous example of PTPLU model output,
choose the highest concentration for stability class B(2) for
both winds constant with height and stack top winds.
4. False
6. The PTPLU is called
a. a refined model.
b. an unproved version of the PTMAX model.
c. a complex terrain model.
d. a photochemical model.
5. Constant winds:
1.61XlO-*g/ms
Extrapolated winds:
1.79xlO-*g/ms
The PTPLU model has three technical options available.
This means that the model is a
a. more detailed screening tool than the PTMAX model.
b. carbon copy of the VALLEY model.
c. complex terrain model.
d. refined model, very similar to CRSTER model.
6. d. an improved version of
the PTMAX model.
8. The PTPLU model requires two physical inputs not called
for in the PTXXX models. These inputs are
a. stack coordinates and distance to property line.
b. anemometer height and mixing height.
c. anemometer height and stack coordinates.
d. thermometer height and anemometer height.
7. a. more detailed screening
tool than the PTMAX
model.
8. b. anemometer height and
mixing height.
4-36
-------
Lesson 4
CRSTER
Lesson Goal and Objectives
Goal
To familiarize you with the CRSTER model—its treatment of
stacks, receptors, and terrain features.
Objectives
At the end of this lesson, you should be able to:
1. state the regulatory use of the CRSTER model.
2. describe the method of terrain adjustment in the model.
3. name one limitation of the model.
4. describe the adjustment to the dispersion curves that are
used in CRSTER to simulate urban conditions.
5. name the computer program that ensures the
meteorological data is in proper format for use in the
CRSTER model.
6. give the total number of receptors that can be treated by
CRSTER and describe their arrangement.
Introduction
The fifth model discussed in this course is the single-source
(CRSTER) model. It is called the single-source model because
it simulates up to 19 different point sources, but assumes they
are collocated at a single plant site. Thus, the model is well
suited for a single industrial facility with several different emis-
sion sources. CRSTER is a refined model in the UNAMAP
package. It requires a large amount of meteorological input
data and has an extensive receptor network. As a result,
CRSTER is a larger computer program than the screening
models discussed previously, and, like all refined models, it
runs in a batch mode only. CRSTER is EPA's benchmark
model for rural areas. This means that any other model
applied to rural areas should use the same dispersion equations
and assumptions as those found in CRSTER.
4-37
-------
Model Assumptions
CRSTER is based on the binomial Gaussian plume equation
(Equation 4-2) and uses Briggs' plume rise algorithm.
Where: (x,y) = receptor coordinates
X = ground-level concentration
Q= emission rate
H = effective stack height
u = mean wind speed
ay,aI = dispersion coefficients
(m)
(g/ms)
(g/s)
(m)
(m/s)
(m)
The Gaussian plume model is modified for limited mixing
heights and incorporates the Pasquill-Gifford dispersion coeffi-
cients. CRSTER is a steady-state model and assumes the source
and meteorological conditions are invariant for the basic time
period of 1 hour. Pollutant concentrations for longer periods
such as 3 hours, 24 hours, and 1 year are produced by
averaging together many 1-hour values. The pollutant is
assumed to be stable with no chemical reactions or deposition
allowed. The wind speed used in the calculations is that at
stack top and is estimated from a ground-based (10 meter)
wind measurement using the power law wind profile*
(Equation 4-3).
(Eq. 4-3)
u = u0(h,/10)"
Follow Example 4-3 to compute the power law formula.
Remember wind speed (u) at stack height (h,) is a function of
wind speed at a lower level.
The adjustment for elevated terrain in CRSTER is a new
attribute not found in any of the screening models. CRSTER
can simulate the effects of "simple terrain"; that is, ground
elevations that are no higher than the top of the lowest stack.
CRSTER uses the "full height" correction illustrated in
Figure 4-20. All receptors are assumed to be ground based.
CRSTER takes account of the rise in terrain by decreasing the
effective stack height (H) used in the Gaussian plume equation.
This is an admittedly simple idea. In reality, terrain can cause
significant changes to wind fields and alter the characteristics
of turbulence and dispersion.
f
Wind speed (m/s)
Given: Class A stability,
£ = 0.1
u,, = 4 m/s
Calculate: u = 5.2 m/s
Example 4-3. Power law wind profile.
'Note: z is used in the power law formula for PTPLU = h, for CRSTER.
4-S8
-------
Figure 4-20. Terrain correction in the CRSTER model.
Input Parameters
The CRSTER model requires three basic types of input data:
source, receptor site, and meteorological. As mentioned, the
model has simulated up to 19 separate point sources which are
colocated at a single plant site. The receptor network in
CRSTER consists of 180 points arranged on five rings that
surround the plant site (Figure 4-21). The receptor points are
fixed at 10° azimuth spacing on the rings, but the ring
distances from the plant are not fixed and must be specified by
the user. The PTPLU screening model is often used to select
rings corresponding to the downwind distances where max-
imum concentrations are expected to occur.
t
Hourly measured
wind direction
Figure 4-21. Receptor rings and
tune-averaged plume for Class B
condition* using the CRSTER
model.
4-39
-------
Meteorological data used in the CRSTER model are actual
hourly observations from a National Weather Service station
for periods ranging from one to five years. These data must
first be prepared by the EPA Meteorological Preprocessor, a
separate computer program. As shown in Figure 4-22, the
preprocessor is driven by a tape of hourly surface
meteorological data (wind speed, wind direction, temperature,
cloud cover, and ceiling height) and card images of twice-daily
mixing heights. The program calculates hourly values of
stability class using Turner's method, interpolates mixing
heights to hourly values, randomizes the wind direction to 1°
increments, reformats other data, and writes all of the
parameters out on a magnetic tape. This output is called a
Preprocessed Meteorological Data Tape and is read directly by
CRSTER and most other EPA refined models. When calm
periods are sensed by the preprocessor, the program uses the
wind direction from the most recent non-calm hour.
Twice
daily
mixing
height
Hourly
surface
^meteorological J
data
Preprocessor
L
Stability wind,
temperature, and
mixing height
by hour
LI
(Meteorological |
data to
model
Figure 4-22. Schematic of meteorological
data preprocessor.
4-40
-------
The tape contains two sets of hourly mixing heights, one for
a rural environment and one for an urban environment.
Although designed principally for rural areas, CRSTER does
have an urban option. When activated, the urban mode uses
different mixing height values and modifies the selection of
stability class. The principal difference between dispersion in
rural and urban environments concerns the occurrence of
ground-based temperature inversions in rural areas on calm,
clear nights. Pasquill-Gifford stability Classes E and F are
associated with these stable conditions. In urban areas, the
heat island effect and numerous roughness elements (e.g., tall
buildings) increase turbulence and preclude the occurrence of
Class E and F conditions. Thus, CRSTER modifies the stability
data for a given hour by converting any Class E or F condition
to Class D (neutral condition).
Limitations
CRSTER is a refined air quality model with several limitations.
Since it is based on the 1-hour steady-state assumption, it is less
valid when emissions or meteorological conditions are changing
rapidly. It is incapable of treating complex terrain where the
ground elevation is higher than stack top, nor can it treat
building downwash, pollutant deposition, or chemical
transformation.
Output
The CRSTER model provides several different types of output
information. Summary tables of highest and second-highest
concentrations are generated for averaging times of 1 hour, 3
hours, 24 hours, and 1 year. Additional time periods can be
specified. In addition, an output magnetic tape can be written
(optional) that archives the full record of dispersion calcula-
tions made in the model run. Figure 4-23 is an example of a
summary table run by CRSTER, in this case the set of second-
highest 3-hour SOj concentrations at all 180 receptors. The
highest, second-highest SOZ concentration is highlighted at the
top of the page: 6.91 X 10~4 g/ms or 691 /*g/m3. Concentrations
for each receptor are given in scientific notation and arranged
in columns for each ring distance. The numbers in parentheses
give the Julian day of the year on which the concentration
occurred and the time period within that day. For example,
(33,4) means Julian day 33 (February 2) and the fourth 3-hour
period within that day (0900 to 1200 local standard time).
CRSTER costs in the range of $50 per year of meteorological
data processed.
4-41
-------
PLANT NAME: EXAMPLE RUN POLLUTANT: soz EMISSION UNITS: OH/SEC AIR OUALITY UNITS: en/two
YEARLY SECOND MAXIMUM 3-HOUR CONC= 6.91*0-04 DIRECTION: ID DISTANCED 3.8 KM OAV=130 TIME PERIOO= 0
RANGE
DIR
1
2
3
4
5
6
7
8
9
10
It
12
13
1*
IS
16
17
18
19
20
21
22
23
21
25
26
27
2B
29
30
31
32
33
31
35
36
2
*
1
4
4
5
4
1
2
5
6
9
6
2
•
3
a
i
3
5
1
7
3
2
7
4
2
4
7
3
1
3
1
1
5
7
SECOND
.9 KM
.8072-05
.6599-03
.1004-04
.7235-05
.6511-05
.9782-05
.8120-05
.0025-05
.3717i06
.6159-06
.1385-06
.93*7-06
.9326-05
.5316-04
.6508-0*
.76*0-0*
.6169-05
.0033-05
.2*93-06
.8065-06
.0076-05
.7991-05
.0518-0*
.12*9-0*
.9995-05
.3966-03
.20«*-OS
.30*8-06
.1851-07
.0534-06
.1108-03
.2*6* -OS
.9298-0*
.8818-0*
.3069-03
.18*1-06
(126.
(126.
1126.
I126>
(1*5.
1111.
<1«3>
(1«2>
(131.
1130.
1131.
(1*2.
11*2.
C142.
11*2.
11*2.
11*2.
ll*2>
(111.
(111.
11*3.
(1*3.
(1*3.
11*6.
11*2.
(1*2.
(1*6.
11*6.
(1*3.
(132.
(132.
(123.
(125.
(125.
1125 1
(1*3.
HI8HEST
*l
41
5)
3)
3)
«)
«l
51
*)
*)
51
3)
SI
SI
SI
*)
41
*l
SI
51
31
31
51
51
41
*l
SI
*l
SI
31
31
31
31
*)
*)
31
1
3
3
6
6
<
3
6
8
1
2
1
1
3
5
2
3
1
5
2
3
3
1
3
3
3
1
2
9
2
1
2
2
1
8
1
3-HOUR CONCENTI
1.3 KM
.3951-0*
.68*2-0*
.3769-0*
.86*8-0*
.*900-0*
.9982-0*
.0616-0*
.7964-05
.2532-03
.3923-0*
.6089-0*
.7920-0*
.1837-0*
.1002-0*
.7291-0*
.6*79-0*
.69*6-05
.3753-03
.5262-05
.5266-0*
.1060-0*
.5513-0*
.9727-0*
.1*06-0*
.6231-01
.5269-0*
.7181-0*
.8258-05
.** 12-06
.2875-05
,»19S-OS
.1*33-05
.1565-0*
.0298-0*
.8*11-05
.0722-0*
(125.
1123.
(123.
(1*2.
(1*5.
11*2.
(1*0.
(1*0.
(131.
(131.
(1*8.
11*8.
11*2.
(1*8.
(1*2.
(112.
(1*2.
(1*1.
(141.
11*1.
11*1.
11*1.
11*3.
(1*2.
(1*6.
(1*6.
(116.
(1*6.
(132.
(IV
(1J2.
(125.
(125.
(125.
1113.
11*3.
SI
• 1
11
51
51
SI
31
11
41
SI
SI
• 1
SI
SI
*l
11
41
1)
41
51
61
51
41
41
61
61
61
61
SI
61
31
SI
61
2.4295-04
2.9218-04
2.7464-04
2.7035-04
2.5587-04
2.0591-04
5. 9<>86-04
3.0023-04
3.0834-04
5.4265-04
2.7182-04
2.5399-04
2.2799-04
2.4299-04
2.7078-04
7.9046-05
3.8366-05
2.7073-04
1.9385-04
3.6800-04
2.7620-04
2.0773-04
1.0913-04
1.4661-04
1.3181-04
3.3562-04
1.4742-0*
1.6761-0*
2.0351-01
2.5798-0*
1.9661-0*
3.1283-0*
2.57*9-0*
2.9WJ5-0*
1.3813-0*
2.0660-0*
(113.
(126.
(126.
(126.
(1*4.
1133.
(13*.
(130.
(1*5.
11*1.
(1*3.
(130.
(1*Z>
(118.
(1*8.
11*8.
(1*2.
(1*6.
11*1.
11*1.
lltli
(1*1.
(111.
11*2.
(116.
11*6.
(132.
(1*2.
(1*6.
(132.
1132.
(132.
(1*3.
1123.
(1*3.
1126.
61
51
51
61
61
51
61
31
71
11
61
31
61
SI
41
II
SI
11
»)
11
SI
31
31
31
31
*l
51
61
61
31
SI
41
61
61
31
61
Figure 4-23. Sample CRSTER output.
Review Exercise
1. True or False? The CRSTER is only a screening model.
The CRSTER model is called a(n)
U.S. EPA for rural areas.
a. benchmark
b. numerical
c. complex terrain
d. industrial source
e. useless
model by
1. False
3. One limitation of the model is that it will not:
a. handle more than one collocated stack.
b. treat urban situations.
c. treat inert pollutants like SOj.
d. treat building downwash.
e. calculate 3-hour averages.
2. a. benchmark
3. d. treat building downwash.
4-42
-------
4. The CRSTER model adjusts the P-G dispersion curves
for the urban situation by
a. using Draxler's dispersion coefficients.
b. using Briggs' urban dispersion coefficients.
c. using Tennessee Valley Authority dispersion
coefficients.
d. using the P-G neutral category Class D for nighttime.
e. using Brookhaven National Laboratory dispersion
curves.
5. True or False? The CRSTER model uses the
Meteorological Preprocessor program to properly prepare
the meteorological data.
4. d. using the P-G neutral
category Class D for
nighttime.
The CRSTER Preprocessor program checks the meteoro-
logical data for calm winds. When a calm is found in the
data, the wind direction used in place of the calm is
a. randomly selected.
b. substituted by the user.
c. taken from the most recent non-calm hour.
d. always north.
e. left blank.
5. True
7.
In Figure 4-23 of CRSTER model output, the highest,
second-highest 3-hour SO, concentration is
and its distance from the source is
6. c. taken from the most
recent non-calm hour.
The CRSTER model is termed
a. the complex terrain model.
b. the Briggs' urban model.
c. the single-source model.
d. the multisource model.
7. 691 ng/m*. 3.8 km
8. c. the single-source
model.
4-43
-------
9. The CRSTER model will handle up to 19 different
sized stacks. The model
a. discards all but the largest stack.
b. collocates the separate stacks at one plant site.
c. increases all stack heights to the height of the tallest
stack.
d. arithmetically averages all stack heights into one
"average" stack.
10. The CRSTER model will handle uneven terrain. In the
event of terrain no higher than stack top, the plume is
adjusted by
a. passing plume centerline above all terrain heights.
b. increasing stack gas temperature.
c. decreasing the effective stack height.
d. decreasing wind speed.
9. b. collocates the separate
stacks at one plant site.
11. How many receptor rings and how many receptors are
available in the CRSTER model?
a. 5 receptor rings and 180 receptors
b. 10 receptor rings and 360 receptors
c. 8 receptor rings and 240 receptors
d. 3 receptor rings and 180 receptors
10. c. decreasing the effective
stack height.
11. a. 5 receptor rings and 180
receptors
4-44
-------
Lesson 5
RAM
Lesson Goal and Objectives
Goal
To familiarize you with the RAM model and the options that
make it useful.
Objectives
Upon completing this lesson, you should be able to:
1. state the plume distribution of the RAM model.
2. describe the terrain correction hi RAM.
3. state the current recommended regulatory use of the
RAM model.
4. identify the dispersion curves that RAM uses to simulate
dispersion in urban areas.
5. identify the number of stacks that can be analyzed by
RAM.
6. describe the method used by RAM to calculate the wind
speed at the stack top.
Introduction
The sixth model we will discuss is the RAM model. It is also
called the urban multisource model. It was originally designed
for both rural and urban applications. Now, however, for
regulatory applications, it is recommended for urban situations
only. A newer model, called MPTER, will be used to treat rural
multisource situations and will be discussed in the next lesson.
Plume Characteristics
The RAM urban model is a Gaussian plume model. The model
assumes that the source and meteorological conditions are
steady state. The basic time period for calculations is one hour.
It will estimate concentrations for averaging tunes from an
hour to a day. The RAM model has a unique feature among
those on EPA's UNAMAP package—it is based on dispersion
coefficients different from those of Pasquill-Gifford. The RAM
4-45
-------
dispersion curves, shown in Figure 4-24, were derived by Briggs
from tracer experiments conducted in St. Louis by McElroy
and Pooler. These are dispersion curves representative of an
urban area. Figure 4-24 compares the RAM urban curves to
the rural curves of Pasquill-Gifford. It can be seen that, in
general, the urban curves represent a greater rate of dispersion
caused by the increased turbulence found hi urban areas. The
RAM model uses Briggs' equations for plume rise. There is no
terrain correction in RAM; the surrounding area is assumed to
be perfectly flat.
10'1 10° 10l
Distance downwind (km)
10° 10'
Distance downwind (km)
10*
• Pasquill-Gifford
- McElroy- Pooler
Figure 4-24. RAM dispersion curves.
Input Parameters
As input, the model requires emission information and hourly
meteorological data. The same preprocessed meteorological
data tape described for CRSTER is used as input to the RAM
model. Meteorological information includes hourly values of
wind direction and speed, temperature, stability class, and
mixing height. Emission information consists of emission rate,
physical stack height, stack diameter, stack-gas exit velocity,
stack-gas temperature, and stack coordinates. The information
required is the same as for other models already discussed,
except for the stack coordinates. RAM is a true multiple source
model and the coordinates are necessary to calculate the
4-46
-------
geometry between arbitrary source and receptor locations. The
model also calculates a representative wind speed at stack
height using the power law formula described in the previous
lesson.
The model will allow the user to input a total of 250
stacks—known as point sources —and 100 area sources. Each
stack may be of a different height and a total of three area
source release heights may be selected. This selection of release
heights allows the model to represent several different types of
area sources for an urban area.
Limitations
The model does not treat any aspects of terrain, chemical
transformations, fumigation, building downwash, or multiple
pollutants.
Options
As in the PTPLU model, the effects of stack-tip downwash,
gradual plume rise, and buoyancy-induced dispersion can
be selected.
Use and Output
RAM is the recommended model for urban areas. The RAM
model is recommended for refined analyses only. The user
should exercise caution in running the RAM model. Because of
internal program calculations, the number of point sources,
area sources, receptors, and number of days of analysis should
be kept at an absolute minimum. The model, depending on
the specific application, can be very expensive to run.
Typically, costs will range from 0.10 to 0.20 of one cent per
source-receptor-day. A typical modeling run might involve 100
sources, 180 receptors, and 365 days (one year) of meteorology
for a total cost of over $6,000.
Figures 4-25 and 4-26 are examples of RAM runs involving
typical site situations. Note that, unlike the CRSTER model,
RAM output does not use scientific notation for concentration
values.
4-47
-------
RUN BtS Eft KICKSHAW, All I HAZARDOUS BATE*. DIV., RECION XV.EPAd JAN 78)
EMISSIONS: TEST CITT, 1973
src MET. DATA: TEST CITT 1973 ; UPPER AIR: TEST CITT 1973
INPUT MET DATA
HOUR TNETA
(•EC)
33.00
23.00
73 /
SPEEB
CH/S)
6.17
4.63
M1XIN6 TEMP
HEI6MT(IO (PEC-K)
429.11
401.70
RESULTANT MET CONDITIONS
VINO DIRECTION- 28.71
AVERAGE NINO SPEED" 5.40
UIND PERSISTENCE* .996
269.82
271.48
STABILITY
CLASS
4
4
RESULTANT UIND SPEED*
AVERACE TEMP" 270.65
MORAL STABILITY- 4
5.38
SICNIF1CANT POINT RECEPTORS
RECEPTOR • EAST
3
4
5
6
7
8
9
10
11
7
7
5
5
8
8
9
9
11
12 j 11
564.43
564.16
579.45
579.40
577.38
577.30
576.67
576.59
582.94
582.89
NORTH
4407.01
4406.52
4403.16
4403.07
4401.21
4401.08
4400.55
4400.40
4400.80
4400.70
PREDICTED MAX CONC.
39.39
839.47
448.58
619.39
427.63
MAX. D1ST
(KM)
.902
1.804
.166
.331
.249
.499
.276
.551
.187
,374
Iff. NT
(N)
156.385
156.385
32.007
32.007
47.506
47.506
52.296
52.296
35.952
35.952
U
-------
DUN 61: 10 ««1»
CRISSItkS: TEST CITT, 1973
SfC HIT. DATA: IfST CIH 197! ; U'PfS •!«: TeSI CUT 1973
SUOPIOI CONCCNTIAT10N TABLE <«I CIOCXHS /H««3 ) 7
i : Maun i
MGUS TKtTA
(Bit)
1 3J.OO
•ECEPTOI kO.
SPf£» MIINt TEHP
MEI6HKH) (tl
(.17 429.11 iS9.t?
EAST NOR1N TOTAL
SISNIf
STABILITY
CLASS
4
ml A MIS: U.i K. t '9.;
MOH TOTAL 'to* TOTAL MOD TOTAL FIOH
POINT ALL POINT SItNlf AI€A ALL AtEA
SOUICES
1 1
2 I
3 f
1 P
5 P
6 P
7 P
e P
9 p
10 P
11 r
1? P
13 A
14 *
IS A
u «
17 »
18 A
19 A
20 A
21 A
22 *
23 H
24 H
25 H
26 H
27 H
2« H
2V H
30 H
31 M
32 H
33 «
3* H
35 h
36 h
37 h
3* h
3« H
tO H
<1 K
r
0
7
7
?
5
t
I
V
9
11
11
i
»
5
*
: 2
10
t
7
13
1?
0
C
c
0
0
C1
c
0
c
0
c
0
c
c
r
0
c
p
c
566.00
564.00
5*4.43
564.16
57V.45
sr«.40
577. 3«
577.10
576.67
576.59
5t2.«4
5e^.»9
578.42
576.43
57t.43
57B.43
574.43
5bC.41
574.43
57H.J7
5«2.4t
5«0.<]
572.00
574.00
5(0.00
571.00
573.00
575.00
577.00
572.00
574. 00
576.00
57C.OO
571. PO
573.00
577.00
572.00
574.00
576.00
578.00
5i-f.no
4405.00
4401.50
4407.61
4406.52
4403.10
4403.07
4401.21
4401. Ul!
4400.55
4400.40
4400. (0
4400.70
43««.«4
439*. «5
4401. V6
4405.95
4399. Vt
440i.V2
4405. tt
4403.94
4403.92
4403. «
4400.*?
4400. t7
440C.S7
4402.60
4402.60
4402. »C<
4402.60
4434.33
4404.33
4404.33
4404.33
4406.06
4406.06
4406.06
4407.79
4407.79
4407.79
4407.79
4407. 7i
35
It
723
161
431
204
710
291
433
194
49
7
.0^00
.0300
.7917
.2026
.7571
.0«7
.7*21
.7)43
.0651
.1613
.3493
.8263
.0*37
.1623
.9795
.0000
.oroo
.0000
.ocoo
.0000
.0000
.ocoo
.cooo
.ocoo
.0000
.0000
.0000
.0000
.0000
.0000
.oroo
.0000
.3000
.oroo
.0000
.ocoo
.0000
.ocoo
.0000
. OPOO
.nroo
souncf J
35
11
723
u«
432
205
71Z
293
433
194
J1
7
13
26
a
19
9
29
r
10
4!
12
15
.0000
.0000
.79*7
.2026
.7571
.04*7
.2024
.1913
.6123
.7612
.3493
.(263
.0437
.6786
.9(03
.7536
.!3»9
.0000
.5625
.0000
.0000
.0000
.2047
.6048
.0000
.0214
.45«1
.4123
.51CO
.00*S
.2110
.96*2
.54(2
.8200
.0001
.9563
.0000
.0000
.9420
.2102
.no«p
SOUICES
.0000
.0000
.0000
.0000
1.4215
1.446S
2.7241
2.8280
2.9602
1.0427
.0000
.0000
3.2543
i.om
1.7745
1.1665
1.6338
.8464
1.0529
.4950
.5493
.5444
.1834
1.2702
.227?
.3706
.2353
.1610
.3822
.5696
.2755
.1248
.4t?1
.3781
.4319
.095*
.1342
.4364
.0000
.4971
.?««5
SOUtCES
.0000
.0000
.0000
.0000
1.4667
1.4929
2.72(1
2.8210
2.9602
3.0427
.0483
.0449
3.4000
3.0««1
1.800*
1.16U5
1.633*
.8464
1.052*
.5171
.6120
.4414
.3421
1.2702
.341*
.48*0
.3536
.1610
.3822
.56*6
.2753
.12(8
.41?1
.3788
.431*
.0»S«
.1342
.4364
.0000
.4971
.2645
j£pAt»Tie» uric t;.t u.
TOTAL mo* CONCEIITIATIOII
ALL IOIMKES «AN»
.0000
.0000
35.79*7
18.2026
725.2238
369.5415
434.9305
20*. 01»J
715.6425
296.4038
433.3*75
1«4.«708
3.4837
54.7674
».7811
1.V200
IS. 4727
.1464
1.6154
.5121
.6120
.6414
26.5468
9.8749
.348*
.5104
1V.80S7
9.5734
29.9002
.5724
7.4*33
11.0931
45.9603
1.19*8
.4320
13.0522
'.1342
.4364
.9420
15.7073
.2641
41
to
11
IS
1
J
3
7
2
6
4
1
24
«
21
2S
17
2*
2*
33
31
30
13
20
37
34
14
22
12
32
23
1*
10
11
36
1*
3*
35
2*
16
31
Figure 4-26. RAM model.
4-49
-------
Review Exercise
1.
2.
3.
4.
5.
6.
7.
8.
9.
The RAM mndel is a ... plume mndel
True or False? The RAM model can adjust for simple
terrain.
The RAM model is currently rernmmenrieH for _ ._
areas only.
The RAM model uses
a. Briggs' recommended urban dispersion curves.
b. surface wind to represent stack-top wind.
c. the preprocessor program to check source data.
d. Holland's plume rise methods.
e. the method of collocated stacks for point sources.
True or False? The RAM model uses a linear formula of
the form Y = a + bx to adjust the wind speed to stack
height.
The RAM model will treat
a. building downwash.
b. fumigation.
c. complex terrain.
d. volume sources.
e. stack-tip downwash.
True or False? The RAM model is termed a multisource
model.
The RAM model requires (x,y) coordinates for each
stack to properly account for the source-receptor geometry.
RAM can handle up to ., point so«rces-
1. Gaussian
2. False
3. urban
4. a. Briggs' recommended
urban dispersion curves.
5. False
6. e. stack-tip downwash.
7. True
8. True
9. 250
4-50
-------
Lesson 6
MPTER
Lesson Goal and Objectives
Goal
To familiarize you with MPTER and its method of treating
terrain.
Objectives
Upon completing this lesson, you should be able to:
1. identify MPTER's plume characteristics.
2. list the three technical options available in MPTER.
3. describe MPTER's terrain adjustment method.
4. identify the limitations of MPTER.
5. describe the MPTER method of adjusting wind speed.
6. describe the MPTER method of plume rise.
Introduction
The seventh model to be discussed is the MPTER model.
MPTER stands for Multiple Point Source Terrain, and this
model is EPA's basic multiple point source model for rural
areas. The model is very similar to the RAM model discussed
in the previous lesson, except that the MPTER model treats
point sources in uneven, rural terrain, whereas the RAM model
is designed for urban areas with flat terrain.
Plume Characteristics
The MPTER model is a Gaussian plume model. Conse-
quently, the model assumes that source and meteorological
conditions are steady state. The basic time period for cal-
culations is one hour. It will estimate concentrations for
averaging times from one hour to one year. The dispersion
coefficients are the Pasquill-Gifford rural values found in
Turner's Workbook of Atmospheric Dispersion Estimates and
illustrated in Figure 4-27.
10*
101
101
(a)
10'1 10° 101 10*
Distance downwind (km)
10s
10*
10'
10°
10'1 10° 101 10*
Distance downwind (km)
Figure 4-27. MPTER dispersion coefficient
values.
4-51
-------
Input Parameters
The MPTER model calculates effective plume rise by Briggs'
method. As shown in Figure 4-28, this model offers three
plume and dispersion options. These options are stack-tip
down wash, gradual plume rise, and buoyancy-induced disper-
sion. Buoyancy-induced dispersion is a concept proposed by
Pasquill in which the rate of vertical dispersion is increased
from the normal Pasquill-Gifford value to account for the tur-
bulent entrainment of air during plume rise (Equation 4-4).
(Eq. 4-4)
-------
Use and Output
EPA recommends MPTER for use in rural, multiple point
source situations. MPTER is a refined analysis model with
approximately the same execution costs as RAM. RAM should
be run with the same care that was used to run MPTER. A
sample of MPTER output is given in Figure 4-30. This model
produces summary tables of the five highest 1-, 3-, 8-, and
24-hour concentrations at each receptor. Figure 4-30 gives the
five highest 3-hour SOi concentrations at each receptor in a
one-year period. The values shown are given in units of /ig/ms.
The highest concentration in each column is indicated with a
star to the left of it.
Maximum
receptor height
Figure 4-29. MPTER receptor height.
i Hi* MIT- rwrtiimi HUM itmini-01 jutm pn, HDUIH
„
11
51
71
91
111
111
191
171
211
7!<
291
9A1
271
29(
101
111
121
111
191
171
181
191
• 01
«J<
• 41
• 91
901
921
*I4,
•11,
.90,
.45.
.49,
.45,
'*"'
.14.
- .14.
- .49.
- .5ft.
-.49.
-.7ft.
- J5.
- _fl4.
-.•9.
- .78.
- .A 9.
-.18.
-.45.
- .11.
- .16.
.10.
.74.
.11.
.75.
.94.
1 .15.
1 .10.
1-41.
-1-411
1 .50.
1 -48*
1 -10.
1.15.
.44.
.79.
.191
• HI
.711
111)
.00)
-— rn
-.in
---.m
• .41)
-.711
-.19)
-.1*1
-.711
-.»«>
-.91)
- .•*!
-.11)
.on)
.11)
.441
.911
.64)
.71)
.85)
.44)
.40)
1 .101
1.151
.96)
.75)
.511
J6J.
.00)
-.26)
- .75)
-.941
-1.151
"^.loT
-1 .411
414 .1 9
tU.lf
411.17
491. M
111 5. HI "' "
I 70 1. 12 1
1179. ts >
1204, 121
1 14 4, 71 1
1207. 11-1
1701,171
I 11 3* 19 1
( 70 6. 14 I
1 >79, 121
1110,171
1 14 4. 11 1
• BS.12I
(701.12)
I 70 9, 15 1
1 19 9. H 1
49| .01 I209.P )
543.11 1207.12)
4 37 .9 7
I10J.41
517.47
530.44
250^1*
J29.21
721 .45
8 17 II 5
-Trrrri-
• 08.76
1 19 7, 11 1
1 141, 121
1 14 1, 12 1
1114, fl
J!*1«J*!
1 15 4. t< 1
174 7. l«)
< If 5. 17 I
in J.11)
t l< 6, f7 )
1203, I' I
517. tj 1199,19)
H9.71 I»0».1SI
179.00 MIS, 111
117.54 1174.121
TTT77* in27T4^-
119.46 1247,171
'7.02 1710.15)
•75.94 11(9,121
314.T7 m3.!Jl
441.91 1147*151
HI, OJ IH/, 111
11.14 HAl.HI
JO. 82 1141. 151
1 1. 95 1) 46 . 1 1 1
212.14 (228.111
4T4.71 1771.151
4«t.5l 1144.191
521.31 tJll.l})
J95.17 007.171
472.^ ILTT»IM
544.14 1204.141
774.32 170* *15L
<4.4? 1204.13)
145.28 II8U.I2I
3KI. IT 1149.12)
717.34 1214.13)
5)4.4* 1175. 121
301.44 1178.151
•91.17 1191.171
411.21 1191.171
3*1.04 H26.I2I
414.14 1125.121
*94.40 1197.141
521.94 1710.121
442.59 1174.121
T22.13 1134,121
7*1.05 (156,121
230.20 005,121
8*4.57 1709.121
7*4. H 170* •!*»
4*7.94 1 91,111
56 T. 11 11 4* ,121
241.44 1741,121
501.*4
111 l»t".57
81.84
14'ill
44.11
I»TOI
~ 43. 11
74.95
42.42
!6»7*iJ"
60J.lt
99 1. IV
119.07
U-2*
21.1*
• 0.0*
151.71
34 »>.t4
•74.7*
9JUJ8
151.44
124.47
111.90
231.9]
91.1?
111,14
304.1*
12*. IT
211.14
224.4*
112.14
IT*. 01
499.10
81T.49
517.21
411.24
3O4.4T
195.92
212.2*
431.33
»OO.JO
477.42
^14.*!
24 3 . 16
(1*9.17)
'(I**. Ill
11*7.12)
tun. ir~
<17«.'M
Ti»*sm
r^m.i*,!
1205.12)
r«7jf5T-
111*. 15)
1115.12)
, . U . .77-
(IH7.I2I
-trmTn-
UflllUL
115(1. 12 >
(177.151
(207,131
1201,121
(271,131
1I19.H»
1179,12)
Ut7.UI
(177,12)
(179,121
11*1,131
(164.IJJ
(214,151
(211.121
(1M.J51
1164.151
(139.121
11*9.12')
1709.131
(144.1 21
1 41.111
1148.12)
1141.12)
1194,12)
(222,151
( 154.15)
1159.151
(JOJ.J7J
(148.151
TJJTTISP
1174,191
• 64.97 I 7" 4, 191
41 1.09 1201
954 J) I 1190,721
443.79 1174,121
!5«. 94 1221,121
Figure 4-30. MPTER output.
4-53
-------
Review Exercise
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
The MPTER allows dispersion estimates to be made at
elevated receptors by using
a. Holland's plume rise.
b. the CRSTER method of terrain adjustment.
c. polar coordinates.
d. terrain downwash.
e. final plume rise.
True or False? MPTER will calculate concentrations in
uneven, rural terrain.
Three technical options available with MPTER are
, anri
True or False? MPTER will allow either Holland's or
Briggs' plume rise methods.
MPTER adjusts wind speed by a
a. linear extrapolation.
b. user estimation.
c. power law formula.
d. ratio of gas velocity to atmospheric stability.
e. Gaussian formula.
True or False? MPTER can treat complex terrain.
The MPTER is a plume model.
True or False? The MPTER model calculates plume rise
using gradual plume rise only.
The MPTF.R model allows a total of point ,
sources and receptor sites per run.
True or False? The MPTER can generate a circular set of
five receptor rings just as the CRSTER model does.
The MPTER model is recommended by U.S. EPA for
modeling pollution sources in
a. rural situations.
b. complex terrain.
c. urban situations.
d. Texas.
1. b. the CRSTER method of
terrain adjustment.
2. True
3. stack-tip downwash,
gradual plume rise,
buoyancy-induced dispersion
4. False
5. c. power law formula.
6. False
7 . Gaussian
8. False
9. 250, 180
10. True
11. a. rural situations.
4-54
-------
Lesson 7
VALLEY
Supplementary Reading
For more information about the VALLEY model and its
methods of treatment, obtain a copy of EPA 450/2-77-018,
VALLEY Model User's Guide, September 1977.
Lesson Goal and Objectives
Goal
To familiarize you with the VALLEY model, its method of
making dispersion estimates at receptor sites located on terrain
higher than stack top, and how it calculates worst-case air
pollution concentration.
Objectives
Upon completing this lesson, you should be able to:
1. identify the VALLEY model plume characteristics.
2. list the limitations of the VALLEY model.
3. describe the worst-case meteorological conditions used
with the VALLEY model for estimating the maximum
short-term concentration in complex terrain.
4. state the regulatory use of the VALLEY model.
Introduction
i
The last model to be discussed is the VALLEY model. As
shown in Figure 4-31, it is also known as the complex terrain
model. It was designed to allow modelers to estimate pollution
concentrations at receptors located above stack height. The
CRSTER and MPTER models handle terrain up to the lowest
stack height. If receptors are located on terrain above the
height of the stacks (shown as "x"), these models cannot be
used. Therefore, a reasonably accurate technique to estimate
air quality in complex terrain is needed, and VALLEY was
developed to fill this gap. Efforts to create an accurate complex
terrain model are far from over, and EPA is in the midst of
developing an extensive complex terrain model. However, a
Figure 4-31. Complex terrain model.
4-55
-------
refined model isn't expected to be available for several more
years. In the interim, VALLEY is an approved screening
technique for complex terrain.
Complex Terrain
Complex terrain influences the trajectory and diffusion of a
plume. The adverse effects of complex terrain are well known.
First, concentrations are increased because of the proximity of
elevated ground to the plume centerline. In extreme cases, the
plume can sometimes directly impact the side of a hill. Second,
drainage flow from mountain slopes at night causes air to pool
and stagnate in the valleys, and high concentrations often
result. Yet, there are physical processes acting that also tend to
lower concentrations. Field studies have shown that winds tend
to follow the terrain instead of going across steep height gra-
dients. This is called channelization, and it reduces the chances
for plume interaction with elevated terrain. In addition, the
increased turbulence from complex terrain will often lead to
lower concentrations at distances farther downwind.
The focus of concern in complex terrain is on the near field
receptors close to a source where very high concentrations can
often occur. Potential flow theory and field studies indicate
that plume impaction will most likely occur under stable
atmospheric conditions (Pasquill-Gifford Class E or F). The
kinetic energy required by a fluid to overcome the temperature
inversion and rise up over the terrain is not available. EPA has
analyzed field data from several sites in the Rocky Mountains
and determined that a reasonable set of worst-case meteoro-
logical conditions for short-term concentrations in complex ter-
rain is Class F stability, a wind speed of 2 to 5 m/s and per-
sistence of the wind direction within a 22 V£ ° sector for 6 hours
in a 24-hour period. These are the meteorological conditions
that EPA recommends be used in the VALLEY model to
estimate maximum 24-hour concentrations in complex terrain.
Plume Characteristics
Since the VALLEY model is a Gaussian plume model, condi-
tions are assumed to be steady state. That is, the atmosphere
and source conditions are constant over an averaging period.
The plume height is calculated by Briggs' method, and the
gradual rise option is recommended for complex terrain
calculations. In addition, the option of buoyancy-induced
dispersion is appropriate in complex terrain. In unstable
atmospheric conditions, the plume height is constant over ter-
rain. In stable atmospheric conditions, the plume height is
4-56
-------
constant above sea-level elevation, and the plume centerline is
allowed to come as close as 10 meters to the surface of the
ground, as shown in Figure 4-32.
Unstable and
neutral categories
—CL-i
Fraction of
ilume remaining
in sector
Stable
categories
Figure 4-32. VALLEY model treatment of terrain.
Input Parameters
The basic model allows a total of 50 point and area sources
that can be assigned to any locations to evaluate impact at a
fixed network of 112 receptors. The receptor network is
defined by 16 radials and seven equally spaced ring distances.
The user must scale the receptor from a known map scale. The
user must have a U.S. Geological Survey topographical map
(1:24,000 scale) to be able to properly assign receptor heights
from stack bases. The meteorological data recommended for
use with VALLEY are the worst case conditions discussed
above. Alternative inputs may be specified using guidance in
the User's Guide.
Limitations
The VALLEY model is designed to simulate a specific worst-
case condition in complex terrain, namely that of plume
impaction under stable atmospheric conditions. During
unstable conditions, it will tend to underpredict concentra-
tions. The model is also not designed to simulate terrain
4-57
-------
down wash, fumigation, changes in wind field trajectories, or
stagnant air conditions. Finally, the VALLEY model results are
only valid for receptors on the front of the first ridge of com-
plex terrain encountered by a plume. Behind the first hill and
farther downwind, the VALLEY model results are unreliable.
Use and Output
VALLEY can be used as a screening model in urban or rural
areas. It is recommended for use in rural areas where concen-
tration estimates are required at receptor sites in complex ter-
rain. The averaging times available are 24-hour and annual
averages. The primary use is to estimate 24-hour averages using
a set of predetermined meteorological conditions: Class F
stability, a wind speed of 2.5 m/s and wind directional per-
sistence for 6 hours in a 24-hour period. The output from
VALLEY is very difficult to read and interpret. An example is
shown in Figure 4-33 and concentration isopleths have been
drawn in as an aid.
/— .b
RCLOCIU i/3 INCH UP-/ ,Q
. U .0 .0
• tl « b
HLtr~C: I.UU HRS. COUCH COHIICTO 10 STO COM) V I« rifTP* I.1M. Ha IOHUP 10. OCC. HOBIM login*
.0 «I« VU.LO VI
SIM CMC DUC to «U. SKS
TEST Ml* — S02
HIND nose ro* c«s» OUPIICITIOX.
»u u*> w\.
?'/* / ° •" s°x'' "r*i I
/ I i.i If Tl.J I I I
M I . C00*0 k 1 I I I
1.1 is.i fc.j '/~*-- !• H-* itu-tl "-•
I i \ I «u>ii> I sen n.n Icoa
0
0
I
I i \ I «u>ii> I sen ILC» Icoaw)
\\ *\ r""*.a/i *V7'"I
-«•' \ \ll V .0 liVill ••16.£
\ V\ \0 J-
\ 11.1 \ setAcesJ '»•« ' i
1.1 .0 .a/ J».s /
V 4) Jl / /
.« \ .0 _/.
\^ .a
"
VLIiriV PHIIIfCD «ILIICS IT
|1.Q<11 TO SCI COfK. Vk UC/m
11.T I 10 S.> It.I
COOROTl Sl« HI OIGHSSCCI till DM
• •••4**M|*^*««II O.OOAO ••••••
DIUX /IOWII ST4B F HZOfM
3* £/ 100.
VV N(W HIND SPDSIMPSI W .11
UbU • OOCOO ^jObllO *0(;uUO UlOCtUb <
•I1CUK PIDtl HKT
....... (TO. .e
_ tic T iu f on« IK * ri.au
0 «» ._
num.. 9iir-K*ii «oot.
.0— /
>.%!«• (.MM* I.«f«>D
••]•••• *• ••*)*•*•* 3
..
SWIM it«*«in concert.
Figure 4-33. Output of test run for SO, using VALLEY.
4-58
-------
Review Exercise
1. True or False? The VALLEY model is a refined complex
terrain model.
2. True or False? The VALLEY model will give valid
concentration estimates behind hills.
1. False
Which of the following is(are) a limitation of the VALLEY
model?
a. It cannot handle more than one source.
b. It underestimates plume rise.
c. It only treats building down wash.
d. It underestimates concentrations for unstable conditions
(Classes A through D).
e. It underestimates concentrations for stable conditions
(Classes E through F).
2. False
EPA has found that maximum short-term concentrations
in complex terrain are most likely to occur under
a. unstable atmospheric conditions.
b. stable atmospheric conditions.
c. very high winds.
d. wind channelization.
e. the old oak tree.
S. d. It underestimates con-
centrations for unstable
conditions (Classes A
through D).
The worst-case meteorological conditions that should be
used with VALLEY are Class stability, a
wind speed of , and wind directional persistance
4. b. stable atmospheric
conditions.
in a 22)4° sector for
period.
hours during a 24-hour
6. The VALLEY model will handle receptors located
a. only at ground level.
b. only up to stack-top height.
c. in urban situations only.
d. from ground level to above stack top.
5. •
F
• 2.5 m/s
• 6
In stable atmospheric conditions, the VALLEY model
allows the plume centerline to come how close to the
ground-based receptor?
a. 100 meters
b. 400 meters
c. 0 meters
d. 10 meters
6. d. from ground level to
above stack top.
8. True or False? VALLEY is the approved EPA screening
model for receptors in complex terrain.
7. d. 10 meters
8. True
4-59
-------
Unit 5
Practical Use
of Point Source
Atmospheric
Dispersion Models
Lesson 1 Receptor Siting
Lesson 2 Roughness Length and Terrain Adjustment
5-1
-------
Lesson 1
Receptor Siting
Lesson Goal and Objectives
Goal
To familiarize you with the procedures involved in siting recep-
tors for determining downwind pollution concentrations.
Objectives
Upon completing this lesson, you should be able to:
1. explain the relationship between an air quality model and
a receptor location.
2. describe the criteria for siting an air quality receptor.
3. explain the difference between guessing where receptors
should be placed and making educated guesses for the
same placement.
4. name one statistical method and its technique for locating
receptor sites.
5. explain the reason the PTDIS and PTPLU models might
be chosen to select receptor sites.
6. explain the procedure that uses PTDIS and PTPLU to
find the distance to maximum ground-level concentrations.
7. explain what is meant by a receptor site being called
semipermanent.
8. state the number of receptor sites found in the CRSTER
single-source model.
9. recognize the reason the RAM model's output can be
meaningless for time periods longer than 1 hour.
Introduction
All air quality models discussed in Unit 4 estimate pollution
concentrations at points downwind from the source, called
receptors. Air quality modelers are interested in determining
what the concentration of a pollutant will be after it is
transported and dispersed by the atmosphere to specific loca-
tions. The locations of interest might be in a city, rural area,
or national park (Figure 5-1).
5-3
-------
Figure 5-1. Areas of environmental concern.
Concern might center on public health or property damage
(Figure 5-2). Whatever the reason for modeling, the user will
want to place the receptors where maximum concentrations are
likely to occur and the general public has access. The locations
for receptors may already be determined before modeling
begins. It may be necessary to find out if the air quality of a
specific area exceeds the National Ambient Air Quality Stan-
dards (NAAQS). In that case, the receptor site is not arbi-
trarily chosen by the air quality modeler. In many instances,
however, the most appropriate sites for receptors are not
known in advance. For instance, the modeler's interest may be
in the location and magnitude of the maximum concentrations
so that air quality samplers might be placed there (Figure 5-3).
As indicated above, any number of air quality decisions might
depend on the outcome of the model. Consequently, receptor
siting is not a trivial matter.
Guessing
Receptor siting may be accomplished by using a number of
approaches. One approach, used at times by the most
experienced air quality modeler, is guessing. Of course, guess-
ing assumes different levels of accuracy, depending on the
individual guessing. For instance, a receptor site chosen by
an individual who has no experience in air quality modeling
can properly be called a guess. An individual's prior experience
in such factors as terrain influence, meteorology, source
characteristics, and specific air quality models increases the
chance that the receptor site chosen is more appropriate. A
selection by this individual is called an "educated" guess
(Figure 5-4). Guessing may at times be the only approach
available to site receptors.
Figure 5-2. Health effects and
property damage.
Figure 5-3. Maximum concentration
location.
Figure 5-4. Guessing vs. educated
guessing.
5-4
-------
Math Procedures
A second approach may appear to be more scientifically sound
because a mathematical procedure is used. The procedure can
range from using a graph to using statistical probabilities in a
complex formulation (Figure 5-5). Obviously, this approach
involves some cost to the user. As the procedure becomes more
complex, the cost increases because of the amount and
accuracy of the required data. Statistical probability methods
such as frequency of occurrence or the Monte Carlo method
are used. A method like the frequency of occurrence depends
on knowledge of historical events such as wind direction and
stability class. Monte Carlo techniques involve the use of ran-
dom numbers to determine the most likely places for maximum
concentrations and, hence, the best places for receptors to be
sited.
Monte Carlo methods (Figure 5-6) are used for applications
where no mathematical solution to a problem exists. Given that
a complex statistical method may be employed in a model for
siting receptors, the model may not be as consistently accurate
as educated guesses by an experienced air quality modeler.
Screening Models
Another approach to receptor siting is to use the output of a
screening model to define the locations for receptors in subse-
quent runs of either a refined or screening model.
The PTPLU model is ideally suited to this task since it gives
the downwind distance of the maximum concentration from a
point source under a variety of conditions. One procedure used
is to find the highest concentration predicted by PTPLU for
each of the six stability Classes A through F, then identify the
downwind distances of the maxima from the PTPLU output
(Table 5-1). These six distances can then be used to locate
receptor rings in a refined model such as CRSTER, MPTER,
or RAM. The PTDIS screening model can also be used to help
select receptor sites. PTDIS allows the user to specify a number
of downwind distances for the purpose of generating the profile
of concentrations from a point source. By running the model a
number of times, the maximum concentration and its location
can be "cornered." U.S. EPA recommends using inexpensive,
simple screening models, like PTPLU or PTDIS models, for
determining maximum ground-level concentrations and the
distances to them before using a more expensive refined model.
Downwind distance, x
Figure 5.5 Graph and complex
formula for plume
dispersion.
Figure 5-6. Monte Carlo method.
Table 5-1. PTPLU output can be
used to select receptor
ring distances.
Stability
class
A
B
C
D
E
F
Max.
cone.
1661
802
611
214
179
98
Distance
(km)
1.15
2.73
4.S4
9.72
12.21
25.40
5-5
-------
Semipermanent Receptor Site Models
Other air quality models have semipermanent receptor sites
that are generated internally for the user. Being semiperma-
nent means that the number of receptors and their direction
from the source are fixed. The distances away from the source
are not fixed and can be varied through several runs to find
the location of maximum concentrations.
The user may select the distances as needed. Because the
wind direction at a specific source is variable throughout the
year, the receptor sites are generally placed in circular rings
around the source. This differs from the PTXXX models,
which site receptors in a straight line. For example, Figure 5-7
shows the CRSTER with 180 receptor sites available on five cir-
cular rings, and Figure 5-8 shows the VALLEY with 112 recep-
tor sites in seven circular rings.
•
•
. . . . o . • •
• • - • •
•....••
• ...»
Figure 5-7. CRSTER receptor rings.
Figure 5-8. VALLEY receptor rings.
For siting receptors, individual runs for these models are
more expensive than runs for the simpler PTXXX models. For
example, if a source and 50 receptors are to run in the PTDIS
model, the output will cost approximately $1.50. However, if
180 receptors, 19 sources, and one year of meteorological data
are run in the CRSTER, the output may cost $50.00. A point
to remember is that cost of model output is directly related to
the model's complexity. Factors such as the amount of
meteorological data, number of sources (area and point), and
volume of output all affect the cost.
5-6
-------
RAM Model Receptor Site Option
Guessing, using statistical probability procedures, or using
internally generated semipermanent receptor sites complicates
receptor siting. These factors complicate siting because, unless
enough runs identify maximum concentrations, the accuracy of
the siting is uncertain. The RAM multisource model has an
option available that can help: program-selected receptors
(Figure 5-9). When used, this option allows the model to locate
receptor points where it predicts the maximum concentration
will occur for a given hour. Since the winds change each hour,
the program-selected receptors will also change locations.
Thus, output for these receptors for averaging times greater
than one hour is meaningless, since the point was not fixed in
space.
Figure 5-9. RAM receptor siting option.
5-7
-------
Review Exercise
1. The relationship between an air quality model and a recep-
tor is that
a. an air quality model always computes a distance to the
receptor locations.
b. receptors are always in urban areas because air quality
models are only concerned with health effects.
c. both are always arbitrarily chosen.
d. an air quality model estimates pollution concentrations
at points called receptors.
2. Receptor sites should be located where
expected to occur and has access.
are
1. d. an air quality model
estimates pollution concentra-
tions at points called
receptors.
3. Choose one statistical method and its technique used in
siting receptors.
a. Monte Carlo, random numbers
b. hypergeometric, exponential decay
c. Weibull, normal distribution
d. poisson, geometric
2. maximum concentrations,
the general public
4. The PTPLU and PTDIS models might be chosen to
select receptor sites because they are
a. refined air quality models.
b. inexpensive and simple screening tools.
c. recommended by U.S. EPA.
d. both b and c
3. a. Monte Carlo, random
numbers
5. The PTPLU model is ideally suited for selecting receptors
because
a. it is a statistical technique.
b. it is a highly sophisticated model.
c. the general public has access to it.
d. it gives the distance to the maximum concentration for
each stability class.
4. d. both b and c
6. When an air quality model's receptor sites are of a fixed
number and direction from the source they are
called
5. d. it gives the distance to
the maximum concentration
for each stability class.
7. The CRSTER single-source model has
receptor sites on circular rings.
6. semipermanent
7. 180,
5
5-8
-------
The RAM model's output may be meaningless for
averages longer than one hour because
a. RAM only produces hourly output.
b. program-selected receptor sites will change location
every hour.
c. RAM's output is only for 24-hour periods.
d. the RAM model divides 24-hour concentrations into
eight, 3-hour periods.
8. b. program-selected receptor
sites will change location
every hour.
5-9
-------
Lesson 2
Roughness Length and
Terrain Adjustment
Lesson Goal and Objectives
Goal
To familiarize you with the effect of small surface roughness
features and large terrain features on wind flow and with how
certain models use roughness length and are modified for
applications in complex terrain. The adjustment methods of
the CRSTER and MPTER models, and the special problems of
the VALLEY model in describing concentrations in complex
terrain will also be covered.
Objectives
Upon completing this lesson, you should be able to:
1. describe the effect of natural and artificial objects on
wind flow.
2. define roughness length.
3. recognize the difference between roughness features and
terrain features.
4. name three UNAMAP models that can be adjusted for
terrain.
5. name the model that was designed for rough,
mountainous terrain.
6. describe the method used by the CRSTER and MPTER
models to adjust for terrain that may extend up to the
height of the stack.
7. state one reason that the VALLEY model may perform
poorly in its attempt to describe pollution concentrations
in complex terrain.
8. describe the method the EPA models use to adjust for
terrain lower than stack base.
Roughness
Roughness is a function of surfaces of objects on the earth such
as buildings, trees, bridges, etc. Each object, whether natural
or artificial, slows and distorts the direction of the free wind
due to its height, shape, and surface characteristics. Roughness
features are usually relatively small, such as grass, trees, and
5-11
-------
houses. These features affect the wind pattern, particularly the
wind speed as it approaches the earth's surface, by causing fric-
tional drag in the lower atmosphere (Figure 5-10).
Roughness Length
Roughness length is defined as the height above ground when
the mean wind speed goes to zero due to frictional effects.
Some roughness length considers the earth's texture (objects in
the path of the wind), it is graded in a manner similar to the
way sandpaper is graded—from smooth to rough (Figure 5-11).
Some typical values for z, are given in Table 5-2. Notice that
the roughness length for a desert is O.OS cm. This means that
the wind speed declines to zero very close to the earth because
a desert is relatively smooth. Since an urban park has objects
extending higher off the ground, the wind is blocked or slowed
to zero farther from the ground, so roughness length is higher.
Grade 0
Sandpaper
Grade 10
Figure 5-11. Sandpaper grade analogy for roughness factors.
Inclusion of Roughness Length
in Air Quality Models
Roughness length is represented in air quality models by the set
of dispersion rates that are used. z» is a measure of turbulence
and dispersion is increased by turbulence. Therefore, the
greater the value of z,, the faster a plume will spread in the
vertical and horizontal directions. The Pasquill-Gifford disper-
sion curves used in most EPA models are based on a few
carefully performed diffusion experiments from the 1950s. The
terrain in these cases was rural, gently rolling, and z, ranged
from 3 to 30 cm. By contrast, the McElroy-Pooler dispersion
experiments, on which the urban RAM model is based, were
performed in an environment downwind of a city where z,,
equalled 100 cm. The larger roughness length reflects the
increased turbulence found in urban areas.
Fast
Wind speed
Slower
Figure 5-10. Roughness effect
on wind flow.
Table 5-2. Typical values of
roughness length.
Surface
Desert
Alfalfa field
Com. field
Urban park
Central business district
Z, (cm)
0.03
2.72
74
127
32 1
5-12
-------
Terrain Adjustments
Terrain features differ in size from roughness elements (Figure
5-12). Terrain is usually considered to be large surface
features, such as mountains and hillsides. Roughness elements,
which are relatively smaller, slow the wind through fractional
drag and usually affect only the edges of the plume. Terrain,
however, affects the entire plume by distorting wind flow
(Figure 5-13). The adjustment to models for terrain provides
information about plume behavior. The information is used in
the model to predict ground-level concentrations downwind.
Terrain adjustment is concerned with the resulting path of the
plume centerline with respect to large terrain features.
Figure 5-12. Terrain and roughnen features.
Figure 5-13. Terrain effect on
wind flow.
CRSTER
As shown in Figure 5-14, the CRSTER model uses simple ter-
rain adjustments. Terrain adjustments are either simple or
complex. The CRSTER model will not estimate concentrations
at receptors on terrain that is higher than the stack top. That
is, the difference between the height of a receptor on terrain
and the stack top is calculated using the base elevation of the
stack. The highest terrain considered can be no taller than the
physical height of the stack. If more than one stack is grouped
Figure 5-14. CRSTER terrain
adjustments for
receptor height!.
5-13
-------
in a location, then the shortest stack of the group is used to
determine the terrain height limit. For example, in Figure
5-15, if the stack is located on terrain that is 400 meters above
sea level and the stack is 60 meters high, the model would not
estimate concentrations at receptors higher than 460 meters
above sea level.
Areas of terrain that are lower than stack base may be
included, as shown in Figure 5-16. Any receptor lower than
stack base is automatically raised by the model to stack base
elevation. For example, if the stack base is at zero elevation, a
receptor lower than stack base would have a negative elevation.
A receptor 10 meters below the stack base would be entered as
a minus 10. The model would raise the elevation of such a
receptor up to zero.
The user should remember that pollutant concentrations-
increase as the elevated terrain approaches the plume
centerline. As discussed earlier, terrain adjustments in the
Gaussian model are made by decreasing the effective plume
height, H. This can be thought of as the plume centerline
remaining level and the terrain rising up toward it. In reality,
the terrain adjustments are made, not by raising the receptors,
but by lowering the plume centerline so that the plume is
moved closer to the receptor. The effect on concentration is
the same either way.
MPTER Adjustment
The MPTER model's terrain adjustment goes beyond
CRSTER's by allowing the user to decide how the plume will
travel over the terrain feature. The adjustment may be chosen
from 0 to 1, or from 0% to 100% (Figure 5-17). If zero adjust-
ment for terrain is called for, then the MPTER model will
keep the plume centerline height constant above the terrain.
This means that the plume will follow the terrain shape. If
100% adjustment is called for, then MPTER will adjust for
terrain as previously described for CRSTER. In MPTER, the
user may elect to use any percent of adjustment between 0 and
100% that is deemed necessary. MPTER, with this adjustment
option, is more complex than CRSTER, but less complex than
the VALLEY model, which is described next.
Figure 5-15. Receptor height
example.
Figure 5-16. CRSTER adjustment
for depreuioni.
0%
100%
Figure 5-17. Variable terrain
adjustment for
MPTER.
5-14
-------
VALLEY Adjustment
The VALLEY model is called the complex terrain model. It
was designed to adjust for terrain by using more complex
methods than the CRSTER. VALLEY is also a Gaussian plume
model. All of the assumptions inherent in using the Gaussian
distribution still apply.
VALLEY was designed for complex terrain—that is, rough,
mountainous areas. It was developed using sparse data from
the western U.S. The model adjusts plume behavior for terrain
(Figure 5-18). It considers plume centerline behavior as a func-
tion of two atmospheric stability situations. These situations are
stable (Pasquill-Gifford stability Classes E and F) and
unstable/neutral (P-G Classes A through D).
For stable atmospheric conditions, VALLEY assumes the
plume centerline is located at the height calculated by Briggs"
plume rise equation plus the physical stack height. This height
is called effective stack height. The plume centerline is then
assumed to stay at a constant height above stack base. This
means that if the terrain increases in elevation downwind from
the source, the plume centerline will approach the ground. In
effect, the distance from the plume centerline to the ground
becomes smaller. The model will not allow the plume center-
line to actually impact the terrain. It maintains a 10-meter
minimum separation between the plume centerline and the ter-
rain beyond the first point where the centerline comes within
10 meters of the terrain. If the terrain continues to increase in
elevation, the plume maintains the 10-meter separation as it
spreads vertically for 400 meters, at which point the concentra-
tion is assumed to have decreased to zero.
For neutral and unstable atmospheric conditions, the model
assumes that the plume centerline remains constant above
ground level. This means that no matter what the terrain
features are downwind of the source, the plume follows the
shape of the terrain.
The user should note, however, that the concentration
estimates should be ignored after the plume first comes in con-
tact with any part of a hill (Figure 5-19). This is because
VALLEY does not incorporate increased turbulence on the
backside of hills, ridges, etc. The VALLEY model may per-
form poorly in making concentration estimations in complex
terrain. This is because the Gaussian distribution concepts have
been considerably modified in the attempt to handle complex
terrain. Unfortunately, the effects of complex terrain have not
been studied thoroughly enough to develop a technique that
performs significantly better than VALLEY in estimating the
highest concentrations that are of concern to regulatory
agencies.
Figure 5-18. Complex terrain,
or VALLEY.
Figure 5-19. Terrain limitations
for VALLEY.
5-15
-------
Review Exercise
1. Define roughness length.
The effect of natural and manmade objects on wind flow
near the surface of the eanh is to
a. increase the wind speed and direction.
b. slow and distort the wind.
c. reduce frictional drag.
d. both a and c
Roughness length is the
height above the ground
where the mean wind speed
goes to zero due to
frictional effects.
The typical roughness length assumed in the urban RAM
model is
a. 100 cm.
b. 100 m.
c. 10 m.
d.' 1 cm.
2. b. slow and distort the wind.
4. Three UNAMAP models that can be adjusted for terrain
such as hills are
a. CRSTER, ELSTAR, VALLEY
b. VALLEY, RAM, CDM
c. CRSTER, RAM, ELSTAR
d. CRSTER, MPTER, VALLEY
3. a. 100 cm.
The one UNAMAP model designed for complex terrain,
such as large mountains, is
a. CRSTER
b. CDM
c. VALLEY
d. RAM
4. d. CRSTER, MPTER,
VALLEY
The simple terrain adjustment method used by models such
as CRSTER is to
a. decrease the effective plume height by the rise of terrain
above stack base.
b. count hills and multiply by that number.
c. lower hill tops to stack base height.
d. raise stack base to hill tops.
5. c. VALLEY
7. One reason VALLEY may perform poorly in estimating
concentrations in complex terrain is that
a. VALLEY does not use the Gaussian distribution.
b. turbulent effects over terrain are well understood.
c. the Gaussian distribution is severely modified.
d. it was designed for flat terrain only.
6. a. decrease the effective
plume height by the rise
of terrain above stack base.
7. c. the Gaussian distribution
is severely modified.
5-16
-------
8. The method used by models to adjust for terrain lower
than stack base is to
a. supply pollution material to the model equal to the area
of the sink holes.
b. lower the stack base until it is level with the bottom of
the hole.
c. lower the stack top by an amount equal to the depth of
the deepest depression.
d. raise the terrain to the level of the stack base.
8. d. raise the terrain to the
level of the stack base.
5-17
-------
UnitG
Case Studies:
Modeling and
Interpreting Results
Lesson 1 Oil Refinery
Lesson 2 Iron-casting Plant
6-1
-------
Lesson 1
Oil Refinery
Lesson Goal and Objectives
Goal
To familiarize you with an actual case of modeling the air
quality surrounding an oil refinery and interpreting the results
of that modeling.
Objectives
Upon completing this lesson, you should be able to:
1. describe the terrain features around the oil refinery.
2. describe the new proposed construction.
3. name the model chosen.
4. identify why the oil refinery was "screened" first, instead
of using a refined analysis.
5. interpret the results of modeling by correctly identifying
the maximum ground-level concentrations given by the
model run.
Introduction
We've looked at several air quality models found on UNAMAP,
Version 4 — particularly the point source dispersion models that
are Gaussian plume models. Each model is designed for specific
air pollution applications. Screening models are designed to be
simple, for fast calculations of pollution concentration
estimates. Other complex models examine the details of
atmospheric and industrial processes and their interactions in
an attempt to estimate pollutant concentrations. Each user
must decide which kind of model to use for any specific
application. Each situation tends to be site-specific. A model
that produces acceptable results in Oklahoma City may not
perform as well in Dallas. The situations are different, even
though both are large cities located on flat terrain.
The user may seek advice from a professional air quality
modeler, air quality meteorologist, or U.S. EPA as to the
appropriate model to use. For example, the single-source
CRSTER model would not be appropriate for widely-spaced
pollution point sources since it locates all sources at a single
plant site. A flat terrain model, such as RAM, would not be
selected for a problem in complex terrain. Again, the PTXXX
6-3
-------
models were designed for flat terrain and Gaussian plumes. If a
large body of water is added to the situation, the models may
not perform well due to shoreline effects.
Oil Refinery Case Study
One case study to be examined is an oil refinery located in
northeastern Oklahoma. The oil company plans to expand the
existing facilities to increase its capabilities. While some
renovation of the existing facilities will take place, the refinery
will not stop operating. The refinery currently operates 24
hours per day. It will continue this schedule after renovation
and expansion of the facilities. Except for a few farmers, most
of the population living around the refinery work directly or
indirectly for this oil company. Factors to be considered are
listed in Figure 6-1.
Description of the Area
The northeastern section of Oklahoma (Figure 6-2) can be
described as fairly flat, with some elevations up to 10 meters
above the surrounding terrain. Most of the area is forested.
Some gravel pits and coal strip mines dot the area. A few oil
wells are also within the area. Small streams cross the area,
and a river runs north to south by the western edge of the
refinery. A major four-lane highway and a railroad serve the
area. A small lake is located south-southeast of the refinery. A
small town is located 1 kilometer east of the plant. A large
urban area is located 24 kilometers southwest of the refinery.
Description of the Plant Site
The plant site has an elevation of 173 meters (567 feet) above
mean sea level (MSL) and is situated in a relatively flat river
valley. At a distance of 1.25 kilometers northeast, the terrain
rises 10.1 meters (33 feet) above stack base. The plant site
occupies about 90 acres of land. An area 1 kilometer to the
southwest is 8 meters (26 feet) lower than stack base.
Physical layout of industry site
Problem
Meteorological situation
Selected model
Reasons for selection
Output
Figure 6-1. Considerations for case studies.
6-4
-------
Figure 6-2. Northeastern Oklahoma.
Description of the Oil Refinery
Figure 6-S includes the refinery with one existing stack that
emits S.28 grams per second (114 tons per year) of sulfur
dioxide (SO*). The stack is 35.0 meters high and has an inside
diameter of 1.56 meters. The stack gas velocity is 13.2 meters
per second, and stack gas temperature is 394 K. The tallest
existing building is 12 meters high. The stack is located on the
west side of the building, adjacent to it.
Meteorology of the Area
The average annual ambient air temperature is 18.3°C, and
the average annual mixing height is 1200 meters. The wind
rose for Tulsa airport is shown in Figure 6-4. Approximately
45.5% of the year, the wind blows from four directions: north,
north-northeast, south, and south-southeast.
d= 1.56m
v,= 13.2 m/s
Figure 6-3. Refinery: existing
conditions.
Figure 6-4. Tulsa wind rose.
6-5
-------
New Proposed Construction
A new stack will be built that will be 35.0 meters high, will
have an inside diameter of 1.56 meters, and will emit 1.5
grams per second (52 tons per year) of SO2. It will be built
next to a building that is 16 meters high. The stack gas velocity
will be 13.2 meters per second, and gas temperature will be
394 K. The new construction will be 20 meters east of the
existing stack (Figure 6-5).
d= 1.56m
h,= 35 m
Q=1.5g/s
v,= 13.2 m/s
T,= 394K
h=16m
Figure 6-5. Proposed construction.
Model Selection and Application
The UNAMAP package has a wide range of regulatory models
available. The choice of models was narrowed to point source
dispersion models available on UNAMAP since these are point
sources. The Guideline on Air Quality Models recommends
that the modeler screen a site before committing to an expen-
sive, refined model. It was decided to use the PTPLU screen-
ing model in conjunction with the conservative time-scaling
factors in EPA's Volume 10 Guideline to model the impact of
the oil refinery. These factors assume that the maximum
3-hour and 24-hour values are 90% and 40% of the maximum
1-hour value, respectively. A conservative scaling factor for
annual average concentrations is 10% of the 1-hour maximum.
Since existing emissions exceed 100 tons per year of SOt, this
is already a major source. The new stack will add 52 tons per
year of SO2 to the atmosphere, which is a significant increase
under EPA PSD regulations. Thus, the new stack at the oil
refinery must be modeled to ensure it does not violate the Class
II PSD increments for SOj. of 20 fig/m3 (annual average), 91
Hg/m* (24-hour maximum) and 512 /tg/m3 (3-hour maximum).
No Class I lands exist within 50 km of the refinery. The stack
6-6
-------
parameters and emission rates for the new stack were entered
into the PTPLU model along with values for temperature and
mixing height at the site. The modeling results are shown in
Figure 6-6 for 1-hour SOZ concentrations. The maximum
1-hour value is 8.7 fig/m3 and occurs at a distance of 442 m
from the stack. Using EPA scaling factors, conservative
estimates of SO* concentrations for longer averaging times are
obtained, and these are shown in Table 6-1. The results
indicate the new stack will not violate the Class II PSD
increments for SO».
Table 6-1. Maximum SOt concentrations from the new stack.
Averaging
time
3-hour
24- hour
Annual
Maximum
impact
(/tg/m«)
8
3
1
daw II
PSD increment
(/tg/m1)
512
91
20
Under New Source Review, the refinery must also demon-
strate that it will not exceed the NAAQS for SOi, given the
new emissions source. To do this, both stacks at the refinery
were modeled with the PTPLU screening model used in con-
junction with the conservative scaling factors. Since the height,
temperature, velocity, and diameter of the new stack are the
same as the parameters for the old stack, the emissions were
combined and modeled as one source (Figure 6-7). The maxi-
mum 1-hour SOj concentration for the entire plant is 28 /*g/ms
and is scaled to appropriate averaging times in Table 6-2. The
results indicate no problems with the NAAQS either.
The EPA PTPLU screening model has been used to make
conservative estimates of air quality impacts. Since the oil
refinery can demonstrate compliance with the PSD increments
and the NAAQS using PTPLU, there is no need to run a more
expensive, refined model.
Table 6-2. Maximum SO* concentrations from both stacks.
Averaging
time
3-hour
24-hour
Annual
Maximum
impact
Oig/m1)
25
11
3
NAAQS
-------
PTPLU (VERSION 81O36I
IN AIR QUALITY DISPERSION MOOEL IN
SECTION 3 MODELS PROPOSED SEP8O fOR «1 GUIDELINES.
IN UNAM1P (VERSION 4) DEC 80
SOURCE FILE 13 ON UNAMAP MAGNETIC TAPE FROM NTIS.
NEW STUCK AT THE OIL REFINERV
•••SOURCE"'
EMISSION RATE
STACK HEIGHT
STACK DIAM
E/IT VELOCITY
STK GAS TEMP
VOLUMETRIC FLOW
1 SO (G/SEC)
35 CO (M)
I 56 (M)
13 2O (M/SEC)
394 CO (K)
•OPTIONS'
>»INPUT PARAMETERS<«
IF • 1. USE OPTION
IF • 0, IGNORE OPTION
IOPTI1)
IOPTC2)
IOPTO)
35 23 (M"3/SEC)
O (GRAO PLUME RISE)
0 (STACK DOWNWASH)
0 (BUOY INDUCED DISP.)
"•METEOROLOGY"'
AMBIENT AIR TEMPERATURE • 391 CO (K)
ANEMOMETER HEIGHT
MIXING HEIGHT
WIND PROFILE EXPONENTS
RECEPTOR HEIGHT
1O CO (M)
12OO.OO (M)
A: .10. B: .IS. C
0: .25, t: .30. F.
O 0 (M)
>»CALCULATED PARAMETERS«<
BUOYANCY FLUX PARAMETER •
2O 59 (M"4/SEC"3)
20
SO
ANALYSIS OF CONCENTRATION AS A FUNCTION OF STABILITY AND WINO SPEED
STABILITY
SFABUII Y
2
2
2
2
2
2
2
2
2
STABILITY
3
3
3
3
3
3
3
3
3
STABILITY
• '
ST'BILITr
5
5
S
5
5
STABILITY
6
e
6
6
f.
WIND SPEED
1 M/SEC)
0 SO
0 BO
1 00
1 50
2 00
2 50
3 00
WIND SPEED
(M/SECI
0 SO
0 80
1 00
1 50
2 00
2 50
3 OO
4 OO
5 00
WINO SPEED
IM/SEC)
2 00
2 50
3 OO
4 OO
5 OO
7 00
10 OO
12 00
15 00
WINO SPEED
(M/SECI
0 50
O 80
t OO
1 50
2 00
2 50
3 OO
4 CO
5 00
7 OO
10 OO
12 OO
15 OO
2O 00
WIND SPEED
(M/SEC)
2 OO
2 50
3 00
4 OO
5 00
WINO SPEED
IM/SEC)
2 00
2 50
3 00
4 OO
5 00
MAX CONC
(G/CU M)
6 4485E-O6
7 3538E-O6
7 7437E-O6
8 31O6E-06
8 5461E-OS
8 6622E-06
8 7172E-O6
MAX CONC
(G/CU M)
3 0371E-06
4 0874E-O6
4 6441E-O6
5 6891E-O6
6 3B86E-O6
6 8S79E -06
7 I672E-O6
7 4755E-06
7 5357E-06
MAX CONC
(G/CU M)
5 4443E-06
6 0127E-06
6 4233E-GG
6 9171E-O6
7 1310E-06
7 1275E-O6
6 7038E-O6
6 3485E-O6
5 8224E-O6
MAX CONC
(G/CU M)
4618E-07
0246E-06
3476E-OS
O864E-06
7177E-O6
2520E-06
6963E-O6
2952E-06
6652E-OS
9979E-OS
9977E-O6
86O2E-OS
5807E-06
03O6E-OS
MAX CONC
(G/CU M)
4 0281E-O6
3 6776E-O6
3 4O51E-06
3 OOI2E-06
2 7096E-O6
MAX CONC
(G/CU M)
3 1294E-O6
2 8777E-06
2 6790E-06
2 3796E-06
2 I602E-06
DIST OF MAX
(KM)
0 916
0 75O
0 684
0 584
0 526
0 475
0 443
DIST OF MAX
(KM)
3 769
1 88O
1 574
1 158
0 945
0 B17
0 729
0 618
O 549
DIST OF MAX
KM)
677
405
225
004
0 874
0 728
O 619
0 577
0 535
OIST OF MAX
(KM)
35 381
17 263
12 201
7 139
4 937
3 766
3 060
2 369
943
495
187
075
OOO
O 962
OIST OF MAX
(KM)
5 749
5 227
4 847
4 318
4 OOO
OIST OF MAX
(KM)
11 953
10 722
9 832
8 633
7 839
EFFECT HT
(M)
449 t(3>
293 8(3)
243 1(2)
173 0
138 5
117 8
104 O
EFFECT HT
(M)
449 1(2)
293 8(2)
242 1(2)
173 O
138 5
1 17 a
1O4 O
86 8
76 4
EFFECT HT
(M)
138 S
117 a
1O4 O
86 8
76 4
64 6
55 7
53 3
48 8
EFFECT HT
(M)
449 1(2)
393 8(2)
343 1(3)
173 O
138 5
117 8
104 0
86 8
76 4
64 6
55 7
52 3
48 8
45 4
EFFECT HT
(M)
99 5
94 9
91 4
86 2
82 5
EFFECT HT
(M)
88 5
84 7
81 8
77 5
74 4
••••EXTRAPOLATED WINDS""
WINO SPEED MAX CONC OIST OF MAX
(M/SEC) (G/CU M) (KM)
O 57 G 6954E-O6 0 868
O 91 7 5777E-O6 O.712
1 13 7 942OE-O6 0 651
1 70 8 433SE-O6 0 557
2 37 8 59S3E-O6 O 495
2 83 8 710IE-O6 O 462
3 40 8 6989E-06 0 422
""EXTRAPOLATED WINDS""
WINO SPEED MAX CONC DIST OF MAX
(M/SEC) (G/CU M) (KM)
0 SO 3 433SE-O6 3 366
0 97 4 S545E-06 1 618
1 21 5 1284E-O6 1 361
1 81 6 1550E-O6 1 O12
2 41 6 7897E-O6 0 834
3 02 7 1754E-OS O 726
3 62 7 3963E-O6 O 653
4 83 7 5373E-06 0 560
6 03 7 4565E-O6 0 503
"••EXTRAPOLATED WINDS'"*
WINO SPEED MAX CONC OIST OF MAX
(M/SEC) (G/CU M) (KM)
2 57 6 O782E-06 1 375
321 6 5595E-O6 1 167
3 85 £ 8661 £-06 1 029
5 14 7 1457E-O6 0 859
6 42 7 1665E-O6 O 761
8 99 6 8710E-06 0 647
13 85 6 1967E-06 O 564
15.43 5 7S26E-O6 0 531
19 27 5 1585E-06 0 499
""EXTRAPOLATED WINDS""
WIND SPEED MAX CONC OIST OF MAX
(M/SEC) (G/CU M) ' (KM)
0 68 8 3903E-O7 22 072
t O9 1 4993E-O6
-------
BOTH ST»CKS *T TH€ OH REFINERY
•••SOURCE'"
EMISSION R»TE
STACK HEIGHT
STACK DIAM
E»I1 VELOCITV
STK CAS TEMP
VOLUMETRIC FLOW
4 78 (G/SCC)
35 00 (M)
1 56 (Ml
13 2O (M/SEC)
394 00 IK)
»>1NPUT P«RA*ETE«S<«
•••OPTIONS'"
IF • t. USE OPTION
IF - 0. IGNORE OPTION
IOPTII) • 0 (GRAO PLUME RISE)
10PTO) • 0 (STICK DOWNWASH)
IOPT(3> • O (BUOY INDUCED OISP )
•••METEOROLOGY'"
AMBIENT AIR TEMPERATURE
ANEMOMETER HEIGHT
MIXING HEIGHT
WIND PROFILE EXPONENTS
RECEPTOR HEIGHT
291 OO (K)
10 00 (M)
120O 00 (Ml
A 10. 8 15. C
D 25. £• 30. F
0 O (Ml
35 33 (M"3/SEC)
>»CALCULATED PARAMETERS""
BUOYANCY FLUX PARAMETER •
20.59 (M"4/SEC"3I
ANALYSIS OF CONCENTRATION AS A FUNCTION OF STABILITY AND WIND SPEED
20
3O
•EXTRAPOLATED WINDS'
STABILITY
STABILITY
3
2
2
2
2
2
2
2
2
STABILITY
3
3
3
3
3
3
3
3
3
STABILITY
J
STABILITY
5
5
5
5
5
STABILITY
6
6
6
e
6
WINO SPEED
IM/SEC)
0 50
0 80
1 00
1 50
2 OO
2 50
3 OO
WIND SPEED
(M/SEC)
0 50
0 80
1 OO
1 50
2 00
2 50
3 OO
4 00
5 00
WINO SPEED
IM/SEC)
2 OO
2 50
3 00
4 OO
5 00
7 OO
10 OO
12 00
15 00
WIND SPEED
(M/SECI
0 50
0 80
1 00
1 50
2 00
2 50
3 00
4 00
5 00
7 00
10 OO
12 00
15 OO
JO OO
WIND SPEED
IM/SEC)
2 OO
2 50
3 00
4 OO
5 OO
WINO SPEED
CM/SECI
2 00
2 50
3 00
4 00
5 OO
MAX CONC
(G/CU M)
2 0549E-05
2 3434E-05
2 4G76E-OS
2 6483E-05
2 7334E-05
2 7603E-05
2 7779E-05
MAX CONC
(G/CU M)
9 678ZE-06
1 3025E-05
1 4799E-05
1 8129E-O5
2 0358E -05
2 I854E-05
2 284OE-05
2 3822E-05
2 4014E-05
MAX CONC
IG/CU M)
1 7349E-O5
1 9 16 IE -OS
2 0469E-05
2 2O42E-O5
2 2724E-05
2 27I3E 05
2 I363E-OS
2 023 IE 05
1 8554E-05
MAX CONC
(G/CU Ml
1 7405E-06
3 2650E-06
4 3943E-06
6 6487E-O6
8 6604E-O6
0363E-O5
I779E-05
3687E-05
4866E-05
5927E-O5
5926E-O5
5488E-05
4597E-OS
2844E-05
MAX CONC
IG/CU Ml
1 2836E-05
1 I719E-05
1 085 IE -OS
9 S637E-OS
8 S346E-O6
MAX CONC
(G/CU M)
9 9724E-06
9 I702E-06
8 5371E-06
7 5829E-06
6 8837E-O6
DIST OF MAX
(KM)
0 916
0 750
0 £84
0 584
0 526
0 475
0 442
DIST OF MAX
(KM)
2 769
1 880
1 574
1 158
0 945
0 817
0 729
0 618
O 549
DIST OF MAX
(KM)
1 677
1 4O5
1 225
1 004
0 874
0 728
0 619
0 577
0 5J5
OIST OF MAX
(KM)
35 381
17 263
12 201
7 139
4 937
3 766
3 06O
2 369
943
495
187
075
OOO
0 962
DIST OF MAX
(KM)
5 749
t 227
4 847
4 318
4 OOO
DIST OF MAX
(KM)
11 953
1O 722
9 832
8 633
7 839
EFFECT HT
(M)
449 1(2)
293 8(2)
242 1(3)
173 0
138 5
117 8
104 0
EFFECT HT
(M|
449 1(2)
293 8(2)
242 1(2)
173 0
138 5
117 8
104 0
86 8
76 4
EFFECT HT
(M)
138 5
117 8
104 0
86 8
76 4
64 G
55 7
53 3
48 8
EFFECT HT
(M)
449 1(2)
293 8(2)
242 1(2)
173 0
138 5
117 8
104 0
86 8
76 4
64 6
55 7
52 3
48 8
4S 4
EFFECT HT
(M)
99 5
94 9
91 4
88 1
83 5
EFFECT HT
(M)
88 5
84 7
81 8
77 5
74 4
WINO SPEED MAX CONC
(M/SEC) (G/CU M)
0 57 2 I336E-OS
0 91 2 4148E-O5
t 13 2 53O8E-05
1 70 2 6875E-O5
3 27 2 7390E-O5
2 83 2 7756E-O5
3 40 2 7721E-O5
"••EXTRAPOLATED WINDS
WIND SPEED MAX CONC
(M/SEC) (S/CU M)
0 60 1 0943E-O5
0 97 1 4514E-OS
t 31 1 6343E-O5
1 81 1 9614E-O5
2 41 2 IS37E-O5
3 02 2 2866E-05
3 62 2 3570E-05
4 83 2 4019E-O5
6 03 2 3761E-05
••"EXTRAPOLATED WINDS
WINO SPEED MAX CONC
(M/SEC) (G/CLJ M)
2 57 1 9369E-O5
3 21 2 0903E-05
3 85 2 1880E-O5
5 14 2 2771E-O5
6 42 2 2837E-OS
8 99 2 1895E-OS
12 85 1 9747E-05
15 42 1 8332E-O5
19 27 1 6438E-OS
•"•EXTRAPOLATED WINDS
WIND SPEED MAX CONC
IM/SEC) (G/CU M)
0 68 2 6737E-O6
1 09 4 7777E-OS
1 37 6 O658E-O6
2 05 8 8505E-O6
2 74 1064E-05
3 42 2696E-05
4 10 3837E-O5
5 47 5244E-O5
6 84 5886E-05
9 57 5S89E-05
13 68 5015E-O5
16 41 4103E-O5
20 53 2S75E-O5
27 36 0744E-O5
••"EXTRAPOLATED WINDS
WIND SPEED MAX CONC
(M/SEC) (G/CU M)
2 91 1 0089E-O5
3 64 9 9739E-O6
4 37 9 I897E-O6
5 83 a ooaat-oe
7 28 7 125OE-O6
'•"EXTRAPOLATED WINDS
WINO SPEED MAX CONC
(M/SEC) (0/CU M)
2 91 8 6385E-06
3 64 7 8889F.-O6
4 37 7 3023E-O6
5 82 6 427IE-O«
7 28 5 7869E-06
THAT THE SAME STABILITY
OIST OF MAX
(KM)
0 868
O 712
0 651
0 657
O 495
0 452
O 422
OIST Of MAX
(KM)
2 366
1 618
1 3GI
1 012
0 834
0 726
0 653
O S60
0 503
OIST OF MAX
(KM)
1 375
1 167
1 029
0 859
0 7GI
0 647
0 5G4
0 531
0 499
....
DIST OF MAX
(KM)
22 0?2
10 641
8 078
4 778
3 393
2 758
2 317
803
520
218
OO7
000
0 955
0 889
....
OIST OF MAX
(KM)
4 9O9
4 483
4 169
4 OOO
3 748
• •»»
DIST OF MAX
(KM)
9 971
9 011
8 319
7 338
7 OOO
IS NOT UKEL
EFFECT HT
(M)
400.4(3)
363 4(2)
217.7(2)
156 8
136 3
108. 1
95.9
EFFECT HT
(M)
378 2(2)
249 5(2)
206.6(2)
149 4
120 8
103 6
92 3
77 9
69 3
EFFECT HT
(M)
115 6
99.5
88 7
75 3
67.3
58.0
51 1
48 4
45 7
EFFECT HT
(M)
337 8(2)
224 2(2)
186 4
135.9
110 7
95.6
85 5
72 8
65.3
5« 6
SO 1
47 6
45 t
42 6
EFFECT HT
(M)
91 9
87 8
84 7
80 2
76 9
EFFECT HT
(M)
83 3
78 8
76 3
73 5
69 8
Y
(II THE DISTANCE TO THE POINT OF MAXIMUM CONCENTRATION IS SO GREAT THAT THE SAME STABILIT
10 PERSIST LONG ENOUGH FOR THE PLUME TO TRAVEL THIS FAR
(21 THE PLUME IS OF SUFFICIENT HEIGHT THAT EXTREME CAUTION SHOULD BE USED IN INTERPRETING THIS COMPUTATION AS THIS
'INFLUENCE ""* "'* N°T EX'ST T° ™'S MEIGHT 4LSO WINO SPEED """"IONS WITH HEIGHT MAY EXERT A DOMINATING
131 N° f°OMUT«E'SOURCE *TTEMPTED FOB THIS HE!GHT *5 THE POINT °F MAXIMUM CONCENTRATION IS GREATER THAN 100 KILOMETERS
Figure 6-7. PLPLU model run for both stacks
at the oil refinery.
6-9
-------
Review Exercise
The terrain that surrounds the oil refinery in terms of the
stacks can be described as
a. fiat.
b. complex.
c. mountainous.
d. indeterminate.
e. land-to-sea interface.
2. The oil refinery plans to
a. build a new office building.
b. build a new stack.
c. build three new stacks.
d. renovate the existing stack.
e. renovate a barbeque pit.
1. a. flat.
The air quality model chosen to estimate ground-level
concentrations around the refinery was
a. DIFKIN.
b. COM.
c. CRSTER.
d. APRAC-IA.
e. PTPLU.
2. b. build a new stack.
4. The reason for first screening the oil refinery, rather
than using a refined model analysis, was that
a. the oil refinery did not wish to obtain an accurate
answer.
b. the Guideline on Air Quality Models allows screening
first.
c. refined models are not available.
d. the oil refinery president tossed a coin.
e. screening provides a more precise analysis than a refined
model. ,
3. e. PTPLU.
The maximum 24-hour SOZ concentration from the new
stack at the refinery was
a.
b.
c.
d.
e.
87
9
1 /ig/m3.
3 /ig/m3.
91 /tg/m*.
4. b. the Guideline on Air
Quality Models allows
screening first.
5. d. 3 jig/m3.
6-10
-------
Lesson 2
Iron-casting Plant
Lesson Goal and Objectives
Goal
To familiarize you with an actual case of air quality modeling
of an iron-casting plant and interpreting the results of that
modeling.
Objectives
Upon completing this lesson, you should be able to:
1. describe the terrain features around the iron-casting
plant.
2. describe the new proposed construction.
3. name the air quality model chosen.
4. explain why the company used a refined model analysis.
5. identify the results of the modeling analysis.
Introduction
The second case study to be examined is an iron company in
northeast Michigan. It has been at its present location for 40
years. Renovations have taken place over the years, but no
projects to build or change the facilities are planned. However,
if modeling demonstrates that the plant is responsible for high
total suspended particulate (TSP) concentrations downwind,
then additional air pollution control equipment will be
required. The State of Michigan requested an air quality
demonstration to determine the impact of emissions on the
urban area where violations of the TSP National Ambient Air
Quality Standards (NAAQS) have been measured.
Description of Area
The northeastern section of Michigan can be described as
having gently rolling terrain with elevations not exceeding 6
meters surrounding the plant. An urban area is located mostly
to the south and west of the plant. The urban area has a
population of approximately 125,000. The area around the
urban center is forested farm land. A major river runs south to
north by the western edge of the company.
6-11
-------
Description of the Plant Site
The company is located next to the river in a flat area. The
area within 5 kilometers is essentially flat. Two casting plants
are located about 1 kilometer apart along the river (Figure
6-8). The buildings are not taller than 24 meters. The grey
iron-casting plant is approximately 305 meters long and 305
meters wide. The nodular iron-casting plant is 30 meters long
and 245 meters wide. A typical iron-melting furnace is shown
in Figure 6-9. The plant has 14 stacks (see Table 6-3) for the
source inventory). The company built nine of the stacks to 70
meters to minimize ground-level concentrations. These are
termed tall stacks. Air pollution controls were installed on the
five smaller (24 to 51 meter) stacks to minimize the ground-
level concentrations.
Figure 6-9. Iron-melting furnace.
Iron-casting
company
Figure 6-8. Iron-casting company.
6-12
-------
Table 6-3. Iron-catting plant eminiona inventory.
Source
identification
G.I.C.P.
Cupola A- 1
Cupola B-2
Cupola C-3
Cupola D-4
Cupola E-5
Cupola G-6
Cupola K-7
Cupola L-8
Cupola H-9
N.I.C.P.
Cupola 1
Cupola 2
Cupola S
Cupola 4
Cupola 5
E
coordinate
(m)
263866
263852
263853
263877
263866
263856
263893
263911
263911
264549
264572
264597
264623
264665
M
coordinate
(m)
481469S
4814717
4814725
4814743
4814750
4814782
4814868
4814929
4814929
4815695
4815695
4815695
4815695
4815695
Paniculate
eminion rate
3245.76
158
158
3245.76
158
3965.98
158
0
100
2419.2
2419.2
2419.2
2419.2
394.1
Suck
height
(m)
70.104
51.21
51.21
70.104
70.104
70.104
51.21
24.38
24.38
70.104
70.104
70.104
70.104
70.104
Temperature
(K)
316.3
321.89
321.89
316.3
316.3
316.3
321.89
505.2
505.2
310.78
310.78
310.78
310.78
310.78
Inner diameter
at top
(m)
1.524
1.37
1.37
1.524
1.524
1.524
1.37
2@2.03 x 12.8(a)
2@2.03 X 12.8(a)
0.91
0.91
0.91
0.91
1.524
Exit
velocity
(m/i)
22.76
17.34
17.34
7.85
17.34
24.1
18.53
0.54
0.54
15.78
15.78
15.78
15.78
19.7
Meteorology of the Area
Michigan is influenced by cold, dry arctic air masses in the
winter and warm, moist Gulf of Mexico air masses in the sum-
mer. A tremendous amount of influence is exerted on the area
by Lake Michigan, Lake Superior, and Lake Huron. The lakes
store vast amounts of heat. They release heat and moisture
into arctic air as the air passes over them. The resulting con-
vection causes large amounts of precipitation (snow in winter
and rain in the summer) to fall. The winter winds are
predominantly from the northwest, turning southwesterly after
the passage of a storm system. The summer winds are generally
southerly.
The company obtained a computerized tape of hourly
meteorological data recorded at the airport closest to the plant.
Since the terrain is flat, the winds, temperatures, atmospheric
stability, and mixing heights are similar at both locations. The
meteorological data were preprocessed.
Model Selection and Application
The objective of this modeling analysis was to determine the
contribution of the iron company's emissions to high TSP levels
measured on the adjacent urban area. A screening model was
not used first in this case because of the large number of very
different sources of paniculate emissions. The PTPLU model,
which estimates the maximum concentration under various
meteorological conditions for a single point source, was inap-
propriate for a situation with 14 different stacks, each having a
maximum impact at a different location downwind. Because of
the uncertainty over where the combined maximum impact
6-13
-------
might be, a refined model was used with a closely spaced grid
receptor network. The recommended multiple point source
model for urban areas is RAM. The RAM model was used in
this case with a receptor grid measuring 3.3 km wide by 3.9
km long, containing receptor points spaced only 300 km apart
(Figure 6-10). A year of preprocessed meteorological data and
the emissions inventory shown in Table 6-3 were used as inputs
for the model.
t
9
!
3
3
T>
n
1
r
IS 26 39 52
12
11
10
9
8
7
6
5
4
3
2
1
25
24
23
22
21
20
19
18
17
16
15
14
38
37
36
35
34
33
32
31
30
29
28
27
51
50
49
48
47
46
45
44
43
42
41
40
3.3 kilometers
65 78
64
63
62
61
60
59
58
57
56
55
54
53
77
76
75
74
73
72
71
70
69
68
67
66
91 104 117 130
90
89
88
87
86
85
84
83
82
81
80
79
103
102
101
100
99
98
97
96
95
94
93
92
110
115
114
113
112
111
110
109
108
107
106
105
129
128
127
126
125
124
123
122
121
120
119
118
43
142
141
140
139
138
137
136
135
134
133
132
131
•jSOOmf
Figure 6-10. Receptor grid used in the RAM model.
6-14
-------
The modeling results shown in Figure 6-11 give the five
highest 24-hour TSP concentrations at each of the receptor
sites. The highest, second-highest concentration is the largest
value in the second column of results in Figure 6-11, namely
352 /tg/ms at receptor number 73, which is 300 m due north of
the company. Since the 24-hour NAAQS for TSP (secondary
standard) is 150 /tg/m8 (see Unit 2/Lesson 1), the impact from
the casting plants alone is sufficient to violate the NAAQS for
TSP in the nearby urban area. Based on these modeling
results, the iron company subsequently negotiated a TSP con-
trol strategy with the State of Michigan that involved retrofit
application of control technology.
FIVE HIGHEST 24-HOUR CONCENTRATIONS (JULIAN DAY MAX OCCURS)
(MICROGRAMS/M**3)
RECEPTOR NO.
1
2
3
4
5
6
7
8
9
10
1
28.94(133.)
56.50(154.)
59.25(192.)
41.47(192.)
86.72(147.)
52.53(147.)
62.62(134.)
40.89(134.)
23.09(203.)
16.74(250.)
2
22.02(154.)
42.74(129.)
41.44(107.)
24.58(107.)
50.94(221.)
50.35( 84.)
44.00( 62.)
35.55(203.)
21.94(174.)
15.66(286.)
5
18.20(109.)
23.71(232.)
23.43(146.)
19.39(106.)
30.00(161.)
34.72(258.)
25.64(151.)
19.40(258.)
17.97( 62.)
12.31(179.)
MEAN CONCENTRATION
42
02
48
43
60
67
20
83
49
1.15
72
73
74
75
76
77
184.74(348.)
415.55(215.)
184.39(205.)
32.00(205.)
36.27(210.)
30.84(145.)
158.43(236.)
351.57(217.)
103.52(294.)
50.29(210.)
34.70(175.)
30.57(347.)
142.10(248.)
293.21(200.)
81.82(272.)
46.23(214.)
33.71(214.)
26.87(181.)
30.15
55.80
17.40
8.61
5.48
3.89
Figure 6-11. RAM model output.
6-15
-------
Review Exercise
1. The terrain surrounding the iron-casting plant can be
described as
a. complex.
b. hilly.
c. orographic.
d. flat.
e. rolling.
2.
The iron company plans to
a. renovate all facilities.
b. build new stacks.
c. tear down the tallest stacks.
d. paint the buildings.
1. d. flat.
e.
possibly add air pollution control equipment.
A refined model was chosen to estimate air pollution
impact around the iron company because
a. screening techniques are considered childish.
b. it has a large number of different emission sources.
c. iron-casting plants may not run screening models.
d. the State of Michigan prefers refined techniques in all
permit cases.
e. the iron company does not have a screening model.
2. e. possibly add air pollution
control equipment.
What was the highest, second-highest 24-hour TSP concen-
tration in this case study?
3. b. it has a large number of
different emission sources.
a.
b.
c.
d.
e.
150
352 ng/ma
80 ^g/ms
400 /tg/ms
1200 /ig/ms
Which air quality model was chosen for the
iron-casting plant?
a. RAM urban
b. RAM rural
c. CRSTER
d. PTPLU
e. VALLEY
4. b. 352 /tg/ms
5. a. RAM urban
6-16
-------
Unit?
Summary
7-1
-------
Slide/Tape Presentation
Cassette no. 3 and keyed slides 3-1 through 3-18.
Unit Goal and Objectives
Goal
To review the major topics about Gaussian point source model-
ing covered in this course.
Objectives
Upon completing this unit, you should be able to:
1. name the two legal documents that require air quality
modeling.
2. list the three air quality programs that require air quality
modeling.
3. list the three basic types of input data to air quality
models.
4. classify the Gaussian plume model.
5. identify the name of the computerized tape package that
makes the air quality models available.
6. identify the screening models discussed that are available
on the computerized tape package.
7. identify the refined models discussed that are available on
the computerized tape package.
8. identify the reason for discussing case studies.
7-3
-------
Air Quality Modeling
Summary
Slide no.
1. Title slide—no narrative
Script
2. In this course we've looked at dispersion modeling in
detail. Here, we will summarize the major points developed
throughout this course. Air quality models such as the
Gaussian plume point dispersion models are used not only
to identify and evaluate existing industrial and urban air
pollution problems, but also to predict future problems,
and, therefore, to help avoid them.
3. The U.S. Environmental Protection Agency has approved
Gaussian plume point dispersion models for use in
regulatory applications.
4. Congress initially passed the Clean Air Act in 1970. It was
subsequently amended in 1977. The amendments required
EPA to conduct a conference on air quality modeling, and
they required the promulgation of regulations that
specified air quality models applicable to the Prevention of
Significant Deterioration program.
5. The Clean Air Act and the Code of Federal Regulations by
themselves do not specify when air quality models will be
required, or how they will be used. To recommend
specific air quality models, the Guideline on Air Quality
Models was published in April, 1978.
Selected visuals'"
Air Quality Modeling
Summary
Gaussian Plume Point
Source Dispersion Models
Clean Ail
Act of
1977
€S^
40CFR
! Code
of Federal
Regulation*
* Illustrations included here, no live shots included.
7-4
-------
Slide no.
Script
Selected visuals
6. Three programs evolved from the regulations that require
the application of models in specific cases where the air
quality may be in question: Prevention of Significant
Deterioration, New Source Review, and Control Strategy
Evaluation. The PSD program was established to limit the
deterioration in ambient air quality beyond that existing
on a specific baseline date. A permit review process uses
modeling to evaluate whether or not potential emissions
from a new source will cause or contribute to a violation
of the National Ambient Air Quality Standards, or exceed
the PSD increments. And, modeling can identify and
evaluate the control strategy required to solve industrial
and urban air pollution problems.
7. An air quality model can be characterized and classified
so that a modeler can choose the proper one for each
specific situation.
Prevention of Significant Deterioration
New Source Review
Control Strategy Evaluations
Characteristics
and Classifications
8. Three basic types of data are input into air quality models:
source factors, site factors, and meteorological factors.
Source factors are related to the location and
characteristics of pollutant emission sources. Site factors
represent the effects of terrain on dispersion and the
location of sensitive receptors. Meteorological factors
include all of the parameters that define transport and
dispersion of pollutant mass, such as wind speed, wind
direction, stability class, and mixing height.
9. Air quality models can be classified as being empirical,
semi-empirical, or numerical. The Gaussian point source
models we have discussed in this course are semi-empirical.
That is, they are derived from scientific principles, such as
conservation of mass, but they also rely on empirically
defined parameters, such as the dispersion rates sigma y
and sigma z.
10. Each model is also designed for a specific distance scale,
such as the Regional Scale. For instance, the distance
between two cities affected by a source may be 200
kilometers.
Basic Model Inputs
• Source Factors
• Site Factors
• Meteorological Factors
Semi-empirical Models
Regional Scale of Models
7-5
-------
Slide no.
Script
Selected visuals
11. An air quality model such as RAM may cover only a
smaller distance, namely 50 kilometers. Therefore, another
model should be chosen to interpret data from this situation.
12. Gaussian dispersion models are available on EPA's
UNAMAP Series, the User's Network of Applied Models of
Air Pollution. These models, available for public use, are
updated periodically. They are also used as a base for
many current regulatory procedures.
13. Some of the Gaussian plume point source models currently
available on UNAMAP are used for screening. These
include the PTXXX series, PTPLU, and VALLEY.
Screening eliminates, with little effort, sources that clearly
will not cause or contribute to a violation of the ambient
air standards. Screening models require only limited input
data and make a number of "worst-case" assumptions.
However, some models use more refined techniques, such
as CRSTER, MPTER, and RAM. Refined models
require more data than screening models. They use actual
meteorological episodes, and actual source positions and
characteristics to assess potential air quality violations.
14. Special adjustments can also be added to the models that
make them more useful. These include receptor siting,
roughness factors, and terrain adjustments. A model will
be run differently depending on the adjustment made.
15. Two case studies were examined in some detail in this
course. The cases, an oil refinery and an iron-casting
plant, demonstrated practical applications of air quality
modeling.
16. In summary, we rely on air quality models to relate the
release of air pollutants from sources to the corresponding
concentrations of pollutants in the ambient air. These data
can help predict the changes in air quality for either the
present or for future years.
-200km-
o-RAM—o
UNAMAP
T«mln
Ad|u«amt
7-6
-------
Slide no. Script Selected visuals
17. Credit: Crew Air Quality Modeling
Summary
Technical Content: Peter Goldberg
Don Billiard
Deeign: Marilyn Petereon
fflmtratloo.: Katfay Ward
Photography/Audio: David Churchill
Narration: Rick Palmer
18. Credit: EPA/NET Contract Lecture development
and production by:
Northrop Services Inc.
under
EPA Contract No. 68-02-3573
19. Credit: NET Northrop
Environmental
training
7-7
-------
Review Exercise
1. The two documents that require air quality modeling by
law are the
a. Constitution and Bill of Rights.
b. Magna Carta and 14th Amendment to the Constitution.
c. Clean Air Act Amendments of 1977 and the Code of
Federal Regulations.
d. States Rights and 17th Amendment to the Constitution.
2. List the three air quality programs that require air
quality modeling.
1. c. Clean Air Act Amend-
ments of 1977 and the Code
of Federal Regulations.
3. The Gaussian plume model is a(n)
model.
2. New Source Review,
Prevention of Significant
Deterioration, and
Control Strategy Evaluation
4. List the three basic types of input data to air quality
models.
3. semi-empirical
5. The name of the computerized tape package that contains
the air quality models is
a. ASCII.
b. UNAMAP.
c. EBCDIC.
d. none of the above
4. source factors,
site factors, and
meteorological factors
6. The computer tape package contains screening models.
The ones discussed in this course are
a. COM, PTDIS, PTMAX, PTMTP, VALLEY.
b. CRSTER, PTDIS, PTMAX, PTMTP, PTPLU.
c. PTMAX, PTDIS, PTMTP, PTPLU, VALLEY.
d. PTMAX, PTDIS, PTMTP, PTPLU, PAL.
5. b. UNAMAP.
The computer tape package contains refined
models. The ones discussed in this course are
a. CRSTER, MPTER, RAM.
b. CRSTER, MPTER, RAM, ISC.
c. COM, MPTER, VALLEY.
d. CRSTER, COM, RAM, MPTER.
6. c. PTMAX, PTDIS,
PTMTP, PTPLU, VALLEY.
8. True or False? The reason for discussing case studies in
detail is that it demonstrates practical applications of air
quality models.
7. a. CRSTER, MPTER, RAM.
8. True
7-8
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1.
4.
7.
9.
REPORT NO. 2. 3. RECI
EPA 450/2-82-007
TITLE AND SUBTITLE 5. REPO
APTI Course SI: 4 10 Ma
Introduction to Dispersion Modeling e. PERF
AUTHOR(S) 8. PERF
Peter Guldberg, Donald R. Bullard, Marilyn Peterson
PERFORMING ORGANIZATION NAME AND ADDRESS 10. PRO
Northrop Services, Inc. Bl
P.O. Box 12313 11-CON
Research Triangle Park, NC 27709 6g
12. SPONSORING AGENCY NAME AND ADDRESS 13. TYP
U.S. Environmental Protection Agency Self
15
Manpower and Technical Information Branch 14. SPO
Research Triangle Park, NC 27711
'lENT'S ACCESSIOWNO.
RT DATE
irch, 1983
ORMING ORGANIZATION CODE
ORMING ORGANIZATION REPORT NO.
GRAM ELEMENT NO.
8A2C
TRACT/GRANT NO.
-02-2374
E OF REPORT AND PERIOD COVERED
"-instructional Guidebook
NSORING AGENCY CODE
. SUPPLEMENTARY NOTES
EPA Project Officer for this Student Guidebook is R. E. Townsend,
EPA-ERC, MC 20, Research Triangle Park, NC 27711
16. ABSTRACT
The Student Guidebook is to be used in taking APTI Course SI: 4 10, "Introduction
to Dispersion Modeling." This Guidebook directs the students progress
through the course material. This Guidebook will assist the student in
learning about dispersion modeling.
This Guidebook is intended for use in conjunction with slide/tape
presentations and other readings. The only required reading is the
Guideline on Air Quality Models, EPA 450/2-78-027.
17.
18.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS b. IDENTIFIERS/OPEN ENO£
Air Quality Modeling Self-instructioiu
Self-instructional Guidebook Guidebook
Dispersion
DISTRIBUTION STATEMENT Unlimited 19. SECURITY CLASS (This I
Available from National Technical Unclassified
Information Service, 5285 Port Royal Rd. 20. SECURITY CLASS ^u;
Snrinsfield. VA 22161 Unclassified
D TERMS c. COSATi Field/Group
il 13B
51
68A
Report) 21. NO. OF PAGES
150
>age) 22, PRICE
EPA Form 2220-1 (9-73)
7-9
------- |