EPA-450/3-75-080-a
November 1975
AIR QUALITY ANALYSIS
WORKSHOP
VOLUME I - MANUAL
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Waste Management
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
-------
EPA-450/3-75-080-a
AIR QUALITY ANALYSIS
WORKSHOP
VOLUME I - MANUAL
by
R. R. Cirillo, J. F. Tschanz, A. E. Smith,
R. F. Freeman, J. E. Camaioni, andV. Rabl
Argonne National Laboratory
Argonne, Illinois 60439
Interagency Agreement No. EPA-IAG-D6-0902
Project No. F 52047
Program Element No. 2AC 129
EPA Project Officer: David C. Sanchez
Prepared for
ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Waste Management
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
November 1975
-------
XX.
This report is issued by the Environmental Protection Agency to report
technical data of interest to a limited number of readers. Copies are
available free of charge to Federal employees, current contractors
and grantees, and nonprofit organizations - as supplies permit - from
the Air Pollution Technical Information Center, Environmental Protection
Agency, Research Triangle Park, North Carolina 27711; or, for a
fee, from the National Technical Information Service, 5285 Port Royal
Road, Springfield, Virginia 22161.
This report was furnished to the Environmental Protection Agency
by Argonne National Laboratory, Argonne, Illinois 60439, in fulfillment
of Interag«ncy Agreement No. EPA-IAG-D6-0902. The contents of
this report are reproduced herein as received from Argonne National
Laboratory. The opinions, findings, and conclusions expressed are
those of the author and not necessarily those of the Environmental
Protection Agency. Mention of company or product names is not to
be considered as an endorsement by the Environmental Protection
Agency.
Publication No. EPA-450/3-75-080-a
-------
-OCX,
TABLE OF CONTENTS
1. INTRODUCTION 1
1.1 SCOPE AND OBJECTIVES 1
1.2 AIR QUALITY ANALYSIS OVERVIEW 2
1.3 EXAMPLE COUNTY DESCRIPTION 4
1.3.1 Existing Conditions 4
1.3.2 Projected Development 6
2. DETERMINING LEVEL OF ANALYSIS DETAIL 8
2.1 OVERVIEW 8
2.2 PREASSESSMENT 10
2.2.1 Gather Minimal Data 10
2.2.2 Estimate Extent of Problem 11
2.2.3 Indicate Ideal Level of Detail 12
2.3 EVALUATE RESOURCES 12
2.4 EVALUATE DATA BASES 13
3. DEVELOPMENT OF BASELINE DATA 14
3.1 AIR QUALITY DATA 14
3.1.1 Uses of Data 14
3.1.2 Estimation Methods and Averaging Times .... 14
3.1.3 Spatial Distribution of Sites 17
3.1.4 Time Distribution of Data 18
3.1.5 Evaluation of Data 18
3.1.6 Method of Measurement 19
3.1.7 Sources of Data 19
3.1.8 Illustration of County X Data 21
3.2 METEOROLOGY 30
3.2.1 Uses of Data 30
3.2.2 Data Required 30
3.2.3 Worst Case Data 33
3.2.4 Representativeness of Data 35
3.2.5 Illustration from County X Data 36
-------
TABLE OF CONTENTS (CONTD.)
3.3 EMISSION INVENTORY 36
3.3.1 Definitions 39
3.3.2 Data Required 39
3.3.3 Sources of Information 40
3.3.4 Special Considerations 42
3.3.5 Updating Procedures 44
3.3.6 Illustration of County X 51
4. ESTIMATING FUTURE EMISSIONS 63
4.1 PROJECTION REQUIREMENTS 63
4.2 PROJECTED VARIABLES 65
4.3 ACTIVITY SCENARIOS 67
4.4 SOURCES OF DATA 68
4.4.1 HUD 701 Planning 69
4.4.2 FHWA 3-C Planning 70
4.4.3 EPA 208 Planning 70
4.4.4 CZM Planning 71
4.4.5 OBERS Growth Projections 71
4.5 PROJECTION METHODOLOGIES 72
4.6 ESTIMATING SOURCE CONTRIBUTIONS 80
4.7 FEDERAL NEW SOURCE PERFORMANCE STANDARDS 86
4.8 SUMMARY OF COUNTY X DATA 87
5. ALLOCATION OF EMISSIONS 105
5.1 DETERMINATION OF GEOGRAPHIC SCALE 105
5.2 ALLOCATION PARAMETERS 107
5.2.1 Population 107
5.2.2 Transportation 109
5.2.3 Commercial/Institutional, Industrial,
and Electric Generation 109
5.3 ALLOCATION PROCEDURES 110
5.4 MASTER GRIDDING 118
-------
TABLE OF CONTENTS (CONTD.)
Page
APPENDIX E - COUNTY X FUEL USE DATA 283
*
APPENDIX F - FUGITIVE DUST CALCULATIONS 292
APPENDIX G - MATHEMATICAL DESCRIPTION OF EMISSION PROJECTION
AND SUBCOUNTY ALLOCATION PROCEDURES 300
APPENDIX H - MASTER GRID MAPPING PROGRAM 313
APPENDIX I - PROBLEM SOLUTIONS 327
ACKNOWLEDGMENTS 381
REFERENCES 382
-------
LIST OF FIGURES
No. Title Page
1-1
1-2
2-1
3-1
3-2
3-3
3-4
3-5
3-6
3-7
3-8
3-9
3-10
3-11
3-12
3-13
3-14
3-15
3-16
4-1
4-2
4-3
Fulton County Georgia. - County X
Determining the Required Level of Analysis Detail ....
Annual SAROAD Frequency Distribution for TSP Data ....
Use of Larsen's Method to Estimate Air Quality
SAROAD Frequency Distribution by Quarter for TSP Data . . .
Isopleths (m x 102) of Mean Annual Afternoon Mixing Heights .
Portion of Stability Wind Rose for Class C Stability . . .
Location of Airport Used for Meteorological Data
Update Procedures for Industrial Process Sources
Update Procedures for Fuel Combustion Sources
Update Procedures for Highway Vehicles
Update Procedures for Electric Generation Sources ....
Update Procedures for Incineration Sources
Update Procedures for Miscellaneous Sources
Baseline County Fuel Use, Table 2.1 from Ref. 7
Sulfur and Ash Content of Coal and Heating Oil,
Table 2.3 from Ref. 7
Apportionment of State Heating Oil Sales Totals to
Consumer Categories, Table 2.4 from Ref. 7
Baseline State Fuel Use, Table 2.5 from Ref. 7
Air Quality Analysis Time Period Requirements
Use of Surrogate Variables ,
Emission Projection Procedures for Industrial Process
Sources
3
5
9
22
23
28
32
37
38
45
46
47
48
49
50
58
59
60
61
64
66
74
-------
LIST OF FIGURES (CONTD.)
No. Title
4-4
4-5
4-6
4-7
4-8
5-1
5-2
5-3
5-4
5-5
5-6
5-7
6-1
6-2
6-3
6-4
6-5
6-6
6-7
Emission Projection Procedures for
Fuel Combustion Sources
Emission Projection Procedures for Highway Vehicles ....
Emission Projection Procedures for
Electric Generation Sources
Emission Projection Procedures for
Incineration Sources
Emission Projection Procedures for
Miscellaneous Sources
Point Source Industrial Process Emissions, Table 3.4-1
from Ref . 13
Industrial Point and New Source Process Emissions by
Process Category, Table 3.4-2 from Ref. 13
Process Emissions by Process Category and Subarea,
Table 3.4-3 from Ref. 13
Industrial Point and New Source Emissions - Subarea
Summary, Table 3.4-4 from Ref. 13
Possible Displays of Spatial Emission Patterns
Master Grid System for County X
Portion of County X for Gridding Problem 5-2
Sample AQDM Output
Sample SYMAP Output
Sample AQDM Culpability List
Regression Analysis for Model Validation
AQEM Regression Analysis for County X
County X Particulate Air Quality in 1975
Under Base Conditions
County X Particulate Air Quality in 1980
Under NSPS Only
75
76
77
78
79
114
115
116
117
119
121
123
131
132
133
135
138
141
142
-------
LIST OF FIGURES (CONTD.)
No. Title Page
6-8
7-1
7-2
7-3
7-4
7-5
7-6
8-1
8-2
8-3
8-4
8-5
8-6a
8-6b
8-7
8-8
8-9
8-10
8-11
County JC Particulate Air Quality in 1985
Under NSPS Only
Possible Confidence Analysis for Air Quality Calculations .
Isopleth of Localized Hot Spot Problem
Isopleth of Widespread Problem
Isopleth of Combined Hot Spot and Widespread Problem
Boundary Problems in an Air Quality Analysis
Temporal Extent of Air Quality Problems
Screening of Strategies for Detailed Evaluation
County X Isopleths and Single Source Footprint
for Problem 8-5
Computed Receptor Concentrations Using Linear
Programming Solution for Emission Density Zoning
Computed Emission Densities Using Linear Programming
Solution for Emission Density Zoning
Emission Density Zoning Solution Generated by
Linear Programming
Existing Emission Densities
Adjusted Emission Densities
County X Particulate Air Quality for Compliance with
Existing Regulations in 1975
County X Particulate Air Quality for Full Retrofit
Strategy in 1980
County X Particulate Air Quality for Full Retrofit
Strategy in 1985
County X Particulate Air Quality for Selective Retrofit
Strategy in 1980
County X Particulate Air Quality for Selective Retrofit
Strategy in 1985
143
152
154
155
156
158
159
176
187
191
192
194
195
195
198
201
202
204
205
-------
XXX.
LIST OF FIGURES (CONTD.)
No. Title
8-12
8-13
8-14
8-15
9-1
A-l
A-2
A-3
A-4
A-5
A-6
A-7
B-l
B-2
B-3
B-4
B-5
C-l
C-2
C-3
C-4
County X Particulate Air Quality for Emission Density
Zoning at 3000 T/km2/yr in 1985
County X Particulate Air Quality for Emission Density
Zoning at 1500 T/km2/yr in 1935
County X Particulate Air' Quality for Emission Density
Zoning at 500 T/km2/yr in 1935
Air Quality Impact of a Single Source at 100 T/day in
County X
Outline of Strategy Selection Process
SAROAD Yearly Frequency Distribution
SAROAD Quarterly Frequency Distribution
SAROAD Yearly Report by Quarters
SAROAD Inventory by Site
SAROAD Inventory by Pollutant within State
SAROAD Listing for Data with Averaging Times Greater
Than or Equal to 24 Hours or Composite Data
SAROAD Listing Comparing Data to Standards
NEDS Point Source Listing
NEDS Area Source Report
NEDS Stationary Source Fuel Summary
NEDS SCC Emissions Report
NEDS Annual Fuel Summary Report
The State: 1973 and 1972
Counties : 1973 - Fulton County
Populat ion- States : 1960 to 1973
Occupancy, Utilization, and Plumbing Characteristics
for the State: 1970
209
210
211
212
215
237
238
239
240
241
242
243
246
247
248
249
250
252
253
261
262
-------
LIST OF FIGURES (CONTD.)
No. Title Page
C-5
C-6
C-7
D-l
D-2
D-3
E-l
E-2
E-3
E-4
E-5
E-6
E-7
E-8
F-l
F-2
H-l
H-2
H-3
H-4
Fuels and Appliances for the State: 1970
Occupancy, Utilization, and Plumbing Characteristics
for Counties: 1970
Fuels and Appliances for Counties: 1970
Alternative Transportation Direction I,
Highway Orientation
Alternative Transportation Direction II,
Transit Orientation
Alternative Transportation Direction III,
No New Highways and Adopted Transit
Sales of Kerosene in the United States
Sales of Distillate -type Heating Oils in the United States
Sales of Residual-type Heating Oils in the United States
Sales of Distillate-type and Residual-type Fuel Oils
for Industrial Use (Excluding Oil Company Use) in the
United States
Sales of Distillate-type and Residual-type Fuel Oils
for Use by Oil Companies in the United States
Sales of Distillate-type and Residual-type Fuel Oils
for Use by the Military in the United States
Distribution of Bituminous Coal and Lignite
Ovuantity and Value of Natural Gas Delivered to Consumers
Map of Precipitation Frequency
Map of PE Values for State Climatic Divisions
Grid Program Overview
Subroutine MAIN Flowchart
Subroutine OUTPUT Flowchart
Subroutines CENPEN and MAKARR Flowchart
263
264
265
272
273
275
284
285
286
287
288
289
290
291
294
296
315
316
317
318
-------
XX.U
LIST OF FIGURES (CONTD.)
No. Title Page
H-5 Subroutine DENSIT Flowchart 319
H-6 Card Input 320
H-7 Program Listing 321
-------
LIST OF TABLES
No. Title Page
3-1
3-2
3-3
3-4
4-1
4-2
4-3
4-4
6-1
6-2
6-3
6-4
6-5
7-1
7-2
8-1
8-2
8-3
8-4
8-5
8-6
National Ambient Air Quality Standards
Emission Source Categories
Emission Source Data Required .
County X Baseline Emissions, 1973
County X Participate Emission Projections
County X Projected Emissions, 1975
County X Projected Emissions, 1980
County X Projected Emissions, 1985
Multi -Source Atmospheric Dispersion Models
Summary of Simulation Model Characteristics
Models Applicable to Specific Pollutants and
Averaging Times
Range of Background Concentrations
Composite of Computed Air Quality for County X from AQDM .
Source Contribution Analysis for County X
Detailed Breakdown of Area Source Contributions
to Calculated Air Quality at Receptor 151
Land Use Control Measures and Implementation Instruments
Land Use Implementation Instruments
Industrial Process Emissions for Retrofit of
Existing Sources to NSPS
Industrial Process Emissions for Selective Retrofit
of Existing Sources to NSPS
Calculated Emission Densities Under Base Conditions
for 1975, 1980, and 1985
Required Emission Reduction to Attain 500 ton/km2 /yr
Zoning Regulation
15
41
42
52
88
89
94
99
126
128
129
139
144
161
162
170
171
199
199
207
207
-------
xux,
LIST OF TABLES (CONTD.)
No. Title
9-1
9-2
9-3
9-4
9-5
9-6
9-7
D-l
D-2
D-3
D-4
D-5
D-6
D-7
D-8
D-9
D-10
F-l
Cost Elements Involved in County X Strategies
Examples of Social Factors
Derivation of W^ in Klee Methodology
Consistency Check
Computation of Evaluation Scores
Summary, Ranking of Evaluation Scores
Washington Environmental Research Center (WERC) Matrix .
Preliminary Non-farm Wage and Salary Employment
Projections, 1960-2000
Preliminary Population Projections, 1960-2000
Preliminary Household Size Projections, 1950-2000 ....
Civilian Non-farm Wage and Salary Employment, 1970-2000
Regional Planning Commission Area Employment (in thousands)
Legend for Computer Printouts
Land Use Projections for 1970
Land Use Projections for 1980
Land Use Projections for 1990
County X Industrial Land Use Change Calculations
OBERS Growth Projections
Control Methods for Unpaved Roads
219
228
229
229
231
231
233
268
269
269
270
276
277
278
279
280
282
295
-------
1. INTRODUCTION
1.1 SCOPE AND OBJECTIVES
The development of an air pollution control strategy designed to
attain and maintain the National Ambient Air Quality Standards (NAAQS) re-
quires an analysis of current and possible future air quality problems.
Previous publications " by the EPA Office of Air Quality Planning and
Standards have been designed to provide guidance to regional, state, and
local air pollution control and planning groups in the development of an
14
analysis conforming to the requirements of the current federal regulations.
It is the objective of this workshop manual to illustrate, through the use of
a ficticious county, some of the quantitative and qualitative procedures used
in developing an acceptable plan.
Since it is not possible to incorporate the multitude of different
situations that might influence the design of a NAAQS attainment/maintenance
plan, this example of control strategy development is necessarily limited in
scope. The major emphasis will be on the development of measures for the
control of particulates since, of the 268 Air Quality Maintenance Areas
designated, ' all but a few have been identified as having particulate
problems. Many of the procedures demonstrated here will be equally adequate
for any of the other criteria pollutants (i.e., S02, N02, HC, CO, Oxidants)
and the similarities and differences in the analysis will be noted accordingly.
This document likewise does not attempt to address the issues of
how states or agencies affect intergovernmental cooperation, how states review
and provide for public hearings on proposed plans, or how the mechanisms of
plan submission and review are administered. The major emphasis in this vol-
ume is on the analytical procedures for the review and development of control
options. In order to achieve a measure of realism in this exercise, the data
for the ficticious county, "County X," have been based on Fulton County, Georgia.
This choice was made due to the large volume of information previously compiled
for this area. In certain instances the data have been adjusted for the
purpose of creating illustrative problems. Wherever possible, however, the
form of the data, as it exists in reality, has been preserved, while only the
numerical values have been changed.
-------
It is not the purpose of this review to present new policy state-
ments or to issue new guidelines for analytical procedures. Rather, it is
designed to illustrate existing guidelines and to clarify the application of
current policy. This document is specifically directed at providing working
level state and local air pollution control engineers and planners with this
guidance.
1.2 AIR QUALITY ANALYSIS OVERVIEW
Figure 1-1 presents a flow diagram of the Air Quality Analysis System
(AQAS) procedure that will be reviewed here. The first step is to determine
the level of analysis that is appropriate for the area under study. The ob-
jective of this step is to assure the most efficient use of available resources
in control plan development. The next step is to compile a baseline data base
consisting of air quality data, meteorological information, an air pollutant
emission inventory, and regional planning data. This will form the foundation
of the analysis. Next, growth and development information must be translated
into indices to estimate future emissions. At this point, it may be necessary
to revise the level of analysis detail to adjust to special problems identi-
fied by these estimates. The baseline and future emissions are then allocated
to portions of the study area for the purpose of improving the spatial resol-
ution of the analysis. Next, the air quality impact of the baseline and
estimated future emissions are determined through the use of one of several
available air quality simulation models. The modeling results are analyzed
to determine the type and extent of the air quality problem present in the
study area. At this point it may again be necessary to revise the level of
analysis to properly treat the identified situations. Should the analysis
indicate a NAAQS attainment and/or maintenance problem, then control strat-
egies must be developed and tested through reapplication of the simulation
model. After establishing a set of technically adequate control strategies,
a final evaluation step should be performed to determine the relative effect-
iveness of each strategy in meeting the air quality goal. This final evalua-
tion must account fdr known economic, social, legal, and institutional con-
straints as well as probable impediments to control strategy implementation.
-------
DETERMINE REQUIRED
LEVEL OF ANALYSIS
DEVELOP BASELINE
DATA BASE
ESTIMATE FUTURE
EMISSIONS
i
ALLOCATE
EMISSIONS
1
MODEL AIR
QUALITY IMPACT
ANALYZE MODEL
RESULTS
I
DEVELOP AND TEST
ALTERNATIVE CONTROL
STRATEGIES
I
EVALUATE AND SELECT
STRATEGY ON THE
BASIS OF EFFECTIVENESS
AND IMPLEMENTABILITY
Fig. 1-1. Air Quality Analysis System
-------
1.3 EXAMPLE COUNTY DESCRIPTION
1.3.1 Existin& Conditions
County X, as shown in Fig. 1-2, is an elongated county, approximately
56 miles long and 14 miles wide with the long dimension oriented in a northeast-
southwest direction. A river forms most of the western boundary, and the
topography characterized by low hills and rolling terrain is not extreme.
Urban development dominates the central third of the county, in
which the major city of the region and its contiguous suburbs are located.
Little usable vacant land exists in the central portion of the urbanized
area. The northern and southern thirds of the county are relatively sparcely
populated, and less than 5% of the land in areas farthest removed from the
center city is devoted to urban land uses (including residential, commercial,
and industrial uses).
The largest concentration of commercial activity in the county occurs
in the central business district (CBD) of the major city, with nearly 30% of
the land area there in commercial use. The CBD is very close to the geographic
center of the county. A few additional areas near the CBD also have appre-
ciable commercial development, but clusters of commercial activity large enough
to account for as much as 15% of the area of any census tract do not exist
farther than about 12 miles from the CBD.
The CBD is also the location of appreciable industrial development.
Additional industrial districts occur along railroads to the north, east, and
southwest of the CBD. One of the major concentrations of large particulate
point sources is in the industrial district about three miles north of the CBD
in which a steel rolling mill, a grey iron foundry, and a lead smelting opera-
tion are located. Northwest of the CBD, between it and the river, is another
significant industrial point source at the site of a brick and structural tile
plant. A steam plant located in the CBD and another to the northwest are also
large point sources of particulates.
Other industrial point sources beyond the boundaries of the county
contribute to the particulate air quality levels over the region. Two coal,
oil, and gas fired electric generating plants are closely spaced along the
western bank of the river to the northwest of the CBD. Seven stone quarrying
operations in neighboring counties form a ring of large particulate emitters
-------
»re«3HIE HICHMTS
mm
Fig. 1-2. Fulton County, Georgia - County X
-------
around County X. These stone quarries account for more than half of the
particulate emissions from point sources in the regional emission inventory.
One large area of predominantly residential development exists
relatively near and to the west of the CBD. Suburbs to the north of the
central city are primarily residential. Southwestern suburbs have a more
diverse mix of urban land uses, but they too are the location of large areas
of residential development.
1.3.2 Projected Development
The change in land use between the present and 1985 is typified by
withdrawal of some land from urban use in the center of the built-up area,
a significant increase in urban development in areas 8 to 10 miles from the
CBD that currently are at the edge of the urbanized area (to the north, this
rapid growth extends nearly 20 miles from the CBD), and a more modest trend
toward urbanization in the northern and southern extremes of the county.
While the total percentage of area devoted to urban land use in areas near
the CBD appears to be in decline, a more significant trend is the increasing
percentage of industrial land use in many of these same areas. Several of
the existing industrial districts, in particular the area of foundries and
smelting to the north of the CBD, show some decrease in activity, but nearby
areas to the northeast and northwest of the CBD have the largest increases in
industrial land use in the county. The moderate industrial activity that
currently exists in the southeast corner of the urbanized central third of
the county is another potential site of industrial expansion.
Residential land use more nearly follows the overall trend in urbanized
land area, with reductions nearly universal in areas within 3 or 4 miles of the
CBD. The actual locations of the growing residential areas are, in part,
determined by the transportation network in the county and changes proposed
for it. The completion of a new rail rapid transit system will result in much
development activity near the terminals of its branches. Without further
transportation investments, which is the assumption made here, this develop-
ment will be clustered more densely about the terminals. The transit branches
within the county extend only to the edges of the central third of the county,
and overall growth, correspondingly, tends to be concentrated near the fringes
of the current built-up area. Clusters of development, including sizable
-------
increases in residential land use will be found by 1985 near the river to the
west and northwest of the CBD and also approximately 12 miles to the southwest.
Existing residential areas in the suburbs to the north will continue to more
steadily northward, although these areas aren't well serviced by the new
transportation system.
In summary, the anticipated increase of about 251 in population of the
county up to 1985 will be accommodated somewhat less expansively than has
been customary for metropolitan growth in the recent past.
-------
2. DETERMINING LEVEL OF ANALYSIS DETAIL
2.1 OVERVIEW
The first step in the Air Quality Analysis System procedure is a de-
termination of the required level of analysis detail (see Fig. 1-1). Ideally,
the level of detail should be commensurate with the severity and extent of the
existing and/or projected air quality problems. The unavailability of re-
sources or data necessary to perform the analysis at the ideal level can limit
the analysis to a less detailed level. The guiding concept should be that all
analyses are to be carried out with the maximum reasonable effort.
During the analysis period it may be necessary to revise the esti-
mates of the required level of analysis detail as the extent of the problems
become more precisely defined and the limitations of available data and re-
sources become clearer as indicated in Fig. 1-1. It is not possible to, a
priori, define the detail required with complete accuracy. The level chosen
must be a balance between the ideal of complete detail and what can be done
within the available time with the available resources and data. This is often
evident only after some experience with the analysis.
The basic steps in determining the level of analysis detail are out-
lined in Fig. 2-1. Three basic inputs determine the level of analysis:
Level of analysis appropriate to the air quality problems,
Limitations due to available resources, and
Limitations due to available data.
Prior to beginning the full AQAS procedure, some preassessment of these three
inputs is necessary so that available resources can be used most efficiently.
It should be emphasized that the description of the process for
determining the level of analysis detail is not intended to imply that this
must be carried out as part of an acceptable Air Quality Maintenance Plan.
It is designed only for the purpose of assisting the states in determining
the best allocation of resources and need not be reported or referenced to
in any plan submission.
-------
GATHER MINIMAL DATA
ESTIMATE EXTENT
OF PROBLEM
t
NO PROBLEM
1
INDICATION OF
IDEAL LEVEL
OF DETAIL
I
EVALUATE RESOURCES
i
EVALUATE DATA
1
DATA AMD
RESOURCES AVAILABLE
FOR IDEAL LEVEL OF
DETAIL?
YES
ANALYSIS CAN
BE DONE AT
IDEAL LEVEL
I
NO
LEVEL OF ANALYSIS
LIMITED BY RESOURCES
NEED:
1. INCREASED FUNDING
2. HELP FROM EPA
3. CONSULTANTS
LEVEL OF ANALYSIS
LIMITED BY DATA
POSSIBLE INTERIM
PLAN WITH PROVISIONS
FOR DATA GATHERING
AND REASSESSMENT
Fig. 2-1. Determining the Required Level of Analysis Detail
-------
10
2.2 PREASSESSMENT
2.2.1 Gather Minimal Data
The preassessment phase consists of the initial completion of first
five steps in Fig. 2-1. If reevaluation of the level of detail becomes neces-
sary as the analysis progresses, the considerations given below would still
apply, but the experience already gained during the analysis could be used to
limit the reevaluation to those areas where a greater level of detail than was
initially indicated would be needed.
The first step is to assemble some minimal data base that must include
air quality data and the variables required for emissions projections. A
minimal set of air quality data is available from the EPA's Storage and Re-
trieval of Aerometric Data (SAROAD) information file. Appendix A contains
sample tabulations of the SAROAD data. Where more detailed air quality data
are available at the state and local levels, it should be used, even in the
preassessment phase, since it is important that the most severe problem be
found. Several methods of making air quality estimates are discussed in
Section 3 and may be used when actual data is unavailable.
The types of data needed to make emission projections are listed in
Reference 7 and discussed in Section 4. The preassessment phase will generally
be performed at a simple level of detail. Knowledge of local conditions may
indicate that intermediate or full detail may be necessary for certain source
categories or specific sources. If this type of a priori knowledge is avail-
able it can be used to avoid the duplication of effort involved in doing a
simple analysis to confirm a result already established by experience. At a
minimum, the emissions data from EPA's National Emission Data System (NEDS)
are available. Appendix B contains some sample tabulations. Growth projec-
tions from the Office of Business Economics and the Economic Research Service
(OBERS) are available at this stage. Appendix D contains this data for the
Metropolitan Atlanta Air Quality Control Region (AQCR). Alternatively, a
dialogue with local planning commissions may generate an insight into the growth
prospects for the study area that would be usable for this preassessment.
Local knowledge and experience can, even in this initial phase, aid
in guiding and limiting the data gathering efforts. Local air pollution con-
trol agencies will usually already have a reasonable estimate, based on
-------
11
experience, what the major problems are and whether they are likely to be due
to a few large point sources, numerous small area sources, or mobile sources.
The preassessment of air quality data can also provide some guidance at this
stage. If the data indicates severe and widespread problems, say areas more
than about 20% above allowable levels, then it is likely that a complete
analysis will be needed of suspect sources, and efforts to obtain the neces-
sary data can begin early in the analysis process.
At this point a minimal data base has been collected and even before
a formal estimate of the extent of the problem has been made, some focusing
of the data gathering efforts has been made possible based largely on local
experience. Efforts to obtain more detailed data can thus be continuing
simultaneously with the first estimates of the extent of the air quality
problem.
2.2.2 Estimate Extent of Problem
The most rudimentary type of air quality analysis is that based on
subjective judgment or generalized procedures and criteria such as those used
in the initial designation of Air Quality Maintenance Areas. During the pre-
assessment phase described here, a preliminary analysis at a simple level
but more complete and reflective of local data is carried out. What is desired
is a refinement of the indications of where problems will arise and what sources
cause them by using readily available data.
Future air quality levels may be projected by using some form of
atmospheric simulation model and rough approximations of future emission levels
from the minimal data available. (Section 6 discusses the use of atmospheric
simulation models.) If a sophisticated dispersion model (e.g., the Air Quality
Display Model) is available and the agency performing the analysis has experi-
ence in its use, then it may be desirable to perform the preassessment with
this tool. Otherwise, a simple proportional model will suffice for this
phase. (Note that the modeling requirements described by the regulations
for the actual plan are not this slack.)
The preassessment analysis to determine the ideal level of detail can
be structured by attempting to provide answers to the questions:
Is there an existing air quality problem?
Is there a maintenance (future) problem only?
-------
12
What is the magnitude of the problem?
What is the spatial extent of the problem?
What types and categories of sources are involved?
How many sources are involved?
2.2.3 Indicate Ideal Level of Detail
With at least a partial answer to some of these questions, it is
possible to determine what the most useful level of detail might be. For
example, if the preassessment shows a particular source category to be creat-
ing a problem, then some effort should be expended in developing a detailed
description of the emission patterns of sources in that category. If the
preassessment shows projected air quality to be substantially in violation
of the NAAQS, then substantial detail is required. If the preassessment shows
air quality levels close to the NAAQS, detailed analysis is required to deter-
mine whether the projected problem is real. If the problem is a spatially
localized "hot spot" problem due primarily to nearby sources, the greatest
detail is necessary for these sources. Problems occurring throughout an
analysis region require detailed information from the entire region. If prob-
lems occur in the future (beyond about 8 years) and more detailed data is
needed, the time to develop the needed data bases is available. If the pre-
assessments show no air quality problems, additional detail will still be
needed to substantiate this projection.
2.3 EVALUATE RESOURCES
The available resources must be evaluated to determine the level of
detail actually attainable. The steps are indicated in the left-hand branch
on Fig. 2-1. The resource evaluation should examine:
The state of the present AQAS effort,
Level and expertise of staffing,
Existing plans affecting air quality,
Existing air quality management programs, and
Modeling capabilities.
Each of these parameters is examined for indications that it limits the level
of detail obtainable to less than the ideal level. For example, there may
not be enough personnel to perform the analysis at the ideal level. Much of
-------
13
the required detail can already exist in the present State Implementation
Plan, transportation control plan, or within other air quality programs such
as new source review and variance programs. In other cases, the lack of
experts in fields like transportation or planning may limit the level of
detail. When the level of analysis detail is limited by resources, aid in
increasing the level of detail obtainable can come from:
An increased level of funding,
Assistance provided by the EPA Regional Office, or
The use of consultants.
If these sources can provide no help, the maximum reasonable effort with the
available resources must be put forth, knowing that the level of analysis
detail will be less than ideal.
2.4 EVALUATE DATA BASES
Air quality projections depend upon several different data bases.
The level of detail available in any base limits the level of detail at which
the analysis can be done. The level of detail of the following data bases
must be considered:
Air quality data,
Emissions inventory,
Growth factors and activity levels,
Transportation data, and
Allocation parameters.
The preassessment analysis will not use the most detailed data. It does
indicate what level of detail is ideally needed in the data bases. If com-
pletely detailed data is available or can be made available within the limi-
tations imposed by available resources, the air quality analysis can be done
at the ideal level of detail. If sufficiently c-3tailed data are not available,
a detailed analysis cannot be done. It may be possible to prepare an interim
plan using the best detail available with provision for gathering more de-
tailed information. This would, of necessity, be followed by a reassessment
of the problems. Otherwise, lack of data will limit the level of analysis
to less than ideal.
-------
14
3. DEVELOPMENT OF BASELINE DATA
The set of information required as a baseline for the development
of an air quality analysis is made up of three parts: air quality data,
meteorological information, and an emission inventory.
3.1 AIR QUALITY DATA
3.1.1 Uses of Data
Ambient air quality data are necessary for the development of attain-
ment and maintenance strategies. Comparison of measured ambient levels or
statistics computed from them with the National Ambient Air Quality Standards
(NAAQS) is the primary means of determining whether air quality goals are
being achieved. Where the NAAQS are being exceeded, the reason for the high
levels must be determined and the need for corrective action evaluated. Com-
parison of data over a period of years for a specific area is an indicator
of developing air quality trends. In areas exceeding the NAAQS, the trend
can be used as a monitor of progress in attaining the standards. In areas
where NAAQS are being met but where growth and development are occurring,
upward trends in measured concentrations may dictate a reevaluation of the
need for specific maintenance measures in addition to existing stationary
source and mobile source programs.
In addition to the identification of current attainment and mainte-
nance problems, projection of potential future problems and assessment of the
effectiveness of present control strategies are based on ambient air quality
data. When simulation models are used to either estimate present air quality
at locations where no monitors exist or to predict future levels, air quality
data should be used to calibrate the model. When proportional techniques
of predicting air quality are used, air quality data are a necessary part
of the calculations.
3.1.2 Estimation Methods and Averaging Times
In order to be compared with the NAAQS, air quality data must be
available for the averaging times specified in Table 3-1. In addition, for
averaging times less than one year, the data must include the second highest
measured value since the NAAQS specify values "not to be exceeded more than
once per year" for other than annual averages.
-------
15
Table 3-1. National Ambient Air Quality Standards
Pollutant
Particulates
Primary
Secondary
SO 2
Primary
Secondary
CO
Hydrocarbons
Oxidants
N02
Concentrations in yg/m3
Averaging Time
1-hr 3-hr 8-hr 24 -hr
260
150
365
1300
40b 10b
160C
160
1-yr
75 (G)
60a (G)
80 (A)
100 (A)
Intended as a guide to meeting the 24-hour secondary
standard (40CFR50).
CO concentrations measured in mg/m3.
C6 AM to 9 AM only. The hydrocarbon standard is
intended as a guide to achieving the oxidant standard.
(A) Arithmetic average.
(G) Geometric average.
NOTE: All averages less than 1 year are not to be exceeded
more than once per year.
When data for the averaging times specified in the NAAQS are not
available, the following situations arise:
For continuous data where the hourly averages are
available, the appropriate averages and second
highest values may be calculated.
For non-sequential data, for incomplete data sets, and
for data where only summary statistics are available,
the appropriate averages and second highs may be
estimated using Larsenrs statistical methods.17>18
-------
16
In using continuous data, References 11 and 19 may be consulted to determine
how to average data and compute second highest values. In particular, it
should be noted that:
The running average beginning at each clock hour should
be used for the short-term standards, and
The maximum second highest non-overlapping value should
be used for standards not to be exceeded more than
once per year.
It should be emphasized that only measured data can be used to evidence viola-
tions of the NAAQS; Larsen's methods can, however, be used to estimate the
severity of existing or future air quality problems.
Frequently aerometric data is only available in summary form giving
some of the following information:
Annual arithmetic mean,
Annual geometric mean,
Standard deviation,
Geometric standard deviation,
The total number of samples and the maximum measured
value, or
Two percentiles and the corresponding concentration
values.
If any pairs of these except the third and sixth or fifth and sixth is avail-
able, Larsen's methods can be used to:
Estimate the concentration likely to be exceeded
more than once per year for comparison with the
NAAQS, and/or
Use data available for one averaging time to estimate
the statistics appropriate to another averaging time.
Larseh also describes a simple graphical method that can be used to estimate
the concentrations that are exceeded with a particular frequency when the
geometric mean and geometric standard deviation are known. This graphical
technique can be used in the simplest level of analysis as a quick means
of estimating the values that are exceeded more than once per year.
-------
17
The validity of Larsen's techniques depends upon how closely the
actual data fit the model's assumption that air quality data are lognormally
distributed. Unless past experience has shown that Larsen's methods are valid
for the pollutant and in the area of interest or unless the individual mea-
sured values are available to test for lognormality, his methods must be used
with some caution. The test for lognormality can be made either graphically
by seeing whether the data fall approximately on a straight line when plotted
on leg-probability paper, or analytically by use of the statistical x2 test.
Larsen's methods were based on data taken at Continuous Air Monitoring
Program (CAMP) sites in urban locations and would be expected to be applicable
in similar locations. They should not be applied in areas dominated by single
strong point sources. If the critical assumption of lognormality cannot be
tested because only summary statistics are available, then a complete set of
individual measurements should be collected to establish the applicability of
Larsen's methods.
If the data prove not to be lognormal or if past experience has shown
Larsen's methods inappropriate, there are nonparametric statistical tests that
do not require the assumption of lognormality. Reference 19 discusses one
simple test that can be used to infer whether a certain quantile of a complete
data set would exceed a specified value when only a sample of the set is avail-
able. The assistance of a trained statistician should be sought when dealing
with these problems.
3.1.3 Spatial Distribution of Sites
The spatial distribution of sites with air quality data is also
important. EPA regulations for State Implementation Plans require a minimal
number of monitors in each Air Quality Control Region (AQCR), but this number
is usually too small to provide a detailed assessment of ambient air quality.
General guidance in assessing the coverage provided by a monitoring network
is provided in References 11 and 19. Ideally, data should be available from:
Hot spots where concentration maxima occur,
Clean areas that can be used to estimate background
concentrations,
Areas with the highest population density or total
population,
-------
18
All distinct subareas of the region of interest,
Areas where rapid growth and development are expected,
Areas projected to have the highest future
concentrations.
Comprehensive data of this type will generally be lacking and modeled estimates
of air quality at various locations will be necessary. In an urban area,
about twenty points provide reasonable confidence in a model calibration.
Frequently, however, fewer will be available and correspondingly less confi-
dence can be placed in the accuracy of modeled air quality estimates. Since
air quality maintenance considerations have not been primary in monitoring
network design, data from areas where rapid growth and development are expected
and areas projected to have the highest future concentrations are likely to
be unavailable. In addition, identification of these areas would emerge only
as the air quality analysis process progresses. Expansion of the monitoring
network to provide more comprehensive coverage is thus likely to be necessary,
especially in areas with anticipated air quality problems.
3.1.4 Time Distribution of Data
When baseline data sets are being gathered, the time period covered
and the number of observations at each site must also be considered. Data
must be available from the appropriate baseline year and all the data must
be from the same year. Since most monitoring programs are expanding, the most
recent available year of data will probably be the most comprehensive. The
air quality data year must also correspond to the emission inventory year
used to calibrate a dispersion model and make projections for attainment and
maintenance planning.
3.1.5 Evaluation of Data
Some evaluation of the data should also be done where possible. In
References 11 and 19, EPA has provided methods of evaluating the acceptability
of air quality data when the full data sets are available. For example, oxidant
data taken at night and carbon monoxide data taken during early morning hours
are likely to be unrepresentative of potential problems, since the maximum
concentrations of these pollutants generally occur at times of high insolation
and peak traffic density, respectively. Where non-sequential sampling schedules
-------
19
have been used, the number of samples can be used to estimate the precision of
20
calculated means by the methods of Hunt. Proposed amendments to EPA's require-
ments for the submission of state implementation plans allow consideration of
the accuracy and precision of data and projections in determining whether NAAQS
will be attained and maintained. The effect on projected air quality concen-
trations of factors such as the following may be considered:
Current air quality concentrations,
The magnitude of past year-to-year variations in
air quality concentrations, and
The degree of confidence in the methods used to
determine air quality concentrations resulting from
projected emissions.
The number of samples in a year's data set also indicates whether
the set can be considered complete. If, for example, there are only 2,000
hourly sulfur dioxide values in a data set for a year, the data set should
probably be rejected in accordance with the guidelines that require 75%
of the possible values to be present before statistics are computed.
3.1.6 Method of Measurement
One further consideration must be given to the data. Only data
collected by certain methods can be used as the basis for air quality analy-
sis and planning. A summary of approved, unapproved, and unacceptable methods
is given in Table 1 in Reference 11. Unapproved methods are currently employed
in many places and for SOa, CO, and oxidants. They may continue to be used
until a federal reference method has been promulgated. For methods not listed
in the table as, for example, sulfation methods of measuring sulfur dioxide,
the EPA Regional Office should be consulted concerning the acceptability of
the data.
3.1.7 Sources of Data
State and local air pollution control' agencies will usually be the
best sources of air quality data, as they generally have the most complete
and recent data sets. A portion of the state and local data is also sub-
mitted to EPA through its Regional Offices. This data is stored in the SAROAD
-------
20
(Storage and Retrieval of Aerometric Data] system from which it can be accessed
in reduced data format through the Regional Offices. Submissions to SAROAD
are filtered at both the regional and national levels. Hence, state and local
agencies usually have more detailed and comprehensive data than is available
from SAROAD. Proposed EPA regulations would require that all applicable air
quality data collected since October 1, 1972, be used in the analysis and
that such data be submitted to SAROAD so that the plan can be evaluated and
future progress assessed.
Agencies at all levels sometimes conduct special short-term sampling
programs aimed at the evaluation of specific problems. When baseline air
quality is being established, data from such programs can often provide
information supplementing the more comprehensive data obtained from other
sources.
In areas where measured air quality data is lacking, two methods are
available to estimate expected levels:
Dispersion modeling, and
Estimates based on measured air quality in similar areas.
Reference 12 discusses the air quality simulation models generally available,
their capabilities, and their data requirements. Anticipated amendments to
Appendix A Air Quality Estimation will provide explicit guidance as to which
models shall be used in developing State Implementation Plan revisions. When
measured air quality data are unavailable, the model is used uncalibrated;
that is, modeled air quality concentrations are not compared to actual measured
air quality data and hence the modeled results must be regarded as approxima-
tions until a monitoring program can be established, data collected, and the
model predictions compared with actual measured values.
If measured data are available from areas with emission densities,
meteorology, topography, and source types similar to those in the area of
interest, these data may be used as indications of expected air quality in
the area of interest. Such estimates would not account for hot spots caused
by large point sources, such as power plants. Any large sources would have
to be modeled separately. If this method is used, it is essential that con-
firmation of the estimates be made through an expanded monitoring program.
-------
21
3.1.8 Illustration of County X Data
Figure 3-1 shows one of the SAROAD printouts for total suspended
particulates from a high volume sampler site in County X. Air quality data
for County X was obtained from similar printouts which give:
The cumulative frequency distribution,
Number of observations,
Measured maximum and minimum,
Arithmetic mean (average),
Geometric mean,
Geometric standard deviation,
Various other information including site name,
UTM coordinates, and other locational information.
The NAAQS for particulates are listed on Table 3.1. Since there are two
24-hour particulate standards not to be exceeded more than once per year and
since the actual second highest value is not given in the data set, some
method must t>e used to estimate whether the short-term standards are being
attained. If the actual second highest measured value is available, it would
be sufficient to use this value in the analysis although such a procedure
usually underestimates the severity of the short-term problem if data is
available for less than about one third of the days (see Reference 11). More
than a sufficient amount of information is available on the printout to employ
Larsen's methods. The graphical and two analytical methods will be illustrated.
Graphical Method In Fig. 3-2, Larsen's methods have been used to
estimate the highest and second highest 24-hour values for comparison with
the standards by the graphical method using log-probability paper. Only the
geometric mean and the geometric standard deviation are needed to plot the
line. The geometric mean plots as the 50% point on the graph. Using the
geometric standard deviation, the value to be plotted at the 16% point (Clfi)
can be found from
C"16 = (geometric mean) X (geometric standard deviation) (3-1)
where C^ is the concentration that is expected to be exceeded 16% of the time.
These two points can be plotted and a "Larsen line" drawn through them as in
Fig. 3-2. As an aid in drawing the line, a third point may be plotted:
-------
22
o o
— O GC
000
o o z o
C3 Jl 3 -• X
•- o o u>
i -O J •• 33 H *
*• »• 2 ••
• ui -• i -* z ar *-
r a ui t- i- o o r
> »- o o < »- t-
z r z: r ui aj u.
o >- f- *- _j _j -•
i _i o 3 o .j ui a
.8
* z
z o
s
« 'T
Q t-
H- >- O
LJ VJ UJ
r z j
O uJ
X => *•
UJ "? —
U aJ
Ul —
u> ar
— »- a:
z o
• o a;
i u
t: or
> O
Z 33 t
o :D
u. — a:
• _i ce
O uJ
CL Z
i X
a: o
z — a IT ji
o a z o o
— . ^- < o --
a. - a:
•H
PH
U •"
7
r Ji
vj « ~
— -j ui
o a.
D O i
a. a.
o o
_l — _1 O
Z) O- O I
U — > I
— Ola-
h- — — M
a. ~ r
r
c
> o
-------
23
o>
o>
o>
op
o>
o>
CO
m
o>
O
CO
O
<£>
O
ID
O
to
-------
24
CR. = (geometric mean) / (geometric standard deviation) (3-2)
where CR4 is the concentration that is expected to be exceeded 84% of the time.
The measured frequency distribution should lie along this line if the
data is lognormally distributed. Since Larsen lines are usually plotted so
that a frequency-concentration pair gives the percentage of data greater than
that concentration while the SAROAD printout pairs give the percentage of
data less than the corresponding concentration, the Pth percent point on the
SAROAD printout plots as the (100 - Pth) percent point for comparison with the
Larsen line.
To estimate the two highest expected concentrations, the frequencies
for the highest and second highest values for a year must be calculated.
Larsen uses a correction term and the formula:
f = ^— - x 100% (3-3)
where f = plotting frequency (%) , r = rank order (highest, second, third, ...),
and n = the number of samples.
For a year of 24 -hour samples, n = 365. The highest expected concen-
tration would occur at a frequency of:
£highest = ^ x 100% * °'16% ™* C3'4)
the second highest or second most polluted (SMP) day would occur at a fre-
quency of:
fSMP = T"X 100% = °'44%- (3"5)
Analytical Procedures The primary annual standard of 75 yg/m3 is
apparently being met at the site tabulated on Fig. 3-1, since the measured
20
average is 58 ug/m3 . Hunt's method can be used to place a confidential
interval around this value. For lognormally distributed data, Hunt gives:
-------
25
'"" (3-6)
M2 = exp(t -2- / Cl-gD ) - 1, (3-7)
/"n
where S = In (geo. std. dev.) = In 1.72 = 0.5423, n = number of samples = 55,
N = number of possible samples = 365, t = the "t-statistic" for n-1 degrees
of freedom and the chosen confidence level, and Ml and M2 are the distances
between the lower and upper confidence limits, respectively, expressed as a
fraction of the measured geometric mean.
If we want 95% confidence limits, the value of t for 54 degrees of
freedom (available in standard statistical tables) is about 2.00. Using
this value:
ML - .1263 and M2 - .1446. (3-8)
With a confidence of 95% it can be said that the true geometric mean
lies between:
57.72(1-.1263) = 50 yg/m3 and
(3-9)
57.72(1+.1446) = 66 yg/m3.
Thus, there is a reasonable assurance that even had a full set of 365 daily
readings been available, this site would not have violated the primary annual
standard, since even the upper 95% confidence limit of 66 yg/m3 is well below
the standard of 75 yg/m3.
The short-term analysis can also be done analytically. The method
to be used depends upon the data available. If the geometric mean and
geometric standard deviation are available, Larsen gives a simple equation
to estimate the second most polluted (SMP) day in Reference 18. His formula
is:
C = MgSg2 (3-10)
where C = the required concentration, Mg = geometric mean, Sg = geometric
standard deviation, and z = distance of C from the mean measured in standard
deviations.
-------
26
Inability to find z often deters the use of Larsen's analytical
formulas. The following simple method permits the calculation of z:
1. Find f from the formula for frequency used in the
above graphical analysis, but express the result
as a decimal rather than as a percentage:
f = r"°-4
n
2. Find U = [In(l/f2)]1/2.
3. Then z = U-(2.52 + .8U + .01U2)/(1.0 + 1.43U + .19U2 + .001U3).
In this analysis for the SMP value, r=2 and there would be n=365 samples in
the year.
= -0043835
U = 3.2954 (3-11)
z = 2.62
The SMP value can then be estimated:
CSMP = 57-72Cl-72)2'62 = 239 yg/m3. (3-12)
This is in good agreement with the 235 yg/m3 value estimated from the graph
and certainly within the limits of expected error in drawing the Larsen line
and interpolating values .
A second frequently available set of information is the actual number
of samples n(=55 in this case), the arithmetic mean m(=66 yg/m3), and the max-
imum measured value C (=181 yg/m3). The frequency appropriate to the observed
I 'let .A.
maximum and the corresponding z can be found as above:
f = I±ii = kOil = 0.019090, and
55 (3-13)
z = 2.2939.
Equation (34) in Reference 17 can then be used to find Sg:
-------
27
Sg = exp(z-[z2-2
= exp(2.2939-[2.29392-2 In (181/66) ]°'5) (3-14)
= 1.6367
Finally, Equation (22) in Reference 17 can be used 'to estimate the
SMP value using the Sg value of Equation (3-14) and the SMP z value of Equa-
tion (3-11):
AAM A7A7-\ (2.62-0.5 In 1.6367) (^ -, 0
= 66(1.6367) (.3-15)
= 213 yg/m3
This result is somewhat less than the estimates obtained by the other two
methods but is still within the expected degree of precision. This last
method depends heavily on the value of the highest measured value, a value
that can frequently be somewhat in error. The method using the geometric
mean and geometric standard deviation is preferable, since it makes full
use of all the data and is computationally simpler.
In conclusion, the area around this site could well be exceeding both
short-term TSP standards but probably not the annual standard. An increase
in the number of samples per year at the site would allow a better definition
of the potential short-term problem and should be undertaken. Modeling should
help to better define the extent of the current attainment problems, if any,
and to determine the contributions of various categories of sources to the
problem.
One further consideration should be given to the data from this site.
High volume samplers are required to run at least every sixth day, giving 60
or 61 samples per year. There are only 55 samples from this site. However,
there are over 75% of the minimum required number of samples and statistics
can be computed. Figure 3-3 shows the SAROAD frequency distribution by quarters
for this site. The five "missing samples" belong in the first quarter of the
year when only 10 samples were recorded. However, every quarter is represented
by an adequate number of samples and the set is probably representative of
the TSP situation at the site for 1974.
-------
28
tn o CM
— o
o
CM * «
X O
• O O 2
x !•* om =>
*«4 «M TV O
o m 1/1 ct
o i^ fv
* W O UJ
o » o >
CC •• O
n — z c? <:
~ z^
«. UJ *• X — Z
UJ O UJ h- t- O
o 3> 2 a: in —
r> K o o •* i-
o
CM in
co or
o =>
o
in in
X O
»- 2 x :
< O K
-» -i P :
r uj uj u.
o
u
~ o
o: >-
»- in
o
UJ 2
X
cr <
x in
<. tr
x o
o-
ui
o
0
Q-
O
in
o-
z »-
< z>
CD CC
< cr
i- t-
< in
o •—
r <*
u —
cr u or
>- z o
U U- UJ
S" Z) 15
O C?
cr uj ••
uj or —
z cr u,
O UJ h-
IT
C
in
o
c o.
o ct cr z
z o - —
o »- in _i
CM f\i t-
— IT 3
C (M U-
D- --
fst •• >- —
(N cr »- -o
— C IT
c z o
h- —
Z L- »- 0:
r »- < u
IT cr
_J 2
C
K-
« Z C
-^ c a
cr z
o z c
• c
cc
t~ _j cr
Z C Uj
•3 n ;
_l Cr
t- a c
— *- z
c a <:
(Si LJ j
in c tL
• IT.
to
I
to
U I
cr ~
K u
UJ V
cr , r
K u
UJ X.
; a _i _i r vr
> ~ z: C Z •
^ Q. 0 -, ^ v
c o e K t-
i L' c- c u
- v c i a 3
i UJ K- LJ <: a. 'I
: is ~ c? a. :? c
[ «3 LJ «£ U, in L
-S O Ct
Or t~ C U.'
Lu Z- X h-
*- 0- C T »- O-
Q — 5
<: C I
*- — _J O — — •
D. — > 1
tt — > 1
-J C
C I
bO
•H
-------
29
PROBLEM 5-1
Given the SARQAD printout (Fig. 3-1) and the outline of Larsen's graphical
method, use log-probability graph paper to determine whether it is reasonable
to expect a short-term particulate problem at this site, ^e there any indi-
cations that Larsen's methods are not applicable at this site? Determine if
the data are lognormally distributed.
-------
30
3.2 METEOROLOGY
3.2.1 Uses of Data
Baseline data defining the meteorology of the region must also be
collected. This data is needed to:
Input to dispersion models,
Identify the frequency and duration of conditions
when short-term, high air pollution levels can exist,
and
Estimate fuel consumption for and emissions from
space heating.
With the exception of proportional techniques, all models used to
relate emissions to air quality require meteorological input whose level of
detail depends upon the specific model employed. Unless scaling or statistical
techniques, such as Larsen's method, are used, comparison of modeled air
quality levels with short-term standards requires the identification of the
frequency and duration of short-term periods when meteorological conditions
conducive to high pollution levels exist. The methods commonly used to esti-
mate fuel usage for space heating require that the number of "degree days"
be known. This parameter is a necessary input for the methods used in Ref-
erence 7 to develop emission inventories.
3.2.2 Data Required
The meteorological parameters required include:
Wind speed,
Wind direction,
Atmospheric stability,
Mixing height,
Temperature,
Pressure,
Solar radiation intensity,
Cloud cover,
Ceiling height, and
Degree days.
Which of these parameters are required depends upon the level of detail of the
analysis employed and the model to be used. A list of the meteorological inputs
required by the various models available from EPA is given in Reference 12.
-------
31
Requirements vary from no meteorological input for proportional techniques
to hourly variations of wind direction, wind speed, stability, and mixing
height.
Wind Speed, Wind Direction, and Atmospheric Stability The National
Climatic Center (NCC) in Asheville, North Carolina has wind speed and wind
direction data available as part of hourly or three-hourly weather records.
This data is also frequently available locally from measurements made by state
or local air pollution control agencies or from airports.
Data for wind speed and wind direction are frequently combined with
atmospheric stability in a joint frequency distribution called a stability wind
rose. Various stability wind roses are available from NCC in tabular form,
on punched cards, and on tape. The tapes include the hourly or three-hourly
observations upon which the stability wind rose is based. Five-year, annual,
seasonal, and monthly stability wind roses are available. These stability
wind roses are the primary meteorological inputs required by the common dis-
persion models for annual averages.
21
Mixing Height Holzworth has provided climatological summaries
of mixing heights based on radiosonde observations. He presents isopleth
maps of the United States giving morning and afternoon mixing heights on a
seasonal and annual average basis. The isopleths for annual average after-
noon mixing height (required for the AQDM model) are presented on Fig. 3-4.
Mixing heights can be calculated from the radiosonde observations available
from NCC by the methods outlined in Reference 10. The calculations are, how-
ever, laborious and it is preferable to use Holzworth. If daily morning and
afternoon mixing heights are needed, it may be possible to obtain them from
a nearby weather station or meteorological data center. Otherwise, the cal-
culations based on other meteorological parameters must be made.
Temperature and Pressure Temperature and barometric pressure data
are needed in the calculation of plume rise. NCC can supply the data. In
addition to the hourly and three-hourly temperature records, various summaries,
some of which are listed in Reference 10, are available. Since variations
in temperature and pressure cause only small changes in calculated ground-
level concentrations, most calculations simply use the annual mean values
-------
32
to
X
§
I
Q)
CM
O
X
J3
I
K1
-------
33
from the nearest weather station. Barometric pressure is also available from
NCC in annual average form as well as the hourly or three-hourly observations.
Solar Radiation Intensity, Cloud Cover, and Ceiling Height If
atmospheric stability data are unavailable, they may be estimated from other
22
meteorological parameters by Turner's method, which is also explained in
Reference 10. The method requires: solar altitude, cloud cover, ceiling,
and wind speed. The solar altitude can be obtained from Table 170, Solar
23
Altitude and Azimuth, in the Smithsonian Meteorological Tables. Cloud cover
and ceiling are available as hourly or three-hourly observations from NCC.
The solar altitude, time of day, cloud cover, and ceiling can be used to index
the solar radiation intensity that, together with the wind speed, determines
24
the atmospheric stability. Turner has also presented another method of
determining stability based on the same meteorological observations but not
requiring the ceiling. When a stability wind rose from NCC can be used, the
stability has already been determined and is available on the tape at hourly
or three-hourly intervals.
Degree Days Heating degree-days are used to estimate fuel consump-
tion for space heating. The number of degree-days is determined from the
number of days the average daily temperature drops below 65°F. For example,
a day with an average temperature (high temperature + low temperature-, Q£
64°F counts as 1 degree-day; a day with an average temperature of 50 degrees
counts as 15 degree-days. The data are available from NCC or Reference 39.
3.2.3 Worst Case Data
It is generally assumed sufficient for predicting annual average
concentrations to use a stability wind rose for a typical year or one repre-
senting an average over several years rather than to attempt to identify those
meteorological conditions causing the highest possible annual average. A
sensitivity analysis comparing modeled air quality concentrations for sta-
bility wind roses from several years, while keeping the emissions inventory
constant, allows an estimate to be made of the variations expected from changes
in the weather alone. Such an analysis could be used in the discussion of
factors affecting the accuracy of the calculations of projected air quality
concentrations.
-------
34
For the 3, 8, and 24-hour standards, the situation is more complex,
since the persistence in time of conditions giving rise to high concentra-
tions must be determined. If short-term concentrations are modeled, two
methods can be used to determine a reasonable estimate of short-term worst-
case conditions:
Consultation with an air pollution meteorologist, and/or
A review of past meteorological records to determine what
conditions have led to high concentrations.
In situations where short-term concentrations are to be modeled by repetitive
application of a model for each hour, consultation with an experienced air
pollution meteorologist, particularly one who has knowledge of the local
situation, offers perhaps the best chance of determining what conditions,
consistent with local meteorology, would constitute a worst case. Otherwise
a search of air quality and weather records to determine the conditions asso-
ciated with high pollution levels must be made. Meteorological records are
generally available over a longer time than air quality records, so after
the weather conditions associated with high pollution levels have been deter-
mined, the search for a worst historical case can be carried back through
the available years of meteorological data. Such a procedure is complicated
by fact that the concentration levels are also dependent upon emissions.
Some sensitivity analysis should be carried out to investigate the effect
of changes in the meteorological input on predicted concentration levels.
If the predicted levels do not change significantly for various estimates
of worst-case conditions, then finding the true worst case meteorology is
probably not important. If, however, the predicted levels are significantly
dependent upon the assumed worst case meteorology, the accuracy of the pre-
dicted air quality results is questionable, since the true worst case can
only approximated by this procedure. In making this assessment, changes in
concentration levels are significant if they are on the order of the changes
expected due to errors in the other input variables. For example, if a
change in the mixing height from 1000 m to 1500 m produces only a one or
two percent change in predicted concentration levels and larger errors, say
8-10%, could be caused by suspected inaccuracies in the emission inventory,
an attempt to determine precisely what series of hourly mixing heights con-
stitutes a worst case would unnecessary.
-------
35
3.2.4 Representativeness of Data
Two problems frequently arise when gathering meteorological data:
The data has been taken at a site remote from the
expected problem area, and
The site for which data exists is not similar* to the
problem area.
Meteorological data is often frequently available only from local airports
that are generally located in rural or semi-rural areas. If the problem area
is urban, there may be differences between the values measured at the airport
and those appropriate to the urban area. This is particularly true of temp-
erature, wind speed, mixing height, and stability, which are affected by the
urban heat island and the difference in surface roughness between rural and
urban areas. As another example, mixing heights are calculated from sound-
ings taken at points on a grid of approximately 400 km, and hence their
spatial resolution is limited. Also, the data may be available from a site on
a hilltop, while the expected problem area is in a valley.
Two courses of action can be followed when representative meteorol-
ogical data is unavailable:
Adjust existing data, or
Use data from the closest similar site.
In both cases, a meteorologist should be consulted either to suggest ways of
adjusting the data or to help choose among candidate sites if more than one
is available. This person could also aid in the assessment of the probable
errors resulting from the adjustment of the data or from use of data at
another site. If there is some limited data available from the problem area,
it may be possible to develop a relationship between values or averages at the
problem site and those measured at the remote site. Such a relationship could
then be used to estimate values for the problem area from the more extensive
records at the remote or dissimilar site. A meteorologist may be able to
suggest alternative ways of adjusting data as, for example, rotating or com-
pressing wind roses to account for the effects of local topography.
In some cases, the data on file with the National Climatic Center
is incomplete or represents only a short observation period (e.g., one or two
-------
36
years). In these cases, the same option of adjusting existing data or using
more complete data from another site would have to be exercised.
3.2.5 Illustration from County X Data
In Fig. 3-5, the part of the stability wind rose for County X for
C stability is presented as it comes in printed format from NCC. The entries
in the table give the relative frequency under conditions of C stability with
which winds from sixteen different directions occurred within six different
speed categories. For example, conditions of C stability with south-southeast
winds between seven and ten knots occur .002260 or 0.2260% of the time in
County X. The stability wind rose data in this format was used to punch
cards as input to the AQDM model used for County X.
The stability wind rose is based on data taken at the Atlanta
International Airport. Figure 3-6 shows the outline of County X and the
location of the airport. The airport is located close to the portion of the
county that is an area of low rolling hills. No attempt was made to change
any of the wind rose data to correct for topographic effects.
Atlanta is located by the dot on Holzworth's map of afternoon mixing
heights on Fig. 3-4. It lies midway between the 1400 m and 1600 m isopleths
and 1500 m was used as the input to the model. The nearest stations where
data upon which the isopleths were based are also shown. Since these are
relatively far from Atlanta, the representativeness of the 1500-m value might
be questioned. However, substitution of 900 m for the mixing height produced
less than a 1% change in modeled TSP concentrations and hence, within the
accuracy to be expected from the model itself, the value assumed for the
mixing height is not critical over this range. From the isopleths it seems
unlikely that a lower mixing height would be appropriate, particularly since
Atlanta is not located in a deep valley where greatly reduced mixing heights
can occur.
3.3 EMISSION INVENTORY
The air pollutant emission inventory forms the basis of making an
assessment of air quality management problems. Substantial guidance in the
7 13 25
development of an emission inventory has already been published. ' '
-------
37
I
r-
o
a
•~t
I—
•
«_>
z
UJ
o
or
ui
ct
z
z
cf
a
en
0
r-
o
0
0
o
o
o
0
0
0
o
o
o
o
0
o
o
o
t-
CM
o
o
o
o
•o
m
o
o
o
en
o
CM
o
o
o
J.
CM
•a
o
o
o
o
z
CM
CM
CM
O
0
0
o
o
o
o
0
0
o
o
o
o
o
o
o
o
o
0
0
o
3
o
o
CO
o
o
o
o
o
r*
fl
o
o
o
o
1—
fl
o
o
o
o
Ul
z
z
m
fl
IM
o
o
o
o
o
o
o
o
o
o
0
o
0
o
o
o
o
o
o
o
o
0
o
o
en
0
0
o
in
00
o
o
o
o
m
o
ft
O
o
o
o
Ul
z
fl
•f
o
o
0
o
o
o
o
o
o
0
o
o
o
o
o
o
o
fl
ft
-f
o
0
o
o
™
o
o
o
J.
•*
fl
o
o
0
en
•o
CM
o
o
o
o
Ul
"y
Ul
CO
o
fl
o
o
0
0
o
0
o
o
o
o
c
o
o
o
0
o
•0
o-
o
Q
o
0
en
o
o
o
o
o
o
o
o
o
CM
o
0
o
o
0
0
o
o
CM
^
rn
9
o
o
fl
o
o
o
o
•o
CM
o
o
o
t-_
CM
m
o
o
o
o
Ul
I/I
Ul
en
CM
tn
o
o
o
o
o
o
o
o
o
o
o
o
o
o
0
0
o
1-
CM
o
0
0
0
*
0
o
o
^
o
fl
0
o
o
r-
^
in
o
o
o
o
Ul
tn
0
en
r-
•*•
o
o
o
0
0
0
o
o
o
o
o
o
o
o
o
o
0
m
o
IM
o
o
o
o
/\
•o
o
o
y
CO
f*
o
o
o
•a
m
en
o
o
o
o
Ul
in
vt
en
en
•o
o
o
o
0
o
o
o
o
o
o
o
o
o
o
o
o
0
f*.
IM
O
O
0
0
*
o
0
o
m
o
CM
o
0
o
„,
K\
in
o
o
o
o
in
CO
30
m
0
o
0
0
0
o
o
o
o
o
o
o
o
o
o
o
o
i-
en
••4
o
o
o
0
CO
o
0
o
_,
ea
•-4
O
0
o
o
o*
(-4
o
o
o
o
a:
i/>
u>
en
O
r-
o
o
o
0
o
o
o
o
o
o
o
o
o
o
0
0
0
r-
•*•
o
0
0
0
l>
0
o
o
^
r-
Q
o
o
o
o
*^
0
o
o
o
^E
in
O
o
-o
o
o
o
o
o
0
0
0
o
0
0
o
o
o
o
o
o
CO
•T
in
0
o
o
CM
o
o
o
o
•o
CM
o
o
o
m
o
m
o
0
0
o
^
in
31
•O
CM
fl
O
o
o
o
o
o
o
o
o
o
o
o
o
o
0
0
o
m
o*
o
o
0
0
«*•
o
o
0
o
en
m
o
o
o
CSJ
ft
r-
o
0
0
o
^t
in
ra
fl
r*1
0
0
CO
•0
0
0
o
o
o
o
o
o
o
o
o
0
^
r-4
o
o
o
m
o
o
0
o
*
en
in
CM
o
o
o
m
f^
•a-
Q
O
o
o
^E
z
31
CM
O
n
I—
o
o
o
o
o
o
o
o
o
0
0
0
0
0
0
o
r-
o
m
0
o
o
o
en
o
o
fi
*»•
°
o
o
0
en
m
-0
0
0
o
0
3
Z
0
IM
fl
0
O
o
0
0
o
o
o
o
o
o
o
o
o
0
0
r-
CM
O
O
0
0
5
o
o
o
0
fl
CM
en
o
o
0
fl
o
en
o
0
0
o
X
z
z
CO
•o
o
o
o
o
o
0
o
0
o
0
o
o
o*
^y
o
o
o
m
CO
0
o
m
o
•o
I-
o
o
CO
,_!
o
•o
o
o
o
^
1—
o
f—
•^
^
fl
fl
o
,
1—
t~J
en
tn
o
„_
o
IU
z
,£
a:
o
o
a
u.
0
V
u
z
~>
o
Ul
OC
u.
Ul
^
**
1—
_J
UI
a.
m
0
CM
0
O
o
1-
l-t
fc
CO
<
to
o
X
X
Ul
o
CO
o
Ul
3
a
c^
1-
VI
Q
in
5
^
o
UL
O
V
z
H)
0
Ul
a:
u.
IU
t_e
1-
_l
Ul
1*
•H
U
Vt
I
I
cn
o
•H
tJ
bO
«H
(X,
-------
38
COUNTY X
AIRPORT
0 5 10 15
SCALE -KILOMETERS
Fig. 3-6. Location of Airport Used for Meteorological Data
-------
39
3.3.1 Definitions
Because of growth, development, and regulatory programs applied tp
pollutant-producing activities, the emission inventory can be expected to
change with time. Three distinct inventory types can be identified for the
purpose of the air quality analysis. They are:
Current inventory,
Updated inventory, and
Projected inventory.
The current inventory is that which exists on file with either the state air
pollution control agency or with the federal EPA in the National Emissions
Data System (NEDS). Most of this information was developed in the course of
the original State Implementation Plan (SIP) development and represents data
valid for the 1969-71 period. In many cases, states have updated portions of
the inventory in the process of conducting enforcement activities or special
studies.
The updated inventory, which will form the basis of the air quality
analysis, represents information that is brought up to the most recent time
period for which adequate data are available. In most cases this will prob-
ably be a 1973 or 1974 inventory. All portions of the inventory should be
adjusted to the same year.
The projected inventory(ies) is a forecast of what emissions will
amount to when growth, development, and regulatory programs have been accounted
for. The projected inventory will be made for several years into the future
starting with 1975. (In the unusual situation where there is adequate inform-
ation available, the 1975 inventory can be developed as the updated inventory
with succeeding years being projected inventories.)
This section will deal primarily with the development of an updated
inventory using the current inventory as a starting point. The projected
inventories will be discussed in the next section.
3.3.2 Data Required
The emission data is generally divided into the following six source
categories:
-------
40
Industrial process,
Fuel combustion,
Transportation,
Electricity generation,
Incineration, and
Miscellaneous.
Table 3-2 lists the subdivisions of each source category and the type of
source included in each. Point sources are any stationary source emitting
more than some designated minimum (usually 100 tons/year) of a pollutant.
An area source is a collection of sources whose individual emission rates
are small but whose collective impact may be large. A line source is a source
that can be geometrically described best as a line (e.g., a highway). The
line source description is used primarily in microscale analyses and need
not be employed in all circumstances.
Table 3-3 lists the type of information desired for each source
type. The list assumes that some form of modeling and strategy testing will
be applied to the emission inventory and so includes parameters other than
emission rates.
3.3.3 Sources of Information
Reference 7 lists numerous sources of information that can be used
to develop an emission inventory. These fall into three basic categories:
Nationally available data - information published by sources
that treat all areas of the country (e.g., Bureau of the
Census; Departments of Interior, Commerce, Transportation,
Treasury, and Army; National Coal Association, etc.).
Locally available data - information published for the
region, state, or local area under study by local agencies
(e.g., transportation, land use, air or water quality,
energy studies, etc.).
EPA data - information published by EPA providing guidance
on translating the above two activity data sets into estimates
of emissions.
It must be emphasized that many forms of information are available and exper-
ience and judgment are the only guides as to which are the most useful.
-------
41
Table 3-2. Emission Source Categories'"
Source Category
Subdivisions
Source
Description
Industrial Process
Fuel Combustion
Transportation
Electricity Generation
Incineration
Miscellaneous
Chemical manufacture
Food/agriculture
Primary metals
Secondary metals
Mineral products
Petroleum industry
Wood products
Evaporation
Metal fabrication
Leather products
Textiles
Inprocess fuel
Other
Internal combustion
External combustion
Highway Vehicles
Light duty gasoline autos
Light duty gasoline trucks
Motorcycles
Heavy duty gasoline trucks
Heavy duty diesel trucks
Off-highway vehicles
Rail locomotives
Vessels
Aircraft
Point
Point, Area
Area, Line
Solvent evaporation
Fires
Fugitive dust
Point
Point, Area
Point, Area,
Line
These categories are based on those described in References 7 and 13.
-------
42
Table 3-3. Emission Source Data Required
Source Type Data Desired
Point Pollutant emission rates
Process activity - type of process, process weight rate
Control equipment - type and efficiency
Stack parameters
Geocoded location
Compliance information
Land areaa
Employment3-
Area Pollutant emission rates
Area geometry
Area geocoded location
Area source type and activity level
Line Pollutant emission rates
Line geometry
Line geocoded location
Line source type and activity level
May be useful for certain strategy considerations.
3.3.4 Special Considerations
There are several special considerations that \/ill influence the
development of the emission inventory.
Level of detail The level of detail in the inventory must be suit-
able to the analysis to be performed. It is not adequate to develop an inven-
tory of a source category, for example, via crude approximation methods if
the analysis will require a detailed consideration of that source's impact on
air quality. Should the inventory be developed in an approximate way and the
analysis show the need for more detail, it will be necessary to revise the
data base. A judicious evaluation of the expected problem areas as outlined
in Section 2 should avoid most of these false starts but even a careful pre-
liminary review may not be able to foresee all the potential problems and
some iteration may be necessary.
Reference 7 provides for several levels of effort in developing the
inventory. The most detailed level (Level 3) should be used on major sources
that are expected to impact significantly on regional air quality. The less
-------
43
detailed procedures (Levels 1 and 2) can be used for minor sources. One
method suggested for determining which are the major and minor sources is
to start with the National Emissions Report on emissions by Air Quality
Control Region. The major sources can be identified as those that contribute
greater than a threshold percentage (e.g., 5%) to the regional emission level.
Sources that contribute less than the threshold level may be treated as minor
sources. The threshold level chosen will vary with the resources available,
the magnitude of the existing air quality problem and the degree of existing
air pollution controls.
In general, it can be said that significant plan revisions based on
the use of only Level 1 analyses throughout will not be acceptable without
prior approval of the EPA Regional Office. This level of analysis will not
give adequate accuracy for a substantial plan revision submission; an analy-
sis comparable to Level 1 will be permitted only in special circumstances.
Subcounty spatial resolution Reference 13 describes methods of
allocating emissions compiled on a countywide basis to subcounty areas for
the purpose of improving the spatial resolution of the analysis. However,
if data are available with a better-than-county resolution, Reference 13
presents techniques that project and allocate emissions directly to subcounty
areas. The procedures outlined in Reference 7 and 13 may be used either in
parallel or in series for all emission source categories. Familiarity with
both references prior to starting the analysis is essential.
Actual v. allowable emissions In the development of the emission
inventory it is important to keep in mind that the desired inventory is one
that represents, as closely as possible, the actual situation. In this light,
the effect of the following things must be considered:
Variances granted,
Non-compliance,
Compliance schedules, and
Improved performance.
The incorporation of stack test data into the inventory whenever
possible will ensure that these considerations are included. For sources not
yet in operation or for the projected inventories, the first approximation of
-------
44
assuming that sources will just meet the emission regulations will be adequate;
however, where information is available to the contrary, it should be included.
Applicability of strategies In developing the emission inventory,
it is necessary to foresee what type of information might be needed to eval-
uate alternative control strategies. If, for example, it is felt that a
strategy based on emission density (i.e., emissions per unit area) will receive
consideration, then it is necessary to know what land area is currently owned
by the sources in the inventory. If a strategy will be directed toward a
certain type of process activity, it will be necessary to know which sources
use that process. Although it will not be possible to foresee all potential
problems, a careful preliminary review will eliminate most of the false
starts.
3.3.5 Updating Procedures
The methods of generating an updated emission inventory as described
in Reference 7 are schematically illustrated in Figs. 3-7 to 3-12. In gen-
eral, the most detailed level of analysis relies on the use of an interview
procedure to determine the necessary parameters directly from the major sources.
The less detailed levels rely on estimates based on county or statewide data
and/or national average values of several key items.
For industrial process emissions, the most reliable form of updating
the inventory is to use the interview procedure. This is due primarily to the
source-specific character of the emission sources. The Level 1 and 2 proce-
dures illustrated by Fig. 3-7 will give only rough estimates of actual emissions.
For fuel combustion sources, the Level 1 and 2 procedures shown on
Fig. 3-8 are somewhat more reliable than for process sources because of the
availability of detailed fuel consumption data. However, as the next exercise
will show, there is still a great deal of approximation that goes into these
procedures.
For highway vehicles, the Level 1 procedures in Fig. 3-9 are very
crude and should be used only in the event of complete lack of any other
information. Data from transportation planning agencies is usually available
to some degree to enable an analysis better than Level 1 to be performed.
-------
45
INDUSTRIAL PROCESS EMISSIONS
(IPE)
NEDS
POINT SOURCE
FILE EMISSIONS
LEVEL 1
AGGREGATE
TO 13 NER
CATEGORIES
X
EMPLOYMENT
ADJUSTMENT
FACTOR FROM
COUNTY BUSINESS
PATTERNS
IPE
LEVEL 2
METHOD 1
USE LOCAL INVENTORY DATA.
METHOD 2
SAME AS LEVEL 1.
METHOD 3
SAME AS LEVEL 1, ONLY AGGREGATE TO SCC CLASSIFICATION AND USE SCC
EMPLOYMENT.
LEVEL 3
METHOD 1
INTERVIEW TOP 95% OF EMITTERS FROM NEDS FILE.
METHOD 2
INTERVIEW TOP 90% OF EMITTERS FROM NEDS FILE.
METHOD 3
INTERVIEW TOP 90% OF EMITTERS OF MAJOR POLiUTANT FROM NEDS FILE.
Fig. 3-7. Update Procedures for Industrial
Process Sources
-------
46
FUEL COMBUSTION EMISSIONS
(FCE)
LEVEL 1,2
ESTIMATE FUEL
CONSUMPTION
FROM STATEWIDE
DATA
PRORATE STATE
FUEL USE TO
COUNTY
Y [EMISSION]
A [FACTORS J
FCE
LEVEL 3
INTERVIEW DEALERS
AND MAJOR USERS
TO DETERMINE
FUEL CONSUMPTION
[EMISSION]
[FACTORS J
FCE
Fig. 3-8. Update Procedures for Fuel
Combustion Sources
-------
47
TRANSPORTATION EMISSIONS
HIGHWAY VEHICLE EMISSIONS
(HVE)
LEVEL 1
DETERMINE COUNTY
GAS AND DIESEL
OIL SALES FROM
TAX DATA
/GAS \
\SOLD;
ESTIMATE VMT BY
/VEHICLE TYPE\
VDISTRIBUTION;
/AVERAGE\
\ MPG /
VEHICLE
EMISSION
FACTORS
HVE
LEVEL 2
DETERMINE VMT
FROM LOCAL
STUDIES
DISAGGREGATE BY
VEHICLE TYPE IF
NECESSARY
VEHICLE
EMISSION
FACTORS
HVE
LEVEL 3
DETERMINE VMT,
SPEED, VEHICLE
CLASS AND AGE
DISTRIBUTION FROM
LOCAL STUDIES
VEHICLE
EMISSION
FACTORS
HVE
Fig. 3-9. Update Procedures for Highway Vehicles
-------
48
ELECTRIC GENERATION EMISSIONS
(EGE)
LEVEL 1,2
DETERMINE FUEL USE,
I SULFUR, % ASH FROM
LOCAL OR NATIONAL
SOURCES
OMISSION!
ACTORSJ
EGE
LEVEL 3
DETERMINE FUEL USE,
i SULFUR, % ASH FROM
LOCAL OR NATIONAL
SOURCES
USE STACK TEST
DATA AND/OR
EMISSION FACTORS
EGE
Fig. 3-10. Update Procedures for Electric
Generation Sources
-------
49
INCINERATION EMISSIONS
(IE)
LEVEL 1
ESTIMATE QUANTITY OF
WASTE INCINERATED
USING NATIONAL
AVERAGES
DISAGGREGATE TO POINT
AND AREA SOURCE USING
EXISTING
DISTRIBUTION
fEMISSIONl
[FACTO RSj
IE
LEVEL 2,3
DETERMINE QUANTITY OF
WASTE INCINERATED FROM
LOCAL STUDIES
["EMISSION!
LFACTORSJ
IE
Fig. 3-11. Update Procedures for Incineration Sources
-------
50
MISCELLANEOUS SOURCE EMISSIONS
GASOLINE HANDLING EVAPORATION EMISSIONS
(GHE)
SOLVENT USE EVAPORATION EMISSIONS
(SUE)
LEVEL 1,2
GASOLINE
HANDLING
DETERMINE GAS
SALES FROM
TAX DATA
[EMISSION"!
[FACTORSJ
GHE
SOLVENT
USE
ESTIMATE SOLVENT
USE FROM NATIONAL
PER CAPITA AVERAGES
Y [EMISSION!
A [FACTORS J
SUE
LEVEL 3
INTERVIEW GASOLINE
AND SOLVENT DEALERS
TO DETERMINE USE
[EMISSION]
[FACTORS J
GHE, SUE
OTHER SOURCE CATEGORIES HANDLED AS NEEDED
Fig. 3-12. Update Procedures for Miscellaneous Sources
-------
51
For electric generation, there are adequate sources of information
for virtually the entire country such that even a Level 1 analysis as shown
on Fig. 3-10 should be fairly accurate. This is probably the source category
that can be analyzed in the most detailed fashion with the smallest effort
needed for information gathering.
For incineration, the Level 1 analysis shown on Fig. 3-11 is crude
since national averages are used. The use of local studies is significantly
more accurate.
For the gasoline handling and solvent use, miscellaneous sources on
Fig. 3-12, the Level 1 and 2 analyses are crude because of the use of national
average consumption figures, but the Level 3 analysis would require a fairly
extensive effort due to the large number of small sources.
3.3.6 Illustration of County X
Table 3.4 summarizes the baseline year emissions for County X tabu-
lated in the National Emissions Report format specified in Reference 7. The
total particulate emissions are 29,524 tons/year. The biggest contributors
are industrial process sources (37%), especially the primary metals and
mineral products industries, industrial fuel combustion (11%), industrial
and commercial/institutional incineration (27%), and fugitive dust from
agriculture, unpaved roads, and construction (20%).
-------
o
a
52
CD
— 1
5
n
n
ti
u
u
0
S
3
j>
H
3
2
J
0)
•H
, 1
rn
0)
;
M)
cc
<
>
cc
LU
cc
LU
CL
CO
Z
o
K
00
z
O
c/3
CO
LU
8
o
I
x
0
z
X
o
00
CC
^
CL
jj
J
r
D
2
ANTHRACITE COAL
^
I—
Z
RESIDE
_]
LU
"D
LL
CO
i— I
BITUMINOUS COAL
UJ
D
LL
Z
g
H
oo
cc
S
O
O
00
•—i
DISTILLATE OIL
(AREA)
_i
Z
cc
II 1
1-
X
LJJ
RESIDUAL OIL
cr
cr
i—
NATURAL GAS
Q
O
o
en
^
-------
53
o
cj
CJ
I
X
o
z
X
o
CO
I—
cc
Q.
LU
o
cc
3
o
CO
LU
cc
_)
0
u
LU
ANTHRACI
_j
<
O
cc
LU
5
O
o
_J
LU
O
LL
Z
o
Q_
<
O
H
D
I—
CO
z
z
0
I-
co
Zl
m
O
CJ
»-H
CD
LU
CC
_J
O
O
CO
BITUMINO
LU
u.
_l
<^
Z
X
LU
X
UJ
Z
o
Q.
Q
LU
^
Z
-—
^
o
CJ
z
o
0-
LIGNITE
LU
CC
O
RESIDUAL
1—
Z
0
Q.
cn
Csl
•
LU
CC
_J
5
LU
DISTILLAT
h-
Z
O
O-
co
CO
•— (
LU
X
CO
NATURAL
Z
s
UJ
X
Q
O
i
H
Z
O
Q.
1-
z
o
0-
co
<
__J
O
X
I—
LIQUID PE
i-
Z
O
0.
OTHER
CO
Cs
Cs
LU
X
TOTAL
1-
0
Q.
Z
0
Q.
X
LU
I
0
Ln
CO
t^
UJ
X
CO
CO
CM
CO
\-
z
0
USTION
m
O
CJ
_i
Z
X
LU
H
X
UJ
_l
0
_,
6
LU
DISTILLAT
CJ
X
^.
U
LU
LU
UJ
LL
CO
0
NATURAL
-*
RATIOf
LU
Z
LU
O
Z
O
(-
^
CQ
O
o
DIESEL
~~.
^
^
CC
LU
Z
OTHER
TOTAL
_j
6
LU
DISTILLAT
;TRIAL
vt
Q
z
CO
O
NATURAL
LU
Li-
GASOLINE
DIESEL
^
OTHER
TOTAL
DIESEL
u
X
UJ
0
o
TOTAL
_,
LU
D
LL
<
0
H
b~
t—
CO
AIRCRAFT
o
z
H
LU
H-
LU
Z
z
LU
z
o
1-
co
D
CO
Q
CJ
_J
z
QC
LU
H
Z
O
in
CO
LU
OC
Z
o
H
co
CO
o
o
LU
^
LL
.
1
CO
CO
CM
CO
t-
z
o
o.
-------
8
o
I
X
o
z
X
0
-------
55
I
I
i
to
0>
o
o
o
X
X
o
z
X
o
CO
1-
cc
Q.
LU
o
cc
D
o
CO
m
t-H
t-
Q
t—
T
0
CO
o
«— i
>
Q
>
LU
1.
1
X
o
X
LL
LL
O
<
o
h-
LU
Z
0
CO
o
H
Q
>
LU
X
I
X
o
X
LL
LL
O
,
CC
<
O
h-
_j
LU
CO
LU
5
CO
LU
Q 2
Z I
< LU
-1 >
cc
1-
_J
i
CJ
ERCIAl
O
O
o
H
<
CC
0
cc
<
_.
o
u
INOUS
^
D
h-
m
LU
D
LL
LU
CO
LU
a
_j
o
RESID
LU
Z
0
CO
O
CO
LU
00
CO
LU
CO
CO
O
0
l—
<^
cc
O
LU
CD
Z
Q
Z
X
CO
a
-------
56
13
I
to
[>-
en
to
0>
r-H
•8
H
o
o
o
I
X
o
z
X
o
CO
h-
oc
<
0.
LLJ
U
DC
D
8
<
LU
QC
«$
CO
LU
O
[INDUSTRIAL SOUR
[DRY CLEANING
z
O
H- <
Z £E
LU O
> 5?
_l <
o >
CO LU
CO
[STRUCTURAL
| FROST CONTROL
| SLASH BURNING
| WILD FOREST
| AGRICULTURAL
CO
LU
OC
LL
CT-
on
J-
i-H
[UNPAVED ROADS
C/j
a.
| UNPAVED AIRSTRI
[PAVED ROADS
CJ
Z
| MINERAL PROCESS
t-^
t^
00
to
1 1 1
| TILLING ACTIVITI!
_j
LU
CC
- <
Q ca H
[STORAGE PILES
[TILLED LAND
Q
Z
<
_l
Q
LU
_!
_l
(-
Z
D
ir
£
3
Q
Q
O
U1
1-\
Bi
nil
§
Total Miscellc
(- W
to Q
D Z
Q > i
S«^
IT Q <
0 LU OC
£3?
- < <
< u z
D
O
LU
z
^
_i
LU <
U LU
CO DC
i <
<
LU
CC
<
.j-
csf
in
J-
•->
(—
Z
O
CL
_!
<
|—
O
i —
Q
Z
<
CC
a
-------
57
PROBLEM 3-2
Update the County X emission inventory for fuel combustion sources using a
Level 2 analysis; that is, use statewide fuel consumption data and determine
the county's share. Follow the procedures in Reference 7 using the copies
of Tables 2.1, 2.3, 2.4, and 2.5 from that report, which are attached here
as Figs. 3-13 to 3-16. All of the necessary data are in the appendixes.
-------
58
CM
0)
en
CD
cn -a
p CD
r-l O
CD g
[L, CO
10
>~> CD
*J i—i
c c
U g
CD Tt
d)
cn
a
m
cn
3
^3
s
o
U
"rt
^
0)
-M-
x
W
6 6 >
^En
LU 1- ~
E <£>
ig
5 "-
MAJOR
FUELS
oiniiisans land
sv aasn)
.nVNb31NI! H3H1O
U3H10
.nvNdaiNii
9NI1S31 3NION3
.nvNaaiNii SNIIOSVD
.HVNH31NII 13S3IQ
3SSVOV9
QOOM
3>tOO
OdT
SV9SS3DOad
.nvNH31NII
svo ivanivN
SVD ivunivN
.I1VNH31NI)
aivniisio
5 sivmiisia
ivnaissa
31INOH
_j
<
g snoNiwniia
31IOVaHlNV
SOURCE
H
2
s
<
UJ
cr
<
INDUSTRIAL
>—
z
£
<
LU
cc
<
COMMERCIAL'
INSTITUTIONAL
<
LU
cc
<
RESIDENTIAL
I
5
^
1000
GALLONS
CO
Z
O
1—
§5
K
LL
D
U
cn
o
1000
GALLONS
CO
Z
O
K
oo
H
z
D
>
LU
_1
Z
O
O
CO
>
cc
<
z
g
h-
<
t—
CO
C
I—
CO
cc
LU
U.
LU
tr
z
O
t—
CO
D
CD
s
o
U
_J
<
z
cc
(Nl
CD
iH
•a
0)
I/)
0)
.a
iH
CD
10
rt
PQ
to
bO
•H
-------
59
C
tJD
a
OJ
C
rt
O
oo O
(N "5
^ C
T3
C
03
3
CO
I
to
1-
z
CJ
QC
LU
Q.
QC
D
LL
'D
CO
1-
z
CJ
QC
LU
-|VCO 31INDI1
ivoosnoNiwnug
1VOO 31IDVUH1NV
iioaivmiisia
noivnaiS3y
1VOD 31INDI1
nvoosnoNiwniia
1VOO 31IDVUH1NV
LU
CJ
DC
D
O
CO
<
or
i-
co
Q
Z
i <
Z
i- o
<-> H
E <
(_ QC
CJ LU
LU Z
_l LU
LU (J
Z
-^
CO
III
CC
Q.
Z
LU
>
O
LU
CC
1-
LJJ
t;
O
CJ
I
CO
" ^
cc Q; Q
13 < 5
O LU <
<" > cc
< LU 13
1- CO LL.
< < -1
Q CD D
CO
_J
O
CC
O
LL
LU
5
_J
O
H
D
QC
[ | |
Q_
_f
0
CJ
QC
0
U-
(-
I
CJ
LU
5
m
0)
•
CVl
"rt
H
•H
O
txO
ri
TJ
y"f
CD
X
Tit
g
1
OJ
O
u
m
0
£
CD
4->
g
u
,-Cl
to
"^
13
I
to
-------
60
(D
tn
CD
tH
•8
ai
U
OJ
s
o
-*->
CO
U
o
+J
w
3
"o
ra
rt
tm
_c
-t->
cti
rt
U
u
o
.p
tfl
i-H
rt
+j
o
H
c/)
rt
CO
I
to
-------
61
in
CO
03 C
CC O
CJ
aj
C
" y >
MINOR
FUELS
MAJOR
FUELS
oiniusans i3rn
sv aasni
.I~IVNH31NII H3H1O
H3H1O
DNI1S31 3NI9N3
.nVNH3iN,,3Nnos«
.nVNH31NU 13S3IQ
3SSV9V8
OOOAA
3M03
9d3
SV9SS300yd
.1 1VNH31NM
svo nvanivN
SV9 IVHOIVN
.HVNH31NI)
31V I1I1SIQ
O
Tvno,S3y
31IN9H
—i
CJ
31IOVHH1NV
SOURCE
Z
2
AREA
INDUSTRIAL
O
CL
LU
07
COMMERCIAL/
INSTITUTIONAL
AREA
RESIDENTIAL
3
1000
GALLONS
CO
z
O
— O
1—
LL
O
CD
0
CO
8§
O -J
0
en
0
c/l
o
2 LU
O O
u <
_J V-
< <
cr
z
O
O
u
z
cr
I
uo
•
CD
rH
'rt
CD^
S
Bn
CD
rt
4-1
CO
3
i—i
CD
PQ
VO
i
to
-------
62
PROBLEM 3-3
Estimate the emissions from fugitive dust sources in County X. The necessary
equations can be found in Appendix F. Assume the following information is
known:
Agricultural tilling: 194,000 acres are under cultivation;
the land is tilled, on the average, once a year; the silt
content (i.e., portion of particles in the surface soil of
size between 2 y and 50 y in diameter as determined by the
Buoyocous hydrometer) is 45%.
Unpaved roads: 30 miles of unpaved roads with an average
traffic load of 60 veh/day, the average speed is 35 miles
per hour, the road silt content is 12%.
Heavy construction: 250 acres are under construction, the
activity continues year-round.
-------
63
4. ESTIMATING FUTURE EMISSIONS
Accumulated experience in the preparation and evaluation of air
quality attainment and maintenance plans has emphasized the need for con-
sidering the effect of growth and development on pollutant emission levels.
This was recognized in the early State Implementation Plan (SIP) efforts but
has only lately become a significant aspect of all SIP revision actions.
This section is aimed at describing the methodologies for projecting future
trends in emission levels.
4.1 PROJECTION REQUIREMENTS
The initial regulations for the development of air quality mainte-
nance plans indicated that the time period to be covered was the 10-year
span between 1975 and 1985. Proposed revisions to these regulations recog-
nize that the uniform 10-year period:
Does not reflect coordination with other on-going federal
planning programs and
Does not recognize the unique problems of each study area.
To address these situations, the revisions include the requirements indicated
on Fig. 4-1.
Where another federally-sponsored planning program is underway (e.g.,
Department of Transportation 3-C plan, Department of Housing and Urban Devel-
opment's 701 comprehensive plan, or EPA's area-wide wastewater treatment plan
under Section 208 of the Water Pollution Control Act Amendments of 1972) then
the air quality analysis must address the time period extending furthest into
the future of these programs. Where there is no additional federally-funded
planning program, then the analysis must address, at a minimum, the 10-year
period to 1985. These same timing considerations must be reflected in any
submitted air quality maintenance plan.
To allow for flexibility in addressing the problems of each study
area, the EPA Regional Administrators are given the authority to allow for
a less detailed analysis for the period beyond 1985. Similarly, discretion
is given to the Regional Administrator to accept air quality maintenance or
long-range air quality management plans covering less than the 10-year period
but not less than 3 years.
-------
64
o
£52
UJtO
o
^i
UJ1
a.
—
V)
: too
to
ZUJQ:
-------
65
4.2 PROJECTED VARIABLES
The projection of future emissions brings the air quality analysis
to the least rigorous portion of the analysis procedures. Forecasting the
future has never been amenable to a tnr.y scientific approach and relies,
in the case of the air quality analysis, on a combination of considered judg-
ment and best^"guesses. Using the six emission categories, the variables that
must be projected are the following:
Industrial process activity,
Fuel consumption,
Transportation activity,
Electricity demand,
Solid waste generation, and
Miscellaneous emission-producing activity.
27
These may be converted to emissions using emission factors. In addition,
the temporal and spatial distribution of these variables (e.g., when new plants
will come on line and where they will be located) is important to the air
quality analysis. In most situations, projections of these parameters are
not available or are available only on a cruder scale than needed for an ade-
quate air quality analysis, especially in attainment/maintenance problem
areas. Reliance must then be placed on surrogate variables such as population,
employment, land use, earnings, and others that are projected with reasonable
accuracy and precision and that can then be transformed into growth factors
for the desired variables. Figure 4-2 illustrates this process. The surro-
gate variables are those most frequently used by planning groups in developing
comprehensive regional plans.
Projections of population, and to some extent employment, have custom-
arily been the starting point for the projection of other variables. Overall
size is the most important factor in population projection, but other charac-
teristics of the population composition can be important for determining
activity levels. The age distribution of the population, distribution of
household sizes, and the income composition of the population affect the
details of the need for residential areas and community facilities. Like-
wise, more than just the total employment figures are of value in projecting
the future industrial and commercial/institutional growth of the region.
-------
66
EMISSION
FACTORS
SURROGATE
VARIABLES
POPULATION
EMPLOYMENT
LAND USE
EARNINGS
EMISSION-PRODUCING
VARIABLES
INDUSTRIAL PROCESS ACTIVITY
FUEL CONSUMPTION
TRANSPORTATION ACTIVITY
ELECTRICITY DEMAND
SOLID WASTE GENERATION
MISCELLANEOUS ACTIVITY
EMISSION
ESTIMATES
Fig. 4-2. Use of Surrogate Variables
-------
67
Separation into basic and non-basic employment, for example, leads to some
indication of the probable land requirements for industrial and commercial/
institutional areas. Further disaggregation of the projected employment on
the basis of Standard Industrial Classification (SIC) categories or EPA's
Source Classification Code (SCC) System will be useful in a. refinement of
these requirements.
Changes in population are accompanied by changes in the area occu-
pied by various land uses (e.g., residential, commercial/institutional,
industrial, transportation, vacant, etc.). New plant locations and industrial
process activity (measured, for example, by earnings estimates) are private
decisions that integrate site-specific characteristics (physical features,
local regulations, access) with regional and, in some cases, extraregional
considerations (demand patterns, regional and interregional transportation,
location of inputs).
There are numerous techniques available to regional planners to make
estimates of the surrogate variables. Reference 4 outlines some of the more
widely used procedures. Appendix D contains some of the background material
and a computer printout of the projected population and land use for County X
based on the regional planning commission's efforts. It is emphasized that,
while the data are based on the Atlanta Regional Commission's work, some
information has been changed to suit the needs of this example. Thus, the
data should not be viewed as corresponding to actual projections for the
Atlanta area and is included here for illustrative purposes only.
4.3 ACTIVITY SCENARIOS
Estimates of future activity are dependent on the assumptions made
concerning levels of growth and growth management policies. Every regional
planning group uses its own methodologies and assumptions in arriving at these
estimates. A set of assumptions leads to one possible picture, or scenario,
of the future. Because it is not possible to fully predict the detailed
decisions that will determine the future nor to foresee the events and dis-
coveries that influence the decisions, a number of futures can be hypothesized.
Several scenarios of the future might be available for air quality management
planning that indicate both the more likely future activity levels and the
range of levels that is readily conceivable.
-------
68
An elusive, but valuable, concept is that of the most probable growth
»
scenario. Most probable growth customarily is taken as projections of current
trends, altered by known constraints and incorporating already apparent changes
in growth-affecting policies and actions. As such, it is a basic reference
scenario for evaluating the desirability of intervention and the effectiveness
of proposed alternatives. Planning agencies often develop such a scenario for
just this purpose.
Often the reason for variations in projected development scenarios
is the mandated orientation of the agencies preparing the estimates. Trans-
portation planning groups base their projections on various transportation
scenarios (e.g., highway development emphasis, transit emphasis, etc.). Com-
prehensive planning groups may use land-use development scenarios as the
driving mechanism. These differences can result in wide discrepancies in
the projection of regional development. One of the problems that the lead
agency in the air quality analysis has is the reconciliation of the various
planning outputs. In order to use information from the various groups, it
will be necessary to insure that they are based on compatible assumptions
and data bases. The need for a coordinated effort on the part of the involved
organizations is reemphasized by this need for a unified projection plan upon
which to base the analysis. The question of how many alternative scenarios
should be evaluated as part of the air quality analysis is answered by the
available resources. Certainly the "most probable growth" situation should
be given detailed review. In the interest of determining the compatibility
of other alternatives with air quality constraints, it may be useful to con-
sider additional possibilities although this would not be required for the
analysis. If, however, there is strong feeling that other scenarios have
a reasonable chance of being implemented and would impact on the region's
air quality, then analysis of their air quality conditions at this stage may
save the necessity of developing a major revision to the air quality manage-
ment plan at a later date.
4.4 SOURCES OF DATA
Long-range planning has been carried out for many years in a number
of areas. The information required for the air quality analysis should build
on the experience and background of the planning groups that have been involved
-------
69
in these activities. There are four federally-funded planning programs that
should be able to provide some framework for the projections. They are:
HUD 701 planning,
FHWA 3-C planning,
EPA 208 planning,
CZM planning, and
OBERS growth projections.
There are, of course, numerous local and regional studies that can be used
to supplement this information.
4.4.1 HUD 701 Planning
The planning assistance program defined by Section 701 of the Housing
Act of 1954 and subsequent revisions is the Department of Housing and Urban
Development's (HUD) main program of support for comprehensive planning. It
has been in existence for a number of years and presently provides funds to
all the states, 70% of the cities of over 50,000 population, more than 80%
of the metro and non-metro area-wide planning organizations, and approximately
1,200 counties and smaller municipalities. Because the priorities are set
by such a broad scope of recipients, a variety of approaches is permitted
to the general requirement that 701-supported activities address the major
growth or no growth decisions facing the jurisdiction; the diversity is rep-
resented by such subjects as controls over strip mining, preservation of prime
agricultural land, government cost cutting, energy saving programs, and review
of the status of local land controls and land use planning.
The Housing and Community Development Act of 1974 contains new
requirements for 701 recipients. The requirement that the planning effort
be comprehensive has been specified to mean, at a minimum, the inclusion of
a housing element and a land use element that reflects consideration of the
principal land use issues facing the jurisdiction. Many of the previous 701-
supported planning programs did not include a land use element, but satisfac-
tory progress toward developing a land use plan must be demonstrated for
assistance to be continued after August 1977. It is also necessary that each
jurisdiction should work toward the establishment of a unified or integrated
system for land use planning and policy making, which includes the principal
-------
70
land use related planning and regulatory process. As part of this coordina-
tion requirement, the 701 activities must include a consideration of the
current status of other planning that impacts land use (e.g., coastal zone,
air and water quality, transportation, and solid-waste planning), related
implementation activities, and the relationship between the 701 work program
and such other planning or implementation efforts.
Three additional requirements are a part of the current 701 program:
(1) The planning must attempt to deal with the complexities of the intergovern-
mental structure; (2) citizens must be involved at significant points in the
process; and (3) the planning should serve the needs of growth management as
well as protection of critical environmental features.
4.4.2 FHWA 5-C Planning
The Federal Highway Administration has been administering the require-
ments of the Federal Highway Act of 1962, which states that all federally-
assisted highway projects be a part of Continuing, Comprehensive transportation
planning process carried on Cooperatively in the state and local communities.
The 3-C process has been carried out with varying levels of sophistication
throughout the country. The basic procedure is divided into four phases.
The data collection phase assembles information on current land use, popula-
tion, dwelling units, employment, traffic volume and patterns, and existing
transportation facilities. In the analysis phase, analytical methods such
as comprehensive land use models and trip generation, distribution and modal
split models are used to develop the pattern of travel demand and form the
basis for developing travel projections. In the projection phase, land use
and travel demand are forecast for a variety of alternatives. The continuing
planning and implementation phase results in the development of the area's
long-range transportation program. Reference 4 gives a more detailed descrip-
tion of the 3-C process.
4.4.3 EPA 208 Planning
EPA is administering the requirements of Section 208 of the Federal
Water Pollution Control Act, which deals with areawide wastewater treatment
management. In regions that require management plans for the attainment of
water quality standards, areawide agencies have been designated to carry out
-------
71
the planning effort. Other agencies may have the responsibilities for imple-
mentation of the plans. In many ways similar to the air quality analysis
process, the 208 planning process seeks to identify water quality problems
associated with construction and growth over a 20-year period and to suggest
appropriate technical and regulatory alternatives as well as growth and
development options, to assure continuing achievement of satisfactory water
quality. Most 208 plans will be due about mid-1977. It is entirely possible
that a single regional planning agency could and should be responsible for
developing projection indices for both air quality and water quality planning.
A critical parameter in attempting to coordinate the goals of these two, or
any other planning programs, is the time scale frame by which results are
required.
4.4.4 CZM Planning
The Coastal Zone Management Act of 1972 establishes a national policy
for the development of a program to manage the land and water resources of
the coastal zone. The act recognizes the multiplicity of uses and values
ascribed to lands and waters within the coastal zone and encourages the
management of these lands through an approved plan that incorporates important
ecological, cultural, esthetic, and economic values. States are encouraged
to rely upon and coordinate their activities with appropriate local govern-
ments and regional agencies in the development of a coastal zone management
plan (CZMP). Incentives to prepare and administer a comprehensive CZMP are
provided, but no specific land or water use decisions are made. States sub-
mitting grant requests for programs that impinge on the CZMP must show that
these programs are consistent with the approved plan. Further, according to
Section 307(f) of the act, requirements of the Clean Air Act as amended and
subsequent federal regulations will be incorporated into the CZMP. An impor-
tant point of the act is that the state land use control authority, when
necessary, can supersede local governmental authority. This is in marked
contrast to the tradition under which all 50 states have delegated land use
regulation power to local city or county governments.
4.4.5 PEERS Growth Projections
The projections of growth based on the work of the Office of Business
Economics (QBE), presently the Bureau of Economic Analysis of the Department
-------
72
of Commerce, and the Economic Research Services (ERS) of the Department of
Agriculture form a last resort to be used in areas where no other planning
information is available. The projections are available by state, water re-
source area, QBE economic area, Air Quality Control Region, and Standard
Metropolitan Statistical Area and include population, employment, and earnings
(disaggregated by 27 industrial categories that are basically 2-digit SIC
classes). The projections are derived by an apportioning of national data to
the smaller subareas and are, as such, only crude estimates. Appendix D con-
tains a tabulation of the OBERS projections for the Atlanta Metropolitan AQCR
that contains County X.
4.5 PROJECTION METHODOLOGIES
The methodology used to forecast emission levels is conceptually
simple, although, in practice, one encounters some difficulty in identifying
the appropriate parameters to insert into the equations. Emissions are gen-
erally estimated by the following equation:
T-, . . Process Emission ,-, Control . /-/i -,-,
Emissions =,.--, x „ . x(l- „,.,-. - ) (4-1)
Activity Factor v Efficiency' v J
The process activity is measured in units such as fuel consumption, raw mater-
ial input or quantity of finished product. The emission factor is an emission
rate per unit of process activity and usually represents uncontrolled emis-
sions. The control efficiency accounts for the use of some form of control
device to reduce emissions and is a percentage removal of pollutant.
When projecting emissions, the established procedure is to forecast
the change in process activity that will result from economic growth and
development. The implementation of air pollution control regulations may
affect the activity (e.g., in a transportation control plan) but will most
often affect the emission factor and/or the control efficiency applicable to
specific sources. Because many regulatory programs differentiate between
existing, modified, or new sources in the application of control regulations,
it is necessary to maintain this distinction in an air quality analysis. Thus,
projected emissions must be disaggregated into three components:
p . , Emissions Emissions from Emissions
oj c e _ from existing + modifications to + from new (4-2)
sources existing sources sources
-------
73
The emissions from each of these types of sources is computed in the standard
way (i.e., as the product of a process activity, emission factor, and control
efficiency factor), but with different values for these parameters. Unfortun-
ately, it is not always straightforward to identify the distribution among
them. For example, planning data generally indicates overall growth for a
given industry (e.g., steel) but does not identify how much of the growth
will be at existing sources and how much will be at new sources.
To accurately estimate emissions, the process activity must be dis-
aggregated into its three components:
. , Activity due Activity due Activity due
Frojected = tQ existing + to modifications + to new (4-3)
ctivity sources of existing sources sources
In the application of air pollution control regulations, the last two compo-
nents are generally subject to more stringent emission controls (either more
stringent state regulations or federal New Source Performance Standards).
Also, the first two components can be used to identify the spatial distribu-
tion of emissions, while the last component represents an unknown in terms
of source location. A further complication to the calculation of projected
emissions is the fact that the activity at existing sources is variable and
dependent upon two important factors: the maximum capacity available at
existing facilities (e.g., if a plant is operating at 901 capacity it can
only absorb a 10% increase without modification), and the rate of retirement
of existing facilities (e.g., the activity at existing facilities is not,
in general, constant over the entire projection period because of the closing
down of obsolete equipment).
In the ideal situation, the agency performing an air quality analysis
would have information available on projected activity disaggregated into its
three components, disaggregated by industry type, and disaggregated spatially.
In reality, only a portion of this information will be available and estimates
of the disaggregation will have to be made. This section will focus on
estimating the component and industry distinctions, while the next will con-
centrate on the spatial distributions.
Figures 4-3 to 4-8 schematically illustrate the emission projection
procedures outlined in Reference 7. For industrial process emissions, as
-------
74
PROJECTED INDUSTRIAL PROCESS EMISSIONS
(PIPE)
LEVEL 1
METHOD 1
TEASELINE"!
[EMISSIONS]
OBERS GROWTH
FACTORS FOR 13 NER
PROCESS CATEGORIES
CONTROL EFFICIENCIES
FOR 13 NER
PROCESS CATEGORIES
PIPE
METHOD 2
SAME AS METHOD 1, ONLY USE SCC CATEGORIES.
LEVEL 2
SAME AS LEVEL 1, METHOD 2, ONLY USE LOCALLY AVAILABLE GROWTH PROJECTIONS.
LEVEL 3
INTERVIEW INDUSTRIES TO DETERMINE GROWTH AND EXPANSION PLANS.
Fig. 4-3. Emission Projection Procedures
for Industrial Process Sources
-------
75
PROJECTED FUEL COMBUSTION EMISSIONS
(PFCE)
LEVEL 1
CONVERT BASELINE
COUNTY FUEL USE
TO BTU
EQUIVALENTS
SCALE RESIDENTIAL
BTU GROWTH BY POP-
ULATION GROWTH, ALL
OTHERS BY OBERS
DISTRIBUTE
BTU GROWTH
BY FUTURE
FUEL MIX
EMISSION
FACTORS
INCLUDING
CONTROLS
= PFCE
LEVEL 2
SAME AS LEVEL 1, ONLY USE LOCALLY AVAILABLE GROWTH PROJECTIONS FOR EACH
CATEGORY.
LEVEL 3
SAME AS LEVEL 1, ONLY USE LOCALLY AVAILABLE GROWTH PROJECTIONS FOR EACH
CATEGORY AND INTERVIEW RESULTS FOR FUEL MIX AND OTHER DISTRIBUTION CHANGES.
Fig. 4-4. Emission Projection Procedures
for Fuel Combustion Sources
-------
76
PROJECTED TRANSPORTATION EMISSIONS
PROJECTED HIGHWAY VEHICLE EMISSIONS
CPHVE)
LEVEL 1
[BASELINE!
L VMT J
POPULATION
GROWTH
FACTOR
VEHICLE
EMISSION
FACTORS
PHVE
LEVEL 2,3
DETERMINE PROJECTED
VMT FROM LOCAL
STUDIES
VEHICLE
EMISSION
FACTORS
PHVE
Fig. 4-5. Emission Projection Procedures
for Highway Vehicles
-------
77
PROJECTED ELECTRIC GENERATION EMISSIONS
(PEGE)
LEVEL 1
DETERMINE PROJECTED
QUANTITY OF
ELECTRICITY TO .BE
GENERATED
DETERMINE FUEL
USE DISTRIBUTION
AND FUEL REQUIRED
PER KWH
COMPUTE
FUEL
USE
C
EMISSION!
FACTORS J
= PEGE
LEVEL 2
DETERMINE PROJECTED
FUEL USE FROM
PUBLISHED DATA
[FACTORS
PEGE
LEVEL 3
SAME AS LEVEL 2, USING INTERVIEW RESULTS.
Fig. 4-6. Emission Projection Procedures for
Electric Generation Sources
-------
78
PROJECTED INCINERATION EMISSIONS
(PIE)
LEVEL 1
BASELINE
SOLID WASTE
GENERATED
OBERS GROWTH
FACTORS FOR
MFG, COMM/
INST, RESID
APPORTION
BY DISPOSAL
METHOD
[EMissiOH]
^FACTORS J
PIE
LEVEL 2
SAME AS LEVEL 1, ONLY USE LOCALLY AVAILABLE GROWTH FACTORS.
LEVEL 3
USE INTERVIEW RESULTS TO DETERMINE WASTE GENERATED AND DISPOSAL METHODS,
Pig. 4-7. Emission Projection Procedures
for Incineration Sources
-------
79
PROJECTED MISCELLANEOUS SOURCES EMISSIONS
GASOLINE HANDLING EMISSIONS
(PGHE)
SOLVENT USE EVAPORATION EMISSIONS
(PSUE)
LEVEL 1,2.3
GASOLINE HANDLING
SCALE GASOLINE
USE BY VMT
PROJECTION OR
INTERVIEW
[FACTORS J
PGHE
SOLVENT USE
SCALE SOLVENT
USE BY
POPULATION
GROWTH
EMISSION!
[FACTORS J
PSUE
OTHER SOURCE CATEGORIES HANDLED AS NEEDED
Fig. 4-8. Emission Projection Procedures
for Miscellaneous Sources
-------
80
shown in Fig. 4-3, the Level 1 and 2 analyses rely on the application of a
growth factor for an industry. This will, in general, mean that no informa-
tion will be available to directly identify new, modified, and existing source
contributions and some estimates must be made. Only the Level 3 analysis can
supply this information, and even there, it may not do so in all cases. For
fuel combustion sources, as shown on Fig. 4-4, the same comments apply as
for process sources; i.e., only Level 3 will give a direct indication of the
source contributions.
For highway vehicles, shown on Fig. 4-5, the "activity" of new
sources is represented by the vehicle age distribution that is available from
local registration data. For electric generation, as shown on Fig. 4-6, there
is a significant amount of information available on power plant expansion
plans to enable a good estimate of future activity to be made. A Level 2 or
3 analysis should be relatively easy to effect.
For incineration, on Fig. 4-7, the Level 1 and 2 analyses provide
no information on new, modified, and existing source contributions. Only
Level 3 provides data to make the necessary distinctions. For the gasoline
handling and solvent use sources, on Fig. 4-8, there is no provision for
making a separation of the emission contributions. The nature of the sources
(i.e., small and generally treated as area sources) renders this situation
more tolerable.
Another possibility for projecting emissions is given in Reference 4.
This procedure relies on the utilization of land-use-based emission factors
rather than source-specific emission rates. This procedure has been tried in
several places with varying success. It is significantly less detailed and
subject to more error than the methods described here.
4.6 ESTIMATING SOURCE CONTRIBUTIONS
When a projection technique other than Level 3 is used to forecast
emissions, then estimates of the contributions from new, modified, and exist-
ing sources can be made. The most frequently encountered situation is one
in which the growth rate for the industry category is specified by planning
data without regard to whether this will entail utilization of existing
capacity, modifications to increase production capacity, or development of
new facilities. The projected activity is then computed by:
-------
81
Projected _ Existing Growth f4-41
Activity Capacity Factor J
A first-order approximation to the source contributions is that all growth will
occur via modifications to existing facilities or via new facilities. The
implicit assumption is that existing capacity is being utilized at its maxi-
mum rate. The calculation procedure is the following:
Activity from New
Sources and _ Projected _ Existing f4-51
Modified Existing Activity Activity ^ '
Sources
In computing emissions, the new and modified source activity would be subject
to federal New Source Performance Standards or more stringent state standards
(if applicable), while the existing activity would be subject to existing
regulations. The calculation is the following:
Activity from Nsps nv
JYIOCllilGCl o+-1
Sources
A further refinement of this estimate can be made if some information
on facility retirements is known. In this case, the existing activity is re-
duced by the retirements. The new and modified source activity is increased
by the corresponding amount and the activity calculation becomes:
Activity from
New and = Projected _ .-Existing _ RptirpTTlpT,tq-i u 7>
Modified Activity Activity RetirementsJ (4 7)
Sources
Note that this assumes that total industrial capacity remains the same.
Further detail to improve the disaggregation would probably not be available
except from a Level 3 analysis. It is important to note that these calcula-
tions are carried out over a series of time increments that span the air
quality analysis time frame. The application of growth, retirement, and new
source standards may take place at discreet points in that span and should be
accounted for. In reality, this amounts to nothing more than repeating the
-------
82
calculation for every year using the previous' year as the baseline and, at
each point, reevaluating the distribution and the applicability of standards.
Where parameters vary continuously or where there is no change for several
consecutive years, the computations can be shortened. Mathematically, the
entire procedure can be expressed as follows:
In the baseline year:
(4-8)
where E are the emissions for the base year, 0; P is the existing activity
for the source category (e.g., steel) in year 0; and (EF) is the emission
\2
factor considering only existing regulations and control equi.pment.
In the first projection year:
Pl = P0 G01
El = Pl (EF)e = P0 G01
where P-^ is the projected process activity in this first year, GQ-, is the
growth factor between the base year and the first year. Note that the emis-
sion factor is unchanged since no change in regulations has been assumed to
take place.
In the second projection year:
P2 = Pl G12 = P0 G01 G12
= PQ G2 (if the growth factor is constant) (4-11)
E2 = P2 (EF)e = PQ G2 (EF)e (4-12)
where G, - is the growth factor from year 1 to year 2. Note the simplification
if the growth rate is constant. Note also that the emission factor is still
unchanged.
-------
83
In any year, k, prior to the promulgation of a more stringent standard
for new source :
Pk = Pk-l ^k-l = VG01 G12 •"
= PO G (if the growth rate is constant) (4-13)
- Pk ™e = P0 CEF)e (4-14)
This is a general equation suitable for all years prior to the setting of a
new source standard. A disaggregation of source contributions is not neces-
xSary since all sources are under the same regulation.
In the year, £, in which a new source standard becomes effective:
(P») = P. , G, , . = Pn GT (for constant growth rate) (4-15)
*• total L'L *-'*->*-
This must be disaggregated into contributions from existing sources and new
and modified sources. Using the assumption that the growth occurs entirely
at new or modified facilities :
= CPJ - (P/.i)
new + total total
modified
= P (G£ - G£"1)
Q
= PQ G' (G-l)
= P£_1 (G-l) (4-16)
Emissions are then computed as:
El = Vl
new +
modified
[(EF)e + (G-l)(EF)n]
= PQ C" [(EF)e + (G-l)(EF)n] (4-17)
-------
84
where (EF) is the emission factor under the new source standard.
In the next year after a promulgated standard (year £+1)
- po
Using the same assumptions that all growth is at new or modified
sources:
new + total total
modified
= P0
= PQ G£"1 (G2 - 1)
= P^ (G2 - 1) (4-19)
Note that the last term of the equation implies that the existing source
activity is frozen at that level that existed in the year prior to the promul-
gation of the standard (i.e., year £-1). The emissions are then computed as:
modified
[(EF)e + (G2 - l)(EF)n]
= PQ G"1 [(EF)e + (G2 - l)(EF)n] (4-20)
At any year after the promulgation of a new source standard (year
£+p) , the general equation is :
= P0
-------
85
With the same distribution assumption:
(P ) = (P ) ~ (Pp )
^+p new + p total total
modified
- :o ^
= P^ (GP+1 - 1) (4-22)
Emissions are computed by:
+ fP 1 fEFl
l+P new + n
modified
+ P£-1
(GP+1 - 1) (EF)n]
= PQ G^'1 [(EF)e + (GP+1 - 1) (EF)n] (4-23)
In summary, the key equations are the following:
Ek = P0 ^ (^EF') £or ^ear k Prior to the (4-24)
setting of a new source
standard
= PQ CT- A [(EF)e + (G^ - 1) (EF)n] (4-25)
for year £+p where a new
source standard is set
in year L
and the assumptions that are implicit are (1) a constant growth rate and
(2) all growth is at new and modified sources. If there are data on capacity
retirements and/or unused capacity it can be incorporated into this framework
easily; the existing capacity is either reduced (in the case of retirements)
-------
86
or increased (in the case of utilization of currently unused capacity) by
the appropriate amounts.
Equations 4-24 and 4-25 produce exactly the same results as the
28
equations given in the supplement to Reference 13 recently issued by EPA.
Some algebraic manipulations will illustrate the compatibility of the two
descriptions. For example, the term RF in the Reference 28 description is
the emission reduction factor for new source regulations. This is equivalent
to (EF) /(EF) in this description. The Reference 28 term NCR is the growth
•II C
rate for sources covered by new source regulations and is equivalent to
(C?+ - 1) in this formulation. The term CGR, which is the growth rate for
increasing production activity up to full capacity, and the term RR, which is
the retirement rate, are not used in the simplified formulation presented here.
A detailed description of the formulation described here is given
in Appendix G.
4.7 FEDERAL NEW SOURCE PERFORMANCE STANDARDS
The implementation of federal New Source Performance Standards (NSPS)
has a marked effect on the computation of future emissions. To date, NSPS
have been promulgated for only a few of the 250 odd possible candidate source
categories. The issued supplement to Reference 13 contains preliminary esti-
mates of the standards that will be set in the time period of the air quality
28
analysis. These estimates are not to be treated as promulgated regulations
and are only to be used as guidelines of the expected emission levels. The
Regional Office should be consulted for the latest available information prior
to conducting the analysis.
In determining the effect of NSPS on an industry or on a specific
facility, Reference 7 gives the following guidelines:
NSPS are applicable to:
Replacement of obsolete equipment at an existing plant.
Additions of new equipment at an existing plant.
All equipment at a new plant.
NSPS are not applicable to:
Existing equipment at a plant whether used or not.
-------
87
In the terminology of the previous section, the replacement of obsolete equip-
ment and the addition of new equipment constitutes a plant modification, while
the new plant is a new source.
4.8 SUMMARY OF COUNTY X DATA
Table 4.1 summarizes the emission projections for County X. Tables
4.2 to 4.4 give the emission projections for 1975, 1980, and 1985, respect-
ively, in the NER format described in Reference 7. The emission projections
account for existing state regulations, for available compliance information,
and for those federal New Source Performance Standards that are expected to
apply prior to 1985. It is evident that countywide emissions are expected
to increase by about 40? between 1975 and 1985. The largest increases occur
in the mineral products industry, industrial fuel combustion, and the primary
metals industry. The incineration and miscellaneous (fugitive dust in this
case) categories sustain smaller increases but still account for substantial
portions of the total emissions.
-------
88
(/)
g
•H
P
O
0>
•i — t
O
f*\
0
•H
t/J
(/)
•H
(D
•P
nj
3
Tj
fn
»
r-H
X
X
•y
o
u
1
LO O O CN]
co LO r- co
CD tO CM O
ft ft
0 ^0
en to LO ^t
t— ( -^j- \O IO
r^ ro CM co
ft
cn LO
rH
rH CD Cn \D
LO tO LO OO
tO tO CN] LO
ft ft
cn LO
rH i— i
CN] vO O VO ^f O
^t LO I>-
LO to
OO O
•^t r- ^j- co
CO rH LO tO
cn to CN] to
ft
CO LO
i-H rH
CN] \O i— 1 ^- rH CD
CN] LO ro
CN] LO
co cn
"3- ^t1 CD rH
rH O LO Cn
OO tO CN] o
ft
r- LO
rH
CNJ \d r-~ (Nj cn CD
cn LO co
CO ^
t~~~ CO
•^- cn *^ •^~
O CNJ (NJ CO
ft
rH
CN] 0 CNJ 0 O CD
r- LO •*d-
LO CO
r— r—
\o cn to co
r— r^. ^- cn
•^3- CNJ CNJ LO
*
uo ^t
rH
rH LO CO r- ^f CD
•^ ^t CD
CNJ o
LO r- cn I-H
CD vO tO LO
tO CN] CN] tO
t-H
rH LO tO LO rH O
CN] ^J- \&
Cn rH
LO -xt LO - -H
C/) 03 rH rH 4-J +->
tn "-H 3 t/i rt o to
CD 3 4-> rH +J 3 Cl
^SO.0) f-i C/) -HCOrHrH
Cu «H S f^-P-i r< -4-J -H rt nj
rHf-t rH CDrH (/] 4-» -H -H
rH nJbpXcdi-HX rt 3 CUS-.
•H -rj "^. OJCrnO O *ET)i)(/!
4-* QJO'rlOCi-H O C/lfcld
| Ufc0-^^3 1 *U~
rH to o cn ^o 1
*t fN] ^ rH
CM CD t~- cn
J
^ cn m"|
LO cn o r~- \o|
r- cn >o rH
cn cn LO en
ft ftl
\D cn LO
[co \o o LO \D!
or- cn rH
r- cn to cn
* * A]
o cn Lol
CN CNJ O *d- vol
^t" LO CN] rH
• g
O tl o rt § 2
H O -H rH rH U
P- rH O rH
W 4-> S
u
1
0
UH
0)
U
4)
«
• H
rt
1
8
4-»
rt
tn
oo
.5
• H
X
(U
T3
01
{/I
5!
o5
rt
-------
O
D.
89
LO
t~^
CTl
I/I
o
E
a
L,
Oi
cu
><
D
ra
oj
m
c
3
O
U
>-
z cc
D <
O ^
CJ >
cr
^
LJJ
V
LU
Q.
CO
Z
0
CO
g
CO
CO
LJJ
L
(.
0
L>
O
O
CJ
-y-
x
o
z
X
o
CO
DC
Q.
U
J
C
D
3
| ANTHRACITE COAL
5
1—
RESIDEN
_i
LU
13
U.
O
CM
| BITUMINOUS COAL
LU
D
LL
2
q
i—
CO
J5
8
o
CM
| DISTILLATE OIL
(AREA)
_i
Z
cc
LJJ
1-
X
UJ
| RESIDUAL OIL
j-
CM
| NATURAL GAS
I WOOD
j-
CM
TOTAL
1 ANTHRACITE COAL
O
ELECTRI
| BITUMINOUS COAL
Z
0
•
GENERA
| LIGNITE
(POINT)
| RESIDUAL OIL
| DISTILLATE OIL
CO
O
cr
D
H-
z
[ PROCESS GAS
UJ
O
u
| SOLID WASTE/COAL
TOTAL
LJJ
cr
<
ANTHRACITE COAL
^
_
IN DUST R
Z
o
Q_
LJJ
D
LL
CD
LJJ
cr
*
BITUMINOUS COAL
J-
m
Z
o
Q.
h-
Z
O
-L
LJJ
Z
CM
LU
LX
<
| RESIDUAL OIL
CO
J-
CD
h-
Z
O
Q.
t— t
CM
UJ
IT
<
6
LU
1—
^
-1
75
D
H
Z
o
CL
a-
LU
o:
<
NATURAL GAS
en
i—
Z
0
ex
UJ
cr
<
PROCESS GAS
^
0
CL
Z
o
Q.
LJJ
O
u
LU
-------
90
I
•H
•M
CJ
O
CJ
CJ
1.
X
O
z
X
0
H
CC
Q-
UJ
CJ
CC
§
UJ
CC-
_J
O
LU
1-
CJ
(T
I
1—
_J
CJ
COMMER
LU
LL
H
z
0
Q_
FIONAL
INSTITU"
O
\—
in
\ COMBU
j-
ID
LU
X
0
CJ
CO
D
O
Z
D
CD
LU
U.
-J
^
-7
\ EXTER
H
Z
0
Q_
LU
_)
| (CONTI
l-
z
o
Q.
LU
0
UJ
CC
6
"5
Q
CO
LU
cr
I—
Z
o
a.
00
UJ
a.
_j
O
UJ
4
CO
O
h-
Z
£
I-H
J-
r-t
LU
CC
CO
O
D
1-
Z
t—
Z
O
Q.
UJ
CC
Q
0
i
H
z
o
Q.
I—
z
o
Q.
CO
O
oc
LU
a.
n
g
i-
z
o
0-
LU
X
0
in
CO
COMMER
1—
O
_i
u.
Z
O
INSTITUT
H
LL
AIRCR
9NI1S3.
L3NION3
g
H-
co
13
CD
o
CJ
TERNAL
TOTAL lf>
OJ
CD
CO
UJ
CC
z
o
t-
C/3
0
CJ
UJ
D
LL
TOTAL
CM
CO
f-.
CO
H
Z
5
Cu
-------
91
o
o
CJ
I
X
o
z
X
0
I—
cc
Q.
SOURCE
•i
O
z
rr
CHEMICAL MANUFACTU
in
FOOD/AGRICULTURE
CO
Csl
cr
PRIMARY METAL
LO
d-
co
IJ
z
o
z
3
5
O
z
cc
CO
z
LU
CL
O
o
f-H
CC
LU
I
0
0
_j
0
GOVERNMENT
(POINT)
z
^
—
LU
1—
CO
z
0
Z
Z
DC
D
CO
z
LU
CL
0
_,
O
1—
RESIDENTIAL
(AREA)
CO
UJ
cc
Z
o
CL
z
U
LU o
h- —
Z CC
O LU
f-H
f-H
LLJ
CC.
z
0
Q_
O
z
z
cc
D
CO
z
LU
CL
O
H
Z
o
Q.
h-
1—
CC
CL
LU
CC
g
O
Q.
cc
LU
I
3
f-H
CM
LU
CC
h-
Z
O
CL
_!
I—
O
H
COMMERCIAL-
INSTITUTIONAL
CO
I— 1
LU
CC
c\
LO
J-
z
0
Q_
z
U
Z
UJ o
Z cc
O LU
CO
J-
co
J-
LU
CC
t—
Z
0
Q_
a
z
z
cc
CO
z
LU
CL
0
h-
Z
0
Q.
Z
CJ
Z
>-
U
o
CO
0
1-
D
1—
Z
O
Q.
n-
LU
I
H
o
J-
CM
o
in
LU
CC
CM
10
J-
Z
O
c-
1-
O
INDUSTRIAL
J- CM
CO ItD
LLJ
CC
Z
0
Q_
TOTAL SOLID WASTE
DISPOSAL
_j
<
LU CO
C- ' ^— O
sii
-------
92
I
•p
B
LO
CTl
CNl
CD
rH
•8
H
O
O
o
X
X
o
z
X
o
co
cc
oc
1—
_J
I
o
1
ERCIAl
5
O
0
o
H
H
LL
oc
o
cc
_,
0
0
IINOUS
h-
m
LU
D
u_
LU
CO
LU
6
_j
0
Q
CO
LU
OC
LU
Z
o
CO
O
t-
o
1—
CO
LU
CO
CO
LU
>
CO
CO
O
Z
o
h-
oc
o
<
LU
G
2
Q
Z
I
CO
O
to
J-
oo
z
O
H
<
OC
O
Q.
CO
Z
tr
H
o
1—
Z
O
\-
t—
CC
o
Q.
l<
-------
93
1
•H
•P
g
CJ
LT>
1^
O1
CM
I
a>
rH
•8
H
O
O
I
X
o
z
X
o
CO
oc
Q.
LU
o
OC
D
o
CO
<£
LU
ir
CO
[INDUSTRIAL SOURCE
| DRY CLEANING
g
H
K <
Z EC
LU O
> °-
_l <
0 >
CO LU
| STRUCTURAL
| FROST CONTROL
[ SLASH BURNING
| WILD FOREST
| AGRICULTURAL
CO
LU
(E
LL
CD
CO
J-
l-t
| UNPAVED ROADS
| UNPAVED AIRSTRIPS
| PAVED ROADS
a
MINERAL PROCESSIN
CO
| TILLING ACTIVITIES
LU
>
QC
O
a"
z
CO
$
\ LOADING CRUSHED RO
o
o
ID
m
[CONSTRUCTION
tr
<^
— LU
a o x
LU < H
CO ^ y_
^ ^ O
<
Q co H
STORAGE PILES
TILLED LAND
UNTILLED LAND
10
i— i
cr>
LO
LO
OO
LO
i — 1
W1)
1
Total Miscellan
, CO
CO 0
D Z
ui > ^
5: CQ _i
rr Q *^
n LU ^
" D H
< 0 Z
CO
D
0
LU
Z
_l
LU <
O LU
co or
i <
AREA
cr>
CM
CO
l—i
POINT
,
,-J
|
O
l—
D
Z
«^
OC
a
-------
O
D.
O
94
-
tr
„
LLJ
Q.
CO
Z
o
J—
CO
Z
0
CO
CO
LU
o
u
o
i
x
o
z
X
o
oo
1—
CC
-------
95
.3
I
o
oo
CTl
CD
T-H
O
u
cj
I
X
o
z
X
o
uo
1-
o.
LU
0
DC
O
CO
LU
tr
<
_,
0
O
LU
I—
ANTHRACI
COMMERCIAL-
,
LU
LL.
h-
Z
O
Q.
INSTITUTIONAL
Z
O
1—
CO
~^
m
O
0
i-H
^
LU
CC
<
_J
o
CO
Z)
BITUMINOI
LU
U_
_|
^
Z
Q^
LU
1-
X
UJ
^
Z
o
CL
_
Q
LU
D
Z
h~
0
CJ
z
o
a.
31INDI1
LU
OC
<
-J
1 ^
RESIDUAL
H-
Z
O
Q.
j.
CO
LU
CC
<
_J
0
LU
DISTILLAT
h-
Z
O
O-
01
J-
I-t
LU
DC
<
CO
o
NATURAL
h-
Z
O
O-
UJ
DC
<
Q
O
i
Z
0
Q_
1—
Z
o
Q.
o
1
o
cc
1 —
LIQUID PE"
H
Z
O
Q.
OTHER
j-
LO
CM
LU
cc
4
TOTAL
H
Z
o
ex
1-
Z
O
Q.
OTHER
LO
r>
LU
OC
**
CO
oc
t-
z
o
Q.
z
o
h-
co
CD
O
o
TOTAL EXTERNA
_,
5
LU
DISTILLAT
ELECTRIC
1
LU
D
UL
CO
O
NATURAL
GENERATION
Z
O
p
CO
QD
8
DIESEL
~r
^
Z
OC
LU
1-
z
OTHER
TOTAL
_j
5
LU
DISTILLAT
INDUSTRIAL
CO
a
NATURAL i
LU
LL.
GASOLINE
DIESEL
OTHER
TOTAL
DIESEL
COMMERCIAL-
TOTAL
_i
LU
INSTITUTIONAL F
AIRCRAFT
ENGINE TESTING
0
CO
D
CO
O
TOTAL INTERNAL
3
r-t
LU
OC
z
o
1-
co
CD
O
O
LU
U_
^
O
CO
in
CO
t-
z
2
-------
96
s
•M
g
U
o
CO
CT)
0
O
o
I
X
o
z
X
o
CO
1—
cc
CL
SOURCE
^
J-
5t^
CO
0
i— 1
MINERAL PRODUCTS
PETROLEUM INDUSTRY
WOOD PRODUCTS
PROCESS EVAPORATION
METAL FABRICATION
LEATHER PRODUCTS
0
TEXTILE MANUFACTURI
INPROCESS FUEL
OTHER/NOT CLASSIFIED
00
en
00
f-H
TOTAL
CC CO
Q°0
? £
J-
H
Z
0
o_
z
o
z
LU Q
-5
z cc
O LU
en
j-
CM
in
LU
rr
vl
1—
Z
0
Q_
O
z
z
cc
CQ
Z
LU
Q-
O
H
Z
n
o_
z
o
>-
o
o
CQ
0
Z)
H
Z
0
Q.
nr
LU
I
1-
o
0
J-
J-
LO
LU
CC
"-
CM
cn
j-
i-
z
n
Q-
_l
O
1—
INDUSTRIAL
J-
CO
CO
LU
CC
-------
97
8
g
U
o
co
cr>
•9
H
o
o
o
I
X
o
z
X
o
CO
(-
CC
<
CL
LU
CJ
QC
D
O
CO
ID
OO
t^
>
1-
D
Q
H
I
f n
CJ
i
cn
i— i
i— i
>
1-
D
Q
>-
^>
<
LU
I
GHWAY ]
X
LL
LL
O
_i
(-
o
1—
LU
Z
_J
O
CO
<
CD
>
(-
D
Q
>~
^>
<
LU
X
GHWAY ~~|
X
LL
LL
O
_l
<
tr
_i
i—
o
H
_l
LU
CO
LU
5
CO
LU
_l
Q y
Z I
< LU
_l >
>
CC
<
H
_l
§
_i
>
o
<
o
cc
LU
^
5
O
<_>
TOTAI
1—
LL
<
CC
CJ
tr
<
_j
0
CJ
CO
D
O
Z
i
D
(-
CD
_!
LU
D
U-
_i
LU
co
LU
5
_l
O
_l
<
D
Q
CO
LU
cc
LU
z
o
CO
<
CD
<
I—
0
I—
CO
_J
LU
CO
CO
LU
>
CO
CO
O
_J
Z
o
H
<
CC
O
Q_
<
>
LU
CD
Z
Q
Z
X
CO
<
CD
LO
O
cn
Z
O
h-
<
i_
cc
O
Q.
CO
Z
<
cc
1-
-J
<
H
O
I-
Z
o
(-
<
K
CC
O
(X
Z <
1-
£<
-------
98
4->
U
o
oo
to
0)
^H
•8
O
o
o
i
X
0
z
X
0
CO
,_
oc
Q.
LU
CJ
oc
D
o
CO
5
LU
CC
s
CO
INDUSTRIAL SOURCE
| DRY CLEANING
0
1—
l- <
SOLVEN
EVAPOR
(STRUCTURAL
[ FROST CONTROL
[ SLASH BURNING
| WILD FOREST
| AGRICULTURAL
CO
LU
tr
LL
CTl
CO
J-
1— 1
| UNPAVED ROADS
| UNPAVED AIRSTRIPS
| PAVED ROADS
O
| MINERAL PROCESSIN
t^
oo
| TILLING ACTIVITIES
LU
>
cr
0
Q"
z
CO
x
O
| LOADING CRUSHED RO
o
o
CO
ro
| CONSTRUCTION
cr
—
^
_ LU
Q CJ I
UJ < 1-
CO ^ u-
^ > <
O co H
| STORAGE PILES
| TILLED LAND
| UNTILLED LAND
u
o
LT
v
?
Q
e
c.
• r-
a
1
, CO
oo Q
D Z
Q >
^_ ;>
^ 03 -1
§ 2 £
< CJ Z
CO
O
z
«J '"'•>
LU <
(J LU
CO DC
i <
to
rH,
LU
cr
OO
J-
co
CM
POINT
,
O
1 —
Q
Z
<3.
X
(J
-------
i,
o
o.
LO
CO
CTl
(A
in
•a
H
/;
n
J
fl
J
u
n
fl
-^
3
)
cc
•
cc
UJ
cc
UJ
O-
00
z
o
h-
00
z
o
GO
00
~
LU
C.
c
C
c
0
CJ
CJ
I
X
O
Z
X
0
CO
H
CC
-------
100
nd
1
•H
•p
g
u
LO
oo
g
•H
10
a
u
0
X
rH
I
0
O
u
I
X
o
z
X
o
CO
h-
CC
0_
UJ
LJ
CC
D
O
CO
LU
CC
"•
—J
u
o
LU
ANTHRAC
,
COMMERCIA
LU
LL
(-
Z
n
Q.
<
INSTITUTE
O
| COMBUST
00
t*^
LU
CC
<
_l
n
o
CO
BITUMINO
LU
LL
_l
| EXTERNA
H
z
0
O-
Q
LU
| (CONTINU
h-
Z
n
Q_
LIGNITE
-
UJ
cc
"•
6
RESIDUAL
K
Z
n
Q.
00
CO
UJ
cc
"•
|
O
LU
DISTILLAT
1-
z
n
o_
in
CO
I-H
LU
CC
**•
CO
NATURAL
1-
z
0
O-
LU
CC
<
D
0
o
z
n
Q_
^
Z
n
CL
(T
O
CC
1—
LIQUID PE
h-
Z
n
Q.
OTHER
P
\i
1U
cc
*
TOTAL
H
Z
O
o.
1-
z
n
Cu
OTHER
CM
UJ
CM
1-1
111
CC
4
cn
1^
cn
LO
1-
Z
0
O-
USTION
O
u
j
z
DC
TOTAL EXTE
.j
5
LU
DISTILLAT
ELECTRIC
UJ
D
UL
co
O
NATURAL
GENERATIOh
O
1 COMBUST
DIESEL
| INTERNAL
OTHER
TOTAL
i
6
LU
DISTILLAT
INDUSTRIAL
CO
CT
NATURAL
LU
LL
GASOLINE
DIESEL
OTHER
TOTAL
DIESEL
COMMERCIA
TOTAL
_,
LL
5-
INSTITUTION
AIRCRAFT
.n
Z
ENGINE TEST
O
fe
CO
o
O
Z
TOTAL INTEF
uJ
CM
l—f
LU
CC
z
O
1-
D
2
O
O
LU
LL
O
-------
101
X
X
CD
0
CJ
CO
z
LLJ
O_
0
i— t
cc
LU
I
0
£
_J
O
1—
GOVERNMENT
(POINT)
z
in
t-
LO
Z
o
(T
Z
Z
cc
-i
CD
Z
LU
Q-
O
_j
O
I—
RESIDENTIAL
(AREA)
CM
i-i
t— I
LU
CC
^
Z
O
CL
z
CJ
z
— z
Z cc
O LU
CO
cn'
o
CN(
in
CC
^
h-
Z
n
CJ
z
cc
D
CD
Z
LU
Q.
O
H
Z
0
Q-
t-
z
LU
1-
APAR
LU
CC
<
t-
z
0
CL
cc
LU
I
o
o
CM
CM
cr>
Ul
CC
11
H
z
0
Q_
_J
o
H
COMMERCIAL-
IMSriTUTIONAL
cn
0
CM
LU
cr
^
o
CM
LO
Z
0
Q-
z
Z
- z
Z cc
O LU
CO
LO
III
CC
Q
O
CO
O
D
Q.
rr
LU
I
\—
0
CM
00
cn
LO
111
cc
^
o
CM
to
z
0
o_
K
o
H-
INDUSTRIAL
CM
0
CM
cn
cc
11
CO
in
Z
n
CL
TOTAL SOLID WASTE
DISPOSAL
_j
*i
LJ | — O
— ' ^ co
c/5 ? 5
-------
102
1
•H
•M
g
U
LO
oo
cr>
0>
rH
I
o
o
CJ
X
X
o
Z
X
o
CO
PART
LU
CJ
CC
O
CO
01
co
oo
I—
D
Q
1—
X
(J
_J
o-
h-
D
Q
f
^
LU
X
GHWAY
j;
LL
U_
O
rf"
M.
1 —
0
LU
Z
_J
o
co
o
(-
Q
^
^
LU
I
GHWAY |
X
u_
u.
O
— 1
CC
_J
<3-
1 —
0
i—
LU
CO
LU
5
CO
LU
Z I
< LU
-J >
cc
t—
^J
i
CJ
j
o
QC
LU
O
CJ
H
O
(-
LL
CC
0
cc
_,
O
o
CO
D
0
Z
D
1—
CD
LU
D
LL
LU
CO
LU
5
_J
6
<
RESID
LU
Z
o
CO
(D
0
h-
CO
LU
CO
CO
LU
CO
CO
O
_J
Z
O
<£
cc
O
<
UJ
0
Z
_J
Q
Z
X
CO
CO
CM
o
r-H
Z
o
£i
\—
cc
O
Q.
CO
cc
r-
_l
r-
O
I—
Z
o
K-
CC
O
Q.
||
-------
103
.a
^
oo
CTl
•s
0
(J
o
i
X
o
z
X
0
CO
H
cc
^
Q_
LU
O
QC
^
O
CO
<
UJ
tr
| INDUSTRIAL SOURCES
| DRY CLEANING
o
t- <
Z DC
UJ O
> 5?
_i <
o >
CO LU
| STRUCTURAL
[ FROST CONTROL
[ SLASH BURNING
| WILD FOREST
[ AGRICULTURAL
CO
UJ
tr
LL
en
CO
J-
i— i
| UNPAVED ROADS
| UNPAVED AIRSTRIPS
| PAVED ROADS
| MINERAL PROCESSING
^
F*.
CO
[TILLING ACTIVITIES
tr
Q"
^
oV
ifi
3
[Total Miscellaneo
L_ in
ts> Q
D Z
Q g
% ffl -J
1 0 <
O m 5
£ 3 H
< 5 z
CO
D
O
LU
Z
_J
UJ <
0 UJ
co tr
I <
CO
o
• J-
I-l
LU
EC
co
CO
CO
CN|
•
1—
Z
O
0-
_J
|_
O
i—
Q
Z
<^
tr
a
-------
104
PROBLEM 4-1
Estimate the future particulate emissions for the clay products industry in
County X in the period 1975 to 1985. Assume the following information is
known.
There are currently two clay products plants in County X with the following
characteristics:
Company ABC - current production is 35,000 tons of finished product per year.
Company DBF - current production is 15,000 tons of finished product per year.
Regional planning data estimates a growth in demand for the product from the
region of 2.51 per year. A federal New Source Performance Standard is expected
in 1983 that will limit emissions to 4.5 Ib per ton of finished product. The
current state regulation is 10 Ib per ton. Also, it is known that Company DBF
is scheduled to shut down its current plant at the end of 1984. Assume the
air quality analysis is being performed for 1980 and 1985 and the baseline
inventory is for 1973.
-------
105
5. ALLOCATION OF EMISSIONS
To this point, the focus of the emission inventory portion of the
analysis has been to generate the emission pattern on a county-by-county
basis for the region under study. Prior experience with air quality manage-
ment has shown the following:
Countywide spatial resolution of emissions is too coarse
for particulates, SOa, and CO.
Countywide spatial resolution of emissions is too fine
for oxidants, nitrogen oxides, and hydrocarbons.
The latter three pollutants are reactive under ambient conditions and their
impacts on air quality can be felt at large distances from the emitting sources.
For this reason, there is no pressing need to develop an analysis with finer
spatial resolution for these pollutants. Rather, they will be treated on a
regionwide basis. The exception to this general guideline is when an analysis
of one or more of these pollutants will be made with a photochemical disper-
sion model. In this case, the model requires, as input, a better spatial
description of emissions than is available from countywide data and the same
allocation procedures that will be described for the other pollutants can be
used.
For particulates, S02, and CO, the effects of individual sources are
much more localized in impact and an adequate analysis and control strategy
development relies on a fine spatial resolution of emissions. The objective
of this section is to describe techniques that can be used to achieve this
resolution.
These procedures should not be viewed as necessarily separate and
distinct from the development of the baseline inventory or the estimation of
future emissions. In many instances these techniques can be used in parallel
rather than in series with the previous methods.
5.1 DETERMINATION OF GEOGRAPHIC SCALE
The identification of the level of spatial resolution is a function
of several variables:
Available data
Anticipated problem areas
-------
106
Anticipated problem sources
Applicability to modeling
Applicability to strategies
Resources available
The principle constraint on spatial resolution is, of course, the available
data. If the information is on too coarse a geographical scale, then an
extensive effort is usually required to improve the data base. In general,
this will be required only in special circumstances and only for special sources.
It is not unusual to have data from different source categories with different
spatial resolutions. For example, population data may be available on census
tract subcounty resolution, while transportation data may be on planning dis-
trict resolution. This presents no special problems since all sources even-
tually will be mapped into a master grid.
The spatial resolution should be fine enough to distinguish special
problem areas and special problem sources. As an illustration, it is not
possible to treat a major metropolitan city in its entirety. The spatial
resolution should be able to distinguish the Central Business District (CBD)
from industrial parks and from residential neighborhoods.
The use of dispersion models dictates certain limits on the size
of sources and hence on the spatial resolution. Although the models can
theoretically treat any size areas, the accuracy suffers if the sources are
too large and computation expense suffers if they are too small. This con-
sideration becomes more significant when the master grid is developed and
will be discussed later.
The applicability of control strategies is an important consideration
and is one of the prime reasons for using a subcounty resolution. In general,
control strategies may not be appropriate for application to an entire county
across the board. From an enforceability standpoint, however, a control
strategy will, most likely, be applied within political boundaries (as opposed
to arbitrary areas) and the analysis should be sensitive to these considerations.
Finally, the resources available will dictate the level of spatial
disaggregation that can be tolerated. The large number of calculations
required for each subcounty area dictates the need for computer assistance
if any significant number of areas are to be considered.
-------
107
5.2 ALLOCATION PARAMETERS
5.2.1 Population
There are a number of sources whose emission distribution may be
linked directly to the population distribution. For each county there are
*
several different geographical descriptions of population distribution, any
one of which may be used. The Census of Population and Housing provides the
most widely available demographic data, although some areas have regional
planning commissions, which normally embellish the resolution of the Census
data.
The types of subcounty areas for which demographic data are normally
available include municipalities, census tracts, Master Enumeration Districts,
regional planning districts, and townships. In every area there are population
data available on at least a municipality basis. This information is tabu-
lated in the Census of Population for all places containing a population of
2500 or more. -The geographical location of these municipalities may be found
either in the census publications or on regional maps. It should be noted,
however, that the municipalities may not cover an entire county, since there
may be extensive unincorporated areas. In these cases, the total of the
municipality populations is subtracted from the total county population (also
tabulated by the Census Bureau) to determine the residual county population
(the number of people living outside of municipal boundaries). This selection
of subareas will generally result in delineating only a small number of dis-
tricts within each county, and population-based allocations can, therefore,
be easily handled without a computer.
For most large, urbanized areas, the region is subdivided into tracts
by the Census Bureau. These tracts, while lying within county boundaries,
do not necessarily lie along municipal boundaries and may overlap several
political jurisdictions. The tracts are large in the less densely populated
areas and small in centers of population concentration. Because tracts are
delineated according to population density, they portray the distribution
of a county's population at a high degree of resolution. Nevertheless, choice
of census tracts as the population-based subareas must be made with careful
consideration. In some regions the number of tracts, though larger than the
number of municipalities, is still small enough to be managed without computer
assistance (the Atlanta SMSA has 238 tracts). On the other hand, some regions
-------
108
have a large number of tracts, which make hand computations unwieldy (the
Chicago SMSA has over 1500 tracts). Census tracts should therefore be used
as subcounty areas only when sufficient computational resources are available
and good detail is needed. In any case, the Bureau of the Census publishes
the tract information both in printed form and on computer-compatible magne-
tic tape.
Master Enumeration Districts are essentially the same as census
tracts in areas that are tracted and have other definitions in untracted
areas. Data for these districts are tabulated on computer-compatible mag-
netic tape, available from the Census Bureau. A set of computer programs
has been written to enable one to process this information and to develop
a grid system for the allocation of area source emissions. These programs
and gridding procedures are documented in Ref. 8.
Areas in which there are active regional planning commissions will
usually be subdivided into planning districts that the commissions use for
displaying data. In some areas, these planning districts lie along municipal
or census tract boundaries. In other areas, they are drawn up to meet the
specific requirements of a particular commission and do not correspond to any
other subarea definition.
In most areas of the country, a political jurisdiction, referred to
as a township, is superimposed on existing municipal jurisdictions. The town-
ships are normally square-gridded with 36 square-mile sections in each. These
townships can be used as regional planning districts, as in Northern Illinois.
The main conclusion to be drawn from this discussion is that there
are a variety of subcounty areas for which population distribution information
is displayed. In any given region, there may be more than one subarea set.
Choice of the appropriate set of subareas for use in allocating emissions is
based on identifying which one contains the most detailed set of information
and yet is manageable within the resources of the planning agency charged
with maintaining air quality. It is recommended that the subareas to be used
for the transportation allocation and the commercial/institutional-industrial
allocation be investigated prior to making a final choice of the data resolu-
tion needed to allow for the possibility that one subarea set may provide
information more appropriate for all stages of the air quality analysis system.
-------
109
5.2.2 Transportation
As with the population-based subcounty areas, there are a number of
subarea sets available for describing the distribution of transportation sys-
tems. In some areas of the country, the transportation planning agency may
use one of the regional planning grids for developing its data base, in which
case the population-based subareas and the transportation-based subareas can
be selected as being one and the same.
In other regions, the transportation planning grid may be developed
separately. This separation presents no unusual problems in the allocation
procedure, and there is no need at this point to try to convert one grid system
to another.
Some regions will have no grid system that is used for transportation-
related data display and all that will be available will be highway department
road maps. In most areas where this situation prevails, vehicle count data
will be available only for the major expressways and busy arterials.
In some instances, however, the state transportation department will
have developed traffic data on a link-by-link basis as part of the Continuing,
Comprehensive, and Coordinated (3-C) transportation planning process. The
links will generally be described by the UTM coordinates for their end points
and have vehicle count data on them.
5.2.3 Commercial/Institutional, Industrial, and Electric Generation
The most definitive form of spatial resolution of these source cate-
gories is the location of point sources in which their specific coordinates
are specified. The display of area source data is generally in one of the
forms previously described, such as census districts, regional planning dis-
tricts, or townships. An additional information display may be a land-use
map of the area on which the various land uses are coded to indicate the dis-
tribution of activity. While this type of presentation does not rely on a
grid network, it can nevertheless provide useful data for the analysis by
describing the spatial development of the land in the area. This data may
be converted to a grid network using the techniques described in Reference 29.
-------
110
5.3 ALLOCATION PROCEDURES
There are basically three procedures for determining spatial alloca-
tions of emission-producing activity:
Locate the activity directly from available data,
Develop a distribution function of activity using an
allocation parameter (e.g., population), and
A combination of the above procedures.
Reference 13 outlines the recommended procedures for these methods as applic-
able to each of the source categories. The sources that can be spatially
located directly are:
Industrial point sources -- existing and some new,
Commercial/Institutional point sources -- existing and
some new,
Electric generation -- existing and most new, and
Limited access highways -- existing and most new.
For all existing sources, the spatial coordinates are available from the
emission inventory. (Although limited access highways are not normally spe-
cifically included in an emission inventory, their coordinates are easily
determined from maps.) For industrial and commercial/institutional new
sources, there is often only limited information available on new plant
locations. Where these data are available, they can be used to identify source
coordinates. For power plants, the location of new facilities is well docu-
mented in References 30 and 31. There may be uncertainty in the location of
some new sites, but, in general, the utility planning horizon is long-range
enough to identify most new installations. In a similar way, the location
of most new limited access highways is sufficiently well documented (at least
in terms of alternative possibilities) to enable their geographic location to
be well defined.
Source activity that cannot be located directly can use a distribution
function to allocate emissions. A simple model would start with the knowledge
of the activity on a countywide basis and then distribute this activity to
each subcounty area based on that area's proportion of some countywide para-
-------
Ill
meter, such as population or employment. For example, if a county is divided
into 20 subcounty areas and the 5th area contains 12% of the county's popula-
tion, then all emission-producing activity that is related to population (e.g.,
solid waste disposal) can have 12% of the countywide emissions allocated to
subarea 5. The types of distribution functions that can be developed include:
population
dwelling units
dwelling units by fuel use
employment
land use
area
An illustration of the effect of using different distribution functions is
found in Appendix B of Reference 13 dealing with residential fuel combustion
emissions. The three orders of allocation procedures illustrate the use of
a population distribution function, a dwelling unit distribution function,
and a fuel-use and building-size distribution function available for each
subarea, respectively. In all three cases, the countywide total fuel use and
emissions are the same. Use of the population and dwelling unit distributions
give only slightly different results among the various subareas. The most
detailed distribution, however, results in a change in the calculated emis-
sions of 191 in the largest subarea (i.e., Atlanta).
The third type of allocation procedure is a combination of the spe-
cific location information and the distribution function. Its aim is to use
whatever specific location data are available for a source category and allo-
cate the remainder by some form of distribution function. Its prime utility
is in allocating activity from new sources that cannot be placed with any de-
gree of certainty. Starting with the computation of emissions on a countywide
basis (with account taken of new source regulations where applicable), the
procedure can take one of three paths with regard to locating new source
emissions:
1. Assume all new source emissions occur at the same
location as existing sources (i.e., growth-in-piace).
2. Assume all new source emissions occur at new sites and
allocate according to an employment or land-use distri-
bution function.
3. Determine the mix between existing and new source
activity and allocate accordingly.
-------
112
The first method is the easiest since the location of the source activities
are already known and are assumed to remain the same. It is also the least
accurate since it does not reflect realistic development in most areas. The
second and third methods are more accurate but require significantly more
data.
In some situations it may be possible to utilize a sophisticated
growth and development model to determine the spatial distribution of emis-
sions. Models that can be used for this purpose are represented by the work
32 33
of Lowry and the Hackensack Meadowlands Project. These models generally
follow the third allocation procedure (i.e., combine existing locational
information with distribution function approaches) but in a much more analyt-
ical way than the simple approaches described here. The experience and
resources of the agency performing the air quality analysis will determine
if this approach is useful.
-------
113
PROBLEM 5-1
Using the previously calculated data for the clay products industry in County
X (Problem 4-1), allocate these emissions to the ten subcounty areas defined
by the planning commission data in Appendix D for 1980 and 1985. Company ABC
is in Superdistrict 16 and Company DBF is in Superdistrict 15. Assume that
both plants are currently operating at 100% capacity and that the allocation
of new source emissions will be based on the growth in industrial land area
in each Superdistrict (see Table D-9 in Appendix D for a summary of this
calculation). Tables 3.4-1 to 3.4-4 from Reference 13 (reproduced here as
Figs. 5-1 to 5-4) may be used to structure the computation. Tables 3.4-1
and 3.4-2 may be filled in directly using the results of Problem 4-1. In
using Table 3.4-3 note that the Employment Allocation Proportion is replaced
by an industrial land use growth proportion.
-------
114
•s
E-
o
•H
Ifl
01
to rt
•H
0> ^i
8
^
Jk. 9
#3
< O)
0}
(j
r?
C
O
in
Ul
t/l
wi
y
S
CL.
§
n)
u
H
c
V
hi
3
fM
S§
^ y
sr R
£S
S ^
s
^^ "P
a
S
£ |
(3)
Process
.,
"a u
H 1
/••^ ^ (
1
V)
1
1
I
3
i/i
1
cS
•
X
?fc
^
*Eraployment
"Growth Facto
to
10
0)
•H
s
03
(J
g
CO
O
On
bO
•H
-------
115
(XJ
t
^t-
K)
0
i—I
•a
H
Industrial Point and New Source Process Emissions by Process Category
A. County
B. Year
C. Allocation Order 2 and 3
1
1
•3
r-orci. ast KMT
xisiing Point Sources
0
Forecast Year - \
1 Basel me Year
Fxisting Point Sources
s i
^ u
~ J
So-
c
i
az I
5s *
c
i
• — - u.
1
t/
. jj g
— a.
c
*
s i
?. 9
- 5
~ q
1 s"
c- S
, — *
5 9K
5 y
E 8
O Q
— (f)
£ |
oo" u.
*
- — . x
£ S
j^ U
S 8
£ a"
" s.
0)
Industrial
Process Category
^\
g
d
* Enploynent
••Growth Factor
Fig. 5-2. Industrial Point and New Source Process Emissions by Process Category, Table 3.4-2 from Ref. 13
-------
116
Table 3.4-3
Process Emissions by Process Category and Subarea
A. County
B. Subarea
C. Year
D. Allocation Order
2 and 3
Industrial
Process Category
(1)
E. SUBAREA TOTAL
Total
Enp.
(2)
Point
Source
Erap.
O)
New
Source
Emp.
(4)
EAP*
(5)
1
New Source Emissions (tons/yr)
(6)
Part.
(7)
SOx
(8)
CO
(9)
1IC
(10)
N0»
•Employment Allocation Proportion
Subarea Category New Source Enyloyment
Total Category
ry New
New S
Source Enjoyment
Fig. 5-3. Process Emissions by Process Category- and Subarea,
Table 3.4-3 from Ref. 13
-------
117
0
rt
g
b
03
g
1
1
0
• H
t/1
1
DCCSS
i-,
CX,
O
I
0
3
'o
ex,
03
in
c
>,
5
3
<
rt
e
EC
U
rt
#
o
r
ha
i
-"
I-
1
5
ffl
8
<
missions
%
s
f
f
i
|
>
y.
E
1
p
Industrial
1 Process liitouon'
!*
|'s
|8
^~
11
ct
= 5
r =
'JE
= *"
~ -
E |
oc C
r =
— C
a?
E =
-K
'— -O
C -"
L-l
§
^a-
i
'*
Kl
0)
i—I
•§
H
«\
^
03
CO
03
8
03
CO
in
(D
8
CO
^
T3
G
•H
O
CX,
to
LO
•H
Pi
-------
118
5.4 MASTER GRIDDING
To this point, the allocations for each source category may have
been done with different spatial resolutions. It is possible that the county
emission pattern may be described as in Fig. 5-5 with a variety of source
presentations. It is now necessary to coordinate all of these results onto
a single master grid system that can be used for modeling and strategy analysis,
Choice of Coordinates There are a variety of rectilinear coordinate
systems that can be used in 'developing a master grid. It is desirable, how-
ever, to use Universal Transverse Mercator (UTM) coordinates, which are uni-
versally available on U.S. Geological Survey maps. The UTM coordinates offer
the widest generality and applicability. The only translation problem occurs
when an analysis area lies between two UTM zones where the abscissas are not
continuous. This is easily treated via careful bookkeeping.
Some states have local coordinate systems that may be considered.
Uncertainty as to the reference points will make these systems difficult to
use for anyone unfamiliar with the area.
Selection of Grids The master grids chosen should, in general,
be square with variations in size depending on the resolution of the subcounty
areas. The smallest grid square chosen should be 1 km x 1 km. Smaller grids
would result in data that does not, in all probability, exist in the original
data set. The largest grid square chosen should be 8 km x 8 km. Larger grids
would cancel some of the subcounty resolution already achieved.
Reference 8 describes a procedure for using Bureau of the Census
computer tapes and the EPA-developed CAASE program (Computer-Assisted Area
Source Emissions) for developing a grid based on population density. The
process involves the interaction of a series of five computer programs and
some manual techniques. This is one way in which the grid configuration can
be developed.
An alternative method of developing a grid system is to rely on a
review of the available subcounty areas. The process starts with an overlay
of the largest grid squares (e.g., the 8 km x 8 km grid) placed on the sub-
county area map. The grids are then subdivided until the smallest chosen
grid size is reached or until the subdivision grid contains primarily one
subcounty area.
-------
119
COUNTY X
a. POINT SOURCE LOCATIONS
b. LINE SOURCE LOCATIONS
c. MUNICIPALITIES d. PLANNING DISTRICTS
Fig. 5-5. Possible Displays of Spatial Emission Patterns
-------
120
When the grid has been developed for one subeounty set (e.g., munici-
palities) , it can then be laid over the other sets and be modified accordingly.
This modification consists of further subdivisions to match the differing
resolutions. Figure 5-6 gives the final grid system chosen for County X.
Subeounty Area Mapping Once the grid system has been chosen, it
is necessary to map the subeounty areas into the master grid. This is done
on an areal apportioning basis; that is, it is determined what fraction of a
subeounty area (e.g., a census tract, municipality, etc.) is within the over-
lying grid square. All emissions from that area are then allocated to the
master grid using this fraction. The computation is:
Contribution of Subeounty Fraction of area of T(jtal emissions
Area i to total emissions = ^SrvL • x in Subeounty (5-1)
in Grid j Grid j Area i
As an example, if 35% of the area of census tract 70 lies in grid square no. 5,
then 351 of the emissions computed for census tract 70 will be assigned to
grid 5. The area fractions may be determined using a planimeter or by using
an "eyeball" estimation procedure.
Appendix H documents a computer code that can be used to map emissions
from the subeounty areas onto the master grid once a complete set of area
fractions is assembled.
Model Inputs With the development of the master grid network, it
is now possible to use the emission data as input to a dispersion model.
The master grids may be used as area sources and the previously identified
point and line sources can be input directly. Alternatively, the point and
line sources can be located in their appropriate master grids and the entire
system modeled as a set of area sources.
-------
121
3790-
3780-
3770-
3760-
3750-
3740-
3730-
3720-
371(1
770
Fig. 5-6. Master Grid System for County X
-------
122
PROBLEM 5-2
Using the grid map of a portion of County X given on Fig. 5-7, develop a grid
system that matches the resolution of the indicated subcounty areas.
-------
123
s
SANDY SPRINGS
(SOUTH)
3750
37 42
37 34
3726
i i i i i I i*-r-h i i i i I i i i i -i i
733 ?4I
Fig. 5-7. Portion of County X for Gridding Problem 5-2
-------
124
PROBLEM 5-3
Using the solution map of Problem 5-2, determine the area fractions for the
Northwest Superdistrict into its overlaying grid squares. The total area of
the district is known to be 77.0 sq km.
-------
125
6. MODELING PROCEDURES
An atmospheric simulation model is used to convert the air pollutant
emission data into ambient air pollutant concentration estimates. The model
is an analytical tool that helps the air quality analys.t determine the effect-
iveness of his control strategies. Like any tool, the model is useful only
if the user understands its strengths and \veaknesses and does not attempt to
apply it beyond its capabilities. Perhaps the most widespread abuse of atmos-
pheric simulation is to regard the model as a mystical "black box" that magic-
ally transforms emissions into air quality. This approach to modeling is most
likely to lead to erroneous results and to an analysis that will not withstand
technical challenge.
6.1 CHOICE OF MODEL
There are numerous simulation models available to the air quality
analyst. These models vary widely in their data requirements, complexity,
accuracy, and applicability. Reference 12 gives concise descriptions of some
of the more widely used multi-source models (i.e., models that can be used to
study the impact of many sources or source categories on air quality), and
these are summarized on Table 6.1. The Rollback and Appendix J models rely
on proportional reduction assumptions; that is, if emissions are reduced by
a given amount, the concentrations will be reduced proportionately. Both
are only useful for rough approximations of air quality impacts. The Miller-
Holzworth model uses a dispersion equation that is integrated over an entire
urban area. It is a step better than Rollback in that it can be used where
there are no observed air quality data available. Both the annual and short-
term versions of the Hanna-Gifford model use a simplified dispersion equation
to compute the concentration over each area in which a receptor is located.
The basic Hanna-Gifford model can be improved with the addition of either
a point source or line source model to provide better spatial resolutions.
The Air Cjuality Display Model (AQDM) and its companion, the Climatological
Display Model (CDM) use a Gaussian-plume calculation to estimate specific
source-receptor contributions on an annual average. These are the most widely
used models in air quality analyses. The Sampled Chronological Input Model
(SCIM) and the Real Time Air Quality Simulation Model (RAM) also use a
Gaussian-plume calculation, but on an hourly basis. The APRAC-1A model is
-------
126
used for mobile source emissions and is a line source formulation for use
with highway links. The SAI model is a photochemical model that incorporates
both dispersion and atmospheric chemical reactions.
Table 6.1. Multi-Source Atmospheric Dispersion Models
Model
Model Technique
1. Rollback
2. Appendix J
3. Miller-Holzworth
4. Hanna- Gif ford
5. Hanna-Gifford with
Point Source
6. Hanna-Gifford with HIWAY
7. Air Quality Display Model
(AQDM) (also Climatological
Display Model (COM))
8. Sampled Chronological Input
Model (SCIM) (also Real Time
Air Quality Simulation Model
(RAM))
9. APRAC-1A
10. SAI
Proportional reduction
Proportional reduction of oxidant con-
centration to hydrocarbon emissions
Integration of Gaussian dispersion
across an urban area
Simplified dispersion equation over an
area
Addition of specific point source
calculations to basic Hanna-Gifford
Addition of line source calculations
to basic Hanna-Gifford
Gaussian plume concentration
calculations for annual averages
Gaussian plume concentration
calculations for 1-hour averages
Line source calculations for highway
and street links
Photochemical reaction and dispersion
calculations
In addition to these multi-source models, there are a variety of
single source models available that can be used to evaluate the impact of
isolated point sources or to develop control measures applicable to only a
few sources in an urban area. These models can be used to conduct microscale
analyses to further clarify and amplify the macroscale results of the multi-
source models. (This has been alluded to in the discussion of the Hanna-Gifford
-------
127
Model.) Descriptions of the available models are given in References 10, 24,
"7/^ TTT
34, and 35. Computerized models are also available. ' The discussion of*
Reference 10 indicates how such models might be used to implement a new source
review program to identify the impacts of large, new emission sources. An
q
indirect source review program can also make use of some of the microscale
analysis techniques.
There are numerous other models available for use in an air quality
analysis. Appendix A in Chapter 40, Part 51 of the Code of Federal Regula-
tions, presents EPA's general position as to the appropriate and acceptable
use of specific air quality simulation models. The EPA Regional Office
should, however, be consulted to determine the acceptability of comparable
models.
The choice of model is a function of many considerations including
the following:
Pollutant Ease of use
Averaging time Availability
Data requirements Reliability
Model output Applicability to air quality analysis
Table 6.2, taken from Reference 12, evaluates each of these considerations
for the ten most comonly used models. In some cases a model that specifies
one averaging time can be used to estimate concentrations for other averaging
17
times by using statistical techniques such as those described by Larsen and
24
Turner.
Proposed regulations require that, at a minimum, AQDM or equivalent
model be used for the air quality analysis of participates and S02 with appro-
14
priate estimations made for short-term standards. Proportional modeling
will be acceptable for CO and N02, while Appendix J will be acceptable for
oxidants. Table 6.3 gives the applicability of each model to an analysis of
each of the NAAQS pollutants. Deviations from these general guidelines are
possible with EPA Regional Office concurrence.
6.2 MODEL OUTPUTS
The output of the models varies considerably as indicated on Table
6.2. The Rollback, Appendix J, Miller-Holzworth, and basic Hanna-Gifford
-------
128
Table 6-2. Summary of Simulation Model Characteristics
Model Name
Rollback
Appendix J
Miller-
Holzworth
Hanna-
Gifford
Hanna-
Gifford
w/PSa model
w/HIWAY
AQEM, CEM
SCIM,b RAMb
APRAC-1A
SAIb
aPoint Source
These models
Pollutant
Specifi-
cation
Any
Ox
S02 ,TSP
S02 ,TSP
CO
S02 ,TSP
S02 JSP
CO
S02 ,TSP
S02 ,TSP
CO
CO,N02,Ox
Averaging
Time
Specifi-
cation
Any
1 hr
1 hr,
Annual
Annual
1-24 hr
1-24 hr
1-24 hr
Annual
1-24 hr
1-24 hr
1-10 hr
Emission
Data
1
1
1
1
2
3
3
3
3
3
2
are currently in a developmental and
Meteoro-
logical
Data
1
1
3
2
5
5
5
4
5
5
S
debugging
Concen-
tration
Estimates
3
3
3
3
2
1
1
1
1
1
2
phase; they are
Ease of
Use
1
1
1
1
2
2
2
3
3
3
3
Avail-
ability
1
1
1
1
1
2
2
2
3
2
3
not available for
Reli-
ability
3
3
1
1
1
1
1
1
2
2
2
Applic-
ability
to AQAS
3
3
3
3
2
1
1
1
1
1
2
general distribution
as computer programs.
Key to Table 6.2
A. Pollutant
Specif icati
on
E
. Concentration
Estimates
Any pollutant
Specific Pollutants (S02, TSP, CO, Ox, N02)
B. Averaging-time Specification
Any averaging-time
Annual Average
1 to 24 hour Average
C. Emission Data
1. Area-wide Emissions Total
2. Total emission distributed as finite area
sources
3. Detailed point, line, and area sources
D. Meteorological Data
1. None
2. Average wind speed
3. Average wind speed and mixing height
4. Frequency distribution of wind direction,
wind speed, stability, and mixing height
5. Hourly variations of wind direction, wind
speed, stability, and mixing height
1. Estimates at any specified point
2. One estimate for each area source grid
3. One estimate applicable to entire AOMA
F. Ease of Use
1. Slide-rule
2. Small computer effort
3. Major computer effort
G. Availability
1. Open literature
2. National Technical Information Service
3. EPA, upon request
H. Reliability
1. Can be verified and calibrated
2. Verification is incomplete, possibility of
calibration is uncertain
3. Questionable; acceptable for crude estimates only
I. Applicability to AQAS
1. Can distinguish between specific source and land
use type
2. Can distinguish between land use types only
3. Considers no distinction between sources or land uses
-------
129
Table 6-3. Models Applicable'to Specific Pollutants and Averaging Times
Pollutant 3nd
Averaging Times
S02, TSP
Annual Average
S02) TSP
24-hr and 3-hr
Averages
Alternative Models
Preferred Model
1. AQEM
2. CDM
1. ADDM*
2. CEM3
3. Single -source
node Is
1.
2.
3.
4.
1.
2.
3.
Model
Hanna-Gifford
Rollback
More sophisticated
model
Miller-Holzworth
SCIM.b RAMb
Hanna-Gifford
w/point source"3
Rollback
Condition of Applicability
Inadequate computer facilities
Complex topography and/or
Meteorology
Complex topography and/or
meteorology
No circumstances currently
envisioned permit its use
Availability
Inadequate computer facilities
Complex topography and/or
4. More sophisticated
model
5. Miller-Holzworth
meteorology
Complex topography and/or
meteorology
No circumstances currently
envisioned permit its use
CO
1-hr and 8-hr
Averages
Oxidants
1-hr Average
N02
Annual Average
1.
2.
1.
2.
1.
2.
APRAC-1A Rollback
Hanna-Gifford w/
HIWAY
SAI Rollback
Appendix J
Rollback
Single-source
models0
Unavailability of other models
Demonstration that Appendix J
is not applicable to region
Statistical conversion of averaging times required.
Repetitious application of model to each hour under consideration is required for averaging times longer
than 1-hr.
Usable only when atmospheric chemical reactions are not significant over the impact area.
-------
130
models result in a single estimate of pollutant concentration applicable over
the entire region under study. These models are, therefore, applied to the
expected worst receptor in the region. The short-term Hanna-Gifford model
can generate u concentration estimate over an area source grid. It can also
give concentrations at any point when coupled with a point source model or
with the HIWAY model. The SAI photochemical model calculates concentrations
over an area source grid. All the other models can compute concentrations
at any specified receptor point. A sample printout of the AQDM model is given
on Fig. 6-1. The receptor coordinates and the computed concentrations are
shown.
Another useful form of output for those models that can compute con-
centrations at any point is the isopleth or line of constant concentration.
If the computations are made at a large enough number of receptor points, then
the isopleths are easily drawn by interpolating between adjacent points. A
38
standardized computer program, SYMAP, is available, which uses a computer
line printer to draw the isopleths. Figure 6-2 is a sample output of the
SYMAP program, which has been coupled to the AQDM model. The printer draws
a symbol to represent the concentration level and also prints the computed
concentration at each selected receptor point. The concentration intervals
can be determined manually or automatically. The use of isopleths is an
extremely valuable tool in visualizing the general air quality situation and
identifying "hot spots," although its use is not critical to the air quality
analysis.
In determining the impact of various sources on air quality at a
given location, it is necessary to develop a culpability list; that is, a
list of the contributions of each source to the calculated concentration at
a given receptor. Since all of the models make use of the principle of super-
position in that the concentration is calculated as the sum of the emissions
dispersed from each individual source, it is conceptually straightforward
to develop the list. In practice, without computer assistance the task would
be extremely laborious for all but a very small number of sources. The AQDM
model has a routine incorporated into it that prepares such a list with no
additional burden on the user. Figure 6-3 demonstrates the output of this
routine. This information is extremely useful in control strategy development
and analysis. If a model is being used that does not have such a routine
built-in, it is highly recommended that the effort be expended to develop and
incorporate one.
-------
131
i RECb°T'jR ! Rhjr^cMHK LJC'TIUM ' fcXHK'tD AKIfHh'TlC
i __ UU^siEa __ i
_______ i
_______
1
!
|
|
1
|
(
>
1
!
1
i
i
1
1
1
i
i
i
j_
i
i
;
1
X
^
|
I2i
122 i
12.3.
12-f 1
i y •_ 1
127 i
123 '
i^i
130
iai '
133 !
134 '
i ^ .j i
130 I
137 !
Ii2i2 1
139 !
140 !
~~~142 ~~~!
143 1
145 '
140 I
~"l4d "" T
149 '
j.5U 1
1?..
ii»l •
~~~13H "
iin
13/ 1
:::ili::::i
/ v, .
7^ j .
23 ^
M 3 .
/?'j.
23. ~
"'33.
73- .
2lt' '
/40.
74 ..
74.'}.
740.
7 4 ,
740.
74J.
7 A ,
740.
^/M!'
740?
740.
"/4j?
VA^'
7'oV
' '• ^ *
24,;^
743.
74: .
.'4 ; .
'43.
,' 4 -j .
::::i:
i
«. 3 •' 3 0
o ' 373-5
u ' — Z ^i U
(_. j 7 (} u
j ' 1770
, ' J ' 25
3 , a -j
( . ' J 7 b 5
Li ' ' '1Q
U ' 3713
1 3720
Li ' - /31'1
0 3 ' 3 b
3.7.30.
J 1 3"'43
j ' 3 ? 3 0
J J ;2J
I ' 37aO
; ^^
C""7~"3 ' /5
j 3/aO
U • 5710
') ' ^/13
I ' 1 >±5
s ' 'i HO
-._'._.i735
u J /'«0
! ^'aU
" ^735
; i ?oO
377)
L' 3/0
.0
.0
Q
.0
.0
fj
.0
.0
*a
. )
. 0
.0
.0
*Q
. 0
.0
n
.0
'a
. J
.0
.0
.0
. ')
.
-------
132
if ----- 4. ---- 1 ---- + ---- 2 ---- + ---- 3 ---- + ---- ^ ---- + ---- 5 ---- + ---- 6 ---- + ---- 7 ---- + ---- fa- --- + ____
I.., ..... , ......................
^QU 3000*2 + + + "2 + + + * + +
1.14 SECONDS l-~iK MA
IS 0.05 hi l-.fcb
Fig. 6-2. Sample SYMAP Output
-------
133
SOURCE Ci'NTP IBUTIP',<; TO i-ivE S E L c r.
ANNUAL P*RT lLLuA[ PS
KlCRuGkit-'S PLR CURiT "tTL«
I SOURCE i f % ,
^ ' u'j"w2
' '*fa ' a'1 Vfc '
I 4 ? | 0 . ., -1 %
46 ; 0 . '. 'i .''•
' 47/ ' C'u-Vfc [
1 5j : °'^~ '
51 ' I'aLs '
1 ^ ' u'I*J :
1 ^3 \ ~.'^g^ '
i i> 4 , 0 . / ? "
i 33 1 J> . ', 3 % :
i 3 ,- 1 4 . i 5 %
i 39 ' -j . . 1 % ,
' 6J i r'J-J ;
I 6 i ! 0 . H ^ ''• |
6 ^ 1 2 . v '" %
i e>» : li.oo '/. .
RcCEPr.
) . 0 0
!) . 0 i
0 . 0 j
- _a*22.
O.Oj
c.oo
0.0 L
0.02
3.01
2*5d:
0.3^
a*2it
a»2oJ
3»4iii
1.97
"1.19
Q «.!&£
0.7d
0.0,
0.00
j»aji
O.Qy
Q^au^
0.62
_ Jill'
J J « i ..
Jf- ! ; tCt!JTbR 1 RhCcPTQ^
_^ 1J2 i 121
,; i U.'.M J<- \ o.oo -4
"- ' y .0" "• 1 0. ")0 7.
ij__i U.uUw2__4._ __u»DD03
v- . J.O-) '•', \ O.C'C %
/' 1 0 . 0 b /i i ? • 0 5 /i
1 0.0? ": i ^.OC %
i:- 0.0' •', j J.OO X
O.OO '- 1 0.00 %
V. 1 0 . 0 i % i 0 . C i %
7, \ o.s« ;s i 0.22 %
2j__i u»jlu2__i Q*l32a
••• \ o . ; i ', i 0.08%
?: ! t.i'> '"- i O.H4 %
i /.&*•"/• 1 i . 6 1 %
11 i_ a^idiid _i U»22£6
i 0.35% 1 1 . /> 8 X
% i A.'.O i 1 f-.r.l X,
J , 0 , /. i "' . 0 i 54
/, ! 0 . .-' / 1 " .00 %
iw__^_ ,_j*u.t2_ i a*caii
K ! 0,0/ '; 1 0 .00 7,
•/, ; I.OP "- i 0.82 'X,
0. y3 % i 1.69 %
I 16.jp % 1 6.p5 %
__1 233. J.
1 0.00 X 1
i o.jo x i
1 u.DO X 1
i U.QQ15 J.
i w.«:4 % 1
! 0.02 X |
i 0.00 X I
1 0.01 % \
1 0.01 % 1
1 7.30 X |
i 0.27 X |
: 0.^1 X I
i 1.J3 X 1
i 3.72 X 1
i 3*1322 i
i 0.56 X 1
1 " 2.27 X 1
; o.oi % i
__,. U*aQi4__l
•J.OO X 1
_i_ __a*aoi2--l
0.00 X !
1 0.91 X 1
1 2.44 X !
I 1 '.^7 X I
Fig. 6-3. Sample AQEM Culpability List
-------
134
6.3 MODEL VALIDATION AND CALIBRATION
6.3.1 Validation Procedures
As with, all analytical tools, air quality simulation models are sub-
ject to errors that cause the computed concentrations to differ from observed
values. ~the errors are of two basic types:
Errors in predicting variations in concentrations, and
Errors in predicting absolute levels.
The first type are systematic errors that indicate that the model is not
accurately accounting for variations in emissions or dispersion. The second
type of error indicates that the model is accounting for variations properly
but the computed concentration is in error by some across-the-board quantity.
This is most frequently explained by a background level that is not accounted
for and is the least serious of the two errors.
There are numerous statistical tests that can be used to evaluate
how well the calculated values compare to the observed. These include skill
scores, contingency tables, comparison of time series and spatial variations,
and correlation analyses. The latter, illustrated on Fig. 6-4, is the easiest
to use and is incorporated into the AQDM model. It uses a graph of observed
versus calculated values of concentration either at one receptor location (for
several meteorological or emission conditions) or at a number of locations
for a fixed set of conditions.
A least-square regression line of observed concentration values on
the calculated values is then obtained. To determine the validity of using
the -model to calculate air quality levels, the coefficient of correlation,
a measure of data-scatter about the regression line, is calculated and com-
pared with the maximum theoretical value that could arise due to chance.
The maximum theoretical value for a 5% confidence level (that is, there are
fewer than 5 chances in 100 that a coefficient of correlation as high as this
value would arise due to random sampling variation) is used as the criterion
for acceptable validation in AQDM. A poor correlation coefficient warns the
user that the input data and the model assumptions must be reviewed.
-------
135
(X) NOI1VH1N30N03 Q3Aa3S80
v
oc:
o
o
O
LU
Fig. 6-4. Regression Analysis for Model Validation
-------
136
6.3.2 Sources of Errors
The most frequently occurring reasons for a poor correlation between
computed and observed concentrations are:
Inadequate emission inventory,
Unrepresentative air quality data,
Complex topography and/or meteorology,
Incomplete description of source variations, and
Unaccounted-for atmospheric processes.
By the nature of how emission inventories are developed, there is
substantial room for error. This is especially true when the lower levels of
emission estimation and spatial allocation are used. Some of the more glaring
errors (e.g., the excessive domination of a single source) are correctable by
further investigation of specific sources. The more subtle errors may elude
a quick solution and may require substantial reinvestigation to identify the
problems.
The unrepresentativeness of the observed data against which the
calculations are compared is another source of a poor correlation. Data
collected in an area dominated by a single source is often not suitable for
comparison to the predicted values obtained from some multi-source models.
Likewise, data collected from areas influenced by unusual emission patterns
or terrain features make for poor validation points. These problems are rela-
tively easy to compensate for by reviewing the locations of all the monitors
and screening out unacceptable model validation data. References 11 and 19
give some guidelines on determining the representativeness of various monitor
locations.
Only a few specialized models are adequate for use in complex terrain
(e.g., mountain or valley locations) or for unusual meteorology (e.g., lake
or ocean breeze circulation). Attempts to use other models in these circum-
stances will lead to erroneous results and misleading conclusions. Advice
from EPA Regional Offices and/or headquarters is available to determine the
suitability of various models.
Errors in calculated concentrations occur when the model is not cap-
able of representing variations in the emission patterns of major sources.
-------
137
If, for example, space heating emissions are significant, then the model must
be able to distinguish between summer and winter emission patterns. Problems
of this type can be identified by reviewing the culpability list to determine
if the major sources are adequately treated in the simulation.
Atmospheric processes such as photochemical reactions, scavenging,
and settling have a dramatic effect on concentrations. Of the models listed
on Table 6.1, only the SAI model takes account of photochemistry. The state
of the art of this type of model is still developmental and widespread applic-
ability of these models has yet to be demonstrated. The removal of pollutants
by scavenging (e.g., by washout, ground absorbtion, etc.) has been treated in
some models by incorporating a half-life term in the dispersion equations.
This term results in a steady reduction in concentrations in addition to the
dispersion effects-. Gravitational settling of particulates reduces concentra-
tions in the same fashion as scavenging. The larger particulates (> 30 ym)
tend to settle relatively quickly and hence should not be part of the observed
concentrations. The emission inventory may be adjusted to reflect this situ-
ation by using emission factors representing source particulate size distribu-
tion. Sources for which this type of adjustment is most applicable are the
industrial process sources (especially things like stone quarrying, rock
crushing, etc.) and some fugitive dust sources (e.g., unpaved roads, agricul-
tural tilling, etc.). Appendix F gives some further expansion of this.
6.3.3 Model Calibration
Once the dispersion model estimates have been validated or determined
to be acceptable, the model may be calibrated. The calibration should account
for systematic errors in the estimates. If the coefficient of correlation
(calculated•in AQDM, for example) is greater than the theoretical value for
a 5% confidence level, the regression line is used to calibrate the calculated
values. The slope of the regression line is used to adjust systematic errors
in predicting variations in concentrations, and the intercept adjusts the
prediction of absolute levels by accounting for missing input background
concentration values.
Figure 6-5 gives the correlation analysis for County X using the
baseline inventory and air quality data. The calculation shows the 51 con-
fidence level to correspond to a correlation coefficient of 0.532. The
-------
138
oo
o
1 — 1
5
C£
H-
Z.
LU
Z.
o
o
,_,
to
1 — 1
X
^^
1
>-
«- -«
Q
LU
>
to
t—
LU
I — i
I •>
1 — I
u.
LU
0
Z
o
I— 1
oo
oo
LU
CC
CD
LU
OL
_J
LU
LU
1
•
LU
—f
^—
o
C_}
0
LO
C\J
oo
LO
o
Q
LU
h-
Q_
s:
o
o
CT>
CM
UD
•
O
to
CQ
O
oo
oo
i—i
X
X
Q
C£
O
00
a:
LU
a:
o_
oo
oo
LU
CD
CL.
O
oo
a:
LU
o o
LU OO
i—i a:
I CD
>- O
O
Q-
CM
CT>
CM
r-H
00
oo
LU
O
l—t
g
a.
LU
a:
cc:
o
o
H-1 *—*•
I- a:
o •
o o
o ^v.
(- <
ZD O
o a:
i—i O
Q
LU
<_}
cot— ic\j'd-c\ic\jf-HOcnoo^i-oco
o
LU
oo
CO
o
-------
139
computed, correlation coefficient is 0.629, thus indicating an acceptable
validation. The y-intercept of 31.2 ^g/m3 is within acceptable levels for
background particulate levels, but the small slope indicates the model to
be relatively insensitive to changes in emission patterns. The calibration
equation is the following:
X i , = 0.199 X -, i + •, + 31.2 (6-1)
observed calculated J
This equation is then used to correct all future concentration calculations.
6.3.4 Background Concentrations
Some pollutants occur naturally in the atmosphere independently of
any human activity. Since simulation models rely on a description of all
emission sources, they are likely to underpredict the absolute concentration
by an amount equal to this background level and must be corrected. In another
context, background concentrations can be interpreted as the material trans-
ported into the region of study from external sources, the nature of which
is unknown. In both cases, the background level is added to all computed
concentrations.
Background levels may be determined from air quality monitors upwind
of major source activities. The levels can be expected to lie in the ranges
given in Table 6-4.
Table 6-4. Range of Background Concentrations
Pollutant Concentration (yg/m3)
Particulates 30-40
S02 ^ 0
CO ^ 0
N0? ^ 0
Oxidants ^ Oa
ihe Appendix J model is based on a zero back-
ground concentration of hydrocarbons and
oxidants.
-------
140
6.4 COUNTY X MODELING RESULTS
The Mr equality Display Model (AQDM) was used to perform the air
quality analysis for County X. Figures 6-6 to 6-8 illustrate the participate
air quality confuted for 1975, 1980, and 1985 as plotted from the SYMAP rou-
tine. Table 6.5 is a composite of the computed air quality from the AQDM
output tables. In making these computations it was assumed that no new con-
trol programs were in force, only existing regulations and federal New Source
Performance Standards were assumed to be in effect, and source compliance
data, where available, were used to determine actual emissions.
It is 'evident from Fig. 6-6 that there are several areas in the
county exceeding the primary NAAQS for participates of 75 ug/m3 and that there
are much wider-spread violations of the secondary standard of 60 yg/m3. By
1980, Fig. 6-7, growth and development has caused significant increases in
the area in excess of both the primary and secondary standards. In the period
between 1980 and 1985 (Fig. 6-8], the growth in NAAQS violation areas has
slowed markedly to a point of little increase. This is the result of the
imposition of New Source Performance Standards that arrest the growth in
emissions. This fact is evidenced by the change in county emissions as
illustrated on Table 4-1.
------- |