Unit"-* States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/4-91-014
May 1991
Air
SEPA
PROCEDURES FOR THE
PREPARATION OF
EMISSION INVENTORIES FOR
CARBON MONOXIDE AND
PRECURSORS OF OZONE
VOLUME H: EMISSION INVENTORY
REQUIREMENTS FOR PHOTOCHEMICAL
AIR QUALITY SIMULATION MODELS
-------
-------
EPA-450/4-91-014
PROCEDURES FOR THE
PREPARATION OF
EMISSION INVENTORIES FOR
CAPJBON MONOXIDE AND
/ A "
VOLUME II: EMISSION INVENTORY
REQUIREMENTS FOR PHOTOCHEMICAL
•-»
AIR QUALITY SIMULATION MODELS
By
Lu Ann Gardner
Lyle R. Chinkin
Jeremy G. Heiken
Systems Applications International
San Rafael, CA
EPA Contract No. 68-DO-0124
EPA Project Officer Keith Baugues
Office Of Air Quality Planning And Standards
Office Of Air And Radiation ~~~
U. S. Environmental Protection Agency
Research Triangle Park, NC 27711
Mav 1991
-------
This report has been reviewed by the Office Of Air Quality Planning And Standards, U. S. Environmental
Protection Agency, and has been approved for publication as received from the contractor. Approval does
not signify that the contents necessarily reflect the views and policies of the Agency, neither does mention
of trade names or commercial products constitute endorsement or recommendation for use.
EPA-450/4-91-014
-------
ACKNOWLEDGEMENTS
The authors would like to express appreciation to the various persons involved in the
planning and review of this document; of particular note are the contributions of Keith
Baugues and David Misenheimer of the Office of Air Quality Planning and Standards, EPA,
and Marianne Causley, Jay Haney, and Ralph Morris of Systems Applications International.
Additionally, the authors wish to acknowledge the work of Tom Lahre (EPA OAQPS), Dr.
Lowell Wayne (Pacific Environmental Services), and Keith Rosbury (PEDCo Environmental)
in preparing the original version of this document (EPA-450/4-79-018), published in 1979.
90098 01" jjj
-------
PREFACE
This document is the second volume of a two volume series designed to provide
assistance to air pollution control agencies in preparing and maintaining emission
inventories for carbon monoxide (CO) and precursors of ozone (03). emission inventories
provide the foundation for most air quality control programs. The first volume of this
series describes procedures for preparing inventories of volatile organic compounds (VOC),
oxides of nitrogen (NOX) and CO on a countywide annual or seasonal basis. The 1990
Clean Air Act Amendments require such an inventory for establishing a baseline in O3
nonattainment areas and also require an inventory of CO emissions for CO nonattainment
areas.
This second volume offers technical assistance to those engaged in the planning and
development of detailed inventories of VOC, NOX, and CO for use in photochemical air
quality simulation models. Such inventories must be resolved both spatially and temporally
and must also be speciated into several classes of VOC, NO, and N02. These inventories
are required of the more serious 03 and CO nonattainment areas only.
This volume has been revised from the 1979 version to include current information
pertinent to gridding, speciation, and temporal allocation of emission inventories of CO and
precursors of 03. This edition includes changes and additions as summarized below:
o Inclusion of an additional section containing a brief overview of the Urban
Airshed Model (UAM) and the DAM Emissions Preprocessor System.
o Inclusion of an additional section regarding techniques for estimating
emissions from biogenic sources.
o Revision of the section regarding highway motor vehicles to provide guidance
for developing spatially and temporally resolved exhaust, evaporative,
refueling, and running loss emission estimates from annual and seasonal
county-level total emissions by vehicle type.
o Discussion of currently available computerized data bases useful for the
inventory development process.
o Inclusion of specific guidance for employment of the UAM Emissions
Preprocessor System.
o Discussion of considerations specific to modeling for CO non-attainment
applications.
90098 01'1 V
-------
EXECUTIVE SUMMARY
This document offers technical assistance to those engaged in planning and development
of detailed emission inventories for use in photochemical air quality simulation models.
It is intended to supplement Procedures for the Preparation of Emission Inventories for
Carbon Monoxide and Precursors^l^zong^^/olumeJ. which outlines procedures for
compiling basic annual and seasonal emission inventories at a spatial resolution of county,
township, cr equivalent level, V'C^.TT? ': -:'C":c:,';: rj:c!'j"c? fcr identifying and
incorporating the additional detail reauirac by cnotccnarnicai air Duality simulation models
into an existing inventory of the type described above, with a special emphasis on fulfilling
the input requirements of the Urban Airched Model.
In order for photochemical simulation modals to accurately pratiict temporal and spatial
variations in modeled ozone and CO concentrations, the emission inventories input to
these models must contain considerably more detail than an inventory generated using the
procedures prescribed in Volume I. The primary adaitional requirements of the
photochemical modeling inventory are summarized below.
o Emission estimates of precursor 'pollutants must be provided for each individual
cell of a grid system within the area instead of at a county or regional level;
I
o Typical hour-by-hour emission estimates must be provided instead of annual or
seasonally adjusted emissions;
o Total reactive VOC and NOX emissions estimates must be disaggregated into
several classes of VOC and NO and N02, respectively; spatially and temporally
resolved emission estimates of CO may also be required (EPA requires that CO
emissions be input to the DAM in ozone attainment demonstrations)
o If the model provides for vertical resolution of pollutants, stack and exhaust gas
parameters must be provided for each large point source.
This document presents detailed methodologies for developing the additional resolution
required for photochemical modeling.
Volume II addresses four basic operations used in development of the photochemical
modeling inventory: (1) planning the inventory development effort, (2) collecting any
necessary data, (3) analyzing this data and using it to develop the additional resolution
90098 01"
-------
required of the modeling inventory, and (4) reporting this data in a format which facilitates
its use (data handling). Each of these operations is summarized below.
Inventory Planning and Design. The requirements of photochemical modeling inventories
necessitate additional planning considerations. This document provides a discussion of the
following design issues:
o Selecting an appropriate modeling region and grid system;
o Evaluating existing emission inventories to assess their suitability as a basis for
the photochemical modeling inventory;
o Planning the data cc!i;~t'cr effort, vvnich includes identifying the data
requirements of the photochemical model and prioritizing specific data needs;
o Special planning considerations related to the development of inventory
projections;
o Special considerations for developing CO nonattainment inventories;
o Coordinating the inventory development effort with other agencies; and
o Developing appropriate data handling systems for the emissions-related data;
Note that this document is not intended to replace existing EPA guidance on topics other
than the development of photochemical modeling inventories. Although discussions of
other issues have been included for informational purposes, the reader is directed to other
guidance documents where appropriate.
Data Collection. Usually, point source, highway motor vehicle, and other area source
emissions-related data are acquired separately. It is assumed that a conventional annual cr
seasonal county-level emission inventory, generated in accordance with the methodolcgias
described in Volume I, already exists, and that additional data must be collected to provide
the degree of detail required of the photochemical modeling inventory. Specifically, data
must be collected which allows the emissions modeler to assign emissions to grid cells, to
determine temporal variations in emissions, and to estimate the proportions of VOC and
NOX to be assigned to the chemical species or classes required in the model.
Preparation of the Modeling Inventory. As mentioned above, the photochemical modeling
emission inventory must contain detailed spatial, temporal, and chemical information. In
this document, separate chapters provide detailed methodologies for incorporating this
additional degree of resolution for point sources, mobile sources, and area sources;
specific data handling considerations are also addressed for each of these source types.
goose ov1 viii
-------
Additionally, specific guidance and examples regarding the application of the Urban
Airshed Model Emissions Preprocessor System to facilitate modeling inventory
development is provided throughout the document. Specific topics covered are described
below.
For point sources, data collection techniques and spatial and temporal resolution
methodologies are discussed in detail. Additional information is provided regarding
projection of point source emissions for both individual facilities and at the aggregate level,
including a discussion of control strategy projections. Specific data handling
considerations are also addressed.
For area sources, generai methodologies for spatial resolution are presented, including
determination of emissions at the grid cell level and the use of spatial allocation
surrogates. Detailed examples of the development of spatial allocation surrogates from
land use data and demographic parameters are provided. Additional sections regarding
temporal resolution methodologies, projection techniques, and data handling considerations
are also included.
For mobile sources, selection of appropriate emission source categories is addressed.
Procedures for adjusting existing annual or seasonal emission estimates to be
representative of modeling episode conditions are discussed, and methodologies for spatial
resolution of mobile source emissions using both link- and nonlink-based surrogates are
presented. Temporal resolution methodologies are also addressed.
In addition to the topics listed above, a separate chapter discusses estimation procedures
for biogenic emissions, focusing on EPA's Biogenic Emission Inventory System (BEIS). An
overview of the BEIS is provided along with a discussion of BEIS input requirements and
the use of user-specified land use data in the BEIS. Special considerations for projection
year inventories of biogenic emissions are also discussed.
Finally, a separate chapter is provided which discusses speciation of VOC and NOX
emissions into chemical classes as required by the photochemical model. This chapter
includes an overview of the Carbon Bond IV Mechanism employed by the Urban Airshed
Model as well as specific methodologies for the identification of appropriate split factors
for both base year and projected inventories. Compatibility of split factors with the
emission inventory data and classification scheme are addressed, and special data handling
considerations are outlined.
90098 OVa
-------
TABLE OF CONTENTS
ACKNOWLEDGEMENTS iii
PREFACE - v
EXECUTIVE SUMMARY . vii
LIST OF FIGURES AND TABLES xv
1 INTRODUCTION
1.1 PURPOSE 1-1
1.2 BACKGROUND 1-2
1.3 CONTENTS OF VOLUME II 1-4
2 INVENTORY PLANNING AND DESIGN CONSIDERATIONS
2.1 SELECTION OF THE MODELING REGION AND GRID SYSTEM 2-1
2.2 DATA COLLECTION _ 2-3
2.2.1 Existing Emission Inventories 2-3
2.2.2 Planning the Data Collection Effort 2-4
2.2.3 Inventories of Pollutants Other than VOC, NOX, and CO 2-5
2.2.4 Elevated Point Source Requirements 2-6
2.3 PREPARATION OF THE MODELING INVENTORY 2-6
2.3.1 Spatial Resolution of Emissions 2-6
2.3.2 Temporal Resolution of Emissions 2-7
2.3.3 Chemical Resolution of Emissions 2-8
2.3.4 Special Considerations for CO Nonattainment Inventories .... 2-10
2.4 EMISSION PROJECTIONS 2-10
2.5 DATA HANDLING 2-13
2.6 RESOURCE REQUIREMENTS 2-14
2.7 OVERVIEW OF EMISSION INVENTORY PLANNING PROCEDURES .... 2-15
3 OVERVIEW OF THE URBAN AIRSHED MODEL (UAM) AND THE UAM EMISSION
PREPROCESSOR SYSTEM
3.1 INTRODUCTION 3-1
3.2 CONCEPTUAL OVERVIEW OF THE URBAN AIRSHED MODEL 3-1
3.3 OVERVIEW OF THE UAM EMISSION PREPROCESSOR SYSTEM 3-2
90098 01"
-------
4 DETERMINATION OF THE GRID SYSTEM
4.1 SELECTING AN APPROPRIATE GRID SYSTEM 4-1
4.1.1 Area Covered by the Grid System 4-4
4.1.2 Grid Cell Size 4-5
4.2 MAP GRIDDING PROCEDURES 4-9
4.2.1 UTM Coordinate System 4-9
4.2.2 Orientation of the Grid System 4-9
4.2.3 Problems in Gridding 4-10
5 POINT SOURCE EMISSIONS
5.1 DATA COLLECTION 5-1
5.2 RULE EFFECTIVENESS 5-4
5.3 SPATIAL RESOLUTION 5-5
5.4 TEMPORAL RESOLUTION 5-6
5.5 POINT SOURCE PROJECTIONS 5-14
5.5.1 Individual Facility Projections 5-14
5.5.2 Aggregate Point Source Projections 5-15
5.5.3 Control Strategy Projections 5-21
5.5.4 Point Source Projection Review and Documentation 5-21
5.6 DATA HANDLING CONSIDERATIONS 5-22
6 AREA SOURCES
6.1 GENERAL 6-1
6.2 GENERAL METHODOLOGY FOR SPATIAL RESOLUTION 6-8
6.2.1 Direct Grid Cell Level Determination of Emissions 6-8
6.2.2 Surrogate Indicator Approach 6-9
6.3 GENERAL METHODOLOGY FOR TEMPORAL RESOLUTION 6-29
6.4 AREA SOURCE PROJECTION PROCEDURES 6-34
6.5 DATA HANDLING CONSIDERATIONS 6-41
7 MOBILE SOURCE EMISSIONS
7.1 INTRODUCTION 7-1
7.2 CHARACTERIZATION OF ON-ROAD MOTOR VEHICLE EMISSIONS 7-2
7.2.1 Vehicle Classes 7-2
7.2.2 Roadway Types 7-4
7.2.3 Emission Components 7-4
7.3 MOBILE EMISSION INVENTORY PROCEDURES 7-6
7.4 MOBILE SOURCE EMISSION FACTORS 7-9
7.5 SPATIAL RESOLUTION OF MOBILE SOURCE EMISSIONS 7-14
7.5.1 Link Surrogates 7-15
7.5.2 Non-link Mobile Emission Spatial Surrogates 7-17
90098 O1'a XJJ
-------
7.6 TEMPORAL RESOLUTION OF MOBILE SOURCE EMISSIONS 7-21
8 B1OGENIC EMISSIONS
8.1 INTRODUCTION 8-1
8.2 OVERVIEW OF THE 8E!S 8-1
8.2.1 Leaf Biomass Factors 8-2
8.2.2 Emission Factors 8-2
8.2.3 Environmental Factors , 8-6
8.3 INPUT REQUIREMENTS OF THE BEiS 8-6
8.4 USER-SPECIFIED LAND USE DATA 8-11
8.5 PROJECTION OF BIOGENIC INVENTORIES 8-11
9 SPECIATION OF VOC AND NO, EMISSIONS INTO CHEMICAL CLASSES
9.1 INTRODUCTION 9-1
9.2 THE CARBON BOND-IV MECHANISM 9-1
9.3 CHEMICAL ALLOCATION OF VOC 9-3
9.4 SPECIFICATION OF NOX AS NO AND NO2 '. . . . 9-6
9.5 PROJECTION OF VOC AND NOX SPLIT FACTORS 9-7
9.6 COMPATIBILITY WITH INVENTORY DATA AND SOURCE CATEGORIES . 9-8
9.7 DATA HANDLING CONSIDERATIONS 9-11
GLOSSARY OF IMPORTANT TERMS G-1
APPENDIX A: CODES FOR EMISSION CATEGORIES A-1
APPENDIX B: DEVELOPMENT OF LOCALE-SPECIFIC EMISSION INVENTORIES FOR USE
WITH THE URBAN AIRSHED MODEL B-1
90098 or3 xiii
-------
-------
LIST OF FIGURES AND TABLES
Figures
FIGURE 3-1. Overview of the UAM emissions preprocessor system 3-4
FIGURE 3-2. Input ancLoutput files used by PREPNT 3-6
FIGURE 3-3. Input and output files used by PREGRD 3-8
FIGURE 3-4. Input and output files used by GRDEMS 3-10
FIGURE 3-5. Input and output files used by CENTEMS 3-12
FIGURE 3-S. Input and output files used by POSTEMS 3-14
FIGURE 3-7. Input and output files used by MRGEMS 3-15
FIGURE 4-1. Schematic illustration of the use of the grid in the Urban Airshed
Model 4-2
FIGURE 4-2. The St. Louis Area with locations of the RAPS surface stations and
4 km X 4 km modeling grid superimposed 4-3
FIGURE 4-3. UAM modeling region for the California South Coast Air Basin 4-6
FIGURE 4-4. Comparison of number of grid cells requirecHor a 100 km x 100 km
modeling region for 2 km and 5 km grid spacings 4-8
FIGURE 4-5. Modeling region encompassing the southern San Joaquin Valley and
Sierra Nevada 4-11
FIGURE 6-1. Conceptual representation oflne grid cell identification process 6-15
FIGURE 6-2. County grid cell assignments for the Atlanta, Georgia modeling
region 6-16
FIGURE 6-3. Segment of land use map for Tampa Bay, Florida 6-18
FIGURE 6-4. Location of block group enumeration centroids for the Atlanta,
Georgia modeling region 6-30
FIGURE 6-5. Sample gridded population data for the Atlanta, Georgia modeling
region 6-31
FIGURE 7-1. Depiction of typical link and inventory grid cell 7-16
FIGURE 7-2. Mobile source link surrogates developed for a UAM application of the
Dallas/Fort Worth region 7-19
FIGURE 7-3. Gridded annual average mobile source emissions for a UAM application
of the Dallas/Fort Worth region 7-20
FIGURE 8-1. Nonmethane hydrocarbon fluxes by vegetation type 8-4
FIGURE 8-2. Standardized biogenic NMHC fluxes 8-5
FIGURE 8-3. Emission factor sensitivity to leaf temperature 8-7
FIGURE 8-4. Schematic representation of forest canoopy types 8-8
FIGURE 8-5. Temperature and solar flux variations by canopy layer 8-9
FIGURE 8-6. UAM stand-alone biogenics processor: overview of the Biogenic
Emission Inventory System (BEIS) 8-10
90098 O1'!
XV
-------
Tables
TABLE 3-1. Standard Modeling Emissions Record Format (MERF) employed by the UAM
Emissions Preprocessor System 3-7
TABLE 5-1. Types of emissions data contained in the AIRS Facility Subsystem
and the level of detail at which each is maintained 5-2
TABLE 5-2. Data fields required by the UAM Emissions Preprocessor System 5-3
TABLE 5-3. Day-specific Modeling Emissions Record Format (MERF) used by the UAM
Emissions Preprocessor System 5-7
TABLE 5-4. Default weekday variation codes used in the Emissions Preprocessor
System 5-11
TABLE 5-5. Default diurnal variation codes used in the Emissions Preprocessor
System 5-12
TABLE 5-6. Industrial groupings for BEA economic projections 5-17
TABLE 5-7. Employment by place of work, historical years 1973-1988 and projected
years 1995-2040, for California (excerpt) 5-19
TABLE 5-8. Example temporal factor file for individual point sources (excerpt) .... 5-25
TABLE 6-1. NAPAP area source categories and inventoried ozone precursor
pollutants 6-2
TABLE 6-2. Additional area source category designations for mobile sources for
use with the UAM EPS 6-5
TABLE 6-3. Comparison of NAPAP area source categories and subcategories used in
Procedures for the Preparation of Emissionlnventories for Precursors of
Ozone. Volume 1 6-6
TABLE 6-4. Example spatial allocation factor surrogates for area source
categories 6-10
TABLE 6-5. Additional sources of information for spatial resolution of emissions
for selected area source categories 6-13
TABLE 6-6. Land use classification system used in USGS land use data bases .... 6-14
TABLE 6-7. Land use categories for Tampa Bay area land use map 6-19
TABLE 6-8. Demographic parameters used in San Francisco Bay Area for making
zonal allocations of area sources 6-24
TABLE 6-9. Excerpt from ABAC cross classification table used in San Francisco
Bay Area for subcounty allocation of area source activities 6-25
TABLE 6-10. Illustrative excerpts from zone-to-grid-cell correspondence table
for determining apportioning factors 6-27
TABLE 6-11. Ozone season adjustment factors for selected area source categories . 6-32
TABLE 6-12. Diurnal patterns for gasoline stations in Tampa Bay, in percent of
daily operation 6-35
TABLE 6-13. Example temporal resolution methodologies for selected area source
categories 6-36
TABLE 6-14. Example growth.indicators for projecting emission totals for area
source categories 6-39
30098 or1 xvi
-------
TABLE 6-15. Example file of grid cell apportioning factors for area sources
(excerpt) 6-42
TABLE 7-1. Vehicle class definitions used by the MOBILE models 7-3
TABLE 7-2. Commonly used road type definitions 7-3
TABLE 7-3. NAPAP road type designations versus Federal Highway Administration
road types 7-5
TABLE 7-4. Required input parameters for EPA's MOBILE models 7-10
TABLE 7-5. Optional input parameters for EPA's MOBILE models 7-11
TABLE 7-6. MOBILE 4.0 modeling parameters used in the NAPAP inventory 7-12
TABLE 7-7. State annual average temperatures used in the NAPAP inventory 7-13
TABLE 7-8. Land-use surrogates recommended for spatial allocation of mobile
sources in the absence of link data 7-18
TABLE 8-1. Leaf biomass factors by forest group 8-3
TABLE 8-2. Biogenic emission factors for each biomass emission category 8-3
TABLE 8-3. Carbon Bond IV speciation for BEIS biogenic species 8-3
TABLE 9-1. Definition of the DAM (CB-IV) species 9-2
TABLE 9-2. Example VOC speciation profiles for the Air Emissions Species Manual . . 9-5
TABLE 9-3. Example "split factor" file (excerpt) 9-12
90098 or1 xvii
-------
1 INTRODUCTION
1.1 PURPOSE
This document supplements Procedures for the Preparation of Emission inventories for
Carbon Monoxide and Precursors of Ozone, Volume I. Volume ! outlines procedures for
compiling annual and seasonal emission inventories, which provide the basis for
LS'--iccrr.-in; or ::re emission inventories required for use with photochemical grid models
-'j':h as the Urb;n Airshed Model. Generally, the basic inventory will contain annual or
seasonal estimates of reactive or total VOC, NOX, and CO, at a spatial resolution of
cc'-nty, :c//nship, or equivalent level.
Volume II describes procedures for identifying and incorporating the additional detail
required by photochemical air quality simulation models into an existing inventory of the
type described above. Because photochemical models can simulate the hour-by-hour
photochemistry occurring over numerous, small subcounty areas, such as grid cells, the
input emissions data must be more highly resolved (i.e., chemically speciated and spatially
distributed by grid cell) than required by source/receptor models. Total VOC and NOX
emissions must be apportioned into chemical classes, and information may be required on
other pollutants such as carbon monoxide. Furthermore, evaluation of proposed control
strategies using photochemical air quality simulation models requires that projected "future
year" inventories which incorporate anticipated changes in emissions levels and temporal
and spatial distribution patterns be constructed at the same level of detail as required for
the base year inventories. This document presents methodologies for providing this
additional detail. In each case, the requirements for projected inventories are equivalent to
those for current or "base year" inventories.
The basic emission inventory requirements for photochemical models and the less data-
intensive source/receptor relationships are in many respects quite similar. For both, much
of the same information must be obtained from the same sources. Additionally, the
resulting inventories are used by air pollution control agencies for the same general
purpose: development of control strategies that will assure the achievement and
maintenance of the National Ambient Air Quality Standards for ozone and CO.
Consequently, for many activities such as data collection and emission calculations, the
same considerations and techniques will apply regardless of whether the inventories are
being developed for a photochemical model or a source/receptor relationship. In general,
procedures which are similar to those already described in detail in Volume I will not be
repeated here. Thus, the reader should be familiar with the contents of Volume I in order
to thoroughly understand the procedures described in this document.
900980V1 1-1
-------
Since the £PA-recommended photochemical model for urban applications is the Urban
Airshed Model (UAM), this document emphasizes methods for preparing emission
inventories that fulfill the input requirements of UAM. The data collection methodologies
discussed, however, can usually be applied to generate emission inventories that are
generally suitable for use in any photochemical grid model.
It is assumed at the outset that an annual or seasonal county-level emission inventory
(such as can be generated following the methodologies discussed in Volume I) is available
as a starting point for the photochemical modeling effort. This basic inventory is useful
both in planning the more detailed emission inventory effort and as a source of certain
data. For most urban areas, some sort of basic emission inventory has already been
1.2 BACKGROUND
As described in Volume I, the emission inventory is essential for the development and
implementation of an effective ozone or CO control strategy. It tells the air pollution
control agency what sources are present in an area, how much-of each ozone precursor
pollutant or how much CO is emitted by each source, and what types of processes and
control devices are employed at each plant. Ultimately, the emission inventory is utilized
in conjunction with a source/receptor relationship of some kind for the development of an
ozone control strategy.
Two basic approaches may be used to relate photochemical ozone to precursor emissions.
The first method involves the use of empirical relationships such as EKMA to relate
ambient ozone concentrations with precursor emissions over fairly broad geographical
areas. These models provide answers to questions such as "what level of overall volatile
organic compound emission control is needed to attain the ozone standard in an urban
area?" or "what reduction in maximum ozone concentration will accompany a specified
reduction in ambient levels of volatile organic compounds?"
The second basic approach for relating ozone to precursor emissions involves the use of
photochemical air quality simulation models. These models, which offer a more
theoretically sound approach for control strategy development than the source/receptor
models mentioned above, attempt to simulate the photochemical reactions that occur over
an urban region during each hour of the day or days for which the model is being applied.
Because of their ability to provide detailed spatial and temporal information on
concentrations of both ozone and precursor pollutants and because they can directly relate
emissions to ozone concentrations, photochemical simulation models offer considerable
potential for use in control strategy design and evaluation.
90098 OV 1 -2
-------
In addition to answering the limited questions that the empirical source/receptor
relationships may address, photochemical models enable strategists to make more
sophisticated determinations relating to control program development. For example,
photochemical models enable control agencies to judge whether it is more effective to
control only certain precursor sources within an urban area rather than all sources, or
where (and when) benefits from various control options are most likely to occur within an
urban area. Another application is the development of environmental impact assessments,
since photochemical models allow an agency to evaluate the impact of new precursor
sources (e.g., a major highway) at various receptor locations. Photochemical models are
also usaful in basic scientific research, such as in validation studies of atmospheric
photochemistry and dispersion mechanisms.
Grid models (also called Eulerian models) calculate pcihjt-?~t cnr-Br^rc-rio-s -:•: MV^
locations in space at specified times. The concentrations estimated at each location result
from interaction among emissions, chemical reactions, r^c -rc.i;.:,;;: .5.-,^ dii'-TG.™
introduced by prevailing meteorological conditions. Pollutant concentrations are caic'Jatsa
for each cubicle of a three-dimensional framework in the sntir2 region cf :nt3rasT. A
cubicle might have horizontal dimensions of 1 to 10 kilometers on a side and ba 50 to SCO
meters deep. Some Eulerian models are designed to provide vertical (as weil as horizontal)
resolution of pollutant concentrations by using a vertical "stack" of cubicles; the Urban
Airshed Model, the photochemical model recommended by EPA for ozone control strategy
development and evaluation in urban regions, provides this sort of vertical resolution.
- -»
In order for photochemical simulation models to accurately predict temporal and spatial
variations in modeled ozone and CO concentrations, the emission inventories input to
these models must contain considerably more detail than an inventory generated using the
procedures prescribed in Volume I. Note, however, that the more detailed inventory will
usually be based on and should be consistent with an existing county-level annual or
seasonal inventory prepared using the guidance in Volume I. The primary requirements of
the gridded photochemical modeling inventory are summarized below.
o Emission estimates of precursor pollutants must be provided for eacn individual
cell of a grid system within the area instead of at a county or regional level;
o Typical hour-by-hour emission estimates must be provided instead of annual or
seasonally adjusted emissions;
o Total reactive VOC and NOX emissions estimates must be disaggregated into
several classes of VOC and NO and NO2, respectively; spatially and temporally
resolved emission estimates of CO may also be required (EPA requires that CO
emissions be input to the UAM in ozone attainment demonstrations); and
90098 0V1 1 -3
-------
If the model provides for vertical resolution of pollutants, stack and exhaust gas
parameters must be provided for each large point source.
1.3 CONTENTS OF VOLUME II
This document offers technical assistance to those engaged in planning and development
of detailed emission inventories for use in photochemical air quality simulation models.
Chapter 1 discusses the purpose of Volume il and its relationship to Volume I; it also
includes an introductory description of photochemical air quality simulation grid models
and their emission inventory requirements. Chapt3,' 2 Gc;5Cf ;;-es v^rjc'js :ecrnic3i
considerations that aid in the planning and design of ;'~/3 jjtjiiad emission i.T/anccry
process. Chapter 2 is intended to provide an overall perspective of the detailed inventory
requirements for those who will actually be utilizing en 3 r'?n~5hc9f of the cccu:r.3r',
Chapter 3 provides a brief overview of the Urban Airshea Model (DAM) and the UAiV!
Emissions Preprocessor System, and Chapter 4 addresses selection of an appro-pr^sta
modeling region and grid system. Finally, Chapters 5 tnrough 9 provide detailed "hew to"
procedures for supplying the additional inventory detail required by the photochemical grid
model.
For the convenience of the reader, the following typographical conventions will be used
throughout this document: "*
»> Text containing specific examples or involved calculations will be indented and '
denoted by an arrow like the one to the left.
Additionally, information pertaining specifically to the UAIV! Emissions
Preprocessor System w$ be enclosed m a gray box Hks this.
9009801" 1-4
-------
2 INVENTORY PLANNING AND DESIGN CONSIDERATIONS
In general, compilation of a detailed emission inventory suitable for photochemical grid
modeling (hereafter referred to as a "modeling inventory") involves the same four basic
operations required to compile a less detailed inventory suitable for use with models such
as EKMA. These steps are (1) planning the inventory development effort, (2) collecting
any necessary data, (3) analyzing the data and using it to develop the additional resolution
required of the modeling inventory, and (4) reporting the data in a format which facilitates
its use (data handling).
Many of the planning and design considerations discussed in Procedures for the
Preparation of Emission Inventories for Carbon Monoxide and Precursors of Ozone. Volume
i (EPA, May 1991) also apply to development of a modeling inventory. (Throughout the
remainder of this text, the above document will be referred to as Volume I). The more
rigorous requirements of photochemical modeling inventories, however, necessitate
additional planning considerations. This chapter identifies these additional requirements
and discusses the additional responsibilities imposed on the agency developing the
modeling inventory.
2.1 SELECTION OF THE MODELING REGION AND GRID SYSTEM
Before any data collection effort begins, the geographical region to be modeled must be
selected based on consideration of available meteorological and air quality data, location of
current and expected major emissions sources, and control strategy evaluation objectives.
The guidance set forth in the Guideline for Regulatory Application of the Urban Airshed
Model (EPA-450/4-91-013, OAQPS, June 1991) should be followed when selecting the
modeling region; the following discussion is included for informational purposes only.
The two main elements of the grid system used to identify the modeling region are (1) the
grid boundary, which outlines the area to be modeled (the "modeling region"), and (2) the
individual grid cells which will be used by the model to subdivide this region. In most
cases, the grid boundary will be rectangular and the grid cells will be equally sized squares.
Generally, the modeling region should be fairly large for the following reasons:
o to include all major emission sources which may affect ozone formation in the
urban area;
90098 02" 2- 1
-------
o to encompass as many ozone and precursor pollutant monitoring stations as
possible (which facilitates model validation);
o to encompass areas of current limited land use activity that are expected to
develop significantly as a result of projected growth; and
o to encompass the effects of meteorology in the modeling region.
Note that CO modeling regions do not need to be as large as ozone modeling regions; CO
modeling regions usually only contain the nonattainment area.
In some cases, the selection of the modeling region will not be finalized at the time when
data collection for the modeling inventory must begin (perhaps due to manpower or
scheduling constraints). In this instance, the emission inventory should be developed for
as large a region as possible, to ensure that any modeling region finally selected will be
encompassed by the region for which emissions data has been collected, preventing any
additional data collection effort.
After the grid boundary has been selected, the size and number of grid cells used to
subdivide the modeling region must be chosen. Generally, the grid spacing (the length of a
square grid cell along one side) should be small to optimize the spatial resolution of
emissions. If the grid spacing is too large, the model may lose precision in estimating
ozone and precursor pollutant levels. Too small o"? a grid spacing, however, will result in
excessive manpower and computer resource requirements, because data must be collected
and compiled for every grid cell in the modeling region.
In most urban ozone modeling applications, a compromise between covering as large a
region as possible with the smallest feasible grid cells usually results in the selection of a
grid boundary between 50 to 100 kilometers on a side, with a grid spacing of 2 to 5
kilometers (larger modeling regions and grid spacing may be required for regional
applications). For CO modeling applications, a finer grid system covering less area than
would be necessary for an ozone application is generally appropriate.
Since a number of factors not related to the emission inventory (e.g., meteorology) must
be considered when defining the grid system, photochemical modeling specialists, local
planning organizations, and meteorologists should be consulted before data collection
begins to ensure that the selected grid system meets the general objectives of the
modefing effort. Agencies involved in modeling adjacent areas (multi-State nonattainment
areas, etc.) should coordinate selection of grid spacing, orientation, and origin.
Chapter 4 discusses the considerations and selection procedures mentioned above in
greater detail.
90098 02" 2-2
-------
2.2 DATA COLLECTION
Once the modeling region and grid system have been defined, collection of appropriate
emission data can begin. Usually, point source, highway motor vehicle, and other area
source emissions-related data are acquired separately. (Maintaining this distinction
throughout the modeling inventory development process will generally prove useful, since
it facilitates quality assurance and the construction of various inventories for evaluating
control strategies and/or analyzing the sensitivity of model-predicted air quality parameters
to emissions.) !t is assumed that a conventional annual or seasonal county-lave! emission
inventory, as described in Volume I, already exists, and that additional data must.be
collected as a basis for assigning emissions to grid cells, for determining temporal
ciscriQutions, and for estimating the proportions of VQC and NOX to be assigned to the
chemical spscies or classes required in the model.
2.2.1 Existing Emission Inventories
Because many of the data requirements of the detailed emission inventory are quite •
resource-intensive, existing data and systems should be "used whenever possible. Existing
inventories, data handling systems, and planning models maintained by local agencies
should be reviewed to determine what framework (if any) has already been established for
handling emissions and related data, and what portions of this framework may possibly be
utilized to develop the modeling inventory".**
First, the existing emission inventory should be reviewed to determine what source and
emissions data are already available. Most urban locations have VOC and NOX inventories
at the level of detail of EPA's Aerometric Information Retrieval System (AIRS), which has
replaced the National Emission Data System (NEDS). If an accurate, comprehensive, and
current inventory exists, then it can provide much of the basic data needed for the
modeling inventory. If the existing inventory does not meet these criteria, it should be
updated prior to or during the initial stages of modeling inventory compilation.
Additionally, the existing inventory should be examined to determine if it contains average
annual emissions or has been adjusted to reflect typical emissions levels for the ozone
season. If only annual emissions estimates are available, the inventory must be seasonally
adjusted. Volume I discusses techniques for seasonal adjustment in detail; these
techniques are also addressed herein in Chapters 5, 6, and 7 for point, area, and mobile
sources, respectively.
The modeling region normally encompasses a number of counties. Because most counties
do not have rectangular boundaries, portions of a county may extend beyond the
boundaries of the modeling region. In this case, the existing county-level emission
90098 02" 2-3
-------
inventory must be examined tc determine the emissions occurring within the modeling
region.
The point source records included in most local emission inventories usually contain most
of the data (including stack parameters and operating information) needed to construct the
modeling inventory. The only necessary point source data not generally provided in such
inventories are those dealing with speciation of VOC and NO, emissions into chemical
classes and detailed hourly emissions information. Likewise, most of the county-level area
sourcs activity !evs!s in existing local inventories can be used as the basis for the modeling
inventory, although spatial and temporal allocation factors will be needed to. apportion
county-level annual area source emissions to grid cells for each hour of the modeling
The only important source category for which the existing inventory does not ordinarily
represant a good starting point is highway motor vehicles. !n annual and seasonal county-
level inventories, highway motor vehicle emissions are often based on either gasoline sales
or total vehicle miles traveled (VMT) for the county. These gross techniques do not
provide the best available spatial resolution for photochemical modeling, so link-by-!ink
traffic data for the area should be obtained from a transportation planning agency if
possible.
2.2.2 Planning the Data Collection Effort
As discussed above, the existing inventories should contain much of the required
information on total emissions for the area of interest. (Note, however, that many existing
inventories have been compiled with respect to ozone precursor emissions; if
photochemical modeling is being performed in support of a CO attainment demonstration,
the existing inventory may need to be re-examined to ensure that all major sources of CO
are included.) The documentation provided in support of the existing point source
inventory (e.g., AIRS data) usually contains additional information on stack parameters for
each source. The data collection effort in support of modeling inventory development
should be directed toward providing the additional information required to (1) define the
spatial distribution and temporal variations in emissions from each source or source
category, and (2) assign VOC and NOX emissions to appropriate chemical classes.
Ideally, emissions data (including VOC speciation information) would be available for each
source for each hour of any day selected for modeling. In practice, however, this degree
of detail is neither necessary nor practical for all sources because of the inordinate amount
of effort required to secure such data and because, for many sources, inclusion of this
data would have little effect on the ozone levels predicted by the photochemical model.
90098 02" 2-4
-------
When planning the data collection effort, the agency responsible for compiling the
modeling inventory must decide which apportioning information should be collected and
which is unimportant (in other words, set priorities). For a typical urban region, emissions
from highway motor vehicles, gasoline marketing and storage, solvent consumption, and
power plants may account for much of the total VOC and NOX in the inventory. The
remaining VOC and NOX emissions probably arise from a number of smaller sources, any of
which individually has only a minor influence on predicted ozone levels even if small errors
are made in allocating emissions from these minor sources to the proper grid cells or in
estimating their seasonal and diurnal variations. By examining the existing inventory, the
most important emission sources in the region can be identified; to minimize resource
requirements, the data collection effort should focus on supplementing ine existing jpaasi
or temporal resolution data for these sources. Many sources err.it -^c~. .~ '<" ,' -, ' ... „
VOC and NOX that little, if any, additional effort is warranted in gathering -.y" ;•;"••;,!" ar^.
spatially resolved data regarding them.
Finally, the agency preparing the modeling inventory must work closely v/r:,-: c^e- iccai
metropolitan planning organization (MPO) or other planning agencies in en-? a^s ro
determine what transportation and land use planning models are currently being srr.pioved
and what data from thes.e models can be directly useful to the inventory compilation
effort. In most urban areas where comprehensive transportation and land use planning are
performed, much of the information needed to determine highway motor vehicle
emissions, to make projections for future years, and to apportion emissions to the grid cell
level will already be available. ""*
2.2.3 Inventories of Pollutants Other than VOC, NOX, and CO
The ozone modeling inventory development effort should be directed primarily to\^'2rd
obtaining accurate emission data for VOC, NOX, and CO since these are the most
important precursor pollutants in the photochemical production of ozone in urban
atmospheres. For CO modeling, the focus is on CO inventories. Some photochemical
models, such as the Urban Airshed Model (DAM), also have the capacity to accept
emissions data and generate air quality estimates for other pollutants. This capability is
provided primarily to allow these models to predict the effects of control strategies on
ambient levels of these pollutants. Although sources of these pollutants should be
included in the photochemical modeling inventory, developing spatially and temporally
resolved emission estimates and projections for these sources is not warranted unless NOX
VOC, and/or CO are also emitted in significant quantities.
90098 02'2 2-5
-------
2.2.4 Elevated Point Source Requirements
Some photochemical models, including the (JAM, assign emissions from point sources to
elevated grid cells if they are characterized by an effective stack height {i.e., the sum of
the physical height of the stack and any plume rise) which is greater than the height of the
grid cell. For these models, the emission inventory must include stack information (e.g.,
physical stack height and diameter, stack gas velocity, and temperature) for the major
point sources in the area. The agency must therefore know whether the model it plans to
employ requires data characterizing individual stacks. If required, the stack height used
should be either the physical stack height or the Good Engineering Practice (GEP) stack
height if the physical stack height exceeds the GEP stack height. As manticnsd
previously, the existing inventory will usually contain this informs:; :r
examined and utilized to the greatest extent possible in order to mini,
to the modeling inventory effort.
2.3 PREPARATION OF THE MODELING INVENTORY
As mentioned in Section 2.2, three separate types of information not usually provided in
the existing inventory will have to be added for modeling purposes. All three involve
additional resolution of emissions, namely spatial, temporal, and chemical. The process of
providing this additional resolution can be described as emissions modeling, since spatial,
temporal, and chemical variations in emissions must be identified and applied to the
existing inventory to fulfill the requirements of the modeling inventory. Accordingly, and
also for reasons of brevity, the person (or persons) responsible for preparation of the
modeling inventory will be referred to as the "emissions modeler" throughout this
document.
2.3.1 Spatial Resolution of Emissions
In order for photochemical models to provide spatially resolved predictions of czone and
various other pollutants at the grid cell level, they must be supplied with emission data
that have the same degree of spatial resolution; in other words, emissions must be
resolved by grid cell. The amount of effort required to implement this resolution will vary
depending on the type of source. Point source locations are typically reported to within a
fraction of a kilometer in the existing inventory; hence, assigning emissions from these .
sources to the appropriate grid cell is simple. This assignment can be performed manually
(by overlaying an outline of the grid system onto a map showing point source locations) or
with the assistance of computerized routines.
By contrast, spatial resolution of area source emissions requires substantially more effort.
Two basic methods can be used to apportion area source emissions to grid cells. The
30098 02" 2-6
-------
most accurate (and resource-intensive) approach is to obtain area source activity ievei
information directly for each grid cell. The alternative approach, more commonly
employed, is to apportion the county-level emissions from the existing annual inventory to
grid cells using representative apportioning factors for each source type.
This latter approach requires the emissions modeler to determine apportioning factors
based on the distribution of some spatial surrogate indicator of emission levels or activity
(e.g., population, census tract data, or type of land use) for each grid cell and apply these
factors to the county-level emissions to yield estimates of emissions from that source
category by grid cell. The major assumption underlying this method is that emissions from
each area source behave spatially in the same manner as the spatial surrogate indicator. In
developing scaxiai aooortioning factors, the emissions modeler should emphasize the
determination of accurate factors for the more significant sources. In most large urban
areas, local planning agencies can provide the emissions modeler with detailed land use,
populat'on or in seme cases, employment statistics at the subcounty level; this data can
be ussd to spatially apoorticn most of the area source emissions in the modeling inventory.
Highway motor vehicle emissions, which usually comprise a large fraction of total VOC
and NOX emissions, should be considered separately from other area sources in the
modeling inventory. Instead of using county VMT or gasoline sales to estimate highway
vehicle emissions (as annual and seasonal inventories sometimes do), urban transportation
planning models should be employed to generate VMT on an individual link basis whenever
possible. The emissions for each link couici then be assigned to the appropriate grid cells.
Planning, land use, and transportation models are already in use in many urban areas, and
can provide the emissions modeler with much of the data necessary to allocate area
source emissions and develop emission estimates by link for highway motor vehicles.
Such models are also generally capable of developing forecasts for future years which can
be utilized in the development of projection inventories. Local agencies (particularly
MPO's) should always be contacted during the inventory planning process to determine
what planning models are being utilized and how the data available from these models can
be used in the emission inventory effort. Obviously, trying to independently develop ail
the necessary information that should be available from the MPO requires much redundant
effort on the part of the emissions modeler. Additionally, any subsequent photochemical
modeling results might likely be challenged because of alleged nonconformity with other
projections available to the public.
2.3.2 Temporal Resolution of Emissions
In order to predict hourly concentrations of ozone and other pollutants, photochemical
simulation models require hour-by-hour estimates of emissions at the grid cell level. The
emissions modeler can employ one (or more) of several approaches to provide the temporal
90098 02" 2-7
-------
detail needed in the modeling inventory. The most accurate and exacting approach is to
determine the emissions (or activity levels) for specific sources for each hour of a typical
day in the time period being modeled. This approach, while sometimes applicable to point
sources, often proves impractical.
As an alternative, the emissions modeler can develop typical hourly patterns of activity
levels for each source category, and then apply these to the annual or seasonally adjusted
emissions to estimate hourly emissions. This approach is commonly employed for area
sources, including highway motor vehicles, and is usually used for all but the largest point
sources.
Usually, the photochemical air quality model is applied for an episode in the season of the
year in which meteorological conditions are most conducive to ozone formation; for most
locations, this means the summer months (i.e., June through August). By contrast, CO
non-attainmant episodes often occur in the winter months. Consequently, emissions must
be adjusted to reflect typical levels for the particular non-attainment season (ozone or CO).
Similarly, emissions are usually adjusted to represent the day of the week on which
polluting activities are at a maximum, normally a weekday. In some cases (such as
validation studies), simulating weekend conditions when automotive and industrial
emission levels are reduced may be useful. For this purpose, additional temporal pattern
information pertaining to weekend days must be used to construct a weekend modeling
inventory. Generally, however, the emissions modeler should not compile a weekend
inventory unless (1) significant reductions or changes in emission patterns are expected;
(2) the same inventorying procedures can be used as for weekdays, so that any resulting
changes in predicted ambient ozone levels can be attributed to actual changes in precursor
pollutant levels and patterns rather than simply to changes in methodologies; or (3) a
significant number of ozone (or CO) exceedances occur on weekends. In many urban
regions, the second will not be possible for highway vehicles, since transportation models
are based on information (e.g., travel pattern surveys) applicable only for weekday
situations. If the model is to be used to estimate ambient concentrations of various
pollutants for time periods other than the ozone season, additional seasonal information
may be required.
2.3.3 Chemical Resolution of Emissions
Because photochemical models are intended to simulate actual photochemistry, they utilize
different chemistry for various types of VOCs and require specific information as to the
proportions of these various types present in the inventory. For this reason, VOC emission
totals must be disaggregated into subtotals for various chemical classes. NOX emissions
may also have to be distributed as NO and N02. (Some models do not require a NOX
breakdown because they assume all NOX emissions to be NO.) Literally hundreds of
90098 02" 2-8
-------
individual chemical compounds typically compose the total VOC emissions in an urban
area. No photochemical model considers each organic compound individually; instead,
VOC emissions are distributed into chemical classes which behave similarly in
photochemical reactions. The DAM employs a carbon bond classification scheme based
on the presence of certain types of carbon bonds in each VOC molecule. Other models
employ different classification schemes, which utilize different numbers and types of
chemical classes.
The standard procedure for allocating VOC to chemical classes is to assume that the VOC
emissions from each type of source contain a fixed percentage of each class of compound.
This is the easiest of severs! ,7;srhcds that can be employed for allocating VOC emissions
because the ssm? VCC "j:r":'rr;r^ is assumed to apply tc each facility or process within a
given source category.
In some instances, scurce-sncific VOC species data may be available for certain individual
facilities (perhaps through source tests or material composition considerations), and the
emissions modeler may prsfar to use these in the modeling inventory instead of an
assumed VOC species distribution. Generally, however, most industries cannot provide
reliable VOC or NOX species data or accurately apportion their emissions into appropriate
classes, in which case generalized VOC and NOX distributions must be assumed for various
source categories. Chapter 7 addresses development of representative VOC and NOX
"split factors" from the literature.
" ft-
A potential problem when using generalized split factors to apportion VOC and NOX into
classes is that the source classification scheme (i.e., source category breakdown)
employed in the inventory will probably not be directly compatible with the available split
factor classification scheme in all cases. For example, the inventory may not distinguish
between automotive exhaust and evaporative emissions, whereas different VOC split
factors are typically available for each of these automotive emission components (and may
be significantly different in some classification schemes.) As another example, many VOC
classification schemes do not distinguish between different types of fuel combustion in
external combustion boilers, yet most inventories do.
Hence, as part of the planning process, the emissions modeler should examine the split
factors available for use and compare the classification thereof with the source
classification scheme of the basic inventory. If serious inconsistencies exist for the more
important VOC and NOX source categories, the emissions modeler may need to consider
modifying either or both of the classification schemes to minimize any resultant error.
Alteration of the inventory source classification scheme may require significant resources
and should be carefully evaluated prior to instituting such a change.
90098 02" 2-9
-------
2.3.4 Special Considerations for CO Nonattainment inventories
The 1990 Clean Air Act Amendments (CAAA) require the development, for each CO
nonattainrnent area, of a CO emission inventory which addresses actual CO emissions
during the peak CO season for the area. For many areas, the peak CO season will occur
during ths winter, necessitating the development of a winter inventory for CO modeling
purposes. Most of the methods described in this document apply equally well to
construction of modeling inventories for both ozorce precursors and CO; a few special
considerations for CO inventories, however, should be mentioned.
Ozone modeling episcdc-s generally encompass severa! consecutive days; by contrast, CO
simulations are usually aoo'iad for much shorter time periods (e.g., 8 to 1 5 hours).
Consequently, accurate nouny allocation of emissions becomes more critical for CO
simulations, and ail temporal data for the inventory should be carefully evaluated for both
accuracy and completeness. With regard to point sources, accurate and complete stack
paramacer cata (see Seccion 2.2.4} are also required to ensure that CO emissions are
appropriately allocated to the vertical layers of the modeling grid. As mentioned in Section
2.2.2, the existing inventory may need to be examined to ensure that all major CO sources
are included. The grid resolution will also differ for CO and ozone simulations, with CO
analyses using a finer grid system. Finally, if the existing inventory has been compiled for
the ozone season, additional adjustments may be required to correct the emissions
estimates to levels appropriate for the peak CO season.
- •*>•
For a more complete discussion of the CAAA requirements for CO and ozone
nonattainrnent inventories, consult Emission Inventory Requirements for CO SIP
Nonattainment Areas and Emission Inventory Requirements for Ozone SIP Nonattainment
Areas (EPA, OAQPS, March 1991).
2.4 EMISSION PROJECTIONS
Regardless of the type of model employed, projection inventories are necessary to
determine whether a given area will achieve or exceed the CO or ozone standard in future
years. There are basically two types of projections: baseline projections and control
strategy projections. Baseline projections are estimates of future year emissions that
account for both expected growth in an area and air pollution control regulations that are
in effect at the time the projections are made. Note that certain provisions in existing
control regulations may take effect only at some future date, and baseline projections
should include the effects of these expected changes. By contrast, control strategy
projections also include the expected impact of revised or additional control regulations.
The concept of demonstrating "reasonable further progress" (RFP) was first introduced as
part of the 1977 Clean Air Act Amendments. The 1990 CAAA have continued this
90098 02* 2-10
-------
requirement, whereby States, in order to monitor incremental air quality progress, must
prepare annual RFP reports documenting estimated regional progress in meeting the
policies, programs, and regulations of the adopted State Implementation Plan. Although
annual BFPs are required, projected modeling inventories must be compiled only for Year 6
after enactment (i.e., 1996) and at three year intervals thereafter until attainment is
demonstrated (the number of required modeling inventories varies with nonattainrnent
status). The projected modeling inventories must be compiled using allowable emission
rates as dictated by regulatory limits; these rates should be consistent with those used in
the RFP tracking inventory for the year in question. Accordingly, the emissions modeler
should confer with those parsons responsible for RFP tracking to ensure consistency of
projections and objection methodologies. EPA will be providing updated guidance on
?rniss;en in"'?!--:.".' ^'o'-jcticn *3cnpic,ues (to be released in July, 1991) and on RFP
preparation (to 'oe leased ,n Ncvambsr, 1391).
In many rasoects. the baseline orojection modeling inventory will be the same as the
baseline projection inventory of annual or seasonal county-level emissions compiled for
ozone nonac'tainmc-nt areas for use in models such as EKMA or to meet reasonable further
progress requirements of the 1990 Clean Air Act Amendments. Both inventories will
emphasize the same source categories and pollutants and utilize the same emission
factors, activity levels, and control device data. Consequently, just as for the base year,
the annual or seasonal county-level projection inventory may serve as a good starting point
for developing the projection inventory used for modeling. However, as discussed in the
preceding sections, incorporation of the spatial, temporal, and chemical resolution required
by the model requires additional considerations on the part of the emissions modeler.
These considerations, as they specifically relate to projection inventories, are listed below:
o The emissions modeler should consider anticipated changes in the spatial
distribution of emissions from the base to projection years. Changes in point
source emissions due to growth or control measures should be associated
with specific locations within the modeling area, i.e., at either new or existing
facilities. In this regard, pinpointing the location of any large point sources of
VOC or NOX that will be coming on line is especially critical. Apportioning
factors for spatial allocation of area sources should reflect future land use
patterns, employment, population, etc; while highway vehicle emission
inventories should reflect changes in highway networks and driving patterns.
o Changes in temporal emission patterns should also be considered. Any
anticipated changes in hourly, daily, or seasonal operating patterns between
the base year and projection years should be reflected in the projection
inventories.
90098 02" 2-11
-------
o To tne extent that VOC and NOX split factors are expected to change between
the base and projection year, such changes should be incorporated into the
projection inventories.
Generally, of these three considerations, information will be most readily available
concerning changes in spatial distribution. This is because local authorities and agencies
should know projected locations of large, new point sources (at least in the near-term),
and because highway vehicle and area source emission patterns will directly reflect
changes in ths land usa, employment, and transportation data supplied by local planning
agencies. With the exceotion of on-road motor vehicles, most of the.temporal patterns
and VOC and NOX chemical class split factors will typically not be changed from the base
v-:-•' !r-.'-£-prory ^-:r-:r 09C"'JC3 PC changes are expected or because no data will be available
"o fc.r3cast 3i.ch crnngss. 7h333 considerations are discussed in more detail in succeeding
chapters.
Of course, in some instances, baseline projection inventories of annual or seasonal county-
wide emissions from the particular year of interest will not be available for use as starting
points for the detailed, photochemical modeling inventories. Or, in other cases, the
projection inventories that do exist may not reflect all of the growth and control scenarios
that the agency may wish to evaluate with the photochemical model. In these situations,
the emissions modeler will have to devote resources to the development of projection year
inventories. Specific recommendations for making baseline projections are discussed in
subsequent chapters. However, the following general considerations should be kept in
mind from the outset of the inventory planning stages.
o To a large extent, projection inventories will be based on forecasts of
industrial growth, population, land use, and transportation. The emissions
modeler should not attempt to make these forecasts, but should rather rely on
the local MPO or other planning agencies to supply them. This course of
action has several advantages. First, duplicating the forecasts made by other
planning agencies would be extremely costly. Second, emission projections
should be based on the same forecasts utilized by other governmental
planning agencies. This consistency is necessary to foster the credibility of
any proposed control programs that are based on these emission projections.
o Control strategy projection inventories should be designed to reflect the
control strategies being considered. This consideration may influence the type
of data collected as well as the structure of the inventory itself. For example,
if one of the modeling objectives is to evaluate the effect of applying Stage I
controls only to service stations above a particular size cutoff, the emissions
modeler may wish to treat these particular stations as point sources rather
than lumping them in with a general service station area source category.
90098 02"
-------
All projection inventories should be based on the same methodologies and
computation principles as the base year inventory. For example, if a traffic
model is used for estimating travel demand for the base year, the same traffic
model should be applied to estimate travel demand for projection years. Using
the same methodology assures consistency between base year and projection
year emission estimates and prevents the possibility of spurious inventory
differences resulting solely from methodological changes.
Projection inventories will always be subject to criticism because of their
somewhat speculative nature. The technical credibility of emissions
projections will bs a function of their reasonableness, the amount of research
2nd dcc'_,)r.3;i:jt:cr: of assumptions, and the procaduros and methodologies
used to make tne projections. Some degree of uncertainty will always
accompany emission projections; this fact should be acknowledged openly.
When d?ve!oci-g projection inventories, the emissions modeler should focus
on minimi-zing instead of eliminating uncertainty. Internal and external review
of the projection inventories will improve their technical quality and enhance
their credibility.
2.5 DATA HANDLING
The large amount of data that must be gatfiered and stored in the inventory development
effort and the complexity of developing spatially and temporally resolved emission
projections generally requires a computerized data handling system. Ideally, as many data
handling functions as possible should be computerized to allow devotion of more
manpower resources to collection and analysis of the inventory data. A flow chart of the
entire data handling operation, from the initial gathering of inventory data to the final
development of a data file that is in the input format of the photochemical model being
used, will prove useful to the emissions modeler in the operation of a computerized data
handling system.
Many of the data handling functions are similar to those required for the existing inventory
(e.g., data storage). Several additional data handling requirements arise during the
development of the modeling inventory because of the additional spatial, temporal, and
chemical resolution of this data.
As mentioned in Section 2.3.1, area source emissions are often spatially allocated to grid
cells using apportioning factors. This step is usually computerized because of the large
numbers of factors and calculations involved. Likewise, point source and highway motor
vehicle emissions are usually assigned to the appropriate grid cells using computerized
routines. For area and mobile sources, temporal variations will often be implemented by
developing typical hourly activity patterns for each major source category. For many point
90098 OZJ 2-13
-------
sources, hourly activity levels can be reasonably inferred from the operating information
supplied in the existing inventory. In either case, relatively simple algorithms can be
developed and computerized to provide the necessary temporal resolution for the detailed
inventory. VOC emissions ace usually disaggregated into species classes through the use
of an appropriate species distribution for each source category. NOX emissions are either
assumed to be a!! NO or are split into NO and N02. The allocation of VOC and NOX
emissions into classes involves straightforward calculations that should likewise be
computerized.
One major data handling function involved in compiling the modeling inventory is the
development of emission projections. Growth is typically accounted for by adjusting base
year emissions in 3ccc'd~rc2 either with projected changes in the emissions themselves or
with changes in appropriate surrogate indicators of growth (e.g., earnings, population, land
use, and employment). Control regulation and strategies can be reflected in an inventory
by adjusting activitv levels, control device efficiencies, or emission factors, as appropriate.
The data handling system used should automate the development of growth and control-
strategy projections as much as possible, thus minimizing the amount of manual effort
needed each time a different scenario is modeled.
The final product of the modeling inventory development process is a file containing hourly
gridded emissions estimates for each chemical class employed by the model. Because
each mode! requires a special computerized format for the inventory data, utility programs
will be necessary to convert the emission inventory file to a model-compatible format.
Of course, certain data handling functions, such as determining area source apportioning
or growth factors, may not be supported by existing data handling systems. Thus, during
the planning stages, the emissions modeler must carefully review the data handling flow
chart to determine which activities can be done by computer and which functions must be
performed manually. Specific data handling requirements and EPA systems that support
these requirements are addressed in subsequent chapters.
2.6 RESOURCE REQUIREMENTS
Staffing and expertise requirements should be considered as part of the planning process.
Depending on the status of the existing inventory and the amount of additional detail
required, compilation of a modeling inventory can require from 500 to 5,000 staff hours.
The low estimate is for a case in which a gridded or very complete county-level inventory
already exists, and the region is dominated by only a few major types of sources. The
high estimate is for a case in which little or obsolete inventory information is available and
the region has many significant sources, requiring detailed analysis of all significant
sources and considerable individual contact with managers of specific sources.
9009802" 2-14
-------
In addition to those staff usually responsible for compilation and maintenance of emission
inventories, the agency should enlist the services of (1) a photochemical modeling
specialist familiar with the operation and the VOC species classes of the particular
photochemical model to be used, (2) a computer programmer or systems analyst to plan
the storage and manipulation of the large amounts of emission data needed, and (3) an
urban or regional planner to analyze transportation and land use data from local planning
agencies and to assist the emissions modeler in making emission projections.
2.7 OVERVIEW OF EMISSION INVENTORY PLANNING PROCEDURES
The remaining sections of this document present detailed "how-to" procedures for
producing a modeling inventory. Specific topics addressed include
o defining the grid system (Chapter 4);
o collecting and compiling data from point, highway motor vehicle, other area,
and biogenic sources, including information regarding spatial and temporal
resolution (Chapters 5, 6, 7, and 8, respectively);
o allocating VOC and NOX emission data into chemical classes (Chapter 9); and
o data handling requirements fofeach of these procedures (discussed
individually in each chapter).
Prior to initiating the data collection phase of the emission inventory effort, the agency
should be able to answer "yes" to the following questions:
o Has the size and orientation of the grid been defined? Has the grid cell size
been defined? Have the decisions been coordinated with agencies involved in
modeling exercises for adjacent areas?
o If a detailed emission inventory has been developed, have hourly emissions
been summed to estimate daily county emission totals (e.g., link based data)?
o Have the time periods (i.e., month, day, year) been specified for which the
emission inventory must apply?
o Has the existing emission inventory been reviewed to determine what data
can be utilized in the photochemical modeling inventory, and what additional
temporal and spatial data must be gathered?
90098 OZ'
2-15
-------
o Does the agency know what VOC and NOX classifications are needed by the
particular photochemical model that will be run?
o Are the source categories required in the photochemical modeling inventory
(in order to reflect specific control strategies and to be consistent with
available VOC split factor information) compatible with those in the existing
annual or seasonal inventory?
o Have the appropriate State and local transportation and planning agencies
been contacted to determine what baseline and projection data on traffic,
employment, population, etc., are available for use in the detailed inventory?
o Will the detailed emission inventory be utilized for modeling pollutants other
than ozone? For other seasons than the summer? Are additional or better-
resolved source and emissions data needed for these other uses?
o Are the existing inventory files and data handling systems capable of
generating and storing the additional temporal, spatial, and VOC and NOX
classification data required by the photochemical model? Can the model-
compatible emissions file be readily generated from the resulting point, line,
and area source and emission files? Are sufficient stack data available to
distinguish elevated point sources from ground level point sources?
- -S*
o Are sufficient resources available to complete both the base year and
projection inventories, considering both growth and control strategy options?
Addressing these issues at the outset of the modeling inventory development process will
limit the number of difficulties arising from poor planning rather than data limitations.
90098 02" 2-16
-------
3 OVERVIEW OF THE URBAN AIRSHED MODEL (UAM)
AND THE UAM EMISSION PREPROCESSOR SYSTEM
3.1 INTRODUCTION
In 1984, the EPA's Office of Air Quality Planning and Standards proposed that the Urban
Airshed Mcd^i (UA.Yi) be a "ricoiru^andad" {i.e., preferred) model for "photochemical
,^ ,->! 1 .•- - r^-t- nn ^ H -^ "•--•'--•-"---- - •' < '" ^ - r* '' ' ', - - - ,~ ~ ,. -, -, ^ " — PAfiriali-^oHfHic
fj L. i, ^ L ^ j '„ f 1i u J c. - j - - - > • - ~* - -~ > ' ^ > • • ,; >" . ' - -. .. ~* • — -- > i '. -' o a o . w f /A i 111 d i t L. c U i 11 i o
recomrnendaticn m ' C^o r,-;;r.] ln-"t r.n3 UAW "Is the most widely applied and evaluated
photochemical mcc'ei in existsnca.'1 Currently, the UAM is the recommended air quality
simulation modi8 r"c' f:=-^ ~ —-7-^ ->- ---i:ty analyses in the preparation of State
Implementation riurr.- (Si^' ac: 'S'rjsrec in ;he 1990 Clean Air Act Amendments (CAAA).
Accordingly, this document addresses the development of an emission inventory suitable
for photochemical modeling in terms of the specific requirements of the UAM. As noted
previously, however, the techniques and types of data necessary to generate the UAM
emission inventory inputs are generally suitable for developing an emission inventory for
use with any photochemical model.
3.2 CONCEPTUAL OVERVIEW OF THE URBAN AIRSHED MODEL
The UAM is a three-dimensional photochemical grid model designed to calculate
concentrations of both inert and chemically reactive pollutants by simulating physical and
chemical processes which occur in the atmosphere. These calculations are based on the
atmospheric diffusivity or species continuity equation, which represents a mass balance in
which all relevant processes (precursor emissions, transport, diffusion, chemical reactions,
and removal processes) are expressed in mathematical terms. For ozone assessment, the
model is usually applied for a 36- to 72-hour period during which adverse meteorological
conditions result in elevated concentrations. For carbon monoxide simulations, the model
is usually applied for shorter time periods (e.g., 8 to 16 hours).
Major factors affecting photochemical air quality include:
o spatial (vertical and horizontal) and temporal distribution of anthropogenic and
biogenic emissions of NOX, VOC, and CO;
o chemical composition of the emitted NOX and VOC;
90098 03" 3-1
-------
o spatial and temporal variations in wind fields;
o dynamics of the boundary layer, including stability and mixing;
o chemical reactions involving VOC, NQX, CO, and other important species;
o diurnal variations of solar insolation and temperature;
o loss of ozone and ozone precursors by dry deposition; and
o ambient background concentrations of VOC, NOX, CO, and other species in,
irrvr 5,J:I;:,;! / •„„:',. ii".J ^r, c.~.u :,bc'. - ths study region.
In a UAM application, thesa processes are simulated for the pollutant of interest (this may
be either sunry;:'-:-1; "C.-3 ccre ""rr.iicns or wintertime carbon monoxide
concentrations). The UAM selves the species continuity equation for each time step, in
each grid cell of the moaeiiny domain; the maximum time step is a function of grid size
and the maximum wind velocity. Typical time steps for urban-scale simulations are on the
order of 3 to 6 minutes. "~
Since the UAM accounts for spatial and temporal variations as well as reactivity
differences (speciations) of emissions, it is ideally suited for evaluating the effects of
emission control scenarios on urban air quality. Irf practice, a historical ozone (or carbon
monoxide) episode is replicated to establish a base case simulation. Model inputs are
prepared from observed meteorological, emission, and air quality data for a particular day
or days. The results of the UAM simulations are examined in the model performance
evaluation. Once the results have been evaluated and determined to perform within
prescribed levels, a projected emission inventory that includes changes in emissions due to
proposed control measures is used with the same meteorological inputs to simulate
possible future emission scenarios (in other words, the model will calculate hourly ozone
patterns likely to occur under the same meteorological conditions as the base case).
3.3 OVERVIEW OF THE UAM EMISSION PREPROCESSOR SYSTEM
To facilitate cost-effective development of the detailed emission inventories required by
UAM, the EPA's Office of Air Quality Planning and Standards sponsored development of a
system of computer programs designed to perform the intensive data manipulations
necessary to adapt a county-level annual or seasonal emission inventory for photochemical
modeling use. This system, the UAM Emission Preprocessor System (EPS), is available to
the public from EPA. Although the UAM EPS was originally developed for use with an
annual average inventory similar in format to NAPAP, many of its functions and procedures
can be adapted for use with similar annual or seasonal emission inventories.
90098 03" 3-2
-------
The UAM EPS consists of six FORTRAN programs which are executed sequentially on a
large computer mainframe system (some of the programs in the EPS can be adapted for
use on PCs or workstations) to generate the emission input files for the UAM. To execute
the UAM EPS, the following steps must be performed:
o Define the modeling region of interest. Identify grid origins (UTM
coordinates), resolution of the grid, cell size, number of cells in the x and y
directions, and the dates to be simulated. Discuss with the air quality modeler
the number of vertical layers to be modeled, the number of vertical layers
below and above the mixing height, and the minimum layer thickness.
o Determine the plume height cutoff, which will be used to determine wm'ci
point sources will receive elevated (i.e., vertically resolved) treatment by i'r^
model. For guidance on selection of an appropriate plume height cutoff,
consult Guideline for Regulatory Application of the Urban Airshed Model.2
o Run the EPA mobile source emission factor model MOBILE 4.1 to estimate
mobile source emission factors based on vehicle fleet mix for the specific area
to be modeled.
o Develop relationships between roadway links and grid cell coordinates
(optional).
- •**
o Develop spatial surrogate indicator data for allocating area sources to grid
cells.
o Prepare an inventory of biogenic emissions suitable for photochemical
modeling as described in Chapter 8.
For more information of accomplishing these steps, refer to the following EPA guidance
documents:
o Guideline for Regulatory Application of the Urban Airshed Model (EPA-450/4-
91-013, June 1991),
o User's Guide for the Urban Airshed Model. Volume IV: User's Guide for the
Emissions Preprocessor System (EPA-450/4-90-007D, June 1990), and
o User's Guide to Mobile 4.1 (Mobile Source Emission Factor Model) (EPA-AA-
TEB-91-01, June 1991).
Figure 3-1 shows an overview of the UAM EPS; Figures 3-2 through 3-7 contain
flowcharts of the input and output files for each program in the UAM EPS. Note that the
90098 03'2 3-3
-------
PREPNT
PREGRD
GRDEMS
CENTEMS
POSTEMS
Griddsd /
Biogenics
MRGEMS
UAM >
Prepnxessor ]
PTSRCE /
FIGURE 3-1. Overview of the UAM emissions preprocessor system.
90093 03"
3-4
-------
UAM preprocessor PTSRCE must be executed subsequent to the UAM EPS to prepare the
final UAM elevated point source input file before modeling. The six routines and their
primary functions are briefly described below. For a detailed description of the UAM EPS
and its input file formats, see the User's Guide for the Urban Airshed Model, Volume IV:
User's Guide for the Emissions Preprocessor System.3
PREPNT. The PREPNT program (Figure 3-2) prepares the annual average or seasonal point
source inventory for chemical speciation by the CENTEMS module. Latitudinal and
longitudinal coordinates for each source are converted to Universal Transverse Mercator
(UTM) coordinates, which are in turn converted to modeling region grid cell coordinates.
Each source is assigned a stack identification code and the emissions allocated by stack.
Temporal distribution profiles are assigned based on the operating information contained in
the annual or seasonal inventory. PREPNT also calculates a plume rise for each stack
based on the Briggs effective height calculation, using the stack parameters contained in
the annual or seasonal inventory and default meteorological conditions; the value of this
plume rise determines if stacks will be treated as elevated or low-level sources by
subsequent programs in the UAM EPS.
PREPNT requires four input files:
(1) a user input file, containing information on the plume rise cutoff for elevated
treatment, default stack parameters, and optional emission control factors for
each pollutant by Standard Industrial Classification (SIC) code;
(2) projected industrial growth factors by two-digit SIC code;
(3) a file defining the modeling region, which specifies the modeling grid (UTM
origin, UTM zone, and number and size of grid cells), number of counties
within the modeling region, FIPS and AEROS identification codes for each
county, and control codes by county indicating the presence of Inspection and
Maintenance and Stage il Vapor Recovery control programs; and
(4) the annual or seasonal point source inventory, including source identification
codes, location, stack parameters, operating schedule information, and
emissions.
In addition to miscellaneous informative outputs, PREPNT produces two files which
undergo subsequent processing in the CENTEMS module: a gridded point source inventory
in Model Emissions Record Format, or MERF (Table 3-1), and a control file containing
parameters for the stacks selected for elevated source treatment.
PREGRD. The PREGRD program (Figure 3-3) reformats an annual or seasonal area source
emission inventory and prepares it for gridding. Emissions are separated into two files,
90098 03'3 3-5
-------
User
Input
Growth
Factors
Region
Definition
NAPAP
Point Source
Data
(13)
PREPNT
Point Source
(merf)
Stack
Parameters
/ Report
Elevated
Stack
List
(19)
Control
File for
Stacks
(20)
Messages
(16)
f Point Sources
without
Coordinates
FIGURE 3-2. Input and output files used by PREPNT. (Numbers in parentheses
refer to FORTRAN units used.)
90098 03"
3-6
-------
TABLE 3-1. Standard Modeling Emissions Record Format (MERR employed by the UAM
Emissions Preprocessor System.
Variable
ISRG
IFIP
SIC
sec
1CL
JCL
IYR
IDYCOD
IWKCOD
FID
FST
FCNTY
VMNTH
CO
CNO
SOX
THC
PM
SLBL
Columns
1-3
4-8
9-12
13-20
21-22
24-26
27-28
29-30
31-32
33-41
42-46
47-52
53-56
57-116
117-126
127-136
137-146
147-156
157-166
167-168
169-175
Type
I
I
R
A
I
i
I
I
I
A
A
I
-
R
R
R
R
R
R
-
A
Description
-••••-
Gridded surrogate code (not used, skipped)
FIPS state/county code (not used, skipped)
Either Scurce Industrial Classification or Area
Source Catecory code '
Either Standard Classification Code or Area
Source Category code
I coordinate of grid cell j
J coordinate of grid ceil
Year, two digits (e.g., 39 for 1989)
Diurnal variation code
Weekday variation code
Facility ID (0 or blank for area sources)
Stack ID (0 or blank for area sources)
AEROS state/county code
(Not used, skipped)
Array of 12 monthly factors
CO episode emissions (kg/day)
NOX episode emissions (kg/day)
SOX episode emissions (kg/day)
THC episode emissions (kg/day)
PM episode emissions (kg/day)
(Not used, skipped)
Scenario label (not used, skipped)
Sample MERF Record:
3812304001501229245 0018 1110380 0 0830 O830.0830.O83O 0830 0830.0830 0830 O830.0830 0830.083 0.00 0.00 000 30248 000
90C98 03"
3-7
-------
Growth ;d3)
Fcrtcr
'NAPAPArea
and Mobile /(14)
Emissions
Inventory
Motor
Vehicle
Factor
->
OS)
Skeleton /
' Area Emissions/
Skeleton
Mobile
Emissions
(merf)
Messages
nGURE 3-3. Input and output files used by PREGRD.
90098 03"
3-8
-------
area sources and on-road motor vehicle sources. Emissions from on-road motor vehicles
are disaggregated into exhaust, evaporative, refueling, and running loss components.
PREGRD requires five input files:
(1) a user input file, containing fractions of vehicle miles traveled (VMT) by
gasoline-fueled vehicle type which are used in allocating refueling emissions, a
flag indicating a weekday or weekend modeling episode, and an optional list
of additional control factors for each pollutant by source category code;
(2) a region definition file (this is the same file used by PREPNT);
(3) projected growth factors by source category code;
(4) a file containing motor vehicle adjustment factors used to adjust annual or
seasonal average mobile source emissions for episodic conditions; and
(5) the annual or seasonal area source inventory, containing county-level emission
estimates by source category.
The area source and on-road motor vehicle MERF files generated by PREGRD for input into
GRDEMS contain incomplete records, with grid cell indexes and other fields left blank.
" •>»
GRDEMS. The GRDEMS program (Figure 3-4) allocates the pre-processed county-level
emissions from PREGRD to the modeling region grid cells based on a gridded spatial
apportioning factor field and optional link data (e.g., limited access roads, railways, etc.).
Temporal distribution profiles are assigned by source category.
GRDEMS requires six input files:
(1) a user input file, containing a description of the run, the year of emissions, a
unit conversion factor, and optional pairings of surrogates and source
categories by county to override the pairings contained in the cross-reference
file;
(2) a region definition file (the same file used by PREPNT and PREGRD);
(3) a cross-reference table of spatial surrogate codes and diurnal distribution
profiles by source category;
(4) the gridded spatial surrogate file;
(5) an optional data file specifying link locations for the modeling region; and
90098 03" 3-9
-------
User
Input
Region
Definition
Cross-
Reference
Table
(10)
r(12)
Gridded /(13)
Surrogates
Link
Data.
(optional)
'(14)
Skeleton
Nongridded /<15>
Emissions
(merf)
GRDEMS
Messages
(17)
Gndkd
Emissions
/ (merf)
(18)
Eired
Records
(merf)
FIGURE 3-4. Input and output files used by GRDEMS.
90098 03"
3-10
-------
(6) the county-level MERF emissions file(s) generated by PREGRD.
GRDEMS produces a gridded MERF emissions file for further processing by the CENTEMS
program. Although the separate area and mobile source emissions files from PREGRD can
be merged before running GHDEMS, processing these files separately through GRDEMS
allows better tracking of emissions totals and quality control.
CENTEMS. CENTEMS (Figure 3-5), the central program of the UAM EPS, creates a
gridded low-level binary emissions file in UAM-format and an elevated point source input
file for the UAM preprocessor program PTSRCE. Annual average daily emissions are
adjusted to account for monthly variations and assigned to the hours of the modeling
episode based on tempera! distribution profiles. Total hycrccarbon emissions are spsciatec
into Carbon-Bond (CB-IV) classes using E?H VQC speciation profiles by either Source
Classification Code (SCO for point sources or EPA Area Source Category Code.
In addition to the MERF emissions files from PREPNT and GRDEMS, CENTEMS requires the
following inputs:
(1) a user-input file, containing information on the type of sources, tine and day
flags for the modeling episode, modeling region information, and control
factors;
(2) the elevated point source parameter control file created by PREPNT;
(3) a glossary file matching SIC/SCC combination (for point sources) or EPA
Source Category (for area or mobile sources) with the activity, process,
control, and inventory source category codes used for tabulation of emissions
totals;
(4) a speciation factors file, containing carbon-bond speciation factors in terms of
moles/gram by EPA VOC speciation profile code;
(5) a speciation profile file, which assigns EPA VOC speciation profiles based on
SCC (for point sources) or EPA Source Category (for area and mobile
sources);
(6) a diurnal variation factors file, containing hourly profiles for the diurnal codes
contained in the MERF files, which are used to allocate the daily emissions to
the hours of the modeling episode; and
(7) a weekday factor file for adjusting emissions based on the day of week by the
weekday code in the MERF record.
90098 03" 3-11
-------
/ . " /
/ inputs /
/ /
/ r?"?01 /"')
/ file for A— -^ —
/ stacks /
/i
Direct access /,
y— , > ' '• " )
t rJ?f1|Torv i
<^"U-3U'/ /
glossary /
'
/Speciation /,,a-
factors /-—
glossary /
/ /
/ Speciation /(U)
/ profile 1
/Diurnal 7(15)
faaors /
/Weekly 1(16)
factors I
/ Gridd£d /IT)
/ emissions /
/ (merf) /
—
fri rFNTFMS
-x»
/Messages, /
emission /
total tables /
/ UAM
(31) ^ IO-A -level /
/ ernissiora /
/ (binary) /
/ PTSRCE /
(32) I rij^^'c- j
( } »/ input file /
/ (ASCII) /
(34) / Binary /
1 yH categorized /
/ totals /
FIGURE 3-5. Input and output files used by CENTEMS.
90098 03"
3-12
-------
CENTEMS produces the low-level DAM emissions file and the elevated point source file
mentioned above as well as a binary file of categorized emissions totals for input to the
POSTEMS module.
POSTEMS. The POSTEMS program (Figure 3-6) merges up to six ground-level
anthropogenic UAM emission files into one fi!e. POSTEMS also provides summary
printouts describing emission totals by activity, process, control, and inventory source
category codes.
The input files for POSTEMS include the following:
(1) a user inouc fits specifying the Julian dais .'or :>'..-; ', ,'•-,', .'.v^cec UAiV! emissions
file, the number of files to be merged. 'j,s^f-33'<=ct5'- options, UAM header
record information, and the types of emissions in tha input files;
(2) five input files defining the codes contained in the binary file of caie-gorized
emissions totals produced by CENTEMS;
(3) a chemical species data file, containing information about the molecular
organic species in the emission inventory; and
(4) up to six low-level UAM emissions files and the corresponding binary
categorized totals files from CENTEMS.
MRGEMS. The MRGEMS program (Figure 3-7) merges two low-level UAM emissions files
into one file. Generally, this program is used to merge the anthropogenic emissions file
created by POSTEMS with a biogenic emissions file (in UAM format) generated by another
program, such as BEIS (described in Chapter 8).
90098 03" 3-13
-------
User
Input
Activity
Codes
'(14)
Process /OS)
Codes
Source /
Category
Codes
Emission
Files
(binary)
Emission
Totals by
Category
(binary)
Control /(17)
Codes
Profiles /(18)
Codes
Species /09)
(11)
(12)
(13)
Messages
Merged UA.M
Emissions
(binary)
Emission
Totals by
Category
FIGURE 3-6. Input and output files used by POSTEMS.
30098 03"
3-14
-------
,rv^L> h°L
Emissions
UAM
Einis-'.cns
(Biogenic)
il)
I Merged
/ UAM
Emissions
(binary)
Message
Output
HGURE 3-7. Input and output files used by MRGEMS.
90098 03"
3-15
-------
References for Chapter 3
1 . User's Guide {w the UrD_an...Air^-h^_Mp,d_ei,.-_yo!ume i_i_User's Manual for DAM (CB-iV),
EPA-450/4-90-007A, U.S. Environmental Protection Agency (OAQPS), Research
Triangle Park, NC, June 1990.
2. Gdjleilne_:f^^ Airshed Model, EPA-450/4-91-013,
U.S. Environmental Protection Agency (OAQPS), Research Triangle Park, NC, June
1991.
3. User's Guide for the Urban Airshed Model,
Emissions Pra^r"censer _S /,'••; •• - , :,"-, •>-•-;"<",,'.?,-£0-0 J 70, !j.S. 'fi-vTOimentai Protection
Agency (OAQPS), Resear:'- T--?,-;^ p~rV, M':, jLre
9009803° 3-16
-------
4 DETERMINATION OF THE GRID SYSTEM
4.1 SELECTING AM APPROPRIATE GRID SYSTEM
Identification of the grid system which will be used to spatially reference emissions in the
modeling inventory influences ai! subsequent phases of the emission inventory process.
This chapter has been included to provide a general discussion of issues of concern in
selection of the modeling region and grid system. For definitive C'jid"PC3 concerning
establishment of the grid system, however, consult che Guideiin^ int n^gviiatory
Application of the Urban Airshed Model.1
The first step in defining the grid svstern is seiection of a grid coundary outlining the area
to be modeled. Once the grid boundary has been chosen, ths region enclosed by the grid
boundary (subsequently referred to as the "modeling region" in this text) must be
subdivided into grid ceils. Figure 4-1 illustrates the concepts of grid boundary and grid
cells. The DAM is a three-dimensional grid model, capable of resolving emissions vertically
as well as horizontally. For purposes of compiling the modeling emissions inventory,
however, emissions need only be resolved over a horizontal grid system; the DAM will
automatically allocate emissions from those sources selected to receive elevated treatment
to the appropriate vertical layer based on the stack parameters for each source and
meteorological conditions. Accordingly, in the following discussion, the term "grid" will
refer to a two-dimensional grid system overlaying the area to. be modeled.
Almost always, the grid boundary will be rectangular and the grid cells will be equally-sized
squares. Selection of an appropriate grid system involves consultation with planning
agencies, air pollution control agencies, meteorologists, and photochemical modeling
specialists to ensure that the chosen grid system meets the general objectives of the
photochemical modeling program.
The modeling inventory must spatially resolve emissions in terms of the individual grid cells
comprising the modeling region. A typical grid system will cover thousands of square
kilometers and contain hundreds of grid cells.
»• Figure 4-2 shows an example grid system. In this figure, a 4 km x 4 km
modeling grid has been superimposed over a map of the St. Louis, Missouri
area. This grid consists of 15 grid cells in the east-west direction and 20 cells
in the north-south direction for a total of 300 individual grid cells, each 16
square kilometers in size. The total area encompassed by the grid boundary is
4,800 square kilometers.
90098 04" ' 4-1
-------
(a) The Area to Be Modeled
(b) Specificalion of the Grid
FIGURE 4-1. Schematic illustration of the use of the grid in the Urban Airshed Model.
90098 04"
4-2
-------
FIGURE 4-2. The St. Louis Area with locations of the RAPS surface stations and 4 km x 4km
modeling grid superimposed.
90098 04"
4-3
-------
Obviously, since emissions must be determined for each grid cell in the modeling region,
an appropriate grid system should be developed at the outset of the emission inventory
effort which defines both the overall size and shape of the grid to be modeled and the size
and number of grid cells that compose the grid. Defining the grid system before beginning
the modeling inventory development process will help minimize redundancy of effort.
4.1.1 Area Covered by the Grid System
The selection of the modeling region should reflect location of sources of meteorological
and air quality data, location of current and expected major emissions sources, and types
of control strategies under consideration. The photochemical reactions resulting in ozone
formation can occur many miles downwind of precursor pollutant sources. Accordingly,
the modeling region must be fairly large to ensure that all major emission sources which
may affect ozone formation in the urban area are included.
The modeling region should contain as many ozone and precursor pollutant monitoring
stations as possible to facilitate model validation. The model validation process entails
simulating an historical ozone episode to determine if the observed ozone and precursor
pollutant concentrations at each monitoring station agree with the concentrations
predicted by the model. In the validation process, ambient air measurements are used to
define pollutant concentrations along the boundary of the modeling region.
Additionally, the modeling region should be large enough to encompass areas of current
limited land use activity that are expected to develop significantly as a result of projected
growth. Since an important application of photochemical models is evaluation of expected
ozone concentrations in future years, the emissions modeler should consult land use
planners to determine what types of growth are expected and in what areas.
The modeling region should also be large enough to encompass the effects of meteorology
in the modeling region. Since peak ozone levels often occur downwind of the urban
center, as much "downwind area" as possible should be included in the modeling region so
that the model can predict where and when these peak levels will occur. Selection of a
large modeling region also minimizes the possibility that receptor locations will be impacted
by air parcels that have left the domain and then re-entered it because of shifting
meteorological conditions during the time interval over which the model is run.
Finally, to reduce the dependence of model predictions on uncertain boundary conditions
(e.g., the pollutant concentrations assumed along the boundary of the modeling region),
the modeling region should extend into areas with little or no emissions. Of course, in
certain areas (e.g., the Northeast United States), pollutant transport from nearby urban
90098 04" 4-4
-------
arsas may prscluds ths possibility of a c!s<2r. bccky'CLi.nd ale-rig cny wOundijj'iss tnat mav 23
chosen.
If the selected modeling region covers too large an area, however, certain problems may
occur in gathering and manipulating data. For instance, the modeling region may extend
into areas for which detailed spatiaiiy and temporally resolved emission estimates cannot
be made due to lack of adequate information. This might be the case if one part of the
domain lies within the jurisdiction of a metropolitan planning organization (MPO) and
another part lies within an outlying, undeveloped jurisdiction. Detailed records and
projections will probably be available for the metropolitan areas, out may not 3xiot for the
outlying area. Technical problems may also ba ancouruarcU n: '.onouj iu«"iouiuc;on5 wkn.n
the modslir.- region maintain 'r\:a:rr,z::z~ ',r, Ciffs.v,:-.- .:, '..::. ":. -, .. . ,' ..-_ ,-.;.-
maintain records for townships and use EPA's Ae<"OP"~;-'C • >':c<-~<3;:;on '~-^:.r'~- =. S "-:: = ir,
(AIRS), whereas another area may maintain records for census tracts are use 2 locally
developed data handling system that is incompat;':!B -.'•',-• * ~"
If the exact area for which the photochemical model wii! os «<-,,).;*,-, ,s not initially known,
perhaps because of uncertainties about future land use or the affect of meteorological
conditions, the emission inventory development process can still proceed. In this case, the
emissions modeler must choose an Emissions grid system for which to compile emissions
data. In general, the area encompassed by the emissions grid should be as large as
possible. A-smaller area can then be selected for photochemical modeling within the
emissions grid with no additional emission'data collection effort required. Thus, the
emissions grid can be larger than the actual grid used for modeling. For most efficient use
of time and resources, however, the emissions grid and the modeling region should
coincide.
The modeling region is normally rectangular. Some models, including the UAM, may
accept an irregularly shaped modeling region. Even if the modeling region is irregularly
shaped, however, a rectangle should be used for the boundary of the emissions area for
the sake of simplicity and ease in locating grid cells. The forcing of a rectangular boundary
around an irregularly shaped city may mean that some of the peripheral grid cells may
contain zero emissions. For example, coastal cities usually have portions of the ocean
included within the rectangular grid boundary, as shown in Figure 4-3.
4.1.2 Grid Cell Size
The degree of spatial resolution of the modeling emission inventory must also be decided
during the initial planning stages of the inventory effort. Choice of an appropriate grid
spacing depends on the overall modeling objectives, the total area of interest, the amount
of manpower available, and the cost of running the photochemical model.
90098 04" 4-5
-------
o
-------
Ideally, the smallest feasible grid spacing should be chosen to accurately represent
emissions from a variety of sources in different locations. The grid spacing, however, will
be determined in part by the size of the modeling region.
> If the region of interest is 100 km by 100 km, a grid spacing of 2 km would
result in a total of 2,500 individual cells for which emissions would have to be
calculated, as shown in Figure 4-4a. For such a large area, such fine
resolution is unlikely to result in appreciable improvement in predicted ozone
levels over the entire region relative to a larger grid spacing, such as 5 km
(Figure 4-4b). The major advantage of the larger grid spacing is the
considerably fewer number of grid ceils (400 as opposed to 2,500) and the
corresponding reduction in both computing and manpower costs.
In urban-scale photochemical modeling efforts, covering the maximum amount of area with
the smallest feasible number of grid cells normally results in grid spacing between 2 to 5
km. Since ozone formation occurs over an appreciable amount of time and space, grid
spacing smaller than 2.km is not recommended. Also, grid spacing smaller than 2 km may
exceed the resolution of both available transportation modeling data and area source
apportioning factor data. On the other hand, grid spacing larger than 10 km usually masks
the effect of individual sources, since emissions are averaged over the entire grid cell area
by the model (i.e., when using 10 km grid spacing, all emissions in any cell, including
individual point sources, are assumed to be uniformly emitted from the entire 100 square
kilometer grid cell area). If the grid spacing is large, this artificial dilution can cause
inaccuracies in the modeling. Users are referred to the Guideline for Regulatory
Application of the Urban Airshed Model1 for further guidance on selection of grid cell size.
If a grid system has not been chosen prior to inventory compilation, the smallest grid
spacing under consideration should be used for spatial resolution of the emission
inventory. Aggregation to larger grid cells is then a simple procedure of combining an
integral number of grid cells. Thus, the initial emission grid spacing can be smaller than
the actual grid spacing subsequently used for modeling. However, resource considerations
may preclude such fine emission resolution if the inventory involves a very large grid area.
Compilation of the inventory for a grid spacing larger than is subsequently considered
desirable for photochemical modeling should be avoided because of the difficulty in
allocating emissions to a finer grid cell network once the inventory is completed
(aggregation of smaller grid cells to form larger ones, however, is relatively easy, as
mentioned above). Optimally, of course, the emission grid cells will coincide in size with
the grid cells used for photochemical modeling.
As an important consideration in determining the size of both the modeling region and the
individual grid cells, the emissions modeler, in concert with the other agencies involved in
this decision, should examine the overall objectives of the photochemical modeling
application. If control strategies are to be evaluated over a large region, then a fairly large
90098 04" 4-7
-------
100 fan
. 100km-
(a) 2 km grid spacing (2,500 grid cells)
100 km
,100km-
(b) 5 km grid spacing (400 grid cells)
FIGURE 4-4. Comparison of number of grid cells required for a 100 km x 100 km modeling
region for 2 km and 5 km grid spacings.
90098 04"
4-8
-------
modeling region should be selected and a fairly coarse emission resolution may be
acceptable. If control strategies are to be evaluated for a fairly small area (e.g., an
individual city, such as the St. Louis, Missouri area shown in Figure 4-2), then a relatively
fine emission resolution may be warranted. Evaluating the air quality impact of a proposed
new source would probably require a relatively fine emission resolution, since very large
grid cells may mask the effect of the individual new source. Thus, before a final grid
system is chosen, a photochemical modeling specialist should be consulted regarding the
effect of emission resolution on modeling predictions.
4.2 MAP GRIDDING PROCEDURES
4.2.1 UTJVI Coordinate System
Once the grid system has been selected, it must be overlaid on an appropriate map to
determine (1) which sources lie within each grid cell and (2) area source apportioning
factors for each grid cell. The recommended coordinate system for this task is the
Universal Transverse Mercator (UTM) system, which is used in the AIRS emission data •
system to reference all point source locations. The UTM coordinate system should be
used from the outset in the development of the grid system, since changing from one
coordinate system to another can be time-consuming. For those urban regions which
encompass more than one UTM zone, all coordinates should be referenced to one zone.
"*>
The most accurate maps normally available for gridding purposes are those produced by
the U.S. Geological Survey (USGS), which provides topographic maps in different scales
for all sections of the United States. The more recent USGS maps have a superimposed
10 km UTM grid system; older USGS maps simply have blue tick marks along the edges
that represent the UTM coordinate system.
The master grid system should be based on a USGS map or data base, since other maps
(e.g., highway maps) may contain considerable inaccuracies. The grid system can be
manually overlaid on a map by positioning a transparent plastic sheet over the map and
drawing the gridded area on the plastic sheet. Alternatively, important features (such as
political boundaries, streets, etc.) on the USGS map can be digitized and incorporated into
a computerized data base.
4.2.2 Orientation of the Grid System
Almost without exception, the grid system should be aligned so that the grid lines
essentially run north-south and east-west. Within the region typically modeled in most
urban areas, a grid system based on UTM coordinates will largely meet this criterion.
North-south alignment is not actually required by the photochemical model, but facilitates
90098 04" 4-9
-------
definition of locations on the grid and enhances compatibility of the inventory with
meteorological data. For the extremely few instances where north-south alignment of the
grid would cause significant modeling inaccuracies, the UAM EPS supports skewed (i.e.,
non-north-south) grid orientations. A photochemical modeling expert can be consulted to
determine if a skewed grid orientation should be used for a particular region. Figure 4-5
shows an example of a skewed modeling region.
Likewise, for the sake of convenience, the grid should be oriented so that the grid cell
boundaries coincide with the UTM kilometer grid lines (in other words, so that the grid
cells are defined by whole UTM kilometer numbers). This simplifies location of particular
grid cells and allocation of point sources to the appropriate grid cells. Obviously, this will
not be possible if grid ceil dimensions are determined in terms of non-metric units, such as
miles.
4.2.3 Problems in Gridding
Often, the master grid system developed using the methods described above must •
incorporate features not available from USGS maps or data, such as detailed street
locations, land use patterns, or population density. Other maps may not be as
dimensionally accurate or on the same scale as the USGS map, which can cause major
problems in combining information from different maps. While attempting to align the
master grid on a land-use map, the emissions mocfeler may notice that certain major
features are located in slightly different grid cells than on the USGS map. If the scale of
the auxiliary map is not quite accurate, it may be possible to extend or decrease the map
grid line dimensions so that most grid cells correspond to those on the USGS map. In
many cases, the best procedure is to align the main urban area as correctly as possible.
Inaccuracies in the outer portions of the grid are less important because fewer emissions
normally occur in the outlying grid cells.
USGS maps, although dimensionally accurate, do not always include enough detail to
locate particular sources. This is partly because of the limited number of available scales
and partly because some of the available USGS maps are old and so do not show current
street locations. Thus, the emissions modeler must usually obtain more detailed street
maps covering the entire area of interest in order to accurately locate specific sources. If
possible, all street maps should be at the same scale so that a number of them can be
combined to show a larger area. Overlaying the grid system on the individual street maps
can be difficult, because street maps rarely have UTM coordinates; the grid system must
be carefully positioned using known reference points, such as major street intersections,
shown on the USGS map. Often, the easiest way to overlay the grid system is to digitize
the desired features and, using the known UTM locations of several features shown on
both maps as reference points, incorporate this data into a computerized data base (such a
90098 04" 4-10
-------
550 600 650 700 750 800 850 900
950
4300 -
4250 "-•
3850
3 BOO 4 :
3?CO
550 600 650 700 750 800 850 900 950
UTM Easting (Zone 10)
FIGURE 4-5. Modeling region encompassing the southern San Joaquin Valley and Sierra
Nevada.
30098 04"
4-11
-------
data base of computerized locations of streets and other features may also prove useful
for apportioning emissions from some area sources).
90098 04" 4-12
-------
References for Chapter 4
1. Guideline for Regulatory Application of the Urban Airshed Model, EPA-450/4-91-01 3,
U.S. Environmental Protection Agency (QAQPS), Research Triangle Park, NC, June
1991.
90098 04" 4-13
-------
5 POINT SOURCE EMISSIONS
5.1 DATA COLLECTION
For most urban regions, a basic annual or seasonal point source emission inventory will
already exist which contains most (if not all) of the information required to develop the
modeling inventory. Basic inventories are often maintained in standardized formats, such
as SAMS, AIRS, or NAPAP, and generally contain the following types cf :nfcrr-,2';.cn:
Source identification: county, facility, and source codes; Standard Industrial
Classification (SIC) of the facility; and location (latitude and .'cngituds or '_':" ,,
coordinates) of each source.
Process information: Source Classification Code (SCO or basic equipment codas for
each-process; stack parameters (height, diameter, gas temperature, and gas exit
velocity or fiowrate); control device information; operating rates and schedules; and
fuel characteristics.
Emissions data: annual or seasonal Estimates of VOC, NOX, and CO emissions for
each process within the facility.
Table 5-1 lists the types of data contained in the Aerometric Information Retrieval System
(AIRS) Facility Subsystem (AFS) and the level at which each is maintained. Table 5-2 lists
the data items currently required by the DAM EPS for processing of the point source
inventory. The existing point source inventory will generally fulfill most of the
photochemical modeling requirements; one notable exception may be the lack of sufficient
operating schedule information to accurately estimate hourly emission rates. Additionally,
VOC and NOX emissions will need to be disaggregated into chemical classes and NO and
N02, respectively. Techniques for speciation of VOC and NOX emissions are discussed
separately in Chapter 9.
. | The UAM EPS program PHEFNT expects point source informatFon (such as
stack parameters and emissions estimates) to be repeated in the units of
^measure indicated in table 3-2. PREPNT contains rotrtines wNch convert this
• data to the units .expected by subsequent programs .in the UAM EPS. If the .
: units used in the existing point source inventory d^fer from those M Table 5-2,
the emissions modeler will have to either develop an auxiliary preprocessing
90098 05" 5-1
-------
TABLE 5-1. Types of emissions data contained in the AIRS Facility Subsystem and the
level of detail at which each is maintained.
LEVEL
TYPES OF DATA
Plant
(data pertaining to the
facility as a whole)
geographic and other address information
year of emission inventory estimated plant pollution
emissions
comment information about the facility
miscellaneous mailing label and permit fee da;3
Stack
(data pertaining to
emissions stacks or
vents within the facility)
map coordinates and physical description
gas flow rate, exit velocity and temperature
estimated and measured emissions by pollutant
stack comment data
Point
(data pertaining to
emissions points within
the plant, frequently
boilers or tanks)
point design capacity^
burner make, type and seasonal throughput data
operating schedule
estimated, measured and state defined pollutant emissions
values
tank descriptive and construction data
comment information about the emission point
Segment
(data pertaining to
activities or
components, such as
fuel combustion and
other processes,
associated with each
point)
segment source classification code (SCO
operating rate, fuel, control equipment and emission factor
data
estimated and measured emissions by pollutant
segment chemical data
comment information pertaining to the segment
source: Reference 2
90093 05"
5-2
-------
TABLE 5-2. Data fields required by the UAM Emissions Preprocessor System.
Variable
Description
STATE
COUNTY
PLAMTJD
OTM_ZONE
PCWTJD
SiC
LAT
LON
W1NTHRU
SPRTHRU
SUMTHRU
FALTHRU
HOURS
DAYS
WEEKS
SJACK_HT
STACK_D!
STACK_TP
FLOW
PTSCONST
SCC
NOXEM1SS
TSPEMISS
COEMISS
SOXEMISS
THCEMISS
NEDS state code
NEDS county code
Plant ID code
UTM zone
Point ID code
Standard Industrial Classification code
Latitude, degrees
Longitude, degrees
Winter throughput, %
Spring throughput, %
Summer throughput, %
Fall throughput, %
Hours/day in operation
Days/week in operation
Weeks/year in operation
Stack height, ft
Stack diameter, ft
Stack temperature, °F
Gas flow rate, cubic ft/min
Range of points with common stack, AAZZ
Source Classification Code
NOx emissions, tons/yr
TSP emissions, tons/yr
CO emissions, tons/yr
SOx emissions, tons/yr
THC emissions, tons/yr
source: Reference 1
90098 05"
5-3
-------
,-':. -program to''reformat the existing inventory to be compatible v^tfr PRHPMT or
, : modify the PREPftT source code to bypass the conversion, cateulatrans.
!n genera!, the point source data collection methodologies described in ProceduresJoMha
Preparation.of Emission Inventories for Precursors of Ozone, Voiurne I3 can aiso be used to
collect any additional data required for the modeling inventory. In short, these procedures
include mail surveys of individual facilities, use of other air pollution agency files (such as
permit applications), use of information from selected publications, and examination of
other available inventories. The emissions modeler should consult Volume I for a detai'sc
discussion of these techniaues. The remaining sections of this chapter focus on v-e
additional data requirements of the modeling inventory and specific data handling
:echniques.
5.2 RULE EFFECTIVENESS
Although past inventories have assumed that regulatory programs would be implemented
with full effectiveness, experience indicates that regulatory programs are less than 100
percent effective for most source categories in most areas of the country. Accordingly, a
"rule-effectiveness" factor should be applied (in addition to the control factors associated
with each measure) to account for less than full dbmpliance.
When estimating the effectiveness of a regulatory program, several factors should be
considered. These include:
o the nature of the regulation (e.g., whether any ambiguities or deficiencies
exist, and whether test methods and/or recordkeeping requirements are
prescribed);
o the nature of the compliance procedures (e.g., accounting for the long-term
performance capabilities of the control);
o the performance of the source in maintaining compliance over time (e.g.,
training programs, maintenance schedules, and recordkeeping practices); and
o the performance of the implementing agency in assuring compliance (e.g.,
training programs, inspection schedules, and follow-up procedures).
The current ozone/carbon monoxide policy states that a factor of 80 percent can be used
to estimate rule effectiveness in base year inventories. Alternatively, states are given the
90098 OS' 5-4
-------
option of deriving local category-specific rule effectiveness factors (within tightly
prescribed guidelines) as EPA deems appropriate.
Rule effectiveness should be incorporated into ail baseline and projected inventories with
the following exceptions: (1) sources not subject to ths regulation; (2) sources achieving
compliance by means of an irreversible process change that completely eliminates solvent
use; and (3) sources for which emissions are directly determined by calculating solvent use
over some time period and assuming that all solvent was emitted from the source during
that time period. The rule effectiveness factor is applied to ths sstirnstcd centre!
efficiency as shown in the following example.
> if uncontrolled emissions from a given source -5f3 cO ibs'cc'/- ""*'- '•"'••
estimated control efficiency of a proposed measure ;o SOsi, :'.".: _:.::•_;.;
controlled emissions accounting for a rule effectiveness facto" of 30% are
calculated to be [50 Ibs/tiay] x [ 1 - {0.90) x (0.30;], or 1 -!• ".;., ,L . .-..; - ...
emissions reduction is thus 72 percent.
5.3 SPATIAL RESOLUTION
Since photochemical models require that all emissions be associated with specific grid
cells, the emissions from each point source must be assigned to the grid cell in which the
point is located. This assignment can either be performed manually (using maps) or with
the assistance of a computer.
Point sources can be manually assigned to grids by locating their coordinates (UTM or
latitude and longitude) or street addresses on a map of the area which is overlaid with the
inventory grid system, as described in Section 4.2.1. As a basic quality assurance
procedure, street addresses should be checked against the coordinates included in the
basic inventory to identify possible errors in coordinate assignments. Usually, each grid
cell is assigned a number according to some modei-specific system, and this grid number
and the source-type category should be entered into the data handling system for each
point source to facilitate subsequent processing.
Alternatively, a computer program can be used to assign grid cell coordinates to each
source based on the location data contained in the annual or seasonal point source
inventory. This approach, which is generally more efficient than manual location of point
sources on a map, is especially attractive if the grid assignment process may have to be
repeated numerous times, as would be the case if the grid orientation or grid cell size were
to change. However, even if the grid assignment is computerized, the emissions modeler
may find it useful to overlay the grid system over an accurate map of the area to assist in
visualizing and checking grid cell assignments, especially for the largest emitters in the
region.
90098 05" 5-5
-------
In-the UAM EPS, gnd ceils -are- identified by the U,J> co0*dif»ate of Eh& upper
right comer of each cell; far example, the grid ceil at th« <3ftgin of the
-------
TABLE 5-3. Day-specific Modeling Emissions Record Format (MEHF) used by the UAJV1
Emissions Preprocessor System.
Variable
ISRG
IFIP
SIC
sec
ICL
JCL
IYR
IDYCOD
IWKCOD
FID
FST
FCNTY
VMNTH
CO
CNO
SOX
THC
PM
SLBL
Columns
1-3
4-8
9-12
13-20
21-23
24-26
27-28
29-30
31-32
33-41
42-46
47-52
53-56
57-116
117-126
127-136
137-146
147-156
157-166
167-168
169-175
Type
1
1
R
A
1
1
1
1
1
A
A
1
R
R
R
R
R
R
A
Description
i
Gridded surrogate code {not used, skipped!
FIPS state/county code (not used, skipped)
Either Source Industrial Clsssificdticn or Area
Source Category code
I
Either Standard Classification Code or Arno Source
Category code
1 coordinate of grid cell J
J coordinate of grid cs!i ;
Year, two digits (e.g., 89 for 1339} |
Diurnal variation code1 jj
Weekday variation code2
Facility ID (0 or blank for area sources)
Stack ID (0 or blank for area sources)
AEROS state/county code
(Not used, skipped)
Array of 1 2 hourly factors3
CO episode emissions (annual average kg/day)
NOX episode emissions (annual average kg/day)
SOX episode emissions (annual average kg/day)
THC episode emissions (annual average kg/day)
PM episode emissions (annual average kg/day)
(Not used, skipped)
Scenario label (not used, skipped)
1 Either -1 or -2, corresponding to the first 12 hours of the day or the second 12 hours
of the day, respectively.
2 Always equal to 0 in day-specific MERF.
3 Factors used to allocate daily emissions to hours of day; correspond to first 1 2 hours
of day if IDYCOD = -1 and second 12 hours of day if IDYCOD = -2.
source: Reference 7
90098 05"
5-7
-------
dlumaily allocate total, daily emissions (note the differences between the day-
specific MEflF and the standard MERF shown In Table 3-1},. ;
"The currently available version of the LfAM EPS ^Version TrO) does not provide
software to generate day-specific MERF records; accordingly, tbts Information
'.must;be incorporated Into the modeling Inventory outsids of ErS using
supplement a! software or by manual editing of the point source ftdfiRF fiis =
produced by PREPNT, . : =" - . • : ; ,-- --. • - :
For many smaller point sources, reasonable temporal resolution can be obtained Irorn ^^
operating data that are typically coded on each basic point source record.
> Consider an operation with annual emissions of 20 tons of VOC, with A0
percent of annual throughput occurring in the summer. This source norrr,?,ii/
operates 12 hours per day and seven days each week. Assuming uniform
hourly emissions over a 13-week summer, the emissions rate is estimated tc
be (20 x 0.4)/(12 x 7 x 13) or 0.0073 tons per hour. Applying the conversion
factor, 907 kg/ton, gives 6.7 kg/hr as the average emissions during summer
operations. In the absence of more specific data, these emissions might be
assigned to the period 0700 to 1900 each day.
The tlFAM EPS can perform these ^temporal adjustments automatically based bn
' 'the operating data contained*- in 'tt>e"&iai^c1nventc^r"Tri& 'I^EF?NT'pr6§rd^ -;T ..... '
converts seasonal fractions of throughput to a monthly variation profile' ;
(assuming uniform variation- throughout the season) and assigns default siay-of-
week and diurnal activity profiles based on the number of -. days per week and
the number of hours per day in operation. :.
".'.;.. tons/year}J/{4S weeks in
-------
.possible overestimatlon rather than underestimation of emissions^ thai
program assumes that any source not operating for. the entires year will be In .
.operation during the. modeling epaode,; This assumption; of i.coitrSB^-wJfl have
• no effect for tho^. sources which indicate'.' no 'activity' during 'the season
month} to fee modeled, -..:.; ,
• Note that if seasonal rather than annual emission
•emissions modeler shodd take care thst:f8dundant\^&sori«if;adl6strA'ents are
not applied. This is readily accomplished by preprocessing the existing point
source inventory, setting ail seasonal-. fractions to 25% and the number of
/weeks m operation per vsar to 52 to ensure that the EPS does not further
adjust emissions which 2?? airsady on a seasonal basis. , •
For many sources, daily operation will be confined to one or two workshifts; thus, hourly
operation during working hours would be determined by dividing the daily operating rate by
8 or 16. If hourly operating information is contained in local agency files, it may supersede
the less detailed information contained in the basic point source inventory.
Often, no operating data will be coded in the existing point source inventory for some
sources. For those sources which are too minor to warrant directly contacting the facility,
engineering judgement can often provide satisfactory estimates of hourly emissions.
> Many commercial establishments will operate all year, but only be open 8 to
10 hours a day and 5 to 6 days a week. Hence, as a good approximation, the
annual operating rate in the basic point source record for these sources can be
divided by 2,080 (i.e., 52 x 5 x 8) to estimate an hourly operating rate
applicable during working hours.
;: . The UAM EPS assigns a flat operating profile (i.e., e
-------
sunshine in the morning, and cease expelling vapors in the mid-afternoon when the heating
process ceases. As an approximation, breathing loss emissions from fixed roof storage
tanks can be assumed to occur uniformly from 8:00 a.m. to 3:00 p.m. (this information
can easily be incorporated into the modeling inventory using the day-specific MERF shown
in Table 5-3). Daily emissions from storage tanks can be estimated using procedures given
in AP-42, Compilation of Air Pollutant Emission Factors.5
Tables S-4"and-' 5-8-"show tile default wseldy and"d?urna!"^'ctivityprofile'codes
provided with the UAM EPS, the values in these tables represent relative •
levels of acth/tty by how of day and day of week, respectively. These temporal
variation codas ars assign ad in PREPNT (for point sources} and GRDEMS (for
area and mobife sources}, Ths weekly profiles listed in Table S-4 are used In
conjunction with monthly activity fractions to compute representative episodic
emissions levels from tha annual (or ozone-season} average emissions '•.
contained in the existing inventory; this calculation is similar to that described
previously for adjusting emissions based on number of weeks, per year in
operation and can be mathematically represented as ,. .
E» » i^'IMp'ta months/yr)-K7days/wk total)/D,| _ :; (5-1)
.where E^, denotes episodic daily emissions, JE^ denotes annual average daily
emissions, M$ is the fraction of annual activity occurring in the episode month,
and i>» is. a 'day:o'f week adjustment factor obtained &y dividing thevalue inlP^..'....:Z'.;;
Table 5-4 for:the :day:bf week of ;the episode by the totaf for the day.'-:/ -•Ji;-;::- ?• -.-
For example^ if the weekly variation in activity for a particular point source
••;'•.' having annual average emissions of 10 tons/day of VOC and no seasonal
variation is be characterised by profiJe code 22, representative daily emissions
. for a Tuesday are calculated to be (10 tons/day VQC* x (0.083 x 12) x {7 / (10
/. S1)J, or 11.S tons/day VOC. Diurnal variations in emission^ are accounted for
. .. by multiplying the episodic daily emissions calculated above^ by the fraction'of
:; total daily activity occurring in & 0iyen hour; if 1^6 diurnal vaFijattori for the p>c«F>t
-/'-• source described above.was: characterized by profile cod^'SSifrom fable 5-5;::
':,- '''. the fraction cf: episodic daily emissions occurring in^the hour between 7'•"&n& 8
'•r';'",,;a,rn,rwould; bd'•'{$• t 182i or9,B percent (corresporjdirtg^t&'-:i;t- tons VOO;":-\v"'..
./If none of the existing profiles listed in Tables 5-4 and. S-S.maJtcH the typical : i
:: temporal variations for a given source or type of source,.additional;profiles Can
:: be created and added to tne appropriate files. Alternatively, existing; profile .
. .;... codes can be redefined to reflect, the desired temporal distributions. Appendix
: S, provides additional guidance on creation of new diurnal and weekday profile
. •; codes and redefinition of existing codes, •" . > : • ' '' •" • ' • "". '\' •.
9009805" 5-10
-------
TABLE 5-4. Default weekday variation code." ':*->•.! >n J,hs emission'? ~
Code
1
2
3
4
5
6
7
21
22
23
Relative Emissions Contribution by Day of W-sek
MON
1
0
1
1
1
1
1
1
10
5
TUE
1
0
1
1
1
1
1
1
10
5
WED
1
0
1
1
1
1
1
1
10
5
THU
1
0
1
1 -
' 'S*
1
1
1
1
10
5
FR!
1
0
1
1
1
1
1
1
10
5
SAT
C
1
0
0
0
1
1
2
7
4
SUM
i Otc.i
'-^ O •;
1
0
0
0
0
1
2
4
4
2
5
5
5
6
7
9
61
33
source: Reference 7
90098 05"
5-11
-------
o
V)
X
to
o
V)
(A
0)
O
o
Q.
w
CL
C
'55
(A
1
V
•c
V)
3
(0
0]
o
5. Default diurnal variation c
in
w
m
1-
- f
i- a
«
2
I
Dudon b
s
u
1
1
S
2-
"5
o
10
«
o
o
M
* y
ooooocooooooo-'-------- ---r'r
-„
oocoooooooooooooooo*- ^-^^-^2
OOOOO5>OOOOOOOOOOOOO"-'-'-I-«-'-N
OOOOOOOOOOOOOOOOOOO — -~^ — — --M
OOOOOOOOOOOOOOOOOO o«->-«- — •-•-"
ooooooooooooooooooo — — *- — — — N
OOOOOOOOOOOOOOOOOOO — •-•- — - — w
OOOOOOOOOOOOOOOOOOO- — " — — OW
fi
" 1
)
I)
f
o
o
.
5-12
-------
5. Concluded.
in
Ol
oa
H
4i
" " T
3
9
o
1
>
5
1
1
i
i
i
X
3
1
4
4
y
a
(A
0
in
IN
o
i
! 0
[
o
o
o
o
o
n
o
o
-
o
A
O
o
o
n
O
n
o
n
0
o
o
-
0
o
o
o
00
o
o
CO
0
-
o
0
o
o
o o o *
0 0 0 3
0005
o o o 2
o o o 2
0 0 0 g
* « o P*
n o o *
O O O
-------
5.5 POINT SOURCE PROJECTIONS
As discussed in Chapter 2, emission projections must account for anticipated growth in
activity levels as well as the effects of any regulations under consideration to control
ozone precursor or CO emissions (be sure to also account for rule effectiveness in
emission projections, as discussed in Section 5.2). The baseline projection inventory
should accordingly include the effects of expected growth in future years and the
reduction in emissions that should occur as a result of existing control regulations. Control
strategy projections, on the other hand, must also consider the reductions in emissions
that would occur if alternative or additional regulatory programs were adopted. Control
strategy projections may also take into account other-than-expected growth patterns
which might result from the alternate control programs. This section presents various
methods for projecting point source emissions; the emissions modeler, however, should
consult the EPA guidance document concerning emission inventory projection techniques
(to be released in July, 1991) for definitive guidance on projection of point source
emissions.
5.5.1 Individual Facility Projections
The most rigorous approach for projecting emissions from major point sources is to obtain
information on each facility. Ideally, this type of information would be determined by
contacting plant management directly or could be'solicited on questionnaires. Generally,
questionnaires would not be sent out solely to solicit projection information; however, this
additional information may be solicited on questionnaires used to periodically update the
baseline inventory. Permit applications submitted to various Federal, State, and local
agencies should also be screened to get information on expected expansion or new
construction. The local Metropolitan Planning Organization (MPO) and other planning
bodies should also be contacted for any information they may have on projected industrial
expansion as well as to comment on the reasonableness of any plans submitted by plant
personnel.
Once the agency responsible for the modeling inventory obtains this type of projected
plant growth information, it needs to determine what regulations will apply, in order to
estimate controlled emissions. Appendix C of Volume I3 summarizes the EPA Control
Technique Guideline (CTG) documents. The baseline projection should incorporate any
existing applicable regulations.
* A fossil-fueled power plant now under construction and expected to start
operation in 2 years would be subject to Federal new source performance
standards for particulates, S02, and NOX. Hence, it would be reasonable to
assume emission levels equal to the standards unless plant personnel indicate
more stringent controls will be applied for some reason (e.g., to meet a more
90098 OS'2 5-14
-------
stringent local standard). Similarly, in control strategy projections, effects of
any alternative standards would have to be evaluated.
When obtaining projection information from plant management, it is important to inquire
whether projected increases in activity will occur at the existing facility or elsewhere (i.e.,
at another existing plant or at a new plant), if occurring at an existing facility, the
emissions modeler also needs to determine whether the growth will be expansion to
existing capacity or will require additional capacity. These considerations are especially
important for major sources, since emissions must be assigned to a specific grid cell. This
information will also help to determine what additional control measures are likely to be
required. The schedule for completion of any expansion or new construction is also
needed, in order to determine in what year the source must be included in the projection
inventory.
> Consider a facility employing a large open-top vapor degreasing operation that
emitted 100 tons of solvent per year in 1987 (based on an annual production
of 10,000 of a certain kind of metal part.) Assume that no control measures
were taken to reduce solvent losses from the process. Now, suppose plant
contacts indicate that 5 percent more metal parts would be produced per year
until 1992 using the existing operation. Then, in order to estimate VOC
emissions from this source for a 1992 projection inventory, one could assume
that since no additional controls would be expected, the current emission level
could be multiplied by the ratre of cumulative growth in metal parts production
(i.e., 5 years at 5 percent/year = [1.05]5 = 1.28, or 128 percent) to estimate
VOC emissions in 1992. In this manner, emissions in 1992 would be
estimated at 128 percent of 100, or 128 tons per year, and the point source
record for this projection year should be adjusted accordingly to take this
growth into account.
As is obvious from this example, even when projection information is available for specific
facilities, certain assumptions will have to be made in order to project emission levels for
some future year. For instance, in the 1992 baseline projection, it was assumed that
emissions would increase proportionately with production. Depending on the nature of the
operation, this may not necessarily be entirely accurate. This underscores the point made
in Section 2.4 that projections are always somewhat speculative in nature.
5.5.2 Aggregate Point Source Projections
In many instances, projection information will not be available for every facility in an area
of interest. Some facility managements will not be willing or able to provide forecasts of
their corporate plans, especially for more distant years. In addition, many plants in certain
source categories (e.g., small industrial boilers) will be too small and too numerous to
30098 05rJ 5-15
-------
warrant the solicitation of projection information individually. In these situations, other
procedures need to be employed to make projections of future emissions. Two possible
approaches are discussed below; in all cases, however, the emissions modeler should
consult current EPA guidance for inventory projection to determine the most up-to-date
databases and applicable parameters for aggregate projection of future-year inventories.
!n one case, projection information may be available on many point sources within a given
category, but for various reasons is not obtainable for one or several facilities. In this
situation, a reasonable approach to projecting growth and emissions at the remaining
facilities would be to evaluate the growth trends in the facilities for which projections are
known and apply them to the facilities for which no information was available.
* Suppose there are 10 paint manufacturing facilities in the area of interest, and
successful contacts may have been made with only eight of these. If
production was expected to expand by 6 percent per year, on average, for the
eight plants, then this rate ccuid be applied to the remaining two plants to
estimate expected growth. Then, knowing the increase in production, the
appropriate control measures would be taken into consideration in making a
baseline projection. In some cases, the emissions could be directly estimated
by applying the average growth rate to a base year emission for each facility.
Good engineering judgment is needed in this practice to screen out any unreasonable
projections that may result. '"
For minor point sources, such as cold cleaning operations, where individual solicitation of
projection information is unwarranted, the rate of growth of activity may be assumed
equivalent to that of some growth indicator category for which projections have been
made. Sources of growth indicator projections include local MPO's and the U.S.
Department of Commerce's Bureau of Economic Analysis (BEA).8 For example, it might be
assumed that cold cleaning operations would grow at the same rate as industrial
manufacturing in general. This rate can be readily estimated from projections of
employment in industrial manufacturing categories. Table 5-6 lists the categories for
which the BEA makes projections at the state and Metropolitan Statistical Area (MSA)
level, along with the two-digit SIC designations associated with each category (in addition
to the state- and MSA-level projections, BEA also publishes projections for BEA Economic
Areas, which are larger than MSAs). Note that MSA-level projections are not available for
most two-digit SIC designations. Table 5-7 shows an example of these projections for the
state of California. The BEA projections of industrial employment are regularly updated
and may be used, in the absence of local projections, as general indicators of growth.
: .... Jrvtfte'UAM EPS, projected growth factors are applied by two-digit SIC. .; :
designation These growth factors are expressed as ratios of future year to
30098 05" 5-16
-------
TABLE 5-6. Industrial groupings for BEA economic projections.
» prcjactK)
ktduntries projactsd for Ziatm and ttw Haitian
1872 SIC coda'
Farm
Agricultural sarvicw, forestry, fisheris*. ml otiwr
Agricultural Mrvioai, forestry, fiaharie*. md othor
Agricultural s«^vic«s, forestry, and fishone*
OthoH
01, 02
07, 08, 09
Mining
CcaJ rrjning
Oil srd gas ax-traction
Matai mtning
Nonmatad pufaoc ut5tia*
Railroad transportation
Trucking and warehousing
Local, suburban, and highway passenger
transportation
Air transportation
Pipeline transportation
Transportation services
Water transportation
Communication
Electric, gas and sanitary services
23
26
28
29
30
31
24
25
32
33
34
35
36
37 except 371
371
38
39
40
42
41
45
48
47
44
48
49
continued
90098 05"
5-17
-------
TABLE 5-6. Concluded.
hidufftri** profaetsd for MSA*
Whotesaia trad*
Jtaal tr*d«
nuno*. kttuww.1. *m$ vwri ^t*Sa
iTMIUAtfMM pfOftfCtKI (Of StltsM 4Otj tTMl nurtlitn
Whot«M»e tnd*
Rfltsd tnde)
Firwrwa. imurmcai, end raal altMe
Banking
Other credit and securities agencies
Insurance
Rss! sitata afid combination offices
1 972 SIC code1
50, 51
S2.-53
60
91, 62 97
64, 94
?5 SS
Goverrtm-frrt srwj government «ntarpm«»
Fsderal, civilian
Federal, military
State and local
HoteJ* and oihaf lodging places
Psraorai, business, and miscellaneous rooair
services
Automotive rsosir services »rd gj'^23
Amusemant and recreation sorvices
Motion picturet
Private households
Health lorvicas
Private educational sarvces
Nonprofit organizations
Wiscallaoeous protemonat sarvicas
Government wxJ government anterpfwxa
Federat, civilian
Fadarsl, rmlitarv
Stata and local
70
72, 73, 78
75
79
79
33
30
32
83, 84, 86
81, 89
' Historical data through 1974 are classified according to the 1 987 SIC definitions; subsequent historical data and projections are classified according
to the 1972 SIC definitions.
' Refers to United States residents employed by international organizations.
J The ordnance classification was discontinued in tha 1972 SIC definitions. Earnings and amploymant previously included in ordnance are now
'nck'd^d in one or more of tha following classes: fabricated metal products (SIC 34); electric and electronic equipment (SIC 38); transportation
equ.oment. exceot motor vehicles (SIC 37, except 371); and instruments and related products (SIC 38).
source: Reference 6
concluded
Note that projections at the MSA level are only available for broad industrial classifications
which include numerous two-digit SIC designations (e.g., the MSA-leve! projections for
Mining include aggregated estimates for SICs 10, 11, 12, 13, and 14); by contrast,
separate projections are provided for these subcategorizations at the state and national
levels.
90098 OS"
5-18
-------
TABLE 5-7. Employment by place of work (thousands of jobs), historical years 1973-1988
and projected years 1995-2040, for California (excerpt).
Manufacturing .
Nondurable good*
Food and kindred products
Chemicals and allied
Petroleum and coal
product*
Rubber and mwcellaneous
plastic product* ...
Primary matal mduatras . .
Electric and electron*:
Motor vehicles and
Stone, clay, and glass
Transportation and pubhc utilities . . .
Ratt>oad Transportation . . ....
Local and mterurban passenger
Electric, gas, and sanitary
services
Wholesale trade
Retail trade
1973
1988.5
5577
170 8
55.6
245
52.5
1128 8
59.7
291 3
42.2
57 1
499.8
38.1
24.7
87.7
4734
1504.8
1979
20679
867.7
190.7
95 8
26.7
a95
1400.2
59.8
321 8
29 9
82.8
579.5
32.8
29 2
705
908.4
19578
1983
20128
851.6
179.9
64 3
30.9
82.4
1381 1
43^
368 0
30.4
54.1
589.5
24.8
28.5
78.3
936 0
2094 7
1988
2237.4
741.8
179.4
78.3
28.1
73.9
1495.8
44.0
390 8
35.8
63.8
862.2
18.8
36 8
893
777.1
2486 7
1985
2332.8
793.5
182.3
79.0
279
81.7
1539.3
43.1
397 1
33.8
68.9
736.0
15.3
41 4
98.4
868.0
2833.0
2OOO
2394.6
826 7
183.8
8O 8
28.0
86. 8
1588.0
430
397 5
32 6
68 5
779.8
13.9
44 3
10S 3
9200
3071.8
2009
2424.9
842.3
182.1
81 5
278
89.8
1582 5
42.5
398 0
31 5
69 7
8075
12.8
45 9
109.3
950 9
3234.4
2O10
24350
8480
1793
81 3
27.1
91 8
1587.1
41.7
398 3
30 8
70 7
825.2
12.1
111.7
992 3
3331.3
2020
2352.7
821 7
188 3
77 9
25 4
905
1531 1
39 1
383 0
28 2
89 1
818 8
10 7
110 5
995 4
333O.O
20*0
2222 8
778 7
153 2
72 6
22 9
880
1443 9
35.2
25 0
7972
8 8
107 7
3293.1
source: Reference 6
90098 05'1
5-19
-------
,y;' Ibaseiyear aci3V4ty levels mdtcators {e.g//numbef of :emp
' .^separate set of growth factors wiil be required lor.each projection year.
Regardless of the indicators used for projections, the basic mechanics of projecting the
emission inventory are the same: the ratio of the value of the surrogate indicator in the
projection year to its value in the base year is multiplied by the aggregated activity level for
the point source category in the base year.
>• For example, if pharmaceutical manufacturing operations in the San Francisco
Bay Area are assumed to correlate with the chemicals and allied products
manufacturing sector in Table 5-7, then the level of this activity in 1995
would be 79.0/76.3, or 1.04 times that in 1988.
In many cases the projection years of interest to the air pollution control agency will not
directly correspond to the years for which growth projections have been made, thus
requiring interpolation of the growth indicators. Consult local authorities to determine if
straight-line or some other interpolation method should be employed.
Once aggregate growth has been determined for a source category in the above manner,
the increased activity must be allocated appropriately to the grid cell level. Often, difficult
assumptions must be made regarding the probable location of the new activity. One way
to apportion growth is to assume that it occurs only at existing facilities in the same
source category.
> If the 10 paint manufacturing facilities in the previous example manufactured
10 million gallons of paint in 1987, and 15 million gallons were projected in
1992, then the additional 5 million gallons could be assumed to be
manufactured at these same facilities. The amount of production assigned to
each facility would be proportional to the quantity currently manufactured
there.
If it seems unreasonable to assume that all growth within a particular source category will
occur at existing facilities, an alternative is to apportion the growth according to the
fraction of increase of industrially zoned land within each grid.
* Using the above example again, if 5 percent of the projected area-wide
increase in industrially zoned land occurred within a given grid cell, then
250,000 gallons per year (or 5 percent of 5 million gallons) of additional paint
production would be assigned to that cell.
90098 OS'2 5-20
-------
in this approach, since the growth is not assigned to existing facilities, hypothetical point
source records will have to be created and added to the inventory. Note that all
information contained in a point source record (including process identification codes and
stack parameters) will need to be estimated for each hypothetical point source. Again,
existing applicable control regulations must be evaluated in order to determine the baseline
projected emissions.
5.5,3 Control Strategy Projections
A control strategy projection is an estimate of emissions for some future year which
considers the effect of proposed control measures. Control strategy projections should be
made for the same years as the baseline projections to facilitate comparison of the relative
effects of each strategy as well as to determine which strategy provides the necessary
control of ozone precursor emissions.
In order to evaluate the relative merits of various control measures, the agency will often
need to develop several different control strategy inventories for each of the projection
years. Basically, a control strategy projection is generated by applying anticipated control
factors to the baseline projected emissions from sources affected by the proposed
measures. Obviously, the first step in this process is to identify all affected sources for
each measure under consideration. Then, the anticipated emissions reduction associated
with each measure (usually expressed as gf percent reduction) is applied on a source-by-
source basis. This procedure is not particularly difficult, but the large number of
calculations required (particularly if several control strategies are under consideration) can
generally be performed more efficiently with the aid of a computer.
• The UAM EPS program CENTEMS applies user-specified reduction factors by
: control category designation. For each control code, a separate control factor
must be supplied for each of the pollutants Included In ^the Modeling E-missJon
• Record Format {THC, N0yr CO, SO..,, and PMK The control categories as they
:. are currently defined are listed In Table A-3 of Appendix A. Each unique
:SiC/SCC pairing Is assigned a specific control code -in-trie category glossary .
i Input ftle to the CENTEMS program; the user should review this file to ensure
.;.. that allapplicable SIC/SCC pairings are included; ••.-•••:•. •.••.•-.••<••••• _:.............. •
5.5.4 Point Source Projection Review and Documentation
Because the projection inventories are so important to control strategy development and
evaluation, they should be reviewed internally by the air pollution control agency and
90098 05'1 5-21
-------
presented to as many other groups as possible for comment before being finalized. A!i
assumptions, procedures, and data sources must be carefully documented. Thorough
review and documentation helps ensure thatthe projections are (1) consistent with other
projections being made by various groups in the area, (2) objective in the sense that they
are not biased in order to promote a particular policy, (3) open, because a!! assumptions,
etc., are clearly stated for public review, and (4) defensible, because of a!! the above
characteristics.
Three key aspects of point-source projections wi!! invite criticism:
(1) the choice of indicators for projecting activity level growth;
(2) when and where this growth will occur, and concomitantly, whether it will ie
accommodated by expansion of existing facilities or new construction; and
(3) what emissions will be associated with this growth, either in the base'ine case
or as a result of various candidate control strategies.
When planning, compiling, and reviewing the point source projection inventory, the agency
should focus particular attention on these issues.
5.6 DATA HANDLING CONSIDERATIONS
As mentioned in Section 5.3, either a manual or computerized approach can be utilized to
assign point source emissions to specific grid cells. Generally, unless there are very few
point sources in the modeling area, a computerized assignment proves more practical. In
this approach, a computer program compares the UTM (or other) coordinates stored in
each point source record with the coordinates of the grid cells and determines in which
specific grid cell the point is located. The appropriate grid cell identifier (or coordinates)
can either be stored in the point source record (if space is available) or in a separate
correspondence file. Subsequently, any time a model-compatible inventory is generated,
this point-to-grid-cell correspondence information can be accessed to assign point source
emissions to grid cells.
. . in tfoe Model Smteaiom Record' Format employed by-'the tfANl BPS. jTabffc 3-1},-•"'_,
•••:•::;. the poJnt-to-§Fid-ce# correspo*idenci & stored m thft emissions record for each
pokfl: source, the program PflEPNT makes this assfgrimiertf, :J ; :,. •..; I '•'._'•:
Although it would be possible to generate the point-to-grid-cell correspondence data during
the creation of each model-compatible inventory, this method would have the
90098 05" 5-22
-------
disadvantage of requiring the coordinate comparison step to be repeated for every model-
compatible inventory created.
If the point-to-grid-cell assignment step is performed manually using the techniques
described in Section 5.3, the resulting correspondence data will still have to be
incorporated into either the point source records or a machine-readable correspondence
file, as described previously, in order to be utilized by the programs that create the model-
compatible inventory. The manual approach, as mentioned previously, has the
disadvantage of being much more time-consuming if numerous point source assignments
are necessary.
SosPc ;:i%-':,-- ,;: v. . ,° -,, .'•*•• oo no: require that eievated ooint sources be assigned to gr1 ;
celi.3 'I-, "."•' •' - -' -;;"—"-", -r.'."* rvjrce locations are identified by UTM coordinates}, in
these mccisis, cracrccasacr programs are available that make this assignment based on the
point _--. .-.-;;:-• - •;.-. ; : r ':•; - ,:r:~ :r~ ann-jal, county-ieve! inventory. If this is tre
case, pcinc-to-grid-cai: correspondences need net be determined for these particular
sources. GQ°er?!lv, hO'*'°v?r, s'"c-2 tr.3 modslars may net know in advance which sourcso
will be considered as elevated, and since computerized assignments wiil be practiced in
most instances, little extra effort will be expended in simply making this assignment for a'i
point sources. Thus, this information wiil always be available in case it is needed at a iatsr
date.
The hour-by-hour point source emissions required by the photochemical model are
estimated by applying seasonal, daily, and hourly operating factors to the annual
emissions, as discussed in Section 5.4. In the data handling system, the temporal factors
and the resulting hourly emissions can be stored either in the individual point source
records (if spaca is avaiiabie) or in a separate file created for this purpose. A potential
disadvantage of storing hourly emissions on each point source record is that a great deal
of file spaca is required. An alternative is to have the program that creates the modeling
inventory compute hourly emissions at the point source level but accumulate hourly
emissions at the grid cell lavel.
i In the UAM EPS> temporal distribution data are stored In eaeri point'source data
: record either: as profile codes (see Tables :§-4 and $-$> pr. as ho.ydy distribution:
......;. .factors.'.{u$fng..the; day-specific Modeling Emissions Record Format sfcown In.
. I Table S-3;, Temporal, adjustments .{except for the-, weeks/yosr adjustment
; discussed In Section 5«31 are not actually! applied to the; point source emissions
^; .I. until the UAM '6PS; program; C£NT£MS,-i C£MTEMSi also bombfnes^ll |dw-idvel
{emissions. #,«», emissions from sources not selected foretev^ted treatmentj for
J each grid celt; by Carbon Bond species, reducing file spsce/equlrements for the .
. ;; flnel modeling Inventory.. =: ... : . : : : . ... ;.
90098 05'2 5-23
-------
Note that when emissions from different sources are combined, the VOC and NOX splits
used to allocate these emissions to chemical classes must be applied before the emissions
are summed for each grid cell in order to maintain the pollutant split identity of each
source category. Hourly emissions and pollutant splits must be recomputed each time a
modeling inventory is created. An example of a file containing temporal factors for
individual sources is shown in Table 5-8. The entries for such a file, which are used to
estimate hourly emission rates from annual emissions, are determined using the procedures
outlined in Section 5.4. Different temporal patterns can be simulated (e.g., in projection
inventories) by simply changing the factors in this file to reflect anticipated operating rate
changes.
: •:. '•, . ,\-i;•,... ^,"-.--".3 ;c( u~,a sp3cia:,on data used to allocate VOC and NCX
•rs T-T- c"""~;c~J ~;^~~'~" ?ra discussed in Chanter 9.
90098 05" 5-24
-------
TABLE 5-8. Example temporal factor file for individual point sourcss (excerpt).
Source
SCO" Ft"
30200000 02
30200000 02
30200000 02
30200000 02
30200000 02
30200000 02
30200000 03
30200000 03
30200000 03
30200000 03
30200000 03
30200000 03
Temporal Factors0
1.43
2.29
2.78
5.23
5.23
1.76
2.99
2.99
5.77
4.28
2.29
3.21
5.23
5.23
2.99
3.63
5.77
4.28
27.0
2.29
3.92
5.23
5.23
32.0
"2.99
4.28
5.77
.4.28
2.29
4.58
5.23
5.23
2.99
4.28
5.77
4.28
2.29
5.23
5.23
5.23
2.99
4.28
5.77
3.63
2.29
— • *•* j~ i
o.^3
5.23
3.76
2.99
5.00
5.00
2.99
Code4
s i|
0
1H
2H
I
3H ;
4H
S
D
1H
2H
3H
4H
8 Plant identification by SCC code (eight digits).
b Point source identification by number within plant. (Note that state, county and
plant IDs are not shown and should be included for complete identification.) Six
successive lines constitute one point source record.
c Definitions: Seasonal, percentage of annual activity occurring during chosen
quarter-year; daily, percentage of seasonal activity occurring on selected day;
hourly, percentage of daily activity occurring on selected hour. Six consecutive
hourly values appear on one line.
d Code: S, seasonal; D, daily; 1H, hourly, 0001 to 0600; 2H, 0601 to 1200; 3H,
1201 to 1800; 4H, 1801 to 2400.
90098 05"
5-25
-------
References for Chapter 5
1 . The 1985 NAPAPJEmissions Inventory (Version 2)i Development _of .the Annual Data
and Modeler's Tapes. EPA-600/7-89-G12a, Air and Energy Engineering Research
Laboratory, U.S. Environmental Protection Agency, Research Triangle Park, NC,
1989.
2. Love, R. A., and Mann, C. 0., The Use of the AIRS Facility Subsystem for the
Management of Emissions Inventory Data, presented at the 83rd Annual Meeting &
Exhibition of the Air & Waste Management Association, Pittsburgh, PA, June 1 9;}Q.
3. Procedures for the_Preparation of Emission Inventor]"^ __c^'^"j"~i'_^2^T" P.f-..P^""?.-
Volume I. EPA-450/4-88-Q21, U.S. Environmental Protection Agency (OAQPS).
December 1988.
4. Breathing Loss Emissions from Fixed-Ro_gf Petrochemical Storage Tanks (Draft', E-A
Contract No. 68-02-2815, Work Assignment No. 6, Engineering-Science, Inc., Juy
1978.
5. Compilation of Air_Pollution Emission Factors, Fourth Edition and Supplements, A"-
42, U.S. Environmental Protection Agency, September 1985.
6. BEA Regional Projections to 2040, Volume 'f: States. U.S. Department of Commerce,
Bureau of Economic Analysis, June 1990.
7. User's Guide for the Urban Airshed Model, Volume IV: User's Guide for the
Emissions Preprocessor System, EPA-450/4-90-007D, U.S. Environmental Protection
Agency (OAQPS), Research Triangle Park, NC, June 1990.
90098 OS'2 5-26
-------
6* AREA SOURCES
6.1 GENERAL
The emissions modeler can often use an existing, county-level area source emission
inventory as the basis for the modeling inventory, with the possible exception of emissions
from mobile sources (which are discussed separately in Chapter 7), especially if ths
county-level inventor1/ was prepared >n accordance with the guidelines given in Procs'j'.jria
for the Preparation of Emission Inventories for Carbon Monoxide and Precursors of Ozone;
Volume I.1 The existing inventory will usually contain collective emissions estimates at the
county level for those sources considered too minor and/or too numerous to ba handled
individually in the point source inventory. In addition to smai! stationary sources, the
county-level area source inventory often includes emissions from off-highway mobile
sources, such as aircraft, locomotives, and off-road vehicles. As an example of a source
classification scheme used to identify types of sources 4n the inventory, Table 6-1 lists the
area source category designations currently used in the NAPAP inventory along with the
photochemically reactive pollutants commonly inventoried for each category.
Table; 6-2: contains additional area source: category Designations which the UAM-.
£PS itsestojd&tth'jjuisri' between the exhaust, evaporative; refueling/and .. . .
running, loss components o£ motor vehicle emissions (this distinction, rrust be
maintained for accurate speciatlon of VOC emissions from these sources Into
chemical classes). Chapter; 7 discusses the use of these additional categories.
The area source categories listed in Table 6-1 include all of the area sources addressed in
Volume I, although Volume I separates several of these source categories into
subcategories, as noted in Table 6-3. If the county-level area source inventory includes
estimates for the subcategories shown in this table, the emissions modeler may wish to
maintain this distinction in the modeling inventory, especially if detailed local data are
available for these sources.
Some of the source categories listed in Table 6-1 may be treated as point sources in the
existing inventory; other source categories may be represented in both the point and area
source inventories, depending on the emissions cutoff level used to make this distinction.
Likewise, a number of other source categories traditionally inventoried as point sources
may, at least in part, be treated as area sources (e.g., industrial fuel use, industrial surface
coating, and gasoline bulk tanks). The emissions modeler should be aware of all such
90098 Off1 Q. 1
-------
TABLE 6-1. NAPAP area source categories and Inventoried ozone precursor pollutants.
Code Description
Stationary Source Fuel U«*
1 Residential Fuel - Anthracite Coal
2 ResidentiaJ Fuel - Bituminous Coal
3 Residential Fual - Di'stillata Oil
4 Rasidentiai Fuel - Residual Oil
5 Rasidentiai Fual - Natural Gas
5 Residential Fuel - Wood
7 Commercial/Institutional Fuel - Anthracite Coal
8 Commercial/Institutional Fual - Bituminous Coal
9 Commercial/Institutional Fuel - Distillate Oil
10 Commercial/Institutional Fuel - Residual Oil
1 1 Commercial/Institutional Fuel - Natural Gas
12 Commercial/institutional Fuel - Wood
13 Industrial Fuel - Anthracite Coal
14 industrial Fuel - Bituminous Coal
15 Industrial Fuel - Coke
16 Industrial Fuel - Distillate Oil
17 Industrial Fual - Rasidual Oil
18 Industrial Fuel - Natural Gas
19 Industrial Fuel - Wood
20 Industrial Fuel - Process Gas
96 Minor Point Sources - Coal Boilers
97 Minor Point Sources - Oil Boilers ""
98 Minor Point Sources - Gas Boilers
99 Minor Point Sources - Other
Solid Waste Disposal
21 On-Sit6 Incineration - Residential
22 On-Site Incineration - Industrial
23 On-Sita Incineration - Commercial/Institutional
24 Open Burning - Residential
25 Open Burning - Industrial
26 Open Burning - Commercial/Institutional
Highway Mobile Sources
27 Light-Duty Gasoline Vehicles - Limited Access Roads
28 Light-Duty Gasoline Vehicles - Rural Roads
29 Light-Duty Gasoline Vehicles - Suburban Roads
30 Light-Duty Gasoline Vehicles - Urban Roads
31 Medium-Duty Gasoline Vehicles - Limited Access Roads
32 Medium-Duty Gasoline Vehicles - Rural Roads
33 Medium-Duty Gasoline Vehicles - Suburban Roads
34 Medium-Duty Gasoline Vehicles - Urban Roads
35 Heavy-Duty Gasoline Vehicles - Limited Access Roads
36 Heavy-Duty Gasoline Vehicles - Rural Roads
37 Heavy-Duty Gasoline Vehicles - Suburban Roads
38 Heavy-Duty Gasoline Vehicles - Urban Roads
voc
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
CO
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
NO,
X
X
X
X
X
X
X
X
X
X
X
x
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
continued
90098 OS"
6-2
-------
TABLE 6-1. Continued.
Coda Description
40 Heavy-Duty Diesel Vehicles - Limited Access Roads
41 Heavy-Duty Diese! Vehicles - Rural Roads
42 - Heavy-Duty Diesel Vehicles - Suburban Roads
43 Heavy-Duty Diesel Vehicles - Urban Roads
Nortiiighway Mobii« Sourcua
39 Off Highway Gasoline Vehicles
44 Off Highway Diesel Vehicles
45 Railroad Locomotives
46 Aircraft Landings and Takeoffs - Military
47 Aircraft Landings and Takeoffs - Civil
48 Aircraft Landings and Takeoffs - Commercial
49 Vessels - Coel
50 Vessels - Diesel Oil
51 Vessels - Residual Oil
52 Vessels - Gasoline
Other Combustion Sources
60 Forest Wild Fires
61 Managed Burning - Prescribed
62 Agricultural Field Burning
63 Frost Control - Orchard Heaters
64 Structural Fires
VOC EVAPORATIVE SOURCES
Gasofin* Distribution
54 Gasoline Marketed
103 Bulk Terminals and Bulk Plants
Stationary Sourc* Solvent Us*
78 Degreasing
79 Dry Cleaning
80 Graphic Arts/Printing
81 Rubber and Plastics Manufacture
82 Architectural Coatings
83 Auto Body Repair
84 Motor Vehicle Manufacture
85 Paper Coating
86 Fabricated Metals
87 Machinery Manufacture
88 Furniture Manufacture
89 Flatwood Products
9O Other Transportation Equipment Manufacture
91 Electrical Equipment Manufacture
92 Shipbuilding and Repairing
93 Miscellaneous Industrial Manufacture
95" Miscellaneous Solvent Use
VOC
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
CO
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
NO.
.X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
continued
90098 08" 6-3
-------
TABLE 6-1. Concluded.
Coda Description VOC CO NO,
101 Cutback Asphalt Paving Operation X
102 Fugitives from Synthetic Organic Chsrnical Manufacture X
104 Fugitives from Petroleum Refinery Operations X
105 Process Emissions from Bakeries X
106 Process Emissions from Pharmaceutical Manufacture X
107 Process Emissions from Synthetic Fibers Manufacture X
108 Crude Oil and Gas Production Fields X
Wasta ManaQamant Prscticss
100 Publicly Owned Treatment Works (POTWs) X
109 Hazardous Waste Treatment, Storage, and Disposal X
PARTICIPATE AND AMMONIA SOURCES
55 Unpaved Road Travel
56 Unpaved Airport LTOs
66 Ammonia Emissions - Light-Duty Gasoline Vehicles
67 Ammonia Emissions - Heavy-Duty Gasoline Vehicles
68 Ammonia Emissions - Heavy-Duty Diesel Vehicles
69° Livestock Waste Management - Turkeys
70° Livestock Waste Management • Sheep
71° Livestock Waste Management - Beef Cattle
72C Livestock Waste Management - Dairy Cattle
73' Livestock Waste Management • Swine
74° Livestock Waste Management - Broilers ""'"
75° Livestock Waste Management - Other Chickens
76 Anhydrous Ammonia Fertilizer Application
77 Beef Cattle Feed Lots
• Category 63 is disaggregated into process categories 78 to 95.
" Formerly "miscellaneous industrial solvent use" (94) and "miscellaneous non-industrial solvent use" (95); now
combined into one category (95).
° These categories formerly referred to as "manure field application."
source: Reference 2 concluded
90098 06" 6-4
-------
TABLE 6-2. Additional Area Source Category Descriptions for mobile sources for use with the.
UAM EPS.
Exhaust emissions:
27 Light-Duty Gasoline Vehicles - Limited Access
28 Light-Duty Gasoline Vehicles - Rural. Roads -
29 Light-Duty Gasoline Vehicles - Suburban Roads
30 Light-Duty Gasoline Vehicles - Urban Roads
31 Medium-Duty Gasoline Vehicles - Limited Access
32 Meaium-DuTy Gasoline Vehicles - Rural Roads
33 Medium-Duty Gasoiina Vehicles - Suburban
34 Medium-Duty Gasolma Vehicles - Urban Roads
35 Heavy-Duty Gasoline Vehicles - Limited Access
36 Heavy-Duty Gasoline Vehicles - Rjral Roads
37 Heavy-Duty Gasol.ne Venicles - Suburban Roads
33 Heavy-Du'y Gasoline Vehicles - Urban Roads
39 Gff-mgnw'jy Gasoline vencies
40 Heavy-Duty Diesel Vehicles - Limited Access
41 Heavy-Duty Diesel Vehicles - Rural Roads
42 Heavy-Duty Diesel Vehicles - Suburban Roads
43 Heavy-Duty Diesel Vehicles - Urban Roads
44 Off-Highway Diesel Vehicles
Refueling emissions:
327 Light-Duty Gasoline Vehicles - Limited Access
328 Light-Duty Gasoline Vehicles - Rural Roads
329 Light-Duty Gasoline Vehicles - Suburban Roads
330 Light-Duty Gasoline Vehicles - Uiban Roads
331 Medium-Duty Gasoline Vehicles - Limited Access
332 Medium-Duty Gasoline Vehicles - Rural Roads
333 Medium-Duty Gasoline Venioiss - 3t.c_i. -.--\
334 Medium-Duty Gasoline Vehicles - '_'Q3n 3oass
335 Heavy-Duty Gasoline Vehicles - Limitec Access
336 Heavy-Duty Gasoline Vehicles - Rural poads
337 Heavy-Duty Gasoline Vehicles - Suburban Roads
338 Heavy-Duty Gasoline Vehicles - Urban Roads
339 Off-Highway Gasoline Vehicles
340 Heavy-Duty Diesel Vehicles - Limited Access
341 Heavy-Duty Diesel Vehicles - Rural Roads
342 Heavy-Duty Diesel Vehicles - Suburban Roads
343 Heavy-Duty Diesel Vehicles - Urban Roads
344 Off-Highway Diesel Vehicles
Evaporative emissions:
227 Light-Duty Gasoline Vehicles - Limited Access
228 Light-Duty Gasoline Vehicles - Rural Roads
229 Light-Duty Gasoline Vehicles - Suburban Roads
230 Light-Duty Gasoline Vehicles - Urban Roads
231 Medium-Duty Gasoline Vehicles - Limited Access
232 Medium-Duty Gasoline Vehicles - Rural Roads
233 Medium-Duty Gasoline Vehicles - Suburban
234 Medium-Duty Gasoline Vehicles - Urban Roads
235 Heavy-Duty Gasoline Vehicles - Limited Access
236 Heavy-Duty Gasoline Vehicles - Rursi Roads
237 Heavy-Duty Gasoline Vehicles - Suburban Roads
238 Heavy-Duty Gasoline Vehicles - Urban Roads
239 Off-Highway Gasoline Vehicles
240 Heavy-Duty Diesel Vehicles - Limited Access
241 Heavy-Duty Diesel Vehicles - Rural Roads
242 Heavy-Duty Diesel Vehicles - Suburban Roads
243 Heavy-Duty Diesel Vehicles - Urban Roads
244 Off-Highway Diesel Vehicles
Running loss emissions:
427 Light-Duty Gasoline Vehicles - Limited Access
428 Light-Duty Gasoline Vehicles - Rural Roads
429 Light-Duty Gasoline Vehicles - Suburban Roads
430 Light-Duty Gasoline Vehicles - Urban Roads
431 Medium-Duty Gasoline Vehicles - Limited Access
432 Medium-Duty Gasoline Vehicles - Rural Roads
433 Medium-Duty Gasoline Vehicles - Suburban
434 Medium-Duty Gasoline Vehicles - Urban Roads
435 Heavy-Duty Gasoline Vehicles - Limited Access
436 Heavy-Duty Gasoline Vehicles - Rural Roads
437 Heavy-Duty Gasoline Vehicles - Suburban Roads
438 Heavy-Duty Gasoline Vehicles - Urban Roads
439 Off-Highway Gasoline Vehicles
440 Heavy-Duty Diesel Vehicles - Limited Access
441 Heavy-Duty Diesel Vehicles - Rural Roads
442 Heavy-Duty Diesel Vehicles - Suburban Roads
443 Heavy-Duty Diesel Vehicles - Urban Roads
444 Off-Highway Diesel Vehicles
source: Reference 3
9C098 OB"
6-5
-------
TABLE 6-3. Comparison of NAPAP area source categories and subcategores u5-?'1
in Procedures for the Preparation of Emission inventories for Precursors.of CiO'-v?.
Voii-me ! (EPA, 1988).
NAPAP Code and Description
Volume I Subcategories
54 Gasoline marketing
Tank Truck unloading (Stage 1,
Vehicle refueling (Stage II)
Underground tank breathing
Gasoline tank trucks in transit
78 Degreasing
Open top and conveyorized
degreasing
Cold cleaning degreasing
39 Off-highway gasoline vehicles,
44 Off-highway diesei vehicles
Off-highway motorcycles
Farm equipment
Construction equipment
Industrial equipment
Lawn and garden equipment
Snowmobiles
90098 09"
6-6
-------
distinctions for the existing inventory and may need to institute certain changes to ensure
that the modeling inventory meets the modeling objectives. The following example
illustrates a case where such changes may be appropriate.
* In order to make a detailed analysis of the effect of controlling dry cleaning
operations, the emissions modeler may choose to treat each facility
individually as a point source rather than collectively as an area source to
facilitate evaluation of distinct control measures for each facility. Conversely,
if the existing inventory treats dry cleaning facilities as point sources, but the
emissions modeler cannot obtain specific information on anticipated growth
at specific locations, the emissions modeler may wish to treat dry cleaning as
an area source in the modeling inventory.
In most cases, however, the same point and area source distinctions employed in the
existing inventory should be maintained in the modeling inventory to minimize additional
resource requirements.
The emissions estimates available from the existing inventory usually represent .annual or
(in some cases) seasonal emissions for a fairly broad geographical area, such as for each
county within an urban area, primarily because the estimates of activity levels used to
calculate area source emissions are generally available at the county level. Generally,
these emissions estimates will not distinguish between different reactive classes of VOC
and NOX.
Nevertheless, the area source emissions contained in the existing inventory can often be
used in the modeling inventory. In order to provide the spatial, temporal, and chemical
resolution required of the modeling inventory, the emissions modeler must perform the
following tasks:
o allocate county-level emission estimates for area sources to modeling grid
cells;
o develop hour-by-hour emission estimates for the episode days; and
o apportion VOC emissions for each source into chemical classes (and, in some
models, distinguish NOX emissions as NO and NO2).
Sections 6.2 and 6.3 describe techniques for providing the necessary spatial and temporal
resolution in the modeling inventory, respectively. Section 6.4 addresses projection of
area source emissions estimates. Procedures for speciating area source emissions into
chemical classes are discussed in Chapter 9.
90098 00" 6-7
-------
Remember that the only emissions to be disaggregated by the procedures described below
are area source totals. If some of the sources in any category have been listed and treated
as either major or minor point sources, the emissions modeler must subtract the emissions
from those sources from the county-level category total before applying the disaggregation
procedures. If available local information allows a major fraction of the emissions to be
treated as minor point sources for both the base year and the projection years, such
handling may be advantageous. However, if the number of such sources is minor and
their aggregate emissions are inconsequential, or if the necessary projection information is
inadequate, subdivision of the area source category into area and point components may
not be worth the additional effort required.
6.2 GENERAL METHODOLOGY FOR SPATIAL RESOLUTION
County-level area source emissions estimates can be apportioned to grid cells using either
of two approaches. In certain cases, determining the activity levels and emissions of some
area sources directly for each grid cell may be feasible. More commonly, the emissions
modeler must apportion county-level emissions by assuming that the distribution of the
area source activity behaves similarly to some spatial surrogate indicator. Both
approaches are discussed below.
6.2.1 Direct Grid Cell Level Determination of Emissions
In limited cases, the emissions modeler may possess sufficient information to calculate
area source activity levels and emissions directly for each grid cell. For instance, enough
data may be available for individual facilities that they could have been considered as
minor point sources in the existing inventory; however, for various reasons, a decision has
been made to consider these sources collectively as an area source. Two examples are
given below:
> A local gas company has information on the quantity of natural gas fired in
every household or commercial establishment, allowing direct calculation of
emissions by grid cell.
>• Survey results are available for a particular type of commercial or industrial
establishment; for instance, a survey may have been conducted resulting in
information on the sales and location of each gasoline service station in the
modeling region. Instead of aggregating gasoline sales and calculating
emissions at the county level (as may be done in the annual inventory),
gasoline sales can be aggregated and emissions calculated for individual grid
cells.
90098 06" 6-8
-------
if survey information similar to that described in the example above is available, the
emissions modeler may wish to "reassign" each facility as a point source in the modeling
inventory. This course might be particularly advantageous if a certain control measure
under consideration for implementation can best be evaluated by treating each facility
within the affected source category as a point source. This method, however, requires
that many additionai point source records be generated and maintained in both the base
year and projection inventories, an obvious disadvantage. In projection inventories,
handling numerous small establishments as area sources rather than as point sources will
usually be easier, especially if the emissions modeler does not have information regarding
the location of each facility in the projection years.
6.2.2 Surrogate Indicator Approach
If the approach described above for directly determining area source activity ievc-,'3 2nd
emissions for each grid cell is infeasible, the emissions modeler must implement some
other apportioning scheme to spatially allocate the emissions in the county-lavs! area
source inventory. The most straightforward approach would be to distribute the total
emissions for each county evenly over ajl grid cells in the" county; this approach, however,
defeats the purpose of using a sophisticated grid model like the DAM. Instead, the usual
method employed to spatially distribute emissions to sub-county regions involves the use
of various combinations of spatial surrogate indicators.
' '*>•
A spatial surrogate indicator is a parameter whose distribution is known at a subcounty
level and which behaves similarly to the activity levels of interest. Commonly used spatial
surrogate indicators include land use parameters, employment in various industrial and
commercial sectors, population, and dwelling units. Different surrogate indicators should
be used to apportion emissions for the various area source categories, of course,
depending on which of the available indicators best describes the spatial distribution of the
emissions. The emissions modeler should employ engineering judgment to select
appropriate indicators for apportioning area source emission totals and should consult local
authorities to verify the applicability of the source category/spatial surrogate indicator
pairings for that particular modeling region.
> For example, fugitive emissions from crude oil and gas production fields in
southern California can be distributed using range land as a spatial surrogate
indicator. In Baton Rouge, Louisiana, however, the spatial distribution of
emissions from these sources may be more accurately represented by using
wetlands as a spatial surrogate indicator.
Table 6-4 lists example spatial surrogate indicators for area source categories, as utilized in
various urban areas. These indicators can be used to spatially apportion emissions from
these source types in the absence of more detailed data; however, the emissions modeler
goose ce" 6-9
-------
FABLE 6-4. Example spatial allocation factor surrogates for area source categories.
Category
ID
1
2
3
4
5
6
7
a
9
10
1 1
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
43
49
50
Surrogate
Indicator
Housing
Housing
Housing
Housing
Housing
Housing
Urban'Land
Urban Land
Urban Land
Urban Land
Urban Land
U'bsn Land
Urban Land
Urban Land
Urban Land
Urban Land
Urban Land
Urban Land
Urban Land
Urban Land
Housing
Urban Land
Urban Land
Housing
Urban Land
Urban Land
Land Area
Land Area
Housing
Urban Land
Land Area
Land Area
Housing
Urban Land
Land Area
Land Area
Housing
Urban Land
Land Area
Land Area
Land Area
Housing
Urban Land
Land Area
Urban Land
Population
Population
Population
Population
Population
Emissions Category
Residential Fuel - Anthracite Coal
Residential Fuel - Bituminous Coal
Residential Fuel - Distillata Oil
Residential Fuel - Residual Oil
Residential Fuel - Natural Gas
Residential Fuel - Wood
Commercial/Institutional Fuel - Anthracite Coal
Commercial/Institutional Fuel -Bituminous Coal
Commercial/Institutional Fuel -Distillate Oil
Commercial/Institutional Fuel - Residual Oil
Commercial/Institutional Fuel - Natural Gas
Commercial/Institutional Fuel - Wood
Industrial Fuel - Anthracite Coal
Industrial Fuel - Bituminous Coal
Industrial Fuel - Coke
Industrial Fuel - Distillate Oil
Industrial Fuel - Residual Oil
Industrial Fuel - Natural Gas
Industrial Fuel - Wood
Industrial Fuel - Process Gass
Incineration - Residential
Incineration - Industrial
Incineration - Commercial/Institutional
Open Burning - Residential
Open Burning - Industrial
Open Burning - Commercial/Institutional
Light Duty j6as Vehicles - Limited Access
Light Duty Gas Vehicles - Rural
Light Duty Gas Vehicles - Suburban
Light Duty Gas Vehicles - Urban
Medium Duty Gas Vehicles - Limited access
Medium Duty Gas Vehicles - Rural
Medium Duty Gas Vehicles - Suburban
Medium Duty Gas Vehicles - Urban
Heavy Duty Gas Vehicles - Limited Access
Heavy Duty Gas Vehicles - Rural
Heavy Duty Gas Vehicles - Suburban
Heavy Duty Gas Vehicles - Urban
Off-Highway Gas Vehicles
Heavy Duty Diesel Vehicles - Limited Access
Heavy Duty Diesel Vehicles - Rural
Heavy Duty Diesel Vehicles - Suburban
Heavy Duty Diesel Vehicles - Urban
Off-Highway Diesel Vehicles
Railroad Locomotives
Aircraft - Military
Aircraft - Civil
Aircraft - Commercial
Vessels - Coal-Powered
Vessels - Diesel
continued
90098 36"
6-10
-------
TABLE 6-4. Concluded.
Category
ID
5!
52
84
56
63
60
61
62
64
66
67
S3
69
70
71
72
73
74
76
77
78
79
80
81
82
83
84
86
86
87
88
89
90
91
92
93
95
96
97
93
99 .
100
101
102
103
104
106
106
107
108
109
Surrogate
Indicator
Population
Population
Population
Land Area
Land Area
Composite Forest
Composite Forest
Agricultural Land
Housing
Lard Area
Land Arsa
'. •.•-•; / ' :
AofisuiJuc :i i-.'nc
Aa" .'- _. - .-. -
Agnc'j'tufJi L^nd
Ar'ir."", •'; L "^
Agrii'^.-'j - < '--PC
Agricultural Land
Agricultural Lara
Agncultura' Lar.d
Population
Population
Population
Population
Population
Population
Population
Population
Population
Population
Population
Population
Population
Population
Population
Population
Population
Population
Population
Population
Population
Population
Population
Land Area
Population
Population
Population
Population
Population
Population
Population
Emission* Category
Vessels - Residual Oil
Vessels - Gasoline
Gasoline Marketed
Unpaved Roads
Unpaved Airstrips
Forest Wild Fires
Managed Burning - Prescribed
Agricultural Field Burning
Structural Fires
Ammonia Emissions - Light Duty Gasoline Vehicles
Ammonia Emissions - Haavy Duty Gasoline Vehicles
. -,--;o.";j c.Ti'csi-ns - naj.'y Duv Dii3al Vaiiit^s
L:Y3'S'oc'< V'/'ist^ Management - Turkeys
w.vio^cc^ V.'jitj ,',<,_., agarnent - Sheap
Livestock Waste Management - Beef Cattle
Lv^st"ck V*':3t3 \i ";P2^ or" ?n* - Djrv Chills
L'V35tOC'< 'vYaGCS '>' J n 2 '" C r'"1 S It - b*VIP,3
Livestock Waste ^.nn3gerrifint - Sro'isfs
AnnydroLS NH3 rsrtifizsr Application
Beef Cattla r-ed Lots
Cegreasir.Q |
Drycisaning
Graphic Arts/Printing
Rubber and Plastic Manufacturing
Architectural Coating
Awto Body Repair
Motor Vehicle Manufacturing
Paper Coating
Fabricated Metals •
Machinery Manufacturing
Furniture Manufacturing
Flat Wood Products
Other Transportation Equipment Manufacturing
Electrical Equipment Manufacturing
Ship Building and Repair
Miscellaneous Industrial Manufacturing
Miscellaneous Solvent Use
Minor Point Sources - Coal Combustion
Minor Point Sourcos - Oil Combustion
Minor Point Sources - Natural Gas Combustion
Minor Point Sources - Process Sources
Publicly Owned Treatment Works (POTWs)
Cutback Asphalt Paving Operation
Fugitives from Synthetic Organic Chemical Mfg.
Bulk Terminal and Bulk Plants
Fugitives from Petroleum Refinery Operations
Process Emissions from Bakeries
Process Emissions from Pharmaceutical Mfg.
Process Emissions from Synthetic Fibers Mfg.
Crude Oil and Natural Gas Production Fields
Hazardous Waste Treatment, Storage and Disposal Facilities
(TSDFs)
source: Reference 2
concluded
90098 08"
6-11
-------
should make a special effort to choose spatial surrogate indicators for the various source
categories which accurately reflect the distribution of activity for those sources in the
modeling region. Specifically, emissions from non-highway mobile sources (such as
railroad locomotives, aircraft, etc.) should be allocated only to those grid cells in which
such activity occurs; this will be discussed in greater detail later in this chapter. Table 6-5
lists specific references which contain useful information for developing spatial resolution
for several source categories; other sources, which will be addressed in detail below,
include land use patterns (from maps and/or computerized data bases) and Census Bureau
demographic statistics bv traffic zone or census tract.
Developing Ag^cr^cr.1".;' ••'" r •" • _ ;;~ ™. '..::•:••' '.'-:i.._°^tttf53. Per most urban araas, ia
data will be available ;z; .'- : :/•:: „-.".: :.,-.;; cc'.arci prc^cticn years; tha emissions modeler
can use this data to develop aoccrrioning factors for "hose area sources whose emissions
will be distributed bo;.;;. ,,"> '._:'j;_j .~.-j o~s ^ijjo:."cc::cr,c. Al^r.cugh spatial apportioning
factors can be develoced rr^m^ily from mgos, ccrnoutenzing as many steps of this
process as possible csn-ra!!-/ ;r.lr,:c<-i\z~3 tha r^q;jir5d -fort. Unfortunately, computarizeci
land use data may be unayailab's for projection years, an obvious drawback. In this case,
the computerized land use data bass can bs used to develop apportioning factors for the
base year emissions inventory, and projected changes in land use patterns accounted for in
projection year apportioning factor files by editing the base year file. One national land use
data base which can be used to determine spatial apportioning factors is described below;
other sources of land use data may also be used, if available.
>• The U.S. Geological Survey (USGS) maintains a 'comprehensive computerized
national data base of land use distribution data, based upon the classification
system shown in Table 8-6. The USGS data files, available in both digital and
character formats, contain data for many regions of the country in terms of
four hectare grid cells (200 meters x 200 meters). Items contained in the
data base for each individual grid cell include UTM zone, UTM Easting and
Northing, land use and land cover attribute code (Table 6-6), political unit
code, USGS hydroiogic code, census county subdivision or SMSA tract code,
Federal land ownership agency code, and State land ownership code. Since a
given modeling region will often contain over 500,000 four-hectare grid cells,
manipulation of such large amounts of data is best accomplished with the aid
of a computer.4
Regardless of the source of land use data, the same basic procedures must be followed to
generate the spatial apportioning factor file. First, the grid cells within each county must
be identified, as illustrated conceptually in Figure 6-1; in this figure, the shaded area in the
upper map has been approximated with shaded grid cells in the lower grid. Figure 6-2
designates the grid cell assignments for each county in the modeling region for the
Atlanta, Georgia area. In this figure, the large numbers along the southern and western
90098 Off3 6-12
-------
TABLE 6-5. Additional sources of information for spatial resolution of emissions for selected
area source categories.
Source Type
References
Aircraft, commercial
FAA Air Traffic Activity Reports (Annual), U.S. Department of
Transportation, Federal Aviation Administration, Washington, D.C.
Airport Activity Statistics for Certified Route Air Carriers (Annual), U.S.
Department of Transportation, Federal Aviation Administration,
Washington, D.C.
Aircraft, general Census of U.S. Civi.l Aircraft (Annual), U.S. Department of Transportation,
Federal Aviation Administration, Washington, D.C.
Aircraft, military Military Air Traffic Report (Annual), U.S. Department of Transportation,
Federal Aviation Administration, Washington, D.C.
Agricultural equipment Census of Aqric'ufture, Volume I, Area Reports (Annual), U.S. Department
of Commerce, Bureau of the Census, Washington, D.C.
Off-highway motorcycles Motorcycle Statistical Annual, Motorcycle Industry Council, Inc., Newport
Beach, CA.
Railroad locomotives Transportation maps of various states, prepared by U.S. Geological Survey
for the Office of Policy and Program Development, Federal Railroad
Administration, United States Department of Transportation.
Vessels (ocean-going, river Waterborne Commerce of the United States, (Annual), U.S. Army Corps
cargo, and small pleasure of Engineers, Washington, D.C.
craft)
Gasoline handling Census of Business Selected Services Area Statistics, U.S. Department of
Commerce, Bureau of the Census, Washington, D.C.
Fuel combustion, Sales of Fuel Oil and Kerosene, Mineral Industry Surveys.
commercial/institutional
90093 06"
6-13
-------
TABLE 6-6. Land use classification system used in USGS land use data bases.
1. URBAN OR BUILT-UP LAND
11 Residential
12 Commercial and Service
13 Industrial
14 Transportation, communication
and services
15 Industrial and commercial
complexes
1 6 Mixed urban or built-up land
17 Other urban or built-up land
2. AGRICULTURAL LAND
21 Cropland and pasture
22 Orchards, groves, vineyards,
nurseries, and ornamental
horticultural groves
23 Confined feeding operation
24 Other agricultural land
3. RANGELAND
31 Herbaceous rangeland
32 Shrub and brush rangeland
33 Mixed rangeland
4. FOREST LAND
41 Deciduous forest land
42 Evergreen forest land
43 Mixed forest land
8.
WATER
51 Streams and canals
52 Lakes evergreen
53 Reservoirs
54 Bays and estuaries
WETLAND
61 Forested wetland
62 Nonforested wetland
BARREN LAND
71 Dry salt flats
72 Beaches
73 Sandy areas, other
75 Strip mines, quarries, and
gravel pits
76 Transitional areas
77 Mixed barren land
TUNDRA
81 Shrub and brush tundra
82 Herbaceous tundra
83 Bare ground
84 Wet tundra
85 Mixed tundra
PERENNIAL SNOW OR ICE
91 Perennial snow fields
92 Glaciers
source: Reference 4
90098 OS"
fi-1 4.
-------
FIGURE 6-1. Conceptual representation of the grid cell identification process.
90098 06"
6-15
-------
•- :Bi
•7lJ»iJ»i3» 13* Jv - yl
as
83 K 1C SI S3 ST
13 ^7 TTrnnTIT-JT J1' ftl M (3 117 S3 » S3
13 47 S7 127 J7 37 37 »7
111113 113113113
15 37 47 57 57 S7 S7 57 51 11 7 1 1 7 1 1 7 1 1 7 1 1 7 1 I 7
S> 37 17 37 S7 37 57 57 371 1 1 7 1 17 1 17 1 17 1 1 7 1 1 7 1 1
3» i i» res>
II' 137 |37 li,1
jT-3 nSl'SHSIij .5 1! 13 IS 13 13 1! !5 1357 37 S7 17 47 57L31 -S1 Ul!l£i1171
g33?i3H5Ij3;a 13 13 13 13 11K3'3 15 UJ 37 37 37 37 «7 * 111 1J1 151 tilTVvi
«7 47 «7 «7 117 <17 —y'\ II
3 »3 M3 233 »3 27 J) «7
13 213 213 213 2 3 233373 213 K3 2-23223 223
33133 !3S1i3l3S1«
:23 213 M3 223 2« 223 213
73 M3 M3 713 773 273 723
2»7»7 f»7 7»7 2»7 ?»7 }»7 2*7 7S
»72»7 7»77»7 ?»7 7»7 797 2»7
»7 IT »7 »7 »7 17 tJJOf 171 121 121 171
»7 »7 »7 f7 »7 »7/21 121 121 121 171 121
7 797 7»7 7*7 »7 ?»
7»72»77»73»71»72il
2117M
7211 JM
43 *S 44 *3 »3 *3 I 7 »7 »7 »7 t? «T 1
45 *3 *5 43 *3 43 *3 45 I 7 »7 »7 »7
1 43 U 43 A3 M 151
43 43 43 43 «b 131
«a 43 43 cijai 151131
43 43
1J1 HI 171 1*11111*1 t
1*11*1 111 1*1 1»1 121 1
1J11211I
4J 43 *3 *3 *3 *4 *1 *5
41 *; 45 43 49 45 *: *3
45 43 43 4S 43 43 43 43 43
43 43 43 43 *3 43 41 4J 43
I47JI7J17J17J17J17II
247 217 JIT717717 717 717 21
7172171!«1A«15*15*
15*15*
15*
14t14»14*1**1**14*\77 77 77 77 77 77 77 77
7777T7T77"»777T77771
14<14*14>14t 7777T7777777T77777
9 1**14« 144 14*14«14*\4* T7T777777777T777T7
514t14»l4»1*»14»J»;iM77 77 T71M77 T7 77OB'S»77;i3iJ51I«rS;2i32iS2S3753
15*154 15*15*15*!
731 243 241 7i: 253 S
1231 731 231 2i325i
71 2O77077O770770
0
10
665
FIGURE 6-2. County grid cell assignments for the Atlanta, Georgia modeling region.
9009S 06"
6-16
-------
boundaries of the modeling region represent grid cell (I,J) modeling coordinates; the large
numbers along the northern and eastern boundaries are UTM coordinates. The small
numbers located within each grid cell of the map itself are geographical codes denoting
each county. Within each grid cell in the county, the fraction of the total county land use
for each land use category must then be calculated. Note that several land use categories
may contribute to the total land use for any given cell; similarly, more than one county can
contribute to the total area within a grid ceil, as shown in Figure 6-2 by the overlap of
numerical codes in those grid cells comprising the county borders.
The following example illustrates a manual procedure for developing gridded spatial
apportioning factors from maps. In general, the procedures outlined below can also be
computerized; likewise, ccrnputenzT'C C-K3 ':•-•:;.• . r,: .::.;: Z'~L:; .iViUrns auch as
geographical information systems (G!£; "z~ :;• .::*- :c ~z.i.zz c,:c..:.! . .;,cr
factors, allowing complete automation of LIS soatiai aiiocaticn process.
> Assume that the existing inv?ntc~y contains an ertirrst? of :ct
from dry cleaning for the entire .-;;,>'Jy araa ar.d :hac nc o^::'
information is available for individual dry cieaning establishments. The
emissions modeler must select a spatial surrogate indicator '.ha: v-/ii! permit
distribution of emissions to the individual grid cells in the study area. Land
use maps, which cartographically characterize each part of the study area in
terms of what kinds of activities are predominant in that area, are often
available from local planning agencies. Figure 6-3 shows a land use map of
part of the Tampa Bay, Florida region. The various areas are identified in
great detail by numbers; Table 6-7 shows the coding system used iri this
application. Other land use maps may use colors or shading techniques to
differentiate areas.
Since dry cleaning is a typical commercial activity, a reasonable assumption is
that dry cleaning area source emissions emanate uniformly from the
commercial areas as shown on the land use map. Thus, the spatial surrogate
indicator will be the area devoted to commercial land use (represented in
Figure 6-3 by codes 12 and 15). In this approach, the area within each grid
cell designated as a commercial area on the land use map must be estimated.
For this purpose, the grid system network must be superimposed on the land
use map, as shown in Figure 6-3. The estimates of land use area in a grid cell
can be fairly rough (e.g., to the nearest tenth of a grid cell). As an example,
consider the grid cell designated (15,15) in Figure 6-3. For this grid cell,
about 20 percent of the area is indicated as commercial (code number 12),
while the remaining 80 percent of the grid cell is designated as single-family
residential (code number 11). If a grid cell contains an area designated by
code number 15 (industrial and commercial combined), such an area may be
weighted at 50 percent in this computation.
90098 OS'2 6-17
-------
9 10 11 12 13 14 15 16 17 18 19 20
FIGURE 6-3. Segment of land use map for Tampa Bay, Florida.
90098 06"
6-18
-------
TABLE 6-7. Land use categories for Tampa Bay area land use map (Figure 6-3).
URBAN OR BUILT-UP LAND
10 Multi-family residential
11 Single family residential
1 2 Commercia! ana service
13 Industrial
14 Transportation, communication
and utilities
15 industrial and commercial
combined
16 Mixed urban or built-up land
1 7 Other urban or built-up land
2. AGRICULTURAL LAND
23 Confined feeding operation
24 Other agricultural land
25 Cropland
26 Improved pasture
27 Specialty farms
28 Orchards, groves, vineyards,
nurseries, and ornamental
horticultural groves
29 Citrus groves
3. RANGELAND
31 Herbaceous rangeiand
32 Shrub and brush rangeiand
33 Mixed rangeiand
FOREST LAND
41 Deciduous forest land
42 Evergreen forest land
43 Mixed forest land
WATER
51 Streams and canals
52 Lakes
53 Reservoirs
54 Bays and estuaries
WETLAND
63 Freshwater forested wetland
64 Freshwater marsh
65 Saltwater forested wetland
66 Saltwater marsh
BARREN LAND
72 Beaches
73 Sandy areas other than
beaches
75 Extractive
76 Transitional areas
source: Reference 5
90098 06"
6-19
-------
The emissions for each grid cell are then estimated as a simple fraction of the
total, as follows:
E, = ET (S, /ST) (6-1)
where E denotes emissions, S indicates surrogate indicator, i indicates the
value in grid cell i, and T indicates the total for the county or region.
The units for the surrogate indicator can be arbitrary (e.g., percent of grid
cell, square kilometer, square mile). For example, assume that the total
commercial area in Figure 6-3 covers an area the size of 26.3 grid cells. Then
the fraction of the total commercial area (S, / ST) for grid cell (1 5,15) will be
0.2/25.3, or 0,0076. (This fraction is known as a "apportioning factor.")
Thus, the emissions for dry cleaning attributed to grid cell (15,15) will be
0.0076 times the totai dry cleaning emissions from the entire region.
Mathematically, this can aiso be expressed by Equation 6-2,
• (6-2)
where flk is the apportioning factor for grid cell i with respect to source
category k, and n is the total number of grid cells.
The emissions modeler can also use other maps, if reasonably current, to develop
apportioning factors for various area sources.
> United States Geological Survey (USGS) maps show the location of oil and
gas wells. By counting the number of wells per grid cell, total oil and gas well
emissions can be apportioned by multiplying the total emissions for each
county by the fraction of the total number of wells in each grid cell. (In this
case, the number of wells serves as the surrogate indicator of product or
activity). Similarly, USGS maps show railroad track mileage, which may be
used to develop apportioning factors for railroad emissions.
Often, the most representative way to spatiallyDistribute emissions from some
. off-highway mobile sources, such as railroad locomotives, aircraft and vessels,
: .is to treat these source*.as.*lfoe".sources-, • Emissions.froin these sources: caq :.
be assumed to occur only in those grid cells that contain railroad track mileage^
• ... airports, or waterways,. The UAM EPS program GftOcMS will dpstrlhute ;. ; ;.
emissions to grid cells using user-specified Imk data. Specificaliy, GRDEMS i.
;:,... allocates the emissions associated with each type o? Imk {&.$,,,.railroad track ;
: mileage) to each grid: cell feased on the fraction of the total county liftk distance
: for the link type occurring irt that grid dell Section 7,6, regarding spatial :
: •• • . . : - - • s_^ : . • . .. . -
90098 06" 6-20
-------
: distribution of highway mobile sourcs emissions, contains additional inlcrmatiar*
on the specification of Imk data foruse with the DAM EPS, '
One disadvantage of developing apportioning factors from maps other than land usa maps
is that the corresponding projection information for allocating future year emissions will
often be unavailable. In these cases, the emissions modeler will either have to (1) assume
that projection year spatial emission patterns for these sources will not change, or (2)
locate additional information that shows what changes are expected in the spatial
surrogate indicator distributions.
T,'"o forage-ing discussion deslt only with the allocation of area source emissions basad on
a single surrogate indicator. In some cases, no one parameter may accurately describe the
o^bcounty distribution of emissions from a particular area source category. In this
situation, apportioning factors can be based on two or mere surrcgsts indicators.
> Since miscellaneous solvent use can be associated with both consumer
(residential) and commercial applications, the emissions modeler may wish to
distinguish between the possible different rates of use in these land use
categories (10, 11, and 12 of Table 6-7).
The emissions modeler can use either one'"of two principal methods to perform this
apportionment. First, solvent emission subtotals can be estimated for the three types of
land use involved, and each of these subtotals apportioned according to the corresponding
subcategories (in effect, this creates three new emission subcategories which replace the
one contained in the county-level inventory).
> One third of miscellaneous solvent emissions may be assigned to multifamily
residences (land use 10), one third to single family residences (land use 11),
and one third to commercial and service use (land use 12). Hence, if county-
level emissions from miscellaneous solvent use are 1 2 tons per day, 4 tens
per day would be apportioned at the grid cell level for each of these
subcategories, based on the distribution of the corresponding surrogate
indicator.
Alternatively, the emissions modeler might decide to estimate the level of activity
associated with each land use category.
> Assuming single-family residential areas have the smallest emission rate per
unit area, the emissions modeler might estimate that the emission rate in
multiple family residential areas is three times as large as in single-family
residential areas, and in commercial and service areas, five times as large. In
9003806" 6-21
-------
this case, the apportioning factors would be calculated using an appropriate
weighting factor for each of the three types of land use. This would be
expressed, mathematically, by the equation
f,k = (I ?„ W,kS,) / [Z ?., (I ?„ W^S,,}] (6-3)
where W)k is the weighting factor selected for land use type j in relation to
source category k, and S,, is the value of the surrogate indicator (i.e., the area)
of land use type j in ceii i.
The summation term appearinq in the numerator above is essentially a
composita surrogate indicator for the entire category. Thus, if solvent
emissions ars weighted according to the previous suggestion (W, = 1, W2 =
3, W2 = 5) and the respective areas in a given grid cell are 0.6, 0.2, and 0.2,
then the value of the composite surrogate indicator for that ceil is (0.6 x 1) -t-
(0.2 x 3) + (0.2 x 5), or 2.2. The entire category is then apportioned as
usual, basad on this composite surrogate indicator.
Developing Apportioning Factors from Demographic Statistics at the Traffic Zone Level.
As part of the transportation planning process routinely performed in larger urban areas,
employment and other demographic statistics are^aggregated at the zonal level. These
statistics can be used instead of (or in addition to) land use patterns to obtain the
information needed to apportion area source emissions to the subcounty level.
In theory, these zonal statistics contain the same data available from land use maps or
data bases; thus, the only difference in using one approach or the other is procedural. In
practice, however, typically available land use data are often less detailed than the zonal
statistics. For instance, zonal statistics in a particular urban area may be compiled for five
or more commercial and industrial subcategories; however, the corresponding land use
data may only identify generalized commercial and industrial land use.
To manually develop apportioning factors from land use maps, the emissions modeler must
code the land use data for each grid cell; this step must be repeated for every growth
projection. Using zonal statistics, however, allows this process to be largely automated
once a set of zone-to-grid-cell conversion factors has been developed. These conversion
factors are discussed later in this section. The data handling aspects of utilizing zonal
statistics are addressed in more detail in Section 6.5.
The following example illustrates the use of detailed zonal statistics for developing
allocation factors as well as the use of multiple surrogate indicators to apportion emissions
from a given area source category.
90098 Off2 6-22
-------
> In the San Francisco Bay Area of Caiifornia, emissions from 58 area sources
are apportioned using combinations of the 19 demographic parameters shown
in Table 6-8. ail of which are compiled at the subcounty leve! by the locai
MPO as part of transportation planning studies. For some area source
categories, a single parameter from Table 6-8 is ussd as a surrogate indicator
of the distribution of emissions. For instance, the source category "farming
operations" is linked with the single employment category "AGRI" from Table
6-8, which includes agriculture production and services. Similarly, the source
category "printing" is distributed with ths variable "MFC!," which includes
printing, publishing, and related industries.
The spatial distribution of emissions from other area source categories,
however, cannot be accurately represented using a single variable. In these
cases, emissions are apportioned based on two or more parameters. Table 6-
9 presents an excerpt from a cross-classification table used in the Bay Area;
this table shows the percentage of each area source emission total that is
apportioned by each demographic parameter listed in Table 6-8.
Assume that area source degreasing emissions in a given county are 42
tons/day of VOC. According to Table 6-9, 10 percent of this total should be
apportioned based on manufacture of electrical and optical machinery and
instruments employment (MFG4), 60 percent based on fabricated metal
product employment (MFG5)/"20 percent based on retail service employment
(RET SERV.), and 10 percent based on other services employment, which
includes local transit and transportation services (OTHER SERV.). Thus, 4.2
ton/day (42 x 0.10) are apportioned according to the fraction, in each grid
cell, of the total number of employees in the "MFG4" category; 25.2 ton (42
x 0.60) are apportioned according to the fraction, in each grid cell, of the total
number of employees in the "MFG5" category; 8.4 ton (42 x 0.20)
apportioned according to the fraction of employees in each grid cell in the
"RET. SERV." category; and 4.2 ton (42 x 0.10) apportioned according to the
fraction of "OTHER SERV." employees in each grid cell. For example, if the i'h
grid cell contains 0.1 percent of the total area-wide "MFG5" employees, 0.05
percent of the "RET. SERV." employees, 1 percent of the "MFG4" employees,
and no "OTHER SERV." employees, then the degreasing emissions would be
apportioned to that grid cell as follows:
ith Grid Cell Emissions = 25.2(0.001) +8.4(0.0005)-f 4.2(0.01 ) + 4.2(0)
= 0.0714 ton/day
The degreasing emissions for the other grid cells would be apportioned
similarly, as would the emissions for the other area sources. An equivalent
formulation of this procedure is simply to subdivide the area source
90098 08" 6-23
-------
TABLE 6-8. Demographic parameters used in San Francisco Bay Area for making zoneal
allocations of area sources.
Variabla* Name
O'vVSLL
A^,
- ,
V~3'
MFG2
MFG3
MFG4
MFG5
iViFG5
TRAN
wVHOL
FIN
SERV 1
SERV 2
GOV
RET
BUS. SERV.
RET. SERV
OTHER SERV.
SIC* Classification
(not applicable)
1, 7-9
10, 13, 14
27
26, 28, 29, 32, 33
20
19, 36, 33
34, 35, 37
22-25, 31, 39
40, 42, 44-46
50, 52
62, 63, 67
73
82, 84, 89
91, 92
53-59
80, 81, 96
70, 72, 75-79
15-17, 41, 47-49, 60, 61,
66, 93-95, 99
Description
Dwelling units
Agriculture, forestry
Mining, quarry, oil and gas extraction
Printing, publishing
Petroleum, chemical, paper, and metal industries
Food and kindred products
Electrical, optical, machinery and instruments
Fabricated metal products
Textiles, apparel, wood, leather
Transportation (non-auto), pipelines
Wholesale trade, building material
Financial, insurance
Business services
Educational services, museums, galleries
Government
General merchandise and food stores
Health, legal, administrative services
Hotels, personal service, repairs
Construction, transit, utilities, banking, real estate,
other
* The variable referred to is the employment, totaled in each zone, for the SIC classifications listed in the
next column (DWELL is an exception, as described in Colume 3).
" Standard Industrial Classification Code
source: Reference 6
90098 06-1
6-24
-------
<0
9 '5
u> a
TABLE
source
TIT
M « jjs
n * I
M « J
s j i
s I
s J
ri (jj
IN
2 S
s)
* 1
r- ;
u.
•a 1
- £
» j
2 ^
n 5
"" X
N 1
" i
•- 5
** X
f>4
0 p
S
• sr
X
- i
*•» "S
<
- 1
i
1
*^ iii to tn in
O U} O O ^j
o
o o o o
,- m » w
0
o
-•«
.n
Ul w^ if)
2 10 » in 2 2
33 ft . . . ^ ia , ,
0 O 3 o
n n ° U5
2 «, « « g | 8 2 2
<•> n
in n
n n
n n
(O v-1
rt rt
r-/ m
^ ui rfi
S § 2
o o ^
-
0 § 0 0
"
2 8
*~
rt en 10
<^ n in
S S 2
iij,,.Hjj i
iiiiniiilll
to
Reference
source:
6-25
-------
degreasing category into four subcategories, namely, (1) degreasing, MFG4;
(2) degreasing, MFG5; (3) degreasing, RET SERV.; and (4) degreasing, OTHER
SERV.' Then, if the emissions modeler has estimated the total county-level
degreasing emissions for these subcategories as 4.2, 25.2, 8.4, and 4.2 ton
per day, respectively, thase amounts will be allocated in the appropriate
subcategories, using the corrssponding demographic parameter as the
surrogate indicator in each case.
The preceding apportioning calculation assumes that apportioning factors are compiled at
the grid eel! level. In actuality, as mentioned at the outset of this section, the spatial
surrogate indicators (such as the demographic parameters shown in Table 6-8) used for
apportioning are initially' compiled at the zonal level for transportation and other planning
purposes rather than at the grid ceil level. For example, the San Francisco Bay area local
MPO develops its population, land use, and employment data for 440 zones, each of
wnich comprises one cu seven census tracts. 3y contrast, there are some 5,000 grid cells
to which area source emissions ara apportioned for photochemical modeling purposes.
Thus, use of zonal statistics to apportion area source emissions requires that the emissions
modeler determine a zone-to-grid-cel! conversion before completing the apportioning steps.
This step is unnecessary when apportioning factors are manually developed from land use
maps, since in that method the grid system is overlaid onto the land use map and the
values of each surrogate indicator are directly determined for each grid cell by visual
means.
To determine a zone-to-grid-cell conversion, the emissions modeler must (1) overlay a map
outlining the grid system over a map showing the zone boundaries (or perform the
computer-assisted equivalent) and (2) determine or estimate fractions of zonal area lying
within specific grid cells.
A zone-to-grid-cell correspondence table like the one shown in Table 6-10 facilitates this
procedure. For each zone, the area falling in each grid cell is estimated in terms of the
fraction (A) of the grid cell covered by that zone; the total of these fractional areas for all
of the affected grid cells (SA) is the total area of the zone (note the exception that occurs
when part of the zone lies outside the emission grid). The following example illustrates the
calculations involved in determining the zone-to-grid cell correspondence.
> For each grid cell, the appropriate fraction (g) of the given zone is obtained by
dividing the area of intersection by the total area of the zone. The
contribution of the zonal emissions to the grid square can be obtained by
multiplying the zonal emissions (in any or all categories) by this fraction.
30098 Q6': 6-26
-------
M
O
«rf
O
O)
o
o
a
a
(B
O)
03
•a
0)
u
«
•o
i
o
u
•a
•c
03
O
N
E
S
•*•
CO
S.
V
u
x
03
03
(O
Ul
m
<
w
*
"3
o
1-
03
C
O
N
CM
in
~ co to
qd 9
^f"
o
p r- co
qd -.
CO
o
— in o
q d *":
CM*
o
^~
p q co
q ^ "~.
CM*
o
•"""* f*^. C*}
n
q o r
,-"
o
CM 9°>
9 •" ^
o
p q en
9 *~ *"•
o
^"^
u
O < 03
r-
°*
CO
cf d9
CO*
o
q d 9
CO*
o
— CM in
qo'9
CO*
o
— to in
q d 1
CM*
O
*™^
CO ° ^
O '- °i
CM*
O
— •* tn /y*i
CM . L:
q o .
CM"
o
~ q to
o ^™ .
^_"
o
o d 9
^_"
o
"~*
u
0 < en
CM
to
d
.
~ to q
r- O «-
in*
0
^"^
U
O < o
£
CO
d
^- CM r^
«— o .
to* °
o
^d ^
»• d
o
^"^
u
C3 < CD
CO
CO
d
in *° r>"
CM d "*
to* °
O
gr co co
3°°'
o
*"^
o
O < o
CO
CM
"S
0
4-1
o
£.
0)
o
b.
*^
§
o
'c
•o
03
|
CO
V)
ro
.2
~
•—
o
03
S
=
0
^
"5 o)
o _
o
T3 *^
'§)«
£ «- .2
0 o c
_ o
~^J 2? Q.
C t_ Q.
C3 < <
•o u u u
c
03
O3 Q
5 O < cn
6-27
-------
Next, the fractions (g) are multiplied by the known zonal values of each
demographic parameter to aggregate the data at the grid cell level.
Mathematically, this process may be expressed as follows:
a* = I, g,,bjk (6-4)
where a,k is the value of the kth demographic parameter, aggregated to grid
cell i; b)k is the value of the kth demographic parameter, as compiled for zone j;
and g,i is the areal fraction of zone j in cell i. Note that the value of g,, is given
by
g, = A, /1, (A,,) (6-5)
where A(j is the fractional area of intersection of zone j with cell i, in terms of
the fraction of cell i covered by zone j.
To calculate the apportioning factors (denoted by flk) for allocating county-
level emissions to grid cells, the grid cell level values of each demographic
parameter must be normalized to the total for the county, i.e.,
f,k = a,k /1, a,k (6-6)
The same normalizing factor can, of course, be obtained by totaling the zonal
values; that is,
I, a,k = I, b|k (6-7)
except for necessary corrections for any zone which falls partly outside the
county. The apportioning factors (flk) are applied in the same way as the (S, /
ST) factors (determined from land use or other maps) in Equation 6-1 that
were determined from land use or other maps.
The most difficult part of the zone-to-grid-cell conversion process as described above is
determining the gM fractions. This step may need to be performed manually because In
the past, the"often irregular nature of zonal boundaries in most urban areas complicated
the computerization of this assignment; algorithms for calculating the area of irregularly
shaped polygons do exist, however, and can be used in conjunction with modern digitizing
techniques to facilitate this process. The rest of the calculations described above are
readily automated.
> The procedures described above are also applicable for developing
apportioning factors from population density data at the census tract level.
Census of Population and Housing data can be extracted in computerized
90098 off1 6-28
-------
format from the Master Area Reference File (MARF), available from the U.S.
Bureau of the Census and gridded based on the location of the centroid of
each block group enumeration district (BGED). Figure 6-4 shows the locations
of the BGED chondrites for a modeling region used in a DAM application for
Atlanta, Georgia; note that some grid cells, particularly in the urban area of
Atlanta /located at the center of the modeling /egion) contain numerous BGED
centroids, while others in the outlying rural areas contain no centroids. Figure
6-5 shows a spatial density plot of the population data after assignment to
grid colls based en the BGED centroid locations. Although the population data
available from the Census Bureau will usually be somewhat outdated because
of the infrequsncy of data compilation, the emissions modeler can still use
-r'- :::; •"•; '-'j"?!cp apportioning factors for t'r.3 bass year modeling inventory,
:.•:•,:_v*.; :'".:: no significant changes in population density distributions have
v'./ren estimating the g,, fractions, which represent the areal fractions by grid cell for each
"~rr,cgraphic parameter, the emissions modeler should keep in mind the implicit
assumption that the distribution of each demographic parameter is uniform within each
zone. In situations where the zones are much larger than- the coincident grid cells, this
assumption can lead to erroneous distributions if most activity within a particular zone
actually takes place in some subportion of that zone. Hence, before using the values in
the zone-to-grid cell correspondence table to apportion emissions, the emissions modeler
should submit the table to review by local "planners or others knowledgeable with the land
use patterns in the urban area. In select cases, the emissions modeler may elect to
distribute more activity to one or more grid cells than would be assigned based solely on
area. Since zones are defined as areas of similar activity, however, this will seldom require
major consideration from the emissions modeler.
6.3 GENERAL METHODOLOGY FOR TEMPORAL RESOLUTION
Since the basic area source inventory usually contains estimates of annual (or perhaps
seasonally adjusted) emissions, the emissions modeler must expend additional effort to
estimate hour-by-hour emission rates for the episode days. Several approaches can be
employed to develop hourly emissions resolution; all involve the use of assumed diurnal
patterns of activity. In addition to hourly patterns, estimates of seasonal fractions of
annual activity will be needed if the county-level inventory is not seasonally adjusted.
Activity profiles by day of week will also be required.
If the county-level inventory contains annual emission estimates, the first step is to
estimate the seasonal components of activity for each area source. Chapter 6 of Volume I
discusses seasonal adjustment in detail. For many sources, activity is fairly constant from
season to season. Table 6-11 lists recommended seasonal adjustment factors for selected
90C9S 06"
6-29
-------
"»"'I . .," » . | i 1 I , V
. ? ' \ •>
• •* .\-, '
FIGURE 6-4. Location of block group enumeration centroids for the Atlanta, Georgia modeling
region.
90098 08"
-------
• i- i -i •> 1 i fi i I i t t J i i i il i I i t i -i\ 1 i
0
10
20
30
FIGURE 6-5. Sample gridded population data for the Atlanta, Georgia modeling region.
90098 06"
6-31
-------
TABLE 6-11. Ozone season adjustment factors for selected area source categories.
Category
Seasonal Adjustment Factors
Gasoline Service Stations
Tank trucks in transit
Tank truck unloading (Stage 1)
Vehicle refueling (Stage II)
Storage tank, breathing losses
Seasonal variations in throughput vary
from area to area. Use average
temperature for a summer day where
appropriate,
Solvent Users
Degreasing
Drycleaning
Surface coatings
Architectural
Auto refinishing
Other small industrial
Graphic arts
Cutback asphalt
Pesticides
Commercial/consumer
Uniform
Uniform
1.3
Uniform
Uniform
Uniform
0
1.3
Uniform
Waste Management Practices
POTWs
Hazardous waste TSDFs
Municipal landfills
1.2
Uniform
Stationary Source Fuel Combustion
Residential
Commercial/institutional
Industrial
0.3
0.6
Uniform
Solid Waste Disposal
On-site incineration
Open burning
Structural fires
Field/slash/prescribed burning
Wildfires
Uniform
Refer to local regulations and practices
Uniform
0
Refer to local fire conditions
Off-highway Mobile Sources
Agricultural equipment
Construction equipment
Industrial equipment
• Lawn and garden equipment
Motorcycles
1.1
Uniform
Uniform
1.3
1.3
source: Reference 1
90098 06"
6-32
-------
area source categories. If local activity distribution data is unavailable, the emissions
modeler can use these factors to seasonally adjust the emissions in the annual inventory.
The UAM EPS program GRDEMS assigns temporal distribution codes and
monthly activity factors by; source category code. The cross-reference fife used;
to make these assignments requires that seasonal adjustment factors be '• .
-defined as monthly fractions of annual activity.. The seasonal adjustment
factors in Table 6-11 can be converted to monthly fractions for use in this file
in the manner described below x " . ; . '
For each month of the ozone season (usually summer), the monthly fraction of
annual activity for a particular source category will be the seasonal adjustment
factor for that category divided by 12, For example, the seasonal adjustment
factor in Table 6-11 for architectural surface coating is 1.3; accordingly, the
monthly fraction will be (1.3) / U2}, or 0.108, For the other months, ;the
monthly fraction can be assumed to be 1/9 of the remaining activity, or [ 1 - (3
x monthly fraction for an ozone season month}) / 9. In the current example,
this corresponds to a monthly fraction of [1 - (3 x 0,108)17 9, or 0.075.
If the Inventory contains seasonal emission estimates, the emissions modeler
should ensure that no additional seasonal; adjustment is applied by using a flat
• •••••'" • ' + ?**• ' :
monthly variation profile for all source categories {i.e., each month factor would
;be W / (I2},!or 0,083. ,: : .;.'...•:....'..:.;. . \ '.:.,.>....*J ..'-L.-.-V-^V*-<•'••'••••••''• '••••
Once the seasonal adjustment is known, the weekly variation must be determined. Again,
some area source activities are fairly constant from day to day, making it a simple matter
to estimate daily activities. For example, gasoline storage losses and natural gas leaks
would be expected to be uniform over the week. Many area sources, on the other hand,
will generally be more active on weekdays. For instance, dry cleaning plants and
degreasing operations will concentrate their activities during Monday through Friday (or
Saturday, in some cases). In these cases, the seasonal activity should be distributed to
only those days on which the source is active, as shown in the following example.
> Suppose dry cleaning emissions for an entire modeling region are 312 tons of
solvent over the 92-day period from July to September, and most plants are
typically open 6 days a week (for a total of 78 operating days). Daily
emissions from dry cleaning would then be 4 tons (312 / 78). This daily
emission rate would not, of course, be applicable to a Sunday. As explained
in Chapter 2, photochemical models are usually run for weekday conditions.
90098 06" 6-33
-------
After the daily activity level has been determined for each area source, the next step is to
estimate hourly emissions. This is generally accomplished by applying a 24-hour operating
pattern to the daily activity level.
>• Table 6-12 shows an example of source-specific hourly activity data for
gasoline service stations. As seen in this table, more gasoline is handled in
the Tampa Bay area in the afternoon than other times in the day. For
instance, 13 percent of the daily operation in large stations occurs from 4 to 5
o'clock; hence, 13 percent of the daily emissions from large service stations
would be assigned to that particular hour in the modeling inventory.
If the emissions modeler is using the UAM EPS to develop tha modeling -
inventory,- thb type of hourly operating information cart be incorporated directly
into the modeling inventory using the day-specific. Modeling Emissions Record
Format described in the preceding chapter (see Table 5-3),
The hourly operating information in Table 6-12 is an example of-a case where a special
survey has been made to determine diurnal operating patterns. Where resources allow,
this approach is preferable for the more important area source emitters. For many smaller
sources, however, engineering judgment can provide sufficiently accurate temporal
factors. Table 6-13 lists some approaches that have been employed for incorporating
temporal resolution for several area source categories into the detailed emissions
inventory; these temporal variations in activity levels for several area source categories can
be used for temporal distribution in the absence of more specific data. For temporal
resolution, local working hours and seasonal activity patterns may differ from those
suggested in Table 6-13. The most general default option is to assume complete temporal
uniformity. However, it is usually easy to determine whether any important emitting
activity takes place mainly in the summer (as opposed to the winter), on weekdays (as
opposed to weekends), or in the daytime (rather than at night). When such information is
available, it should be utilized, especially if important emission categories are involved.
The development of hourly area source emission estimates from annual emissions requires
a great deal of repetitive data handling, and should generally be computerized. Specific
area source data handling are discussed in Section 6.5.
6.4 AREA SOURCE PROJECTION PROCEDURES
Two approaches can be used for making growth projections of area source emissions. The
more accurate approach involves projecting the activity levels themselves. The more
common approach, however, involves the use of growth indicators to approximate the
increase or decrease of each activity level. As mentioned in Chapter 5, the emissions
90098 08" 6-34
-------
TABLE 6-12. Diurnal patterns for gasoline stations in Tampa Bay, in percent of daily
operation.
Hour
6 - 7 am
7 - 8
8-9
9 - 10
i 'i 0 - i I
1 ' -12 noon
1
12-1 pm
1 - 2
2 - 3
3 -4
4-5
5-6
6-7
7-8
8-9
9 - 10
10-11
11-12 midnight
Type of Gasoline Station8
Small
5
6
6
5
6
6
5
5
7
7 ~
9
9
6
6
5
5
1 •
1
Medium
4
4
5
5
7
7
7
7
6
7
8
8
8
7
3
3
2
1
Large
8
3
ii
M
8
I
7
2
"9
3
9
5
6
13
13
4
4
2
1
' Separate diurnal distributions were analyzed for three classes of gasoline stations:
(1) small, below 200,000 gal/yr throughput; (2) medium, between 200,000 and
500,000 gal/yr throughput; and (3) large, abaove 500,000 gal/yr throughput.
Data are based on 1,133 gasoline stations in the Tampa Bay area.
source: Reference 5
90098 06"
6-35
-------
i
1
1 ";
,~
^
^L i
Cj '
v •
Si
3 i
O
!/)
(0
3)
BJ
£
O
39
3)
O
U)
3)
*ss«
o
iral resolution methodol
o
D.
E
4)
3)
"5.
E
ra
X
01
C*5
TABLE 6-'
;
2
j
a
U
o
?
$
5
S
o"
8
0
O
o
5
-
^
J
"
X
'
?
3
•^
o
S
wa*io» from area to area. Uae average
temperatures for a summer day whera
appropriate
Gaaolina handling
Q
?
t
2
o
o"
§
0
i
c
c
r-
Z
•^
*0
5
Ji
f1
1
i
i
| uniform
If Dry cleaning
O O
0 0
o o
S 9
O '-
£ ;•
_J- _^
3* *
C 0
s.
-^
5
yj
1
-
orm Monda1
1
uniform
i
a
I
9
J
t
*
i
5
o"
o
OS
0
O
o
i
~
•;
3
^
j.
*
s
V*
f
5
1
orm Monda'
1
1 uniform
f
0
a
3
1
!
]
o
j
S
i
5
3
0-
*
o
c
o
o
e
o
—
-_
3
j
j
^
•day througl
8
2
| uniform apting through fall
|| Cutback Mphalt
—
§
*
o
^
2
5
2; i
2
~o
8
1
| coincide! with growing aeavon
S
|f PeaticKJa applicati
-
SO
0
ffl 'T
O 0
O C
o c
o -
C 1
^ .X
* **
O C
E
5
1
E
1
i
I
|! Miscellaneous ioJ
a
*
M
S
it
i
o
o"
o
M
O
i
o
e
a
—
J
t
i j
0 T3
» «
i j
g i
o o
V o
E
!
It Aircraft. g*nerat
..„
•S
8
ealimate on *n individual baaia; contact
airport •utho'inet. Federal Aviation
Admimttration, »iid appropriate military
agenciM
5
1
i
u
f .
•1
0
S
«
»
J
£
5
8'
o
i
r*
O
C
—
^
J
5
E
S
c
o
| uniform throughout the •grjcu'tufil MM
)
i
s
i
s
^ -Q
1 1
G O
§ ®
0 O
si
0 C
o o
— -
e =
5 o
11
•u
5
*
V)
a?
e
2
20% O«c«mt>«r-F«t>ruary
2611 Mvch-Miy
30% Jun«-Augi«l
26% S«plomber-Nov«mb«r
c
i
Construction *qui
O O
o o
- s
2 S
o o
•i ta
0 -
S 5
C ,7
* J
31 ^
^
.5
s:
w
1
5
i
o
I
1
20% December -February
26% March May
30% June- August
26% September-November
H
|| Industrial •quiprrn
o
u
o
s
J
J
o
0
O
o
o
o
b
5
—
~
5
;
>
1
!i
b Monday-F
f> Saturday
SO
U9
mtlform through montha which hava *n
•wereg* temperature of 38°F or higher
c
6
Q.
1
Lawn and gardan
>.
-5
!i
u!
0 0
baaa on monthly off highway fuel uae
|
S
Off-highway mot<
,1
i
J
>.
U
1 i
if
o o
n f»
bate on monthly off highway fuel uae
|| Snowmobile*
-a
(D
C
o
u
6-36
-------
V
•o
_3
U
O
o
u
UJ
_j
CO
"=
=
^
I
•
toutc* C*i«oory
t
I
1
o
0
s
o
2
o
o
^
E
5
1
f
!
!i
ts
il
st
* *
o o
i|
* o
' a"
i!
If
7 ?
5 C
— —
jf
: E
Smatl pleaaura craft
8§
09 pv
«- 0
O O
o o
o o
o -
e E
2 °
# /,
o o
E
o
1
i
5
Httlrotd locomotiv**
§§
a f.
•- o
o o
So
0
O '-
0 1
# JB
^ (N
E
o
1
E
s
Ocean-going and river cargo veaaeli
£
5
5
§
o
1
: i
fjj
O - B
ilf
o 3 &
111
155
E £ *
,! * s
obi
I
j
1
1
1
1C
o o
S 0
o o
Ii
O (N
E E
0 6
f Ji
O O
0) —
f
•o
5
i.
i|
o ^
# W
OB m
* £
l!|
fM
2 1 5
E E J
355
111
' '• 1
£E"
ii ?
S !£ £
J
I
i
c
o
1
i
E
o
u
§§
03 *
o 2
§o
0 o
^ m •-
o - S
e s •
i r j
»* # 5
o o •£
•
|
V)
1
J
1
E
S
1
S
1
Induatrial fu«l combuation
8
0
O
O
1
^
c
3
3
|
t „
i 1
i i
# ^
A 0>
E
S
1
C
a
|
I
1
•6
i
* ?
§
(N
O
§
o
c
3
c
3
5
E
S
1
s ? S
c c £ _
J * • *
o o # o
c
S
A
e
Ftrec* managed burning, agricultural ft
freat control (orchard heatara)
2
3
E
2
1
E
5
•3
S
5
i
5
±
£
;
E
0
1
E
1
_
O
(n
K
WMta management practice* IPOTWi
-------
modeler should consult current EPA guidance documents on projection of future year
emission inventories when identifying appropriate growth indicators for the various source
categories; an updated guidance document concarning emission inventory projection
techniques is scheduled for release by EPA in July, 1991.
The first of the above-mentioned approaches is generally employed when a local survey
has been made or local estimates are available for projecting growth in specific areas.
> If a survey of dry cleaners has been performed and the aversca estimated
growth in the modeling area is 5 percent per year, then ir r v-;:.r^, :!ry
cleaning activity would be projected to increase by a f.•?---•• -- '' •""T1= •• '
(a 28% increase). As another example, a local asphalt trac? -~.i::.; . ;/ ~. .
be able to project cutback asphalt usage.
When considering such estimates, the inventorying agency must recogr,u~5 >•.'>. .,:-•„•> ., - /
of deliberate or inadvertent biases due to wishful thinking or self-serving rr<:,' :•:, ,_:. :,
should strive to obtain opinions which are as objective as possible. The agency ;••' -v,.,;
also be careful to determine whether or not such estimates of future activity levels refiner
the effects of anticipated control measures, an important consideration since sorru-; sue ;
estimates may be more appropriately used in control strategy projections than in the
baseline inventory. Most importantly, any such projections should be consistent with
projections made by other planning agencies.
- '**
A common alternative to directly projecting activity levels is to use surrogate growth
indicators. Use of surrogate indicators was discussed in Section 6.2 with respect to
spatial allocation of area source emissions. In the context of projections, a surrogate
growth indicator is one whose growth in the future is fairly certain and is assumed to
behave similarly to the activity level of interest. The most commonly used surrogate
growth indicators are those parameters typically projected by local MPO's, such as
population, housing, land use, and employment. In ths absence of local projections, the
BEA economic indicators described in Section 5.5 can be used to develop growth
indicators for area sources. Table 6-14 lists example growth indicators for selected area
source categories. The following example illustrates use af a surrogate growth indicator to
project emissions.
> The quantity of miscellaneous solvent use in a projection year might be
assumed to grow proportionally with population. Hence, if population
increased in an area by 10 percent from the base year to the projection year,
miscellaneous solvent usage would be assumed to increase by 10 percent, as
well.
Regardless of what variables are used as growth surrogates, the basic calculation is the
same: the ratio of the value of the growth indicator in the projection year to its value in
90098 W 6-38
-------
TABLE 6-14. Example growth indicators for projecting emission totals for area source
categories.
Sctifca Cit^-iry Growth iitScatora Information Sourcas
Gasohna handling
Dry cleaning
Decreasing
Nc-cu,™ ,,.««.--,""
Cu'.bacK asi'njit
Pssticida 3cpl'C3T'on
M.c^-oc 3 •;; - - :-
Aircraft, military
Agnc'jltur?! equipment
Construction equipment
Industrial equipment
Lawn and garden equipment
Off-highway motorcycles,
snowmobiles, and small pleasure crart
Ocaan-going and river cargo vessels
Residential fuel combustion
Commercial and institutional fuel
combustion
Industrial fuel combustion
Solid waste disposal, on-site
incineration, open burning
Fires: managed burning, agricultural
field burning, frost control {orchard
heaters)
Fires: forest wildfires, structural fires
Gasoline demand, vehicle use (VMTJ, or
peculation
Peculation, retail servica arrpioymwnt
Ind'jstfal tp-cioynflnt
=m-- =>:i->n ~r •esc'jnti3( aweilmg 'jn/Ts
Con^^,: industry and iccai road d6partrn*ni3
Agricultural operations
?:^u 5t en
PrciscT'ons should ba dona cass-by-casa,
crojacTsd lana uso maps may Da useful
Case-by-casa
Agricultural land use, agncult'jra'
amoioymant
Heavy construction employment (SIC ccda
161
Industrial employment (SlOxodes 10-14, 20-
39, and 50-51} or industrial land use area
Single-unit housing QT population
Population
Cargo tonnage
Residential housing units or population
Commercial and institutional employment,
population, or land use area
lndu*tnaJ employment (SIC codas 10-14, 20-
39, and 50-51} or industrial land use
Based on information gathered from local
regulatory agencies
Based on anticipated local regulations as
indicated by information sources
Forest wildfires can be assumed to remain
constant between base and projection years;
project structural fires based on anticipated
population growth
U.S. Department of Transportation, state transoortation
agency, state tax 3gon~y, local MPO
Soivsnt aucolfsr, fsan di-scciation
Tr^da assccor en
Lcszi V^O
Consul! '^austry ana tocai roa^ dgparrments
State depart/ran* of sgr-cuitura, 'oca' MPO
Local MPO
Local MPO
Local MPO
Local MPO
Local MPO
Local MPO
,
U.S. Army Corps of Engineers
Local MPO
Local MPO, land use projections
LocaJ MPO, land usa projections
Local MPO
Local MPO
90098 06"
6-39
-------
the base year is multiplied by the area source activity level in the base year to yield the
projection year activity level.
A major difference between making area source projections for the county-level inventory
and for the detailed modeling inventory is that, in the latter, emission estimates must be
resolved at the grid ceil levei. This adds a dimension of complexity to the projection effort,
since changing growth patterns may require determination of different apportioning factors
for projection years. Fortunately, in most large urban areas where photochemical models '
are employed, the local MPO will be able to provide land use maps as well as detailed
zonal projections of employment, population, etc., for future years. Hence, these
projections can be jssd directly, as described above, to determine changes in spatial
If the surrogate indicators used for apportioning certain area source emissions are not
cro;ected ^t a sc'^ccunty level, engineering judgment must be used to dscide whether
spatial distributions of various activities will change enough to warrant the effort of
identifying new patterns. Changes may be warranted in rapiciiv growing areas for the
more important area source emitters. For regions where little growth is expected, and
especially for minor area sources, the same apportioning factors can be used in baseline
and projection inventories without incurring appreciable error.
In special cases, temporal factors and VOC split factors may change between the base and
projection years. Temporal factors may change as lifestyles and working patterns change,
or as a result of governmental policy.
> • If a 4-day workweek is expected in a projection year, daily emission patterns
from sources such as small degreasing operations may be altered. Likewise, if
gasoline sales are prohibited on certain days or during certain hours, daily
emission patterns may change.
Generally, however, the temporal patterns for most area sources will remain constant;
hence, for these sources, the same daily and hourly apportioning factors can be used in
the base and projection years. VOC split factors are discussed in Chapter 9.
When making area source emission projections, the emissions modeler will have to
consider the effects of control measures for certain source categories. The effect of
controls on area sources can generally be represented by changes in either emission
factors or activity levels, depending on the source and the nature of the control measure
under consideration.
As with point source projection, the area source projections should be carefully reviewed
by the inventorying agency in light of all the points (i.e., objectivity, openness, etc.)
discussed in Section 2.4. In particular, the emissions modeler should verify that
90098 06" 6-40
-------
consistent methodologies were utilized for the base and projection years to estimate and
apportion emissions for each source.
> If emissions from gasoline evaporation at service stations in a base year are
estimated and distributed as a result of a special study (e.g., questionnaires to
individual service stations), it would be inconsistent to estimate such
emissions for a future year based on projected VMT and to apportion these
emissions based on the number of miles of road within each grid.
This type of methodological inconsistency will likely lead to changes in the emissions
inventory that are not due to n-cwtr. or control measures but, rather, to changes ;n the
•nventory procedures •". •/••• \- ••--..
As a test to determine whether di^srant base and projection year methodologies ara
rnutualiy con.--:---. .•:, -; : / .~: ,..;._' . :;.jn >-sr methodology iz cr.2 bass- ysar case and sea
if the results are identical, if important discrepancies exist, then one methodology should
ba chosen for i:r3 for both y-crs. Ger.c-rslly, any methodology which applies growth
factors to basa year estimatss to estimate projection year emissions (or activity levels) will
meet this consistency criterion.
6.5 DATA HANDLING CONSIDERATIONS
The major difference between area source data handling and point source data handling
concerns the way emissions are estimated at the grid cell level. Since point source
locations are typically known to the nearest tenth of a kilometer, it is easy to assign them
to specific grid cells. Area source emissions, however, are typically only resolved to the
county (or equivalent) level in annual inventories and, thus, must be disaggregated to the
grid cell level using the apportioning procedures described in Section 6.2.
Area source apportioning factors can be stored in a special file; Table 6-15 shows a
sample excerpt from such a file. This file basically consists of a matrix of apportioning
factor values by grid cell. In Table 6-15, the surrogate indicators are designated along the
top and the grid coordinates along the side. The values in the table represent the fraction
of the county-level total of each variable located within each particular grid cell. (Such a
table would have to be prepared for each county for which area source emissions are
resolved in the annual inventory.) In order to determine emissions from a particular area
source in a given grid cell, the calculation program (1) determines what surrogate indicator
is appropriate for the source in question (this information would be written into or supplied
to the emission calculation routine), (2) accesses the apportioning factor file to determine
what fraction corresponds to the grid cell/surrogate indicator combination in question, and
then (3) multiplies that fraction by the county-level emission total for the particular area
source.
90098 06" 6-41
-------
TABLE 6-15. Example file of grid ceil apportioning factors for area sources (excerpt)
Grd C~'.l
Coordinates"
272,734
Apportioning Factors for j
. i , . * •> . .'» o c* < o ^ r* ' is ^
w j i Ji«i Oi,S £)<-*
1 1 I
.001
.001
274,764 .00 I | .001
274,784
274,784
280,784
252,786
254,786
256,786
258,786
260,786
.001 .001
.001
.001
.001
.011
.013
.001
.001
.001
.001
.002
.011
.014
.001
.001
.001
.001
.001
.001
.001
.002
,012
.015
.001
.001
.0
.0
.0
.0
.0
.0
.0
.0
.0
.100
SJ51 j 6io'J
.004
.004
.004
.004
.003
.004
.0
.0
.004
.004
•° i
.0
1
1!
,0 !
.0
.0
.0
.045
.270
.009
.0
' UTM coordinates of grid cell, SW corner, Km.
b Apportioning factors for this example are based on the following surrogate
indicators: SI1, employment; SI2, commercial employment; SIS, dwelling units;
SI4, general aviation; SI5, open burning; SI6, vehicle miles traveled.
The entry in each case is the fraction of the total indicated activity which occurs in
the grid cell.
9C093 06"
6-42
-------
Ths UAM EPS program GRD5MS performs ths calculations dasefi&ad above.' A
cross-re ferenca tabfe of area source categories srai spatial surrogate Indicators
determines which indicators are used to allocate ftmlsstorts-ftom wNcIl source-
categories; -& usef-input option atfows 'paHn<| *f differs^ Sttfr'&gates wtb
source categories by county, As »n ex^rnptey-the ussr'caft redefine ths
.surrogate code designations for a rareiy used isnd. use surrogate (such as.
"barren rocky with lichens} to incorporate Specsal spatial apportioning '•' _ ^
Inf erratic* for Cris'Sr n~or.3 counties, sucfi 33'ii«t3?ted lobatiori data for dry
cleamnn est^htisnrn-ems,. Consult the ilsfl'S'jSill^SJS
fi.nn." for
,.."»', ".'"., ~',, '.',•„'.. '~ ' j.',~.~.^~ " j-cj^rfimsnts for tnsspstrGl
The sequerc'-? of s'.-5p::, f'^°c"bed above applies in cases where each area source category
is 3ppo'ficr:r3;i :.:,.;/':' ;;,v/ one surrogate indicator. If more than one surrogate indicator
must be usea to accurately represent the spatial distribution of emissions for a particular
area source category, the same procedure can be followed by creating new subcategories
corresponding to the level of activity to be apportioned by each indicator as discussed
previously. Consider the example from Section 6.2.2:
*• Assume that total area source^degreasing emissions of 42 tons per day of
VOC are estimated to result from activity in four different sectors: MFG4 (10
percent), MFG5 (60 percent), retail service activities (20 percent), and other
1 service activities (10 percent). The area source degreasing category can
accordingly be partitioned into four subcategories, having respective totals of
4.2, 25.2, 8.4, and 4.2 tons per day. Each subcategory would then be
apportioned to the grid cell level by its appropriate surrogate indicator as
previously described. These subcategories would appear in the area source
apportioning factor file, but not necessarily in the emissions inventory
provided by the operating agency.
If area source emissions are spatially allocated based on zonal statistics on population,
employment, etc., instead of land use data, a data handling procedure will be required to
convert the zonal level apportioning information to the grid cell level. As discussed
previously, the first steps in this process are to overlay a map showing the grid system
boundaries over a map showing the zonal boundaries (the equivalent task can be
performed with computer assistance), and then determine or estimate fractions of zonal
areas lying within specific grid cells. These areal fractions are incorporated into a
computer data file to serve as zone-to-grid-cell correspondence values. This file, in turn,
can be used to generate grid cell apportioning factors by (1) multiplying the surrogate
indicator values available at the zonal level (e.g., from forecasting models) by the areal
90098 OS'2 6-43
-------
fractions for each zone, (2) summing over ai! zones, and (3) normalizing, as shown in the
equations given in Section 6.2.2. The latter steps should be computerized because of the
great amount of data handling involved when hundreds of zones and grid cells are
invoivsd.
Estimating hourly area source emissions requires essentially the same data handling
procedures as are descnoed in Section 5,6 for point sources. Basically, a file of seasonal,
daily, and hourly temporal factors must be created (similar in format to Table 5-8) that can
be multiplied by the annual area source emissions to generate hourly emission estimates.
T/picaily, one set of tsrr.porai ccerating factors will be assigned for each area source
cst-g-rv, which a,-;: ",:;:!.,;;:;:^ fcr ;h.-; snti/e modeling area. Determining appropriate
>•-•-•—•-' '-•I*?'" ::;- ':' •? -•'- :->••••;--• -.-->.--;•-{ f3~tcr file is a manual procoduro, as
1,-t..<-•<•- :;.-.'..( -I;•'.:;, •;,:• • ^of-si distribution profiles- for area (and mobile} sources are
;^::j.',-. ', ~, • . oivJi/vjcJ pro^rsm, based on assignments by source category
cc':r,;2;r,:-o m o cross-reference file. The temporai variation codes contained In
«ii3 tUe correspond to those shown m Tables 5-4 and 5-5; additionally., monthly
variation factors are assigned explicitly in this file. The cross-reference fife ;
provided with the UAME EPS contains temporal distribution profites by NAPAP
source category which reflect national average or default data; accordingly, the
default temporal profile assignrnents in this file should be reviewed for local:
, especially for slgmltcant Sources,..;: "•.-...'•--•:••- :•'':' ..- ; ; ^;
90098 08" 6-44
-------
References for Chapter 5
1 • El^cMyj^^JgT-tLgJ^spJ-LajiQn^LglDJssion Inventories for Precursors of Ozone,
Volume I. EPA-450/4-S8-021, U.S. Environmental Protection Agency (OAQPS),
December 1988.
2. NAPAP_( National Acid Precipitation Assessment Program) Emissions Inventory:
Overview of Allocation Factors, 1985, EPA-600/7-89-010a, Alliance Technologies
Corporation, October 1989.
3 . User's GuLd_e_fgr_the IJrpqn Airsnej3jylodeL Volume IV: User^s Guide for the
^Di^il^22.J^^QL^^3-^ilL..^.21^.Ll- = -:'--'--'-.30;'-i--30-CQ7D, U.S. ''Envreri"-'^!?^ Protection
Agency iQACPS', P- --r^c- ":'•*.-?* °'-^. NC, June 1SSO.
• " ".r-^t'.ra Area Sourc? -.rr:~ -..-ps cf -"> \r 7: '"::
Workip.g_Draft:_._Aoger;d]x_A, cPA Contract Mo. 68-02-4254, Work Assignment No.
105, Versar, Inc., Sprn-^'d, Virginia, Mzrcn 1939.
5. L.G. Wayne and P.C. Kochis, Tampa Bay Area Photochemical Oxid^nt Study:
Assessment of the Anthropogenic Hydrocarbon and Nitrogen Dioxide Emissions in
the Tampa Bay Area, EPA-904/9-77-01 6, U.S. Environmental Protection Agency,
Region IV, Air and Hazardous Materials Division, Atlanta, GA, September 1978.
- -st
6. M.C. MacCracken, User's Guide to the LIRAQ Model: An Air Pollution Model for the
San Francisco Bay Area Lawrence Livermore Laboratory. UCRL-51983, Livermore,
CA, 1975.
90098 06'2 6-45
-------
7 MOBILE SOURCE EMISSIONS
7.1 INTRODUCTION
Mobile sources of emissions include moving vehicles such as automobiles, trucks, boats,
and trains. For most urban areas, emissions from mobile sources comprise a significant
portion of total VOC, NOX, and CO emissions for the region. Mobile sources are typically
categorized by the following vehicle types:
o on-road vehicles;
o off-road vehicles;
o aircraft;
o railroad locomotives; and
o vessels.
-»
On-road vehicles represent the registered vehicle fleet used in travel and transport on all
road surfaces and include light duty automobiles and trucks as well as medium and heavy
duty vehicles. Off-road vehicles include all recreational vehicles and machinery used in off-
road situations, such as farm equipment, construction equipment, snow mobiles, off-road
motorcycles, etc. Aircraft, railroad locomotives, and vessels represent all vehicles used in
air, rail, and water transportation, respectively.
As mentioned in Chapter 6, an existing annual or seasonal area source emission inventory
generally contains adequate estimates of emissions for all sources except on-road motor
vehicles. The Urban Airshed Model will be applied for specific episode days. Mobile
source emissions must be computed specifically for those days or adjusted to reflect
conditions on those days. Accordingly, this chapter will focus on the special
considerations necessary to develop a modeling emission inventory for on-road vehicles
(other mobile sources are addressed in Chapter 6). In general, the emissions modeler may
employ either one of two methods to develop the on-road vehicle portion of the modeling
inventory:
(1) compiling an episode-specific on-road vehicle emission inventory using the
methods described in Procedures for Emission Inventory Preparation, Volume
IV: Mobile Sources3 (hereafter referred to as Volume IV); or
90098 07" 7-1
-------
(2) adjusting an existing annual or seasonal inventory to reflect episodic
conditions, as discussed in Section 7.3.
7.2 CHARACTERIZATION OF ON-ROAD MOTOR VEHICLE EMISSIONS
The emission factors used to estimate emissions from on-road motor vehicles vary non-
linearly with a variety of parameters, including vehicle type, vehicle speed, fuel volatility,
vehicle fleet characteristics, ambient temperature, diurnal temperature variations, and
vehicle fleet inspection program characteristics. Accordingly, computer models such as
the MOBILE series of mobile source emission factors, available from EPA's Office of Mobile
Sources (EPA OMS), are commonly employed to accurately determine on-road vehicle
VOC, NOX, and CO emission factors. These emission factors (which are usually reported
in terms of grams pollutant/vehicle mile traveled) are then used with an activity level (e.g.,
VMT) to generate on-road vehicle emissions estimates; ideally, link-specific traffic volumes
and speeds will be used to generate the emission estimates. Various inventory
classification schemes may then be employed to aggregate these emissions into a
manageable number of categories, such as vehicle class, road type, etc., and emissions for
each category will typically be reported as a county total in annual or seasonal inventories.
To facilitate accurate spatial allocation, speciation of mobile source VOC emissions, and
analysis of detailed control strategies, emissions from on-road mobile sources should be
reported by both vehicle type (i.e., light-duty gasoline automobiles, light-duty gasoline
trucks, heavy-duty gasoline trucks, heavy-duty diesel trucks, etc.) and roadway
classification (i.e., local streets and expressways). In addition to categorization by vehicle
type and road class, on-road mobile source emissions should be disaggregated in terms of
component emissions (exhaust, evaporative, etc.). These three types of categorization are
discussed below.
7.2.1 Vehicle Classes
The registered vehicle fleet can be divided into sub-groups, or classes, such as autos, light-
duty trucks, and diesel vehicles. The emission factors associated with each vehicle class
will vary because of differing emission certification standards and pollution control
equipment. The MOBILE models distinguish nine vehicle'classes, as listed in Table 7-1,
based upon gross vehicle weight (GVW) and fuel consumption type (gasoline or diesel
fuel). Inventories will typically use some combination of these nine vehicle classes to
report emissions.
>• For example, the NAPAP inventory utilized four vehicle classes: light-duty
gasoline vehicles (LDGVs), light-duty gasoline trucks (LDGTs), heavy-duty
gasoline vehicles (HDGVs), and heavy-duty diesel vehicles (HDDVs). The
90098 07" 7-2
-------
TABLE 7-1. Vehicle class definitions used by the MOBILE models.
Light-duty Gasoline Vehicles (LDGV)
I ;^» j.,», , o -
LJMI « ^."0 u )./ O
Light-duty G,
Heavy-c',..: • •
Light-outy Oi
Heavy-dutv Oiese' Vef
iVioecrcycias (iV'iC)
*Gross Vehicle Weight
source: Reference 2
not applicable
less than 6500 I'os.
: JCC to 3500 Ibs.
; ••-. ;r :nan 8500 ibs.
not applicable
less than 8500 Ibs.
more than 8500 Ibs.
not applicable
TABLE 7-2. Commonly used road type designations.
Rural and Urban Interstate
Rural and Urban Other Principal Arterials
Other Freeways and Expressways
Rural and Urban Minor Arterials
Rural and Urban Major Collector
Rural and Urban Minor Collector
Rural and Urban Local
90098 or1
7-3
-------
LDGT category is the combination of the LDGT1 and LDGT2 categories listed
in Table 7-1. Combined, these four vehicle classes total 97.6% of total on-
road VMT based on national average data. The remaining vehicle classes
(motorcycles and light-duty diesel trucks and vehicles) were not included in
>•;••),£ iYAPAP inventory, but should be inciudsd in inventories prepared for
7*^ ** r* -,,~ Jl. . - *.* , ~»V , ~,,- -.
.^,,.i. ! lOaUITV tsy * ¥},)i;',"..9
Qn-road mobile source Tnin^irr^ sh.-yin a.'-jo be ^''t'P^u'shed bv road type in the
~v^.".t:"v. P.cod ';';,:-..; .:.' . . -.• ..'.; ,:,;..,,; :.; ,•,:.-,. ,u> Ac,-; .-niotration (FHWA) maintaifs
statistics ors i.ocij .-. ,"_.;:: .'-11, .. -.-- . , c-y^o ->3 o^n-.inonly used in mobiia emission
inventories. Emission factors vvii; vnry by rone tyoe because of the variance in parameters
sucn 33 spccJ snj ,. ,- ... •. , ,.-....". .::~. -.;.- :=o vviin each different road types.
cor e::?r-;:^, :'"-, ' " ,"; •-,? irvaptory reported emissions for three road types
derived from the list of roadways given in Table 7-2, namely urban, rural, and
limited access (separate source categories for suburban roadways are also
provided, but not currently used). Table 7-3 shows the division of the road
types shown in Table 7-2 into the urban, rural, and limited access categories.
In the development of on-road mobile source emissions estimates, NAPAP
used distinct input parameters for ear£h road type.
7.2.3 Emission Components
In addition to the categorization of on-road mobile source emissions by road types and
vehicle classes, as discussed above, emissions from these sources should be
disaggregated into their components. The individual components of the on-road vehicle
emissions are defined below:
o Exhaust emissions: vehicle tailpipe VOC, NOX/ and CO emissions which occur
during the operation of the vehicle.
o Evaporative emissions: VOC emissions which include diurnal emissions and
hot soak emissions. Diurnal emissions result from ambient temperature
changes which occur when the vehicle is not in use. Hot soak emissions
consist of the evaporation of emissions immediately following the end of a
trip.
o Running loss emissions: evaporative VOC emissions which occur during the
operation of the vehicle typically at warm temperatures and low speeds.
30098 07" 7-4
-------
TABLE 7-3. NAPAP road type designations versus Federal Highway Administration
(FHWA) road types.
NAPAP Road Type
Corresponding FHWA Road Types
Limited Access
Rural
Urban
Rural and Urban Interstate
Rural and Urban Other Principal Arterials
Other Principal
Other Freeways and Expressways
Rural'aTid Urban Minor Arterials
Rural Major Collector
Rural Minor Collector
Rural Local
Urban Major Collector
Urban Minor Collector
Urban Local
source: Reference 6
90098 07"
7-5
-------
o Refueling emissions: VOC emissions resulting from vapor displacement and
spillage associated with vehicle refueling.
Exhaust and evaporative emissions must be differentiated because of the different
characteristic VOC speciation profiles for these two categories. Additionally, emission
certification standards and emission controls vary between all four groups, necessitating
separation of the four components.
The Inventory classification scheme provided with the tiAM EPS categories
on-road vehicle emissions by four road typQ designations (limited access, urban,
suburban, and rural}, low vehicle classifications {LQGV, IDGT, HDGT, zr.d
HQDV), and four emission components (exhaust, evaporative, refusing Io3sssr
and running lossesK Consequently, sixtyfour ^four road types x four vehicle
classes x foof emission components) source categories are dsfined for crv-road
vehicles. By contrast, the four non-on-road vehicle types are represented by
erght source categories for off-road vehicles (two fuel types x !
-------
Some primary reasons for developing the mobile source modeling emission inventory by
adjusting an existing annual or seasonal inventory for episodic conditions are
o If emissions from mobile sources do not contribute significantly enough to the
total inventory to warrant development of an inventory from original data (this
is usually not the case);
o If time constraints prevent development of an original inventory; or
o Unavailability of data such as locale-specific VMT data.
Detailed procedures for development of mobile source emission estimates are presented in
Volume IV; this document should be referred to as the definitive guidance for constructing
a mobile source inventory!
If the emissions modeler decides to adjust an existing annual or seasonal mobile source
emissions estimates to reflect episodic conditions, the change in emissions can be
summarized by the following equation:
MEE = MEB-(EFE/EF9)-(VMTE/VMTB) (7-1)
where ME refers to mobile emissions. The subscript "B" signifies the variables associated
with the existing base (i.e., annual or seasonal) inventory and the subscript "E" indicates
the episode-specific variables. In Equation 7-1, the actual episode VMT and the base VMT
need not be determined, but instead can be replaced by a single factor termed the "VMT
factor". The VMT factor represents the VMT change between the episode and the base
inventories. Note that this factor would also incorporate VMT growth if the modeling
inventory is being compiled for a different year than the base inventory. Equation 7-1 thus
becomes:
MEE = MEB-(EFE/EFB).(VMT Factor) (7-2)
Equation 7-2 can then be incorporated into the following steps which are required to
complete adjustment of the mobile inventory. The appropriate sections of this chapter
containing the details of each step are indicated in parentheses:
o Determine the scenario and base inventory emission factors following the
methodologies described in Volume IV for each road type and on-road vehicle
class using (1) the county-level inputs used to generate the existing inventory;
and (2) inputs reflecting actual episodic conditions.
o Determine the VMT factor for Equation 7-2 for each road type and vehicle
class.
90098 0712 7-7
-------
Apply equation 7-2 to generate annual average mobile emissions for each
county within the modeling region.
Spatially allocate emissions (Section 7.5) to produce a gridded inventory.
Temporally adjust the mobile emissions (Section 7.6) to reflect seasonal,
daily, and hourly diurnal variations.
In a iUAM £PS- application, the emissions modeler should; be aware of some .
additional adjustments assumed by the EPS; <1) the on-rGad refueling
cc-m portent of the VOC emissions is- assumed to be included in the gasoline:
'marketing area source category; snd <2) the. on-read VOC emission total for
each vehicle class is- assumed to not be disaggregated into exhaust,
evaporative arsd running Josses, ' ....
Gasoline marketing emissions are adjusted by first removing the percent of the
gasoline marketing total associated with refueling emissions. :lh the WAPAP:
inventory, this percent is 47,4 %, The mobile source refueling emissions are
estimated from the exhaust VOC' emissions using the following equation:
i; Refueftng Cmtsstons - Exhaust VOC
-...':- ••'••• .. • : . •..-'' •• ''•** ... • ; '••'-,. .''••'.
EF In thisi equation refers to the episodic; emission factor for refueling $£B and
' ' ' • " '" "'" """' " '""
The other VOC components {exhaust, evaporative and running losses) are
determined by applying the following fraction to the total VOC mobile source-
emissions: : ; •'-.-'.•' ' .•" ;
the subscript:*!* refers to. In turn, the exhaust, evaporative,; and running1
tosses compor>ents; the subscript- wTtf refers. to the totaf VOC emissions and; ;:
.emission factors-* From the component totals in Equation. 7-4,; Eo^ato'on 7-2.can.
then be. used to determine the scenario, emission totals.. . ,..;.,. ....... :...:.. • ...... .;,..:;...;...
The. PHEGB0 module of the UAM EPS adjust^ the annual or seasonal mobile . . .
source inventory for episodic conditions.. The emission factor ratios and the
VMT growth factor: shown in equations 7-2, 7-3,. and 7-4 must be calculated
outside of £PS to construct the .motor vehicle factors table (also known as the
*mvf acs* tabled which is required :input to PHEGRO. - .
90098 or2 7-8
-------
7.4 MOBILE SOURCE EMISSION FACTORS
The EPA's MOBILE series of models calculate VOC, NOX, and CO emission factors for the
on-road vehicle fleet. The current version of this model is MOBILE 4.0; an updated
version, MOBILE 4.1, is scheduled for release in May 1991, and should be utilized after
that date. These models incorporate the data from the EPA Surveillance Program designed
to quantify and characterize the emission factors encountered in the on-road vehicle fleet.
MOBILE 4.0 and 4.1 are run-with a single input file containing both model parameters and
user supplied information. Table 7-4 lists required input parameters. In addition to these, a
list of optional parameters is given in Tabie 7-5, The complete definitions of the input
parameters and the required format of the input .:i!o ohcuid be obtained from the user's
guide for thy 'nod;:, "vaiiobla from E?,A. CcrvJ: V: ;•_."'.: V 'cr guidance on der^'moinc
appropriate vaiuas for these inputs.
If the srrrssicrs rrods'sr is adjust'*"'" "<~ 3":::-t;r;7 'nvsn1'?"'/, hs must ducliccte ths e.TJasicn
factors used to gansrate the existing nmbiia scurcs inventory. It is important to get the
most accurate and complete information available from the documentation of tne data
base; if necessary, contact the developer of the existing inventory.
Version 2 of the NAPAP inventory was compiled using MOBILE 3.9 with
supplemental adjustment factors made to produce emission factors equivalent
to MOBILE 4.0. The input parameters used are shown in Table 7-6. For
temperatures, the NAPAP inventory utilized the annual average ambient
temperature of the state with a 20 degree diurnal temperature spread. The
list of annual average temperatures used is shown in Table 7-7.
Considerations For Future-Year Emission Factors. The MOBILE 4.0 and 4.1 models can be
utilized for calendar years up to 2020; however, due to changing regulations (federal, state
and local) for the on-road vehicle fleet, some modifications are generally required for the
development of future-year emission factors. The following lists some commonly
encountered situations and recommendations for their incorporation into the emission
factor calculations.
Fuel Oxygenate Additives. The addition of oxygenates such as MTBE and ethanol have
been used for CO. reductions in winter months; however, their use has become common
enough that fuel oxygenates are beginning to be used year round. With respect to ozone
modeling, oxygenates generally lower exhaust VOC emissions and raise evaporative VOC
emissions. Thus, it is important to determine fuel oxygen content for both current and
future-year inventories. CO reductions due to oxygenates can be incorporated using a
specialized version of the MOBILE mode! called OXY4. OXY4 is available from EPA QMS,
goose or2 7-9
-------
auirod input
Calendar year
"STiVi volatility clas
Minimum and maximum daily temperature
Base RVP
In-use RVP"
" -J%
In-use RVP start year"
Region altitude
Speeds
Operating modes"
'MOBILE default values are recommended.
"Not always used by MOBILE, but input record is required.
3C093 07" 7-10
-------
TABLE 7-5. Optional input parameters for EPA's MOBILE models.
Alternate basic emission rates"
Alternate vehicle tampering rates"
Fleet Characterization Data:
Fleet registration distribution"
Fleet mileage accumulation"
Diesel penetration rate"
Vehicle class distribution"*
Inspection & Maintenance Programs:
start year
model years inspected
compliance rate
frequency of inspection
test type
special mechanic training
Anti-Tampering Programs:
start year
vehicle classes inspected
frequency of inspection
list of equipment inspected
Refueling Programs (Stage II):
start year
LDGV % system efficiency
stringency %
waiver rates
program type
vehicle classes inspected
alternate credits
model years inspected
program type
compliance rate
alternate credits
phase-in period
HDGV % system efficiency
'MOBILE default values are highly recommended.
"MOBILE default values available (national averages) but local or regional data
recommended.
90098 07"
7-1 1
-------
TABLE 7-6. MOBILE 4.0 modeling parameters used in the NAPAP inventory,
Parameter
Value or Source Used
Calendar year:
ASTM volatility class:
Ambient temperature:
Minimum temperature:
Maximum temperature:
Base RVP:
In-use RVP:
In-use RVP start year:
Region altitude:
Basic emission rates:
Vehicle tampering rates:
Inspection & Maintenance Programs:
Anti-Tampering Programs (ATP):
Stage II Refueling Programs:
Onboard VRS (Vapor Recovery Systems):
Speeds:
Urban:
Rural:
Limited access:
Operating modes:
Urban:
Rural:
Limited Access:
Fleet Characterization Data:
Fleet registration distribution:
Fleet mileage accumulation:
Vehicle class distribution:
1985
See Volume IV
See NAPAP documentation
13.7°F below ambient.
6.3°F over ambient
11.5 psi.
not used.
not used.
See Volume IV
MOBILE 4.0 defaults.
MOBILF4.0 defaults.
County level I/M data.
not modeled.
n/a in 1985.
not modeled.
19.6 mph.
45.0 mph.
55 mph.
20.6, 27.3, 20.6.
7.0, 5.0, 7.0.
0.0, 0.0, 0.0
Used county level registration
data from R.L. Polk.
MOBILE 4.0 defaults.
MOBILE 4.0 defaults.
90098 07
7-12
-------
TABLE 7-7. State annual average temperatures used in the WAPAP inventory.
Stats
Alabama
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
District of Columbia
Florida
Georgia
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Annual
Average
£ i ~ i
67.5 !
/-,«.->
W L-< , O ;
65.1
53.9
52.2
56.2
55.9
67.3
60.7
54.5..
51.8
53.6
50.2
54.1
56.9
65.4
49.9
56.6
53.0
51.4
48.3
63.2
56.2
50.0
- " „ „, , *
It *l i -J VJ ' VJ '-t ' * — ".
Nevada
I* ' ; v - •' , r, '.~ 2 ^ I r 3
New j-arssy
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
Annual"™"1
Avera?^
Tom p. I
C
-------
but only adjusts CO emissions. To determine VOC adjustments, the emissions modeler
should consult the EPA guidance document on fuel blends.4
Clean .Air ..A c t. A me n d me n t s. The Clean Air Act Amendments of 1990 address changes in
the on-road motor vehicle fleet for ozone non-attainment areas. Included in the
Amendments are discussions of the following:
o inspection and maintenance programs implementations and upgrades;
o stage !! refueling programs;
o fuel reformulation Jirsc'urlinq ^V3 ccr."/"•'';, fuel ?dd;t:ves and fuel
composition); snd
o new vehicle emission certification sto^'l^rds.
The first two of these modifications can ca croc—led directly with the MOBILE rnodais. For
fuel reformulation effects, RVF limits are moceied; consult Reference 4 for fuel additivo
effects. Fuel composition changes generally will not affect emission factors, but vvn!
affect the VOC speciation profiles associated with the on-road vehicles, as discussed h
Chapter 9. New emission certification standards can be incorporated into the model;
however, the EPA QMS should be consulted for recommendations on the proper procedure
for their incorporation.
Transportation Control Measures (TCMs). TCMs consist of a wide array of control
measure which are designed for the reduction of traffic congestion. TCMs are being
studied and implemented in many CMSAs. For the incorporation of TCMs into the
emission inventory, a VMT reduction or speed adjustment needs to be estimated for each
TCM.
7.5 SPATIAL RESOLUTION OF MOBILE SOURCE EMISSIONS
Mobile sources differ from most other area source categories in that their spatial variation
is more accurately described using a link-based rather than a surrogate-based gridding
procedure. Link-based spatial allocation results in distributing emissions only to those grid
cells that contain transportation pathways (i.e., limited access roadways, railways,
airports, shipping channels, etc.); this approach is usually used in conjunction with
surrogate allocation to complete the spatial resolution of the mobile source inventory. The
methodology necessary to incorporate links into the emission inventory is given in the
following section (7.5.1), and the allocation of mobile emissions through surrogates is
discussed in Section 7.5.2.
90098 07" 7-14
-------
7.5.1 Link Surrogates
Links are used in the spatial allocation of on-road limited access vehicle, railroads, airports,
and vessels emissions because the transport routes used by these vehicles are easily
identifiable and can be modeled using a series of lines or links. Using links provides a more
accurate allocation of emissions for these vehicles than the allocation of these emissions
through other particular surrogates such as population, commercial land-use, etc.
Figure 7-1 illustrates a typical link with respect to 3 grid cs!! cf an ir.vsntcrv. Each link is
associated with a start and end point, and the length of the link in each nrioi ceU needs to
be determined. In Figure 7-1, the link shown will have three lengths c"!cu!atsd, the length
within the grid ceii, the i3r;;-.n 'v;:r':n -r.i csil above th? £,••;; •::•:, :~ : . ••; ;".2:*: ::: :hs link
in the grid cell to ths ri^ht cf :r~ jrid c3ll.
Aftsr locating all of the iir,!-.; ./ ..'.3 inventory, :ha c.icc,,::','/. c. c~,.:,:, .:;^, .,/r.,c3ic~£ tc
gridded emissions is performed by the following equation:
MEcall = MEcoumy • (LINK,,,, / LINKcoun^) (7-5)
where ME is the mobile emission totals, the subscript county refers to county totals, the
subscript cell refers to grid cell totals, and LINK refers to the total length of each link in the
subscripted domain.
- '&
In the generation of links for modeling region, the following parameters need to be
determined for each link:
o Type of link: The different types of links correspond to the different vehicle
types using the link, thus the different types are limited access roadways,
railroads, airports (i.e. runways), and vessels (rivers, sea ianes, and/or ports).
o County designation: Links need to be classified by county so that county
totals can be determined. Digitized link segments should end at county
borders.
o Begin and end point coordinates: The coordinates of the link need to be in the
same units as the grid of the modeling region.
The first step in the generation of these parameters is to obtain a map with the proper
links identified. The U.S. Geologic Survey (USGS) maps are a good source which clearly
indicate railroads, airport runways, rivers and ports. USGS maps also use the coordinates
of UTMs and latitude and longitude. It may be difficult, however, to determine motor
vehicle roadway types from USGS maps, requiring the use of a more detailed street map.
Note that no standard coordinate system is usually identified on street maps; reference
90098 OT1 7-15
-------
v
Sy
S4
Sx
S3
S2
(BxBy)
Sx+^S
Grid size (eel sJcia)
FIGURE 7-1. Depiction of Typical Link and Inventory Grid Cell.3
90093 or'
7-16
-------
points on the etrest rncp ~vj3t ba ;d-~ntih3d (gccc! rsfsrsnca pcints are county line
intersections) whose coordinates can be determined from a USGS map to enable accurate
conversion of the street map locations to the modeling coordinate system.
To facilitate the determination of the links coordinates, it is highly recommended that a
electronic digitizer be used to inap out the start and end points of each link. A digitizer is
an electronic sensor that can translate any position on a digitizer board into numerical
coordinates. Thus, any coordinates of a map can be converted into the numerical
coordinates of the digitizer by simply attaching the map to a fixed location on a digitizer
board and moving the sensor to each link start and end point icc.^'on. In addition, two
rsfefsncs points ars neadad for i;;j cc;r, arsicn o; if a JlyiJj^: ^^^,^l,'i^l~ j^^m anc :f;e
cc'"'d:"3t'2 systam cf tu9 "ir-l ~::-;; r;c:c:".
7.5.2 Men-link Mobile :;r:-::::
Non-link surrogates are smplovsd for the soa;;^, a,;occ,:;on o(' rnooiio emissions for the
following situations:
o Links are too i.._merous to define and process. This is typically the case for
on-road rural and urban vehicles and off-road vehicles.
o Emission totals are too insignificant to warrant the development of links; this
occurs in some applications for railroad locomotives and vessels.
o Use of land-use based spatial surrogates provides a more accurate allocation
of vehicle emissions. Such is the case when vessel transport occurs
uniformly over a wide waterway.
The procedure for the allocation of emissions according to land-use data is outlined in
Section 6.2.2 and is directly applicable to non-link surrogate situations.
Tfre UAM EPS will use the surrogates shown in Table 7-S irt the absence of link
data, and in most UAM applications a combination of link and tend-u?g
: surrogates are used.for the spatial aliooation of mobilejsourcs emiasfons. As an
example, Rgure 7-2 shows the development of links for the Dallas/Fort Worth
; area;. In this- figure, 131 links were developed for tile allocation of UmltetJ
: . access roads and oommerbial airports,. Figure 7-3 sho\ivs the gridded mobile
; inventory from tNs alfocatfort, using- the swrogatee Indfcated In Table 7-6 for
tiiose categories for which links were not developed, . . '-...-
90098 OTJ 7-17
-------
"ABLS 7-3, Land-u-73 sur^or*";"*:';" "~i-
'.rr.iscicns In the abssns* nf iin'c '!rt?,
Mobile Source Category*
On-road vehicles (urbc-,:
On-road vehicles (rural)
Off-road vehicles (a!! cypes)
Airplanes (civil and military)
Railroad locomotives
Vessels (all types)
L^nu-'j?^ Surroaats
Rural
Rural
County area
Urban
Water
For on-road vehicles on limited access roadways and commercial aircraft, no surrogates
other than links are recommended.
source: Reference 5
90098 07'
7-18
-------
00 640 680
T ! 1 i 1 1 T I 1 if"! I 1 I ! !~
720
760
30
20
10
j f
i /;
~^a-««-«». !--_..
3721
~\
il 3541
\i I t-'; i
0
I I ! I 1 t I X I I 1 .'I | [ |
3601
10
20
30
40
^3561
FIGURE 7-2. Mobile source link surrogates developed for a UAM application of the
Dallas/Fort Worth region.
90098 07"
7-19
-------
720
760
3720
H 3680
HCJ600
40
3560
DAL1AS WEEKDAY MOTOR VEHICLE EMISSIONS
NOx (kg/day)
Total - 320702.60
FIGURE 7-3. Gridded annual average mobile source emissions for a UAM application of
the Dallas/Fort Worth region.
SCG98 07"
7-20
-------
7.6 TEMPORAL RESOLUTION OF MOBILE SOURCE EMISSIONS
Temporal adjustment of the mobile source inventory into monthly, daily, and hourly
specific totals is not significantly different than the treatment of other area source
categories. Accordingly, consult Section 6.3 for recommended procedures for temporal
adjustment of mobile source emissions. As a special consideration for weekend
inventories, note that diurnal variations in weekend driving activity generally differ
markedly from weekday patterns (which typically exhibit a "double-peaked" profile, with
the most activity occurring during the morning and afternoon commute hours).
Accordingly, the emissions modeler should be careful to select a diurnal variation pattern
for on-road motor vehicle emissions which is appropriate for the modeling episode. If
hourly vehicular soeeds and ViViT distributions are available from the local Metropolitan
Planning Organizations (MPOs), these should be utilized in estimating hourly mobile source
emissions.
The UAM EPS daily and hourly variation codes are shown In Tsblas 5-4 and 5-5
respectively. In Table 5-4, the default assignment for dafly variation of mobile
emissions is code #23, In Table S-5, the default assignment for diurnal
variation of all on-road source categories is code #43 for weekends and code
#50 for weekdays. It Is extremely important In any UAM application to check
; these values with any locale-specific data to determine their applicability. In
• j '. addition, the:UAM EPS assumes a flat profile for monthly variation of on-road
- • :j • mobile emissions. This is not characteristic of, any. particular region,. ...but. is an
\ Indication of the wide variance of monthly factors depending on location.
. . Monthly VMT adjustments should be determined and directly Incorporated Into .
the VIVTT calculations of Section 7,5.
90098 or2 7-21
-------
References For Chapter 7
1. Compilation of Air Pollution Emission Factors, Fourth Edition and Supplements, AP-
42, U.S. Environmental Protection Agency, September 1985.
2. User's Guide to MOBILE4, EPA-AA-TEB-89-01, U.S. Environmental Protection
Agency, February 1989.
3. Procedures fo_r_lnventory.^ret3aratio_nf Volume IV: Mobile Sources, EPA-450/4-81-
026d (Revised), U.S. Environmental Protection Agency, July 1989.
4, Guidance on Estimating Motor Vehicle Emission Reductions From the Use of
Alternative Fuels and Fuel Blends, EPA-AA-TSS-PA-87-4, U.S. Environmental
Protection Agency, January 1983.
5, User's Guide for the Urban Airshed Model, Volume IV: User's Guide for the
Emissions Preprocessor System, EPA-450/4-90-007D, U.S. Environmental Protection
Agency (OAQPS), Research Triangle Park, NC, Juns 199Qr
6. Anthropogenic Emissions Data for the 1985 NAPAP Inventory, EPA 600/7-88-022,
Alliance Technologies Corporation, November 1988.
30098 or2 7-22
-------
8 81GGEN1C EMISSIONS
8.1 INTRODUCTION
In rec6nt y6ars, B\T cju^'ity rnod*?i?rs h?v/e bsoun to
(naturaily occurring emissions from vegetation) can cont/'ibu'ce sionificantiy ro the io?si
emission inventory, even in predominantly urban regions. Some or tn^ss "nt-jr^iiv
occurring orga":c ;.;?.; -„;• ars quire pho^cc,". •;•"" ':-,.rv r~~r;:;- •> - , •
the modeling inventory must include sorr? est;"n~r= -~* '••:•; ~
completeness.
In a collaborative effort, researchers at Washington State Univar^cv :ro "-.-'--. . . •• ;
developed a computerized system to sctirrvcf? hc-jriv gric^ci ";- ----- 1 - -------- ::•"-. ~rr~,:~
system, called the Biogenic Emissions Inventory System (BEiS), 13 avaii:;b!3 co r.h^ public
from EPA. Much of the following overviev of the BEiS is taken from the paper
"Development of a Biogenic Emissions Inventory System for Regionai Scale Air Pollution
Models"1 (this paper is reproduced in its entirety in Appendix D of the EPS User's
Manual2).
8.2 OVERVIEW OF THE BEIS
The BEIS estimates biogenic emissions based on various biomass, emission, and
environmental factors. In general, the basic equation for these calculations can be
expressed as
ER, = Z,[BF, • EF, • F(S,T)] (8-1)
where ER is the emission rate (in grams/second per model grid ceil), i is the chemical
species (such as isoprene or monoterpene), j is the vegetation type, BF is the leaf biomass
factor (in grams/square meter), EF is the emission factor (in micrograms/gram-leaf
biomass/hour), and F(S,T) is an environmental factor accounting for solar radiation (S) and
leaf temperature (T).
The BEIS produces one output file: a binary UAM-format low-level emissions file. This file
contains hourly gridded biogenic emission rates for olefins, paraffins, isoprene, aldehydes,
NO, and N02.
90098 Off2
-------
8.2.1 Lear Biomass Factors
The leaf biomass data base used by 8513 was dsnved from Sand use data in the Oak Ridg
Laboratory's Geoecoiogy Data Base. The land use data base is resolved at the county
level and includes acreages for forest types, agricultural crops, and other areas such as
urban, grassland, and water. Each of the forest types in the land use data base is
assigned to either oak, other deciduous, or coniferous. The leaf biomass for each forest
group is partitioned into four emission categories: high isoprene deciduous, low isoprene
deciduous, non-isoprene deciduous, and coniferous- T?»h|o 8-1 show the b>o
for each forest group.
BEiS 3'3£ioC<"!u'i v .iJJLoCS QiGmj3.3 0 iC'r'J -'^ ..--j , I'Toi 'I '.,•!, .;,".".
steo function. For each month, cio-ciduous ve^-tDT.c" ' -—--,- _.•-- --..-
either full biomass or no biomass. Sines most high ozcns episodes occur :
summer months, this is no: u,:u^i:y 3 cr1":--:- am.rr:;:1 ~
8.2.2 Emission Factors
The emission factors used in BEIS are based largely on Zimmerman's study of biogenic
emission rates in the Tampa/St. Petersburg Florida area.3 These factors for the three
forest groups are shown in Table 8-2. Emission factors are given for four hydrocarbon
species: isoprene, a-pinene, monoterpene, and unidentified.
The Carbon Bond IV speciations for these four species are shown in Table 8-3 (consult
Chapter 9 for a discussion of the Carbon Bond IV mechanism). Multiplying the biogenic
emission factors in Table 8-2 by the biomass factors in Table 8-1 results in the emission
fluxes shown graphically in Figure 8-1 for each of the vegetation types. Figure 8-2 shows
the spatial distribution of standardized biogenic non-methane hydrocarbon emission fluxes
for the contiguous United States.
Some natural sources also emit quantities cf NOX; these sources include biomass burning,
lightning, microbial activity in soils, and ammonia oxidation. Although these natural
sources will generally be much smaller than anthropogenic source emissions in urban
regions, concerns about air quality in rural regions with fewer anthropogenic emissions
suggests that NOX emissions from natural sources should be considered in the modeling
inventory. Because of the lack of sufficiently detailed emission factors for the other
natural source types, the BEIS currently estimates NOX emissions from grasslands only.
90098
-------
TABLE 8-1. Leaf biomass factors (g/m2) by forest group.
Forest Group
Oak
Other deciduous
Coniferous
High isoprene
deciduous
185
60
39
Low isoprene
deciduous
60
185
26
Non-isoprene
deciduous
-60
90
26
IMon-isoprene
coniferous
70
135
559
source: Reference 1
TABLE 8-2. Biogenic emission factors (jjg/g/h) for each biomass emission category,
standardized for full sunlight and 30°C.
Chemical Species
Isoprene
or-Pinene
Monoterpene
Unidentified
High isoprene
deciduous
14.69
0.13
0.11
3.24
Low isoprene
deciduous
6.60
0.05
0.05
1.76
Non-isoprene
deciduous
0.00
0.07
0.07
1.91
Non-isoprene
coniferous
0.00
1.13
1.29
1.38
source: Reference 1
TABLE 8-3. Carbon Bond IV speciation for BEIS biogenic species (moles CB-IV species/mole
chemical species).
Chemical Species
Isoprene
a-Pinene
Monoterpene
Unidentified
Carbon Bond Species
OLE
-
0.5
0.5
0.5
PAR
-
6
6
8.5
ALD2
-
1,5
1.5
-
ISOP
1
-
-
-
Non-
Reactive
-
-
-
0.5
source: Reference 1
90098 Off1
8-3
-------
O
O
o
o
o
o
"
o
CO
o
§
~
O
X
o
o
o
cn
O
O
CO
13
O
Z)
g
o
LU
Q
CC
LU
X
<0
O
CE
LU
O
o
m
oc
LU
O
cc
CO
LU
Q.
LU
GC
13
1
CO
CO
o
LU
CD
CC
o
CO
I
Q
LU
U.
>-
z:
LU
Q
LU
-z.
LU
Q_
CC
LU
O
-z,
o
n
Ul
z
LU
cn
0.
O
CO
03
O.
c
O
"•p
ra
+-
v
OJ
03
X
C
O
£
to
O
O
•o
>.
(B
c
(U
c
o
5"
CO
U4
cc
D
g
U.
CD
o
8-4
-------
o o o o o o
o o o o o o
O LT) O U"> O U~,
O — —
-------
8.2.3 Environmental Factors
Studies indicate that biogenic emissions from most plant species are strongly temperature-
dependent; isoprene emissions also vary with solar intensity. The emission factors used
by BEIS are standardized for full sunlight and 30° Celsius. The BEIS adjusts these
emission factors to account for the effects of variations in ambient conditions using
relationships derived by Tingey.*'5'8 The emission factor sensitivities to leaf temperature
for isoprene and'monoterpene are shown graphically in Figure 8-3.
BEIS also simulates the vertical variation of leaf temperature and sunlight within the forest
canopy. The canopy model employed by BEiS assumes that sunlight decreases
exponents'.!1/ thro-j^h the hypothetical forest canopy; the rate of attenuation depends on
the assumed biomass distribution. Figure 8-4 shows a schematic representation of the
assumed canopy types for deciduous and coniferous forest groups. Visible and total solar
radiation are calculated for eight levels in the canopy and used to compute the leaf
temperature at each ievel; Figure 8-5 presents the assumed temperature and solar flux
variation by canooy layer for deciduous and coniferous forests.
8.3 INPUT REQUIREMENTS OF THE BEiS
The BEIS uses three types of data files: UAM preprocessor data, user-supplied data, and
data supplied to the user by EPA. Figure 8-6 shows a flow chart of the BEIS flow of
information. Each of the input data files is briefly described below. Appendix D of the
User's Guide for the Urban Airshed Model: Volume IV: User's Guide for the Emissions
Preprocessor System2 contains detailed format descriptions of the various input files.
UAM Preprocessor Data. Two of the UAM preprocessor files are also used by BEIS. The
first is the WDBIN file, produced by the UAM winds preprocessor and containing hourly,
gridded surface wind speed data. The second file is the TPBIN file, which is produced by
the UAM temperature preprocessor program. This file contains hourly, gridded
temperature data. These files should be available from the photochemical modeler.
User-Supplied Data. The user must supply two types of data: meteorological and county
allocation data. The meteorological data consists of surface meteorology information on
relative humidity, cloud coverage, and cloud height for one station within the particular
UAM domain. The county allocation data is comprised of records containing the percent
of a given county that is in a given grid cell.
EPA-Supplied Data. The EPA provides two data files for use with the BEIS. The first of
these files is a leaf biomass distribution data base. This file contains, at a county level,,
the following data for the contiguous United States: (1) hectare values for canopy, non-
canopy, and urban tree areas; and (2) canopy biomass density (kg/ha) for oaks species,
90098 OS" 8-6
-------
2..O -
O
,
C2 1 .5 H
cr
cr>
__».,_>
C_J>
O
I . C
H
0.5 -|
3
•1--
o
ISOPRENE
BOO u£/m*/n
— — -3-OO u
200 uE
1 OP uE/ma/h
5 TO
Leaf
1 5
2Q
e ra t LJ re
2.3
2.O -
CD
CJ>
O
.5 -
i .a -
0.5 -
O.O
MONOTERPENE
Alpha — pinene
Mo note rpene/
u n id en tif iecJ
5 TO 15 2O 25 3O
"Leaf temperature (C)
90098 08"
FIGURE 8-3. Emission factor sensitivity to leaf temperature.
8-7
-------
2:
o
o
C/)
:D
O
ID
Q
o
in
0
E
o
o
c
o
ra
c
tO
o.
0>
o
"*3
(0
-------
UJ
a
oo
•*•
iO
CM
c.
V
'.£>
r>uj
au-
00
o
o
-CN
_
U.
o
O
00
Iri i i i i i i i [ i M i i i i i i | ii i i ii i i i | ii i iiiiii MiiiiiiIT
O US ID •* CN O
a
o
c
(0
.a
w
c
_o
'^
.2
M
X
in
ii
2 ii
<=
O -
II
cc
en
CN
II
<£>
KJ
| I I I I I I I I I | I I I II I I I I | II I II I I l I [ I l l I l I I I l | l l i l | l; | i
O 00 IO •* CN
hSCL
^>
Ld
c
(0
-------
UAM PREPROCESSOR
DATA
TPBIN WDBIN
USER -SUPPLIED DATA
METEOROLOGY COUNTY ALLOCATION
(RAWMET) (CNTYALO)
DATA SUPPLIED
TO THE USER
BY EPA
Binary Low-Level UAM Emissions File
FIGURE 8-6. UAM stand-alone biogenics processor. Overview of the Biogenic Emission
Inventory System (BEIS).
90098 08"
8-10
-------
other deciduous species, and coniferous species, by month. The second data file consists
of actinic (spherically integrated) flux data. This data is provided for ten zenith angles for
different wavelengths of the solar spectrum, ranging from 290 nm (ultraviolet) to 800 nm
(near infrared) in increments of 10 nm.
8.4 USER-SPECIFIED LAND USE DATA
One of the major limitations of the BEIS as it is currently implemented is the Sack of
subcounty spatial resolution in tha land use data base used to grid biogenic emissions.
(Note, however, that the finai modei-compatibie emissions file produced by BEiS may
show spatiai variations in emissions at a subcounty levei because *•-:;• fvi •„";->'• ?;:^-:^ •.
gridded relative humidity and temperature data in its environments! 'jej:o.' c?^3c::c^
algorithms.) Two upcoming modifications to the BEIS, scheduled for completion by May
1991, will rectify this limitation, however; these modifications •""•; .'I.::" ;.,,: ':: .' , ;;:
(1) The BEIS source code will be modified to allow the U3sr to -•—:;•/ .jcd:.;;- '_r,_!
use data on a county basis using existing land use categories for both canopy
and non-canopy (i.e., crop) data.
(2) An option will be added to allow the user to use gridded land use data in the
BEIS, utilizing the current land use categories.
8.5 PROJECTION OF BIOGENIC INVENTORIES
In general, the same emissions factors will be used to estimate biogenic emissions for both
basa and projection years. However, the agency may wish to incorporate the effects of
anticipated changes in land use patterns on spatial allocation of biogenic emissions into the
projection inventories if this type of data is available. Changes in land use may also affect
the amount of biogenic emissions contained in the modeling inventory, since the acreage
of forest or agricultural land in each grid cell is used to estimate the biogenic emissions for
that ceii.
*• One case where land use patterns might be expected to 'change dramatically,
thus affecting the amount and spatial allocation of biogenic emissions, would
be if major land development projects (such as new housing developments or
industrial parks) are planned. Consult local planning agencies to determine if
situation exists.
For most applications, however, the assumption that biogenic emissions will remain
constant between the base and projection years will not be a significant source of error.
90098 08" . 8-11
-------
References for Chapter 8
1. Developm_en_t_.of_a. 3io_qenic_E.P"iiss.iof!S inventory SyMeni for R_e,gion?J Scaje Air
P o 11 u t i o n_M_odjis, T. E. Pierce, B. K. Lamb, and A. R. Van Meter, Paper No. 90-94.3,
presented at the 83rd Air and Waste Management Association Annual Meeting at
Pittsburgh, Pennsylvania, June 1390.
2- User's Guide for the Urban Airshed Model, Volume...IV: User^s_Guide for the
Emissions Preprocessor....System, EPA-450/4-90-007D, U.S. Environmental Protection
Agency (OAQPS), Research Triangle Park, MC, June 1990.
3, P. Zimmerman, Determination oi Emission Rates of ..Hydrocarbons..'X'V:lJH:^r:"!l:?'•*.
Species of Vegetation in the Tampa Bay/Petersburg, Fiorica A/sa, c.-'A-oCs-, c-7 / -
028, U.S. Environmental Protection Agency, Atlanta, GA, 1979.
4. D, Tingey, Atmospheric Biogenic Hydrocarbons, J. Bufaiini and ". Arn';.-5, -^ ',.:,., A.'n
Arbor Science Publication, Ann Arbor, Ml, 1931, pp. 53-7S.
5. D. Tingey, R. Evans, and M. Gumpertz, "Effects of Environmental Conditions en
Isoprene Emissions from Live Oak," Planta (1 52): 565 (1 981).
6. D. Tingey, M. Manning, L. Grothaus et al., "Influence of Light and Temperature on
Monoterpene Emission Rates from Slash Pine," Plant Physiol. (65): 797 (1980).
90098 OS13 8-12
-------
9 SPECIAT10N OF VOC AND NOX EMISSIONS
INTO CHEMICAL CLASSES
9.1 INTRODUCTION
Most photochemical models, including the UAM, require that VOC emissions be expressed
in terms of designated groups or "classes" of compounds. Additionally, in some models,
" <''?'••' h^ve to £-2 S"3C!"' = J ~s ftO =nd NQ-. Each mcC'-i's c'-JC'i::cr.t'cn
:'~ v ' '" •""iai/.'h2 1' this 'jroc'i~r focuses en sceciation cf srrvcr,! -r2 £cc.~, ."~ii, — ' tc ti"3 C2"~c."'
3crd IV Mechanism employed by tha UAM.
Ar^GOW 3GND-IV MECHANISM
Tha currently available version of the UAM uses version IV of the Carbon Bend Mechsnisr-
(CB-iV). Since every reaction of all of the organic species found in an urban atmosphere
cannot 03 considered, these pollutants must be grouped to limit the number of reactions
and species to a manageable level while permitting reasonable accuracy in predicting
ozone formation. Each carbon atom of an organic molecule is classified according to its
bond type. As implemented in the UAM, tne CB-IV contains over 80 reactions and more
than 30 species. These reactions and species are tabulated in Appendix A of the User's
Guide for the Urban Airshed Manual, Volume I.1
The differential equations that describe the CB-IV contain wide variations in time (reaction
rate) constants. The UAM uses quasi-steady-state assumptions for the low-mass, fast-
reacting species and a more computationally efficient algorithm for the remainder cf the
state species. Table 9-1 lists the carbon bond species used in the CB-IV version of the
UAM.
ta the UAM EPS, each- carbon atom of total VOC emissions is assigned to one
o? the following ten species listed In Table 9-1: otefinic; carbon bond (OLE},
paraffirifc carbon bonds {PAF?)r toluene TT0U, xytenejXYLh formaldehyde
, ethanol {ETON}, and isoprene (JSOPJ, NO* eirin'ssions are partitioned
into. NO and NQ^ Emissions of CO can also be included in- tile UAM modeling
inventory. = .=. " ;
9009809'' 9-1
-------
TABLE 9-1. Definition of the UAM (CS-IV) species.
UAM Species
NO
N02
03 '
,
Species Name
]
Nitric oxide
Nitrogen dioxide
Ozone '
PAH
TQL
XYL
ALD2
ETH
CRES
MGLY
OPEN
PNA
NXOY
PAN
CO
HONO
H202
HNO3
MEOH
ETOH
ISOP
O-^f^-ffinir^ c & r r^ o n ~ f n H {C* -C*\
, - J J -^ i . , i . i ^ V-f -J i ^, v> ! w « t • \ji \ w W /
Toiusns (C5HS-CH3)
Xyiene (CSH6-(CH3)2)
rcrmaldshyds (CH2 =0}
High molecular weight aldehydes (RCHO, R>H)
Ethene (CH2 = CH2)
Cresol and higher molecular weight phenols
Methyl glyoxa! (CH3C(0)C(0)H)
Aromatic ring fragment acid
Peroxynitric acid (H02N02)
Total of nitrogen compounds (NO + N02 + N2O5 4- NO3)
Peroxyacyl nitrate (CH3C(0)02N02)
Carbon monoxide
Nitrous acid
Hydrogen peroxide
Nitric acid
Methanol (optional)
Ethanol (optional)
Isoprene
source: Reference 1
90098 03"
9-2
-------
3.3 CHEJVIICAL ALLOCATION OP VCC
Generally, the basic annual inventory will contain estimates of either total VOC or non-
methane VOC, depending on what emission factor information is used for computing
emissions. The basic approach for allocating VOC into the classes needed by a
photochemical model is to employ a set of "split factors" that distribute a certain fraction
of the VOC total into each class. A simple example demonstrates this concept:
»• Assume a source emits 10 tons of VOC per day; the so'it factors for this
particular source are 0.2 tons OLE/ton VOC, 0.5 tons PAR/ton VCC, and 0,3
tons ALD2/ton VOC. Sicnpie multiplication of each vector ay Lne cocc'i tonnage
of VOC yields I:h3 '~oir.:i<./ of VCC in each cfiico.". .,., ..• ..>•:.;.-, .*
5, and 3 tons oar day, rsspactiveiy.
This allocation step would, of course, have to be performed ':z: .3:-::;" 3rV :.;• .;
developed in the inventory using different split factors appropriate for each so^re.; or
source category. Please note that the example above is simplified; the U. For example, because of the importance of gasoline evaporation in VOC
inventories, local gasoline composition data should be obtained corresponding
to the summer season in the modeling area (note that liquid composition data
would have to be adjusted to best reflect vapor composition). Additionally,
source tests could be performed to determine VOC species data for each
major facility in the region (refineries, chemicai manufacturers, etc.), and
solvent composition data could be solicited from smaller commercial and
industrial establishments (dry cleaners, degreasers, etc).
From this type of information, a photochemical modeling specialist could determine
appropriate split factors for each source. See Appendix B for additional guidance on
construction of source-specific speciation split factors.
Because of resource limitations and unavailability of solvent composition data, however,
collecting source specific speciation data is generally impractical for all but a very few
large point or area source emitters. An alternative method employs generalized VOC
speciation data from the literature to develop VOC split factors by source type. To
90098 09" 9-3
-------
develop CB-IV split factors from generalized speciation data, the individual chemical
compounds typically present in the emissions from each source type (and their rndecdar
weights and weight fractions of the emissions mixture) must first ba identified. Then,.
each of the chemical compounds present in the modeling inventory must be classified
according to the CB-IV mechanism.
Table 9-2 contains a sample EPA VOC speciation profile; this particular profile provides an
estimate of the composition of VOC emissions resulting from storage of petroleum
products in fixed roof tanks. This table is tsken from the ^
Volume 1: Volatile Organic Compound Species Profiles.2 The
Manual contains over 250 "emission profiles" like the example in Taah 9-2 for vcri-,i:r.
point and area source cate^or^s1 rno manual also conri'rv: ,:•<•: •••2-j for :•"••
motor vehicles and aircraft. !n each profile, individual chemical compos: _:: - : ' • . .
their corresponding molecular weights and weight percentage of tha rnixturj. iTi-:- ; "
speciation profile codes and descriptions are listed in Ap,;snd:x ,-\i,
The type of information contained in Table 9-2 can
assignments for individual chemical compounds from Gu.iris^nes for ijdnn G'-!!';'.j-4 \-/ -h
CBM-IV or QptionaLMechanisms3 to calculate composite ?^!it factors by speciaticn profile.
For each profile, the weight percentages associated with each organic compound are
summed for each carbon bond classification and the average molecular weight of the
mixture computed. The split factors are expressed in units of (moles of carbon bond
species)/(gram total VOC) and represent a weighted composite of the carbon bond class
assignments for each of the chemical compounds present in the mixture. Mathematically,
this can be expressed as
for each i, I, [(WF, / MW,) • (moles of i / mole j)J (9-1)
where i is the CB-IV species, j is each chemical compound in the mixture (e.g., carbon
tetrachloride), WF, is the weight fraction of j in the mixture, and MW; is the molecular
weight of chemical compound j.
If source-specific data is unavailable, the emissions modeler can use the speciation profiles
contained in the Air Emissions Species Manual (with the CB-IV species assignments for
individual chemical compounds) to generate appropriate split factors by source type in the
manner described above. Whenever possible, however, speciation profiles should be
reviewed to ensure local applicability, especially for the major point sources and important
area sources in the modeling region.
. .... '. The Split factors file provided witfr the UAM EPS coma«iS-€B-(Y split factors {in
'••..;• units.of moles of CS-lV species/gram totsf VOC) for sacfc of tfts profiles
. in Appendix A. In addition to ttte split factors for the t«r* VOC carbon kond
90098 09" 9-4
-------
TABLE 9-2. Example VOC speciation profile from ths Air.Em!SSjgn_s__SsecjesJManual (EPA,
1988).
Profile Name: Fixed Roof Tank 0 Crude Oil Production
Profile Nunber: 0296
Control Device: Uncontrolled
Data Source: Engineering evaluation of test data and literature data \
3,-\,~GAD
Number
431 15
43118
43122
43201
43202
43204
43212
43214
43220
43231
43232
43233
45201
CAS
Number
74-48-8
74-48-0
74-49-6
106-69-8
75-52-5
109-96-0
110-05-3
142-28-5
1 11-16-9
71-14-2
Species Name
C-7 Cycloparaffins
C-8 Cycloparaffins
Isomers of Pentane
Methane
Ethane
Propane
N-Butane
Iso-Butane
N-Pentane
Hexane
Heptane
Octane
Benzene
i •* i W i .
Weight
98.19
112.23
72.15
16.04
30.07
44.09
58.12
58.12
72.15
86.17
100.20
114.23
78.11
SUM TOTAL
^.. ^..c
t/y^i.ih-r ;
1.30 ;
0,50
1.50
6.20
5.60
17.60
27.10
1.50
14.60
7.90
9.20
6.90
0.10
100.00
source: Reference 2
90098 09"
9-5
-------
cfessas u$«d'tflf£PSr this file contains factors to split out the
portion- of VOC and disaggregate NO, emissions into MO an-d ?*02, 'Thess split
factors we?0 calculated on a basis of total VOC;. II iha intvd'mctfy conteifis -
estimates of reactive VOC, th>& split factors for each profile' must • bs . ..'...."
fonormstfeed £y removing the non-rsaefcive faerdw. Seetfoft S,w of this chapter
addresses conversion. of split factors to be compatible wfth the inventory. -
spectation profile file, one of the inputs to tha £PS program CENTSMS>
assigns profile codes based on Source Classification Code iSCQ for point ,
sourcss and area source category {A 30} cede fa*" ara^s unit uiwciia ^ureas,
Emissions from sources with SCC3 c; A3Cs ;~c; Hstsd ?n tits sp;o;ct;c<'/ ^".-
ftls are assigned to CS-I'V spacjea usi-ng the overall average soeci5u-!:-!i v*'v"'=i
IEPA speciation protlte cods 0). The default source caragory/spscfatlon cr
pairings in this fits shouid otways ba roviewsd for gpprovn;v;-n~33 f-;r ".".-•
modsting region.
Whether or not the agency intends to employ a model that incorporates the carbon bor.d
mechanism, a photochemical modeling specialist should be consulted to review a!!
procedures, algorithms, and VOC species/split factor data prior to initiating any data
collection or allocation effort. The modeling specialist can also provide valuable advice on
how to deal with other types of VOCs. In no cas*e should the agency develop split factors
or carry out such an allocation without knowing what photochemical model will be run or
what classification scheme is needed to meet the model's reactivity requirements.
9.4 SPECIFICATION OF f^Ox AS NO AND NO2
Soma photochemical models do not require that nitrogen oxides be distinguished as either
nitric oxide or nitrogen dioxide, instead, these models assume that all NOX is NO, which is
the predominant form of NOX.emitted from combustion processes (the primary source of
NOX emissions).
For models (such as the UAM) that do require this split to be made, however, split factors
are applied in the same manner as are VOC split factors. That is, for each source or
source category emitting NOX, two percentages (totaling 100 percent) need to be defined:
one corresponding to the fraction of NOX emitted as NO and the other corresponding to the
fraction emitted as NO2. In this sense, allocating NOX into NO and N02 is analogous to
utilizing a 2-c!ass scheme for allocating VOC.
The mode of expression of the necessary split factors may vary from one model to
another. NOX emissions are ordinarily expressed "as N02", which means that a molecular
90098 09'2 9-6
-------
weight of 46 is attributed to NO as well as N02, even though the true molecular weight of
NO is 30. Many models, including UAM, take account of this convention by accepting
split factors totaling 100 percent for NQX "as NQ2", but some may require NO emissions to
be expressed in terms of the true weight of NO. The true value for NO is 30/46 or 0.65
times the conventional vaiue of NO "as N02.B If the inventory maintains NO totals as NO,
then care should be taken that a molecular weight of 30 is used when computing moles of
NO for use in the photochemical model. It is important, during the planning stages, to
review the annual inventory to determine how NOX is reported and to consult with a
modeling specialist to find out how the photochemical model accepts NOX emissions data.
> Consider a power plant that emits NOX equivalent to 1,000 kg of NO2 per
hour. Given split factor o.{ 35 and 5 percent by weight "as NO:", then NO
and N02 emissions would bd equivalent to 950 and 50 kg "as N02" per hour,
respectively; however, the actual emissions of NO would be only 30/46 of
950, or 620 kg per hour.
At present, few references are available that define split factors for allocating NO, into NO
and N02. Two sources of such data are References 4 and 5. As a rough average, 97
percent (by weight as N02) of the NOX emitted from most boilers will be NO.
Hie C8-IV splits factor file provided with: the .UAMI-EPS assumes default
for NO,,'emissions from all source typesjof 90% an.d,tp;% by'.w6*grit of MO (as
•:N02)'.a?ul NQj, respectively. :^The.actual::•values;fcoritairiedin the splits factor file
'. 'are. determined as
NO: ; {0.90} x | {30 grains MO/mote* / (46 grams MOx/mo!e* I
; =* 0,59 grams NO / gram NO* -...-•'';"•'.
(0. 1 0* x C C46 gram NOa/mote> / {$& grama N0x/mote} | :
• -"'«* 0.10 grams N02/ gram
9.5 PROJECTION OF VOC AND NOX SPLIT FACTORS
Just as the quantity of emissions may change in an area from the base year to any
projection year, the composition of these emissions may change, as well. To reflect this,
different VOC and NOX split factors may need to be used for each projection year, at least
for important sources for which such projected compositional changes can be estimated.
One source for which this may be an important consideration is motor vehicles. Changes
in emissions control technology and use of alternative or reformulated fuels may result in
significantly different VOC split factors for these sources in projection years; emissions
90098 Off3 9-7
-------
modelers should consult EPA for the latest guidance. Similarly, if significant changes are
expected in the compositions of petroleum products transported and stored in the
modeling area, such changes should be reflected in the projection year split factors. Of
course, for many sources, no changes in emission composition will be expected. For
instance, no change would be needed for any sources that will use the same solvents in
the base year and projection years {e.g., dry cleaners using perchloroethyiene). Likewise,
since no evidence suggests that N0/N02 ratios in combustion emissions will change in the
near future, the same NOX split factors could also be used in projection years.
In any case, different split factors for the base and projection inventories should only be
used to reflect anticipated changes in the composition of future emissions. Any other
changes in the split factors used for the base and projection inventories may cause the
photochemical model to predict changes in ozone concentrations that are due simply to
differences in methodology and are unrelated to expected real effects of composition
changes.
9.6 COMPATIBILITY WITH INVENTORY DATA AND SOURCE CATEGORIES
Two major types of compatibility with the emissions inventory need to be considered
when determining appropriate split factors. First, the split factors must be calculated in
units compatible with those used to express VOC totals in the basic inventory. Ordinarily,
this means they should be given in terms of total'VOC, including methane and any other
organic compounds considered unreactive, if such compounds are present in the emissions
for the category under consideration. However, if the basic inventory has been compiled
in terms of non-methane hydrocarbons (NMHC) or reactive volatile organic compounds
(RVOC), the split factors should be given in terms of these totals rather than of total VOC.
The split factors provided with the:UAM iP*S are applicable to emissions
as total VOC {including methane) , l These splits were tabulated on a total VOC
basis so as to be compatible with the EPA VOC spectator* profiles, which ;
.include methane. ff: they, were to he applied {erroneously} to MMHC emissions
estimates, the resulting emission estimates In each VOC class- would be , . ; ,.
'underestimated -since & specified nonreective fraction of the emission^, would :;
'ir the, reactive portion to; earhon bond: classes* .
. In this lease,." the given split factors would have. tdbe adjusted to make them :
applicable to non-methane VOC emissions. This can he readily accomplished.
.by recalculating the average molecular weight of the mixture without methane;
the methane-free molecular weight ts then used with the CS-iV
30098 O9'a 9-8
-------
assignments by chemical.compound to generate non-m«lhans VOC split factor
the method described ir* Section 3-3.
If the existing inventory is not in terms of VOC or non-methane VOC, and ins^au iss
divided into some sort of classification scheme that is incompatible witn tne cnernistry
employed by the photochemical model, it is questionable whether such an inventory will be
useful as input to the photochemical model. Consult a photochemical modeling speciaiist
if this situation exists.
Another important consideration is compatibility c? ^~ ',:"": r~T°rs w'"^ :'r: :v •""
classification scheme. The source categories and succctsgories chosen re: ':':.: :_;..:;c
inventory may fail to distinguish between sources having substantially ciffef-jn: emission
compositions, requiring different sets of solit factors. There are several cc:.:l'~'~ ,::-;T! ~:
to this problem, as shown in the example below.
> Area source degreasing may be considered as a single category in r,-;
emission inventory, but different degreasing solvents are used in ::'i;f'~reot
plants.
First, any individual plants which emit significant amounts of solvent vapors
from degreasing operations can be treated as point sources, in which case the
solvents used at each should be identified and entered in the point source
inventory.
Second, if there are many degreasing operations, each of which emits cn!y
relatively small amounts of solvent vapors, it may be possible to determine
from solvent suppliers how much of each solvent is used in the region.
Lacking any information to the contrary, the agency may then assume that
the emissions of each solvent are uniformly spread throughout the grid ceils
containing degreasing emissions. In this case, degreasing car be traced as a
single area source category, with a composite sat of split factors r3;;act;ng
the proportions used of each solvent; alternatively, it can be subdivided into
several area source categories, one for each solvent, with the emission totals
as determined from the suppliers (Appendix B describes various methods for
incorporation of additional source categories into the modeling inventory).
Third, if there is no locally available information, state or national supply
statistics can be used to provide estimates of the subcategory totals, to be
used in the manner previously explained.
90098 03" 9-9
-------
A sirniiar situation may be encountered in deaiing with motor vehicle emissions; di'farsr.t
speciation profiles are available for exhaust and evaporative VOC emissions, whereas !he
basic inventory may only provide a single lumped estimate combining thsse emission
components. Since mobile sources comprise a significant portion of the anthropogenic
inventory in most urban areas, these emissions should be recalculated in tsrrns of tha two
subcategories and the appropriate split factors applied for each category. This enhances
the accuracy of the VOC allocation process and facilitates the use of the modal to evaluate
control strategies that may affect exhaust and evaporative emissions differently.
ff the emissions rnotfeler is using the UAM £PS to assist msxtefteg i?
development, emissions from on-rosu ^"t-;--.,' "",-"v->;:-s vvi;; na"/:- •:; -r
dissg^regatad into exhaust *rK? e^-'epcriKr/o '!sc—sssttn-g o? dlurr:':?,- :
running Jess components} by the PReGiRD prcs="sm.
As a special consideration when speciating emissions from on-road rnc'or *, .'•
effects of RVP, oxygenated fuel blends, and alternative fuel use on VOC 32-,c:
also be quantified. Specific guidance on the evaluation of these effects i: ?r3
EPA Technical Memorandum Motor Vehicle VOC Speciation for SIP Development.''
Sometimes, compositional information will simply be unavailable in sufficient detail to
permit determination of split factors for many recognized VOC sources. For instance, a
petroleum refinery may be represented in the inventory by a large number of point sources
(having, for example, different NEDS source classification codes) while only a single set of
split factors is available for the entire point source category.
For less significant VOC sources, such as area source fuel combustion, the need for
accuracy in assigning split factors becomes correspondingly less important, since moderate
errors in these small contributions will result in only very small errors in the individual VOC
class totals. A single set of split factors is, therefore, adequate for aii external combustion
sources operating on a given type of fuel, and no subcategories would tie necessary or
useful in this case.
In general, the source category list should be reviewed during the early stages of the
planning effort to ensure that all subcategories essential for proper allocation of emissions
to VOC classes have been recognized and established for data collection. This is
especially important if special surveys or questionnaires are to be utilized, because failure
to retrieve all the necessary information in the initial contact can seriously impair the
productivity of the effort.
90098 09" 9-10
-------
9.7 DATA HANDLING CONSIDERATIONS
From a data handling perspective, allocation of VOC and NOX into chemical classes is
similar to the allocation of annual emissions into hourly increments discussed in previous
chapters. Basically, as described in Sections 9.3 and 9.4, the VOC (and NOX) emissions
from each point source or area source category (including highway mobile sources) are
multiplied by "split factors" to allocate them into classes. A separate file of split factors
like the example shown in Table 9-3 for point sources can be created for this purpose;
alternatively, the split factors can be stored as part of the emissions data records
(molecular weights should be stored similarly).
Note that the file shown in Table 9-3 gives split factors by source category (i.e., at the
SCC level) instead of for individual point sources. This file is very similar in format to the
split factor files which might be used for area and highway mobile sources (for these
sources, the SCC codes in Table 9-3 could be replaced with area source category codes).
Compilation of split factors by source category instead of for individual sources is generally
recommended, since (1) specific split factors will not be known for most individual
facilities, and (2) considerably less file space will be required. One disadvantage of this
approach is the difficulty of representing any available source-specific split factor
information for individual operations.
Estimates of VOC (and NOX) by class can either be (1) computed prior to the generation of
the model-compatible inventory and storeo^i'n the emissions data records or in a separate
file, or (2) computed during the creation of the model-compatible inventory to conserve file
space. A disadvantage of the latter approach is that VOC (and NOX) emissions by class
have to be recomputed each time a model-compatible inventory is created.
In projection inventories, new VOC and NOX split factors can be reflected by changing the
split factor files and applying them to the projected VOC and NOX emission totals.
The method for allocating area and mobile source VOC and NOX into chemical classes is
similar to that used for point sources; as mentioned above, an area source split factors file
similar to the file shown in Table 9-3 can be created for this purpose. Alternatively, the
split factors can be stored in the area source emissions records if space is available. The
split factors are multiplied by the VOC (and NOX) totals to estimate emissions by class.
For mobile sources, separate split factors should be used for each vehicle category for
which an emission total is carried along by the network emission calculation model.
Ideally, VMT and emissions will be calculated separately for each major vehicle type (e.g.,
LDGV, HDGV, etc.), in which case vehicle-type-specific VOC and NOX split factors can be
applied. In some instances, however, the model used to generate mobile source emissions
estimates may only supply composite emissions for all vehicles, in which case composite
split factors will have to be applied based on the fraction of travel by each vehicle type.
90098 09" 9-11
-------
TABLE 9-3. Example "split factor" file (excerpt).
Source
Category"
sec
30600801
30600802
30600803
40300106
40300107
40300152
40300205
30000606
30000608
Pollutant
Code"
HC
HC
HC
HC
HC
HC
HC
NX
NX
Class 1°
SFd MW
34.5 46
13.8 46
5.0 52
13.9 58
13.9 58
1.1 38
1.1 38
85.0 30
85.0 30
Class 2°
SF MW
56.9 61
75.0 81
84.0 87
73.5 61
73.5 61
57.3 68
57,3, 68
15.0 46
15.0 46
Class 3°
SF MW
7.9 71
3.9 72
6.6 72
11.2 72
11.2 72
37.0 31
37.0 31
Class 4C
SF MW
.7 96
7.3 92
4.4 96
1.4 92
1.4 92.
4.6 101
4.6 101
* Source category by SCC code (eight digits)
b Code: HC = VOC, NX = NOX
c Classes are defined as follows:
for VOC: class 1 - nonreactive
class 2 - paraffins
class 3 - olefins
class 4 - aromatics
for NOX: class 1 - NO
class 2 - N02
d Split factor, percent of total, by weight
• Average molecular weight
f Each line constitutes a record, either for VOC or NOX, for one source category
90098 09"
9-12
-------
Since the VOC compositions of the exhaust and evaporative components of mobile source
differ significantly, this distinction (if present in the inventory) should, be maintained
through the VOC allocation process. As discussed in Chapter 7, the MOBILE 4.0 and 4.1
models calculate separate emission factors for exhaust, evaporative, refueling, and running
loss emissions from highway motor vehicles; in the absence of additional data, running
loss emissions can be speciated using the evaporative loss split factors.
90093 03'J 9-13
-------
References for Chapter 9
1 - Usei;s__Guide,,fQr th^,_yrban A for UAM (CB-jVl,
EPA-450/4-90-007A, U.S. Environmental Protection Agency (OAQPS), Research
Triangle Park, NC, June 1990.
2. Air Emissions Species Manual, Voiume I: Volatile Organic Compound Species
Profiles, EPA-450/2-88-Q03a, U.S. Environmental Protection Agency (OAQPS),
Research Triangle Park, NC, April 1988.
3. H. Hogo and M. W. Gery, Guide!ine_g.for Using QZIPM-4 with CBM-iV or Optional
Mechanisms, Volume__1, EPA Contract Mo. 83-C2-4136, Systems Applications
Incorporated, 1935,
4. R. J. Milligan et al., Review of NOx_Emissign Fdcrors for Stationary, Cornousiiori
Sources, EPA-450/4-79-021, U.S. Environmental Protection Agency, SaptGrnbar
1979.
5. T. W. Tesche and W. R. Oliver, Technical Basis for Refinement of SA.I Airshed Model
Inputs for the 26 and 27 June, 1974, Multiple-Day Simulation, Systems Applications
Inc., November 1981.
6. C. L. Gray, Jr., Motor Vehicle VOC SpeciatrQn for SIP Development. Technical
Memorandum to John Calcagni and William Laxton, U.S. EPA Office of Air and
Radiation, Ann Arbor, Ml, March 23, 1989.
90098 09'- 9-14
-------
GLOSSARY
Activity leva! - Any variable parameter associated with the operation of a source of
emissions which is proportional to the quantity of pollutant emitted.
Anthropogenic emissions - Emissions from man-made sources; commonly subdivided into
are«, mobile, and point source emissions.
- - - ••"V'*? .*<•">-•-;..•>->.;,; - •trT~'-:r:ons which era assumed to occur over a given are~ '?"•'-,•
tnan ai a specified poire; often includes emissions from sources considered too
•- "•-.•;•'':--; ' - • -•;':: handled individually in ;h= point "I."? . - - : ,
,-. i ,-><--! .2 r>i.-; «.i.n!~-.;i.->ns - JP ^ if '~sii»< '"vc1, "T'n/"! smissicHS from vecetrjtion.
Carbon bond machsf;J:;rn - A chemical kinetics mechanism in which various hvdroc-rcons
are grouped according to bond type (e.g., carbon single bonds, carbon double
bonds, carbonyl bonds, etc.). This lumping technique categorizes the reactions of
similar chemical bonds, whereas the molecular lumping approach groups the
reactions of entire molecules. "•"
Concentration background - The concentration of a pollutant in the ambient air of a region
as measured by monitors unaffected by sources within the region (i.e., by "upwind"
monitors); also referred to as "ambient background concentration."
Effective stack height - The sum of the actual stack height plus the plume rise. It is
defined as the height at which a plume becomes passive and subsequently follows
ambient air motion.
Emission faciot - A factor usually expressed as mass pollutant/throughput or activity level,
used to estimate emissions for a given activity.
Emission inventory - A list of the amount of pollutants from all sources entering the air in a
given time period. Often includes associated parameters such as process
identification and stack parameters.
Emissions modeler - The person or persons responsible for compilation of an emission
inventory suitable for photochemical modeling purposes (i.e., spatially, temporally,
and chemically resolved).
90098 10" Q-1
-------
Evaporative emissions - Emissions resulting from the volatilization of gasoline and solvents
due to rising ambient temperatures or engine heat after motor vehicle shutdown,
Exhaust emissions - Emissions resulting from the combustion processes associated with
the operation of motor vehicles.
Grid ceil - The three-dimensional box-like cell of a grid system.
Grid layer - The horizontal layer of grid cells. The grid modal may consist of a number of
:•-. -:; - <-\r 3ir quality sirruiatson nnodei that provides estimates of pollutant
concentrations for a gridded network of receptors, using assumptions regarding the
exc-anne o* r.;* b"v"!;en hveothetical box-like cells in the atrncsch?'? "'cc//-- ~^,
'•, Mdiiiema;icai!y, this is known as an "Eulerian" rncc~i icf.
Growth surrogate • A quantity for which official growth projections are known and v/hose
growth may ce assumed similar to that of activity for some source category for
which projections are needed.
Julian date - A method of referencing dates in which days are numbered consecutively
from an arbitrarily selected point (normally^January 1). The form of the date is
YYDDD, where YY is the year and ODD is the day; for example. May 3, 1990 =
90123 in Julian notation.
Land use - A description of the major natural or man-made features contained in an area o
land or a description of the way the land is being used. Examples of land use
categories include forest, desert, cropland, urban, grassland, or wetland areas.
Line sourca - An emissions source whose spatial distribution is best characterized by
assuming emissions occur along a linear path (rather than over an area or at a
specific point). Examples of line sources include on-road motor vehicles, railroad
locomotives, and shipping vessels.
Link - A surrogate generated to model allocation of line source emissions (see "Line
source"). It takes the form of a line, or a group of lines; spatial allocation is
performed on the basis of link length per grid cell.
Lumping - In chemical mechanisms, the stratagem of representing certain compounds by
surrogate or hypothetical species in order to reduce the assumed number of
elementary reactions to a manageable number.
90098 10" G-2
-------
Mobile source emissions - Commonly used to designate emissions from on-road motor
vehicles (as opposed to "other mobile" sources; cf. "Line source"). This general
category includes emissions from different operational modes (e.g., cold start, hot
stabilized, hot start, hot soak, running losses, and diurnal evaporative emissions).
Nitrogen oxides - With respect to air pollutants, nitric oxide (NO) and nitrogen dioxide
(N02) together comprise nitrogen oxides (NOX).
Point source emissions - Emissions which are inventoried as occurring at a specified
location from a specific process.
Plume rise - The height above a stack at which exit gases rise as a result of the buoyancy
effects of the emissions (due either to a temperature higher than the ambient air or
the momentum of the emissions as they leave the stack).
Reactivity - Measure of the tendency of a chemical species to react with other species.
Receptor - A hypothetical sensor or monitoring instrument, usuaUy a unit of a hypothetical
network overlaid on a map of the-area being modeled. Eulerian grid models usually
assume one receptor at the center of each grid cell.
Seasonal adjustment - Adjustment of emissions from an annual to a seasonal level,
" '»
generally based on seasonal variations in activity levels or temperature.
Source - A process or activity resulting in the release of pollutants to the atmosphere.
Source/receptor relationship - A model that predicts ambient pollutant concentrations
based on precursor emission levels. Photochemical models are one type of
source/receptor relationship.
Spatial allocation surrogate - A quantity whose areal distribution is known or has been
estimated and may be assumed similar to that of the emissions from some source
category whose areal distribution is unknown.
Spatial resolution - Allocation of emissions to grid cells based on facility location or the
distribution of some surrogate indicator. (1) The process of determining or
estimating what emissions may be associated with individual grid cells or other
subcounty areas, given totals for a larger area such as a county. (2) The degree to
which a source can be pinpointed geographically in an emission inventory.
Speciation - Disaggregation of total VOC and NOX emissions into the chemical species or
classes specific to a chemical mechanism, such as the Carbon-Bond Mechanism,
employed in a photochemical simulation model.
90098 10" Q-3
-------
Speciation profile - Characteristic mix of chemical species in the emissions from a
particular activity or group of activities, such as natural gas combustion in an
external combustion boiler.
Split factor - The factor by which total VOC or NOX emissions must be multiplied to give
emissions by chemical species or class (e.g., Carbon-Bond species) as required for
use in a photochemical simulation model.
Stack parameters - Characteristic parameters of a stack and its associated plume, as
required for input to some photochemical simulation models. Typical stack
parameters include stack height, inner diameter, volumetric flow rate, and gas exit
velocity; stack parameters are required to calculate plume rise.
Temporal resolution - Disaggregation of annual or daily emissions into hourly emissions.
(1) The process of determining or estimating what emissions may be associated
with various seasons of the year, days of the week, or hours of the day, given
annual totals or averages. (2) A measure of the smallest time interval with which
emissions can be associated in an inventory.
Throughput - A measure of activity, indicating how much of a substance is handled,
produced, or consumed over a given time period.
Trajectory - The path described by a hypothetical'parcel of air moved by winds. The air
parcel is identified as being at a given location at a given time; the trajectory
connects this hypothetical position at any given time with both earlier and later
hypothetical positions.
Trajectory model - An air quality simulation model that provides estimates of pollutant
concentrations at selected points and times on the trajectories of hypothetical air
parcels that move over an emission grid system. Mathematically, this is known as
a "Lagrangian" model (cf. "Grid model").
Vertical resolution - Allocation of emissions to vertical layers based on plume calculations.
In regard to meteorological parameters and concentrations of pollutants in ambient
air, the provision (in a model) of a means for taking into account various values at
different heights above ground.
Volatile organic compounds - Any hydrocarbon or other carbon compound present in the
gaseous phase in the atmosphere, with the exception of carbon monoxide (CO),
carbon dioxide (CO2), carbonic acid, carbonates, and metallic carbides.
90098 10" G-4
-------
APPENDIX A
CODES FOR EMISSION CATEGORIES
1"A 3 U™ A-i. .4ciiviiy cedes.
TA8UE A-2. Process codes.
TABLE A-3. Control codes.
TABLE A-4. Scarce category codes.
TABLE A-5. Hydrocarbon speciation profile codes.
A-1
-------
TABLE A-1. Activity codes used in the emissions preprocessor system,
Code Description
100 Resource Development & Agriculture
110 Agricultural Production
111 Agricultural Crops
112 Agricultural Livestock
113 Agricultural Services
120 Forestry •
130 Mining
131 Metal Mining
122 Coal Mining
133 Stone & Clay (Mining)
1 311
uo
141
200
210
211
212
213
214
215
216
217
220
230
231
240
241
242
243
244
245
260
261
262
263
270
271
280
Chemicals 4 Fertilizer Mineral
Oil 4 Gas Extraction
Liquid uas Production
Manufacturing, 4 Industrial
Food & Kindred Products
Fruit- 'Vegetable Preservation
Grain Mill Products
Bakery Products
Vegetable Oil^
Sugar Mfg. /Refining
Malt Beverages
Wines 4 Brandy
Lumber 4 Wood Products
Paper 4 Allied
Pulp 4 Paper Mills
Chemical 4 Allied
Rubber & Plastics Manufacturing
Drugs
Cleaning/Toilet Preparations
Paint Manufacturing
Agricultural Chemicals
Petroleum Refining/Related
Petroleum Refining
Paving 4 Roofing Materials
Petroleum Coke/Briquette
Mineral Products
Glass/Glass Products
Metallurgical
continued
5 0 0 0 8 23
A-2
-------
Continued.
Code Description
281 Iron/Steel Production
282 Iron/Steel Foundry
283 Nonferrous Metals
290 Misc. Manufacturing
291 Textiles 4 Apparel
292 Furniture & Fixtures
293 Fabricated Metal
294 Machinery -
295 Transportation Equipment '
296 Rubber 4 Plastics Fabrication
297 Tobacco Manufacturing
298 Instruments
-/*- V
321
322
323
330
331
332
333
334
335
336
400
410
420
430
440
500
510
520
600
610
611
612
620.
630
- - ;-...-
G^vlcj 3'cc.ticn3
F,.>; LUI--J
Aisc. S^r-.ic^s
Sce-ia Supply
Printing i Publishing
Laundry & Drycleaners
Sanitary & Water
Health Services
Educational Services
Transportation
On-road Travel '
Rail Transport
Wacer Borne
Air Transportation
Domestic
Residential
Recreational
Misc. Activities
Construction
Building Construction
Road Construction
Natural Sources
Government
continued
90008 28 A-3
-------
TABLE A-1. Concluded.
Code Description
631 National Security
801 Seeps/Biogenic
802 Channel Shipping
803 OCS And Related Sources
804 Tide land Platforms
900 Unspecified Activities
-------
TABLE A-2. Process codes used in the emissions preprocessor system.
Code
Description
100
110
111
112
113
114
120
130
131
140
141
142
200
210
211
220
221
222
223
300
310
320
330
340
350
400
410
420
430
440
500
510
520
530
540
550
551
552
553
Fuel Combustion
Boilers & Heaters
Boilers
Space Heaters
Orchard Heaters
Process Heaters
In-process Fuel
Stationary I.C. Engines
""ursine - Combustion Gases
Equipment-
Util:r7 Equipment
Mobi e Equipment
Waste Burning
Incineration
Conical Burner -
Open Burning
Agricultural Debris
Range Improvement
Forest Management
Solvent Use
Dry Cleaning
Degreasing
Surface Coating
Asphalt Paving
Printing
Liquid Storage & Transfer
Tanks
Tank Cars & Trucks
Marine Vessels
Vehicle Refueling
Industrial Processes
Chemical Processes
Food 4 Agricultural
Petroleum & Related
Mineral Processes
Metal Processes
Primary Metal
Secondary Metal
Metal Fabrication
continued
-------
TABLE A-2. Concluded.
Code Description
560 Wood & °aper Processes
570 Rubber 4 Plastics
600 Misc. Processes
610 Pesticide Application
620 Solid Waste Land Fill
621 WastP Disposal
630 Farming Operations
6^0 Construction & Demolition
650 Roa'-' Travel
651 " ''npaved Road
652 , ;*•: Road
560 !'np -inned Fires
661 to: -1 Fires
662 Structural Fires
700 Vehicular Sources
710 On-road Motor Vehicles
720 Off-road Motor Vehicles
540 Mineral Processes
550 Metal Processes
551 Primary Metal
552 • Secondary Metal
553 Metal Fabrication
560 Wood 4 Paper Processes
570 Rubber & Plastics
600 Misc. Processes
730 Trains
740 Ships
750 Aircraft
801 Seeps/Biogenic
802 Channel Shipping
803 OCS And Related Sources
804 Tideland Platforms
900 Unspecified Processes
-------
TABLE A-3. Control codes used in the emissions preprocessor system.
Abbreviations are defined at the end of the table.
Code Description
99 Unspecified
101 . Utility Boilers - Liquid Fuels
102 Utility Boilers - Gaseous Fuels
103 Refinery Boilers & Heaters - Liquid Fuel
104 Residential Space Heaters - Natural Gas
105 "ssidential Water Heaters - Natural Gas
107 ^on-utility l.C. Engines - Gaseous
103 utility Reciprocal - Liquids
109 Industrial Soilers
111 Glass Melting Furnace
112 Marine Diassl Engines
113 Non-farm Equipment (Diesel)
114 Sulfur in Fuel
116 Utility Turbines - Liquids
117 Refinery Boilers & Heaters - Gas. Fuel
118 Steam Generators - Liquids
121 Pipeline Hea-t€rs
122 Marine Vessels - Combustion
124 Utility Turbines - Gaseous
125 Cogeneration
126 TEOR Steam Generators - Gaseous
127 Non-utility l.C. Engines - Liquid
128 Resource Recovery
129 Boilers-Space Heaters - Liquid Fuel
130 Boilers-Space Heaters - Gas Fuel
131 Utility Reciprocal - Gaseous
201 Flares
301 Architectural Coatings - Oil Based
302 Architectural Coatings - Water Based
303 Architectural Coatings - Solvents
304 Auto Assembly Line - Surface Coating
305 Auto Assembly Line - Solvent Use •
306 Can 4 Coil - Surface Coating
307 Can 4 Coil - Solvent Use
308 Metal Parts 4 Products - Surface Coating
309 Metal Parts 4 Products - Solvent Use
310 Paper - Surface Coating
311 Paper - Solvent Use
continued
-------
TABLE A-3. Continued.
Code Description
312 Fabric - Surface Coating
313 Fabric - Solvent Use
314 Degreasing - nonsynthetic & misc. (Industrial)
315 " Degreasing - nonsynthetic & misc. (Corrmercial)
316 Cutback Asphalt Paving Materials
317 Dry Cleaning - nonsynthetic
318 Dry Cleaning - synthetic & misc.
3'3 Graphic Arts - Except Litho/Leccerpr^s
320 Wood Furniture - Surface Coatings
321 Wood Furniture - Solvent Use
323 Auto Fefinishing - Surface Coatings
325 Ships - Surface Coating
326 Snips - Solvent Use
327 Aerospace - Surface Coating
328 Aerospace - Solvent Use
331 Degreasing - Synthetic (Industrial)
332 Degreasing - Synthetic (Commercial)
333 Flatwood Products
334 Graphic Arts - Litho/Letterpress
398 Other Industrial Surface Coating
399 Unspecified Industrial Solvent Use
401 Gasoline Working Loss - Bulk Storage
402 Gasoline Working Loss - Tank Trucks
403 Gasoline Working Losses - Underground Tank
404 Gasoline Working Losses - Vehicle Tank
405 Fixed Roof Tanks at Refineries
406 Floating Roof Tanks at Refineries
407 Marine Vessel Operation - Evaporative
410 Oil Production Fields Storage Tanks
411 Marine Lightering
412 Gasoline Breathing Loss - Underground
413 Gasoline Breathing Loss - Aboveground
501 Refinery Valves, Flanges, & Seals
502 Petroleum Coke Calcining
503 Sulfur Recovery Units
504 Sulfuric Acid Plants
505 Fluid Catalytic Cracking Units
506 Gas-Oil Production - Valves, Flanges, Connectors
507 Small Relief Valves
continued
-------
TABLE A-3. Continued.
Code Description
508 Non-refinery Valves
510 Vegetable Oil Processing
511 Paint Manufacturing
512 Rubber Products Fabrication
513 Chemical Manufacturing
514 Pharmaceutical Manufacturing
515 Rubber Products Manufacturing
518 Oil Production Steam Drive Well
519 Winer .es
520 Carbon Black Manufacturing
522 Pumps & Compressors
523 Refinery Sewers & Drains
524 Refinery Pumps 4 Compressors
526 Refinery Vacuum System
530 Oil Production - Pump and Compressors
531 Oil Production - Heavy Oil Test Station
532 Oil Production - Cyclic Well Vents
533 Oil Production - Pseudo-cyclic Well
534 Oil Production^ - Sumps and Pits
535 Natural Gas Plant Fugitives
601 Construction 4 Demolition
602 Waste Solvent Disposal
603 Pesticides - Synthetic
604 Roofing Tar Pots
605 Pesticides - Nonsynthetic
606 Aerosol Propellant - Synthetic
607 Aerosol Propellant - Nonsynthetic
608 Waste Disposal Landfill
609 Domestic Solvent Use
610 Aerosol Consum Prod Propellant
611 Aerosol Consum Prod Solvent
612 Non-aerosol Consum Prod Solvent
620 Agricultural Pesticide - Synthetic
621 Agricultural Pesticide - Nonsynthetic
622 Other Pesticide - Synthetic
623 Other Pesticide - Nonsynthetic
651 . Unpaved City/County Road Dust
711 LDA - Exhaust
712 LDA - Hot Start
continued
-------
TABLE A-3.
Code
713
714
715
716
717
7 IS
719
720
721
722
723
724
725
726
727
731
732
733
734
735
736
737
738
741
742
743
744
745
746
747
748
751
753
757
759
761
762
763
764
765
766
767
Continued.
Description
LDA - Hot Stabilized
LDA - Evaporative
LDA - Running Losses
LDA - Crankcase Blowby
LDA - Tire Wear
LD - Refueling
Off-road Gasoline Exhaust
Off-road Gasoline Evaporative
LDT - Cold Start
LOT - Hot Start
LDT - Hot Stabilized
LDT - Hot Soak Evaporative
LDT - niurnal Evaporative
LDT - Crankcase Blowby
LDT - Tire Wear
MDT - Exhaust
MDT - Hot Start
MDT - Hot Stabilized
MDT - Evaporative
MDT - Running Losses
MDT - Crankcase Blowby
MDT - Tire Wear
MDT - Refueling
HD - Exhaust
HD - Evaporative
HDG - Hot Stabilized
HDG - Evaporative
HDG - Running Losses
HDG - Crankcase Blowby
HDG - Tire Wear
HDG - Refueling
DDD - Exhaust
HDD - Hot Stabilized
• HDD - Tire Wear
Off-road Diesel
MCY - Cold Start
MCY - Hot Start
MCY - Hot Stabilized
MCY - Hot Soak Evaporative
MCY - Diurnal Evaporative
MCY - Crankcase Blowby
MCY - Tire Wear
-------
TABLE A-3. Continued.
Code
Description
801
802
803
804
805
806
807
« i™1 Q
W w w
0 r*\ .'"!
-wU J?
811
8'2
813
81^
815
816
817
821
822
823
824
825
827
831
832
833
837
841
842
843
844
845
846
847
851
852
853
854
855
857
861
862
863
"Jon-farm Equipment (Gasoline)
Farm Equipment; (Diesel)
Lawn & Garden Equip (Utility)
Off-road Motorcycles
Pleasure Craft (Boats)
Railroad Line Haul Operations
Commercial/Civil Piston Aircraft
Ccirjr.3rciai Jet Aircraft
Farm Equipment (gasoline)
LDA - Neat - Cold Start
LDA - Neat - Hot Star;
LDA - Neat - Hot Stabilized
LDA - Meat - Hot Soak
LDA - Neat - Diurnal
LDA - Neat - Crankcase __
LDA - Neat - Tirewear
LDA - Cat - Cold Start
LDA - Cat - Hot Start
LDA - Cat - Hot Stabilized
LDA - Cat - Hot Soak
LDA - Cat - Diurnal
LDA - Cat - Tirewear
LDA - Dsl - Cold Start
LDA - Dsl - Hot Start
LDA - Dsl - Hot Stabilized
LDA - Dsl - Tirewear
LMDT - Neat - Cold Start
LMDT - Neat - Hot Start
LMDT - Neat - Hot Stabilized
LMDT - Neat - Hot Soak
LMDT - Neat - Diurnal
LMDT - Neat - Crankcase
LMDT - Neat - Tirewear
LMDT - Cat - Cold Start
LMDT - Cat - Hot Start
LMDT - Cat - Hot Stabilized
LMDT - Cat - Hot Soak
LMDT - Cat - Diurnal
LMDT - Cat - Tirewear
LMDT - Dsl - Cold Start
LMDT - Dsl - Hot Start
LMDT - Dsl - Hot Stabilized
continued
-------
TABLE A-3. Concluded,
Code
Description
867
873
874
875
876
877
883
sa'-i
gflcc
/- 1*1 fj
001
591
892
c 9 i
89*
902
903
999
LMDT - Dsl - Tirewear
HDT - Neat - Hot Stabilized
HDT - Neat - Hot Soak
HDT - Neat - Diurnal
HDT - Neat - Crankcase
HDT - Neat - Tirewear
HDT - Cat - Hot Stabilized
HDT - Cat - Hot Soak
HH~ - '"it - niurnal
HDT • \.at - Tirswear
S-^ps -'Bioger. ; c
f-irv - . Sh : ; r: ng
nC ai : Related Sources
TiJeiand Platforms
Forsst Mnnsgsnzent Control Burning
Wild Fires Control Burning
Livestock Waste
Misc. Control Tactics
Cat = catalytic
Neat = noncatalytic
Dsl - diesel
LDA = light-duty auto
LOT = light-duty truck
LMDT = light-tnedium-duty truck
MDT = medium-duty truck
HDG = heavy-duty gas
HDD = heavy-duty diesel
MCI' = motorcycle
A-12
•3 0 0 0 i 2
-------
TABLE A-4. Source category codes used in the emissions
preprocessor system.
Code Description
100 Fuel Combustion
110 Agricultural
120 Oil and Gas Production
130 Petroleum Refining
140 Other Manufacturing/Industrial
150 Electric Utilities
160 Other Services and Commerce
170 Residential
199 Other
200 Waste Burning
210 Agricultural Debris
220 Range Management
230 Forest Management
240 Incineration
299 Other
300 Solvent Use
310 Dry Cleaning
320 Degreasing
330 Architectural"Coating
340 Other Surface Coating
350 Asphalt Paving
360 Printing
370 Domestic
380 Industrial Solvent Use
399 Other
400 Petroleum Process, Storage 4 Transfer
410 Oil and Gas Extraction
420 Petroleum Refining
430 Petroleum Marketing
499 Other
500 Industrial Processes
510 Chemical
520 Food and Agricultural
560 Mineral Processes
570 Metal Processes
580 Wood and Paper
599 Other
600 Misc Processes
continued
A-13
-------
TABLE A-4. Concluded.
Code Description
610 Pesticide Application
620 Farming Operations
630 Construction and Demolition
6^0 Entrained Road Dust - Paved
650 Entrained Road Dust - Unpaved
660 Unplanned Fires
680 Waste Disposal
685 Natural Sources
699 Other
700 On Road Vehicles
710 Light Duty Passenger
720 Light and Medium Duty Trucks
730 Heavy Duty Gas Trucks
740 Heavy Duty Diesel Trucks
750 Motorcycles
799 Other
800 Other Mobile
810 Off Road Vehicles
820 Trains
830 Ships
850 Aircraft - Government
860 Aircraft - Other
870 Mobile Equipment
880 Utility Equipment
891 Seeps/Biogenic
892 Channel Shipping
893 DCS and Related Sources
891* Tideland Platforms
900 Unspecified Sources
90008 28 A-14
-------
TABLE A-5. Hydrocarbon speciaticn profile codes used in the emissions
preprocessor system. (Based on EPA, 1938)
Code Description
0001 External Combustion Boiler - Residual Oil
0002 External Combustion Boiler - Distillate Oil
GG03 External Combustion Soiler - Natural Gas
0004 External Combustion Boiler - Refinery Gas
0005 External Combustion Boiler - Coke Oven Gas
'•/,'/ Natural Gas Turnir.e
0008 Reciprocating Diasel irusi Engine
0009 Reciprocating Distillate Oil Engine
C011 By-Product Coke Cven Stack Gas
0012 Blast Furnace Ore Charging and Agglomerate
0013 Iron Sintering
0014 Open Hearth Furnace with Oxygen Lance
0016 Basic Oxygen Furnace
0023 Asphalt Roofing - Spraying
0024 Asphalt Roofing - Tar Kettle
0025 Asphaltic Concrete - Natural Gas Rotary Dryer
0026 Asphaltic Concrete - In Place Road Asphalt
0029 Refinery Fluid Catalytic Cracker
0031 Refinery Fugitive Emissions-Covered Drainage/Separation Pits
0035 Refinery Fugitive Emissions - Cooling Towers
0039 Refinery Fugitive Emissions - Compressor Seals Refinery Gas
0047 Refinery Fugitive Emissions - Relief Valves, Liquified Petroleum Gas
0051 Natural Gas
0066 Varnish Manufacturing - Bodying Oil
0068 Manufacturing - Plastics - Polypropylene
0072 Printing Ink Cooking
0076 General Pesticides
0078 Ethylene Dichloride - Direct Chlorination
0079 Chemical Manufacturing - Flares
0085 Perchloroethylene - Drycleaning
0087 Degreasing - 1,1,1-Trichloroethane
0088 Degreasing - Trichlorofluoromethane (Freon 11)
0089 Degreasing - 1,1,2-Trichloroethane
0090 Degreasing - Toluene
0100 Fixed Roof Tank - Commercial Jet Fuel (Jet-A)
0121 Open Burning Dump - Landscape/Pruning
0122 Bar Screen Waste Incinerator
0127 Surface Coating - Varnish/Shellac
— continued
VI
soooa aa
-------
TABLE A-5. Continued.
Code Description
0166 Printing Press - Letterpress Inking Process
0182 Printing Press - Gravure General Solvent
0183 Printing Press - Gravure Printing Solvent
0195 Residential Fuel - Natural Gas
0197 Solvent Use - Domestic Solvents
0202 Solid Waste Landfill Site - Class II
0203 Solid Waste - Animal Waste Decomposition
0217 Coke Oven Blast Furnace Gas
0219 Surface Coating Paint Solvent - Acetone
0220 Paint Solver.-, - Ethyl Acetate
0221 Paint Solvent - Methyl Ethyl Ketone
0222 Surface Coating - Enamel - Cellosolve Acetate
0223 Surface Coating - Varnish/Shellac Solvent - Xylene
0225 Surface Coating - Primer- Mineral Spirits
0226 Surface Coating Solvent - Ethyl Alcohol
0227 Surface Coating Solvent - Isopropyl Alcohol
0228 Surface Coating Solvent - Isopropyl Acetate
0229 Surface Coating Solvent - Lactol Spirits
0230 Fixed Roof Tank - Hexane
0271 Degreasing - Trichloroethylerje
0272 Automotive Tires - Tuber Adhesive
0273 Automotive Tires - Tuber Adhesive White Sidewall
0274 Automotive Tire Production
0275 Degreasing - Dichloromethane
0277 Degreasing - Trichlorotrifluoroethane (Freon 113)
0282 Surface Coating Primer - Naphtha
0288 Surface Coating Solvent - Butyl Acetate
0289 Surface Coating Solvent - Butyl Alcohol
0290 Surface Coating Solvent - Cellosolve
0291 Surface Coating Solvent - Methyl Alcohol
0292 Surface Coating Solvent - Dimethylformamide
0296 Fixed Roof Tank - Crude Oil Production
0297 Fixed Roof Tank - Crude Oil Refinery
0299 Fixed Roof Tank - Cyclohexane
0301 Fixed Roof Tank - Heptane
0304 Printing Press - Flexographic, n-Propyl Alcohol
0305 Fixed Roof Tank - Crude Oil Marine Terminal
0307 Miscellaneous Burning - Forest Fires
0316 Pipe/Valve Flanges
continued
A-16
-------
TABLE A-5. Continued.
Loae .-escriptic:
0321 Pump Seals - Composite
0332 Printing Press - Lithography Inking and Drying
0333 Lithography - Inking ana Drying-Direct Fired Dryer
1001 Internal Comoustion Engine - Natural Gas
1002 Chemical Menufacturing - Carbon Slack Production
1003 Surface Coating Operations - Coating Application - Solvent-base Pair.t
1004 Plastics Production - Polystyrene
1005 Plastics Production - Polyester Resins
1006 Phthalic Anhydride - o-Xylene Oxidation - Main Process Stream
1007 Mineral Products - Aspnaltic Concrete
1008 Rubber and Misc. Plastics Products - Styrene/Butadiene
1009 Plastics Proauction - Acrylonitrile - Butadiene - Styrene Resin
'010 Oil and Gas Production - Fugitives - Unclassified
'011 Oil and Gas Production - Fugitives - Valves ana Fittings - Licuii
Service
1012 Oil and Gas Production - Fugitives - Valves and Fittings - Gas
service
1013 Surface Coating Operations - Coating Application - Water-base Pair.t
1014 Gasoline - Summer Blend
1015 Gasoline - Winter Blend
1016 Surface Coating Operations - Thinning Solvents - Composite
1017 Surface Coating Operations - Coating Application - Lacquer
1018 Surface Coating Operations - Coating Application - Enamel
1019 Surface Coating Operations - Coating Aoplication - Priser
1020 Surface Coating Operations - Coating Application - Adhesives
1021 Degreasing - Open Top - Chlorosolve
'022 Printing/Publishing - Ink Thinning Solvents - Methyl Isobutyl Xetcr.s
'023 Terephthalic Acid/Disethyl Terepnthalate - Crystal-, Separat-, Dryirr
1024 Terephthalic Acid/Dinethyl Terephthalate - Distillation and Recovery .'
1025 Terephthalic Acid/Dinethyl Terephthalate - Product Transfer Vent
1026 Surface Coating Operations - Thinning Solvent - Hexylene Glycol
1027 Ketone Production - Methyl Ethyl Ketone (MEK)
1028 Acetone - Light Ends Distillation Vent
1029 Acetone - Acetone Finishing Column
1030 Aldehydes Production - Formaldehyde - Absorber Vent
1031 Surface Coating Operations - Thinning Solvent - Ethylene Oxide
1032 Aldehydes Production - Acrolein - Distillation System
1033 Aldehydes Production - Acrolein - Reactor Blowoff Gas
1034 Chloroprene - Butadiene Dryer
contir.uec
A-17
90008 23
-------
rABLE A-5. 3cr.tir.uec.
Code ^escriotion
1035 Chloroprene - Chlorocrene Stripper ana Brine Stripper
1036 Secondary Aluminum - Pouring and Casting
1037 Organohaiogens - Ethylene Dichloride - Direct Chlonnaticn -
Distillation Ven
1038 Organohalogens Production - Ethylene Dichloride - Via Oxycniorinaticn
1039 Organohalogens Production - Ethylene Dichloride - Caustic Scrucoer
1040 Fluorocarbons/Chlorofluorocarbons - General
1041 Fluorocarbons/Chlorofluorocarbons - Distillation Column
1042 Fluorocarbons/Chlo':f1 lorocarbons - Fugitive Emissions - General
1043 Acrylic Acid - Quencn Absorber
1044 Organic Acids Production - Formic Acid
1045 Organic Acids Production - Acetic Anhydride - Distillation Column Vent
1046 Esters Production - Acrylates - Ethyl Aerylate
1047 Esters Production - Butyl Aerylate
1048 Cumene Proauction - Cumene Distillation System Vent
1049 Cyclohexane - General _
1050 Cyclohexanone/Cycionexanol - Phenol Hydrogenation Process -
Distillation Vent
1051 Vinyl Acetate - Inert Gas Purge Vent
1052 Vinyl Acetate - C02 Purge Vent
1053 Vinyl Acetate - Inhibitor Mix Tank Discharge
1054 Vinyl Acetate - Refining Column Vent
1055 Organic Chemical Storage - Methylanyl Ketone
1056 Ethylene Oxide - Oxygen Oxidation Process Reactor - C02 Purge Vent
1057 Ethylene Oxide - Oxygen Oxidation Process Reactor - Argon Purge Vent
1058 Ethylene Oxide - Stripper Purge Vent
;059 Methyl Methacrylate (MMA) - Hydrolysis Reactor, Lignt Er.cs.
Distillation Unit
:060 Methyl Methacrylate (MMA) - Acid Distillation and MMA Purification
1061 Nitrobenzene - Reactor ana Separator Vent - Washer ana Meutraiizer /en:
1062 Benzene
106^ Olefins Production - Ethylene - Compressor Lube Oil Vent
1065 Propyiene Oxide - Chlorohydronation Process - General
1066 Styrene - General
1067 Styrene - Benzene Recycle
1063 Styrene - Styrene Purification
1069 Organic Chemical Storage - N-PropyL Acetate
1070 Alcohols Production - Methanol - Purge Gas Vent
1071 Alcohols Production - Methanol - Distillation Vent
contir.uea
A-18
90008 23
-------
:A3LE A-5. Zcntir.uec.
-ode Descrioticn
1072 Chlorooenzene - Tail Gas Scrubber
1073 Chlorooenzene - Benzene Drying Distillation
1074 Monochlorooenzene
1075 Chlorooenzene - Vacuum System Vent
1076 Chlorobenzene - QiChlorobenzene Crystallization
1077 Chlorobenzene - Dichlorobenzene Crystal Handling/Loading
1078 Hailcar Cleaning - Low Vapor Press., High Viscosity Cargo (Ethylene
Glycol)
1079 Railcar Cleaning - Low Vaoor Press., Medium Viscosity Cargo
(o-Dichlorooenzene)
'080 Railcar Cleaning - Law Vaoor Pressure. High Viscosity Cargo
(Creosote)
'08' "Tanx Trucx Cleaning - u.ec. ;aocr Press., M.ea Vise. Cargo *Methyi
Metr.acryiate)
'082 TanK True* Cleaning - Low Vapor Pressure, Low Viscosity Cargo (Pher.c.,
'083 "an* True* Cleaning - Lew Vapor Press., High Vise.Cargo (Prcpyier.e
Glycoi)
1084 Residential Wood Comoustion (C1-C6)
1085 External Coranustion Boiler - Coal-Slurry Fired
1086 Printing/Flexograohic
1087 Organic Chemical Storage/i-Butyl i-Butyrate
1088 Surface Coating Operations - Adhesive Application
1089 Secondary Metal Production - Gray Iron Foundries - Fouring/Casti.-.z
1090 Fluorocarbon Manufacturing - CF 12/11
1091 Plastics Production - Polyvinyl Chlorides and Copoiymers
1092 Synthetic Organic Fiber Proauction - Nylon Batch Proaucticr. Process
'093 Fiuorocaroon Manufacturing - CF 23/22
'094 Paint Manufacture - Slenaing Kettle
1095 Textile Products - General Fabric Operations - Dyeing and Coring
1096 Textile Proauccs - General Fabric Operations - Tenter Frame
1097 Aircraft Landing/Takeoff (LTD) - Military
1098 Aircraft Landing/Takeoff (LTO) - Commercial
1099 Aircraft Landing/Takeoff (LTO) - General Aviation
1100 Gasoline Refueling
1101 Light Duty Gasoline Vehicles
1103 '-Pentene
1104 Acetaldehyce
1105 Acetic Acid
1106 Acetic Anhydride
1107 Acrolein
— cor.tir.uec
A-19
3000! 23
-------
ABLE A-5. .cr.tir.uea.
Code Descris:
1108 Acrylic Acid
1109 Acrylorutnle
1110 Adipic Acid
1111 Aniline
1112 Benzyl Chloride
1114 Butyl Aerylate
1115 Butyl Carbitol
1116 Butyl Cellosolve
1118 Carbitol
1119 Carbon Tetrachloride
1120 Acetylene
1121 Chloroform
1122 Creaol
1123 Cumene
1124 Cyclohexanol
1125 Cyclohexanone
1126 Cyclopentene
1127 Diethylene Glycoi
1128 Diisopropyl Benzene
1129 Dipropylene Glycoi
1130 Dodecene
1131 Epichlorohydrin
1132 Ethanoiamines
1134 Ethyl Aerylate
1135 Ethyl Benzene
1136 Ethyl Ether
1137 Ethyl Mercaptan
1138 Ethyl Dibronuae
1139 Ethyleneamines
1140 Formaldehyde
1141 Formic Acid
'142 Furfural
1144 Heptenes
1145 laobutyraldehyde
1146 Isobutyl Acrylate
1147 laobutyl Alcohol
1148 Isoprene
1149 Methanol
1150 Methyl Acetate
concir.uea
A-20
30008 28
-------
'ABLE A-5.
Code Description
1151 Methyl Acrylate
1152 Methyl Carbitol
1153 Methyl Cellosolve
1154 Methyl Styrene
1155 Methylallene
1158 Methyl t-Butyl Ether
1159 m-Xylene
1160 Nitrobenzene
1162 N-8utyraldehyde
1163 M-Decane
1161 N-Dodecane
1165 o-Xylene
1166 Pentaaecane
1167 Residential Wooa Concussion
1168 Piperylene
1171 Propionaidehyde
1172 Propionic Acid
1173 Propyiene Oxide
1174 p-Xylene
1175 Tert-Butyl Alcohol
1176 Toluene Diisocyanate
1178 Coal-rired Boiler - Electric Generation
1185 Coal-Firea Boiler - Industrial
1186 Heavy-Duty Gasoline Trucks
1187 Citrus Coating
1188 fermentation Processes
1189 ?ulp and Paper Industry - Plywooa Veneer Dryer
1190 jasoiine Marketed
1191 Graphic Arts - Printing
1192 Degreasing
1193 Drycleaning
1194 Auto Body Repair
1195 Degreasing Composite
1196 Drycleaning Composite
1197 Isooctane
1198 Pentane
1199 Isopentane
1200 Cyciopentane
1201 Light-Duty Diesel Vehicles
contir.uea
A-21
30008 23
-------
E A-5. Contir.uea.
Code Description
1202 Primary Aluminum Production
1203 Light-Duty Gasoline Vehicles - Exhaust Emissions
1204 Light-Duty Gasoline Vehicles - Evaporative Emissions
9001 External Comoustion Boilers - Industrial - Average
9002 Internal Combustion - Average
9003 Industrial Processes - Average
9004 Chemical Manufacturing - Average
9005 Plastics Production - average
9006 Synthetic Organic Fibe' Production - Average
9007 Alcohols Production - Average
9008 Food and Agriculture - Average
9009 Primary Metal Production - Average
9010 Secondary Metal Production - Average
9011 Mineral Products - Average
9012 Petroleum Industry - Average
9013 Pulp and Paper Industry - Average
9014 Rubber and Miscellaneous Plastics Products - Average
9015 Oil and Gas Production - Average
9016 Textile Products - Average
9017 Drycleaning/Degreasing - Average
9021 Surface Coating Operations - Average
9022 Solid Waste Disposal - Average
9023 Thinning Solvents - Average
9024 Petroleum Product Storage - Average
9025 Bulk Terminals - Petroleum Storage Tanks - Average
9026 Printing/Publishing - Average
9027 Transportation ana Marketing of Petroleum Products - Average
9028 Organic Chemical Storage - Average
9029 Organic Chemical Storage - Fixed Roof Tanks - Alcohols - Average
9030 Organic Chemical Storage - Fixed Roof Tanks - Allcanes - Average
9031 Organic Chemical Storage - Fixed Roof Tanks - Alkenes - Average
9032 Organic Chemical Storage - Fixed Roof Tanks - Amines - Average
9033 Organic Chemical Storage - Fixed Roof Tanks - Aromatics - Average
9034 Organic Chemical Storage - Fixed Roof Tanks - Carboxylic Acids -
Average
9035 Organic Chemical Storage - Fixed Roof Tanks - Esters - Average
9036 Organic Chemical Storage - Fixed Roof Tanks - Glycol Ethers - Average
9037 Organic Chemical Storage - Fixed Roof Tanks - Glycols - Average
contir.usc
A-22
90008 23
-------
TABLE A-5. Ccnciusea.
Code .Description
9038 Organic Chemical Storage - Fixed Roof Tanks - Halogenated Crganics -
Average
9039 Organic Chenicai Storage - Fixed Roof Tanks - Isocyanates - Average
9040 Organic Chemical Storage - Fixed Roof Tanks - Ketones - Average
9041 Organic Chemicai Storage - Floating Roof Tanks - Aldehydes - Average
9042 Organic Chemical Storage - Floating Roof Tanks - Alkanes - Average
9043 Organic Chemical Storage - Floating Roof Tanks - Ethers - Average
9044 Organic Chemical Storage - Floating Roof Tanks - Halogenateo.
Organics - Average
9046 Organic Chemica. Storage - Pressure Tanks - Alkenes - Average
9047 Organic Solvent Evaooration - Miscellaneous - Average
A-23
30008 23
-------
APPENDIX B
DEVELOPMENT OF LOCALE-SPECIFIC EMISSION INVENTORIES
FOR USE WITH THE URBAN AIRSHED MODEL
B-1
-------
Technical Memorandum
DEVELOPMENT OF LOCALE-SPECIFIC
EMISSION INVENTORIES
FOR USE WITH THE
URBAN AIRSHED MODEL
SYSAPP-90/109
29 October 1990
Prepared for
g*
David C. Misenheimer
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Researcn Triangle Park, North Carolina
Prepared by
LuAnn Gardner
Marianne C. Causiey
Lyle R. Chinkin
Systems Applications International
101 Lucas Valley Road
San Rafael, California 9^903
(
-------
Abstract
This memorandum discusses methods for incorporation of locaie-specific information
into a photochemical modeling emission inventory. The emission inventory require-
ments of the Urban Airshed Model (UAM) and the UAM Emission Preprocessor Sys-
tem (EPS), Version 1.00, developed for EPA by Systems Applications International,
are briefly reviewed. Types of locaie-specific data incorporation addressed include
the following:
Spatially allocating emissions using distribution surrogates and/or link
data;
Temporally distributing ana adjusting emissions;
Generating or adjusting mooiie source inventories using episcce-specific
parameters;
Speciating VOC emissions;
Projecting future year inventories;
Adding source categories; and
Incorporating biogenic emissions.
For eacn of the topics listed above, we provide a detailed discussion of modification
of EPS input files to accommodate locaie-specific data and supplemental software
requirements.
3 o i
-------
I.. INTRODUCTION
This memorandum briefly discusses tne emission inventory input requirements of tr.e
Urban Airsnea Model (LJAM) ana identifies metnods by which locale-specific cata can
be incorporated into a moaeling inventory. The UAM Emission Preprocessor System
(EPS), Version 1.00, available from EPA, conveniently packages a series of programs
that perform the intensive data manipulations required to develop an emission inven-
tory suitable for (JAM modeling with minimal additional resource requirements. To
avoid unnecessary duplication of these manipulations by the user, this document
addresses the incorporation of locale-specific data from the perspective of modifying
EPS and its inputs.
The UAM Emission Preprocessor System (EPS) contains a set of default inputs repre-
sentative of national average parameters. If these defaults are inappropriate for the
region or modeling episode in question, they can be modified or supplementary code
to the existing EPS modules can be developed to incorporate locale-specific aata into
the modeling inventory. Specific topics discussed include spatial allocation, tem-
poral adjustment, mobile source parameters, VOC and NOx speciation into carcon-
bond classes, projection of future year inventories, and inclusion of biogenic emis-
sions.
0. BACKGROUND
The UAM (the photochemical model currently recommended by EPA) is a three-
dimensional grid model employing carbon-bond chemistry. Concentrations of ozone
and ozone precursor emissions by hour and by grid cell are calculated by simulating
the various physical and chemical processes which take place in the atmospnere.
This spatial and temporal resolution of the concentration field requires that a
detailed emission inventory of hourly emissions by grid ceil of photocnermcaiiy
reactive species (CO, NOx, ana VOC) be usea as input to the UAM. VOC emissions
must 3e further disaggregated into individual chemicals, which are then groucea into
carbon-bond classes; NOx emissions must be separated into NO and NO2. Adci-
•innsily, stack parameters (including height, diameter, gas temperature, ana exit
velocity or flow rate) are required so that point source emissions can be correctly
allocated to vertical layers.
Unaer contract to EPA, Systems Applications International developed the software
pacxage known as the UAM Emission Preprocessor System (EPS), Version l.CO, to
facilitate development of the detailed photochemical modeling emission inventories
required by the UAM. Emissions from a county-level annual inventory such as
NAPAP are spatially allocated using a combination of specified location aata If or
point sources), spatial allocation surrogates such as population and lancuse (for area.
sources), and optional link data (for mooile sources such as motor vehicles, aircraft.
30127 I
3-i
-------
railway locomotives, ana vessels). Emissions are temporally aajustea from annual
totals (tons per year) to episoae-specific leveis ana diurnaiiy distributee using
source-soecific operating information or typical activity profiles by source cate-
gory. Finany, emissions are disaggregated into carbon-oona classes using EPA VOC
speciation profiles oy source classification code. Figure 1 provides an overview of
EPS; Tabie 1 indicates primary functions of each of the EPS modules.
The EPS package includes input files containing default parameters such as temporal
distribution and speciation profiles by source type. The user must supply (1) a
gridded fie id of spatial allocation surrogates for the modeling domain, (2) optional
link data for spatial allocation of mobile source emissions, (3) factors to adjust on-
road motor vehicle emissions from annual-average to episodic leveis, and (<+) a set of
projection factors for development of future year emission inventories.
In addition to the user-specified inputs listed above, other types of locale-specific
data can ce incorporated into the modeling inventory to make it more representative
of the region and episode of interest. This memorandum describes the types of
locaie-speciiic information which may be collected for both baseline ana future year
inventories ana discusses how to compile this information into data bases appropriate
for interfacing with EPS or otherwise assimilate such data into the pnotocr.emical
modeling inventory. For detailed descriptions of the data formats required by EPS,
see the User's Guide for the Urban Airshed Model. Volume IV; User's Guide for :r.e
Emission Preprocessor System (SAI, 1990).
• /•
Table 2 shows the information required by EPS for point and area source annual
emission data bases. Since EPS was developed specifically for use with the NAPAP
inventory, area source category designations compatible with those used in NAPAP
should be maintained. Otherwise, additional modification to EPS may be recuirea.
including cnanges in program code as well as input files (these modifications are cis-
cussed furtner below). A list of the area source categories currently suoportec by
EPS is oroviced in Table 3.
m. SPATIAL ALLOCATION OF EMISSIONS
The annual point source inventory generally induces location information for eacr.
point source (reported as either latitude and longitude or UTM coordinates), allow ire
direct assignment of emissions to grid ceils. If locations are available in UTM
coordinates, minor code modifications to the EPS module PREP NT will be necessary
(EPS was designed specifically for use with the NAPAP emissions data oase, .n wmcr.
point source locations are designated by latitude and longitude).
Unlike point source emissions, area and mobile source emissions are often reported
as county-ievei totals in an annual inventory; emissions from these sources must con-
sequently ce disaggregated to grid ceils. Generally, spatial allocation surrogates are
5-:
-------
Emission Inventory Databases
SAMS
NAPAP
FREDS
NEDS Others
Local Agencies
PREPNT
PREGRD
GRDEMS
CENTEMS
POSTEMS
BEIS
^
MRGEMS
•M Preprocessor
PTSRCE
UAM
FIGuSE 1 . Overview of EPS program nodules.
3-6
-------
TABLE '. Primary functions of EPS r.oauies.
Mooule _ "unctions _
PREPNT • spatially allocate point source emissions
• assign temporal distribution codes basea on operating
information
• identify stacks to be treated as elevatea sources basea.
calculated plume rise
PREGHD • separate area and on-road motor vehicle emissions into
different files
disaggregate mooile emissions into exnaust, evaoorative.
running losses, ana refueling losses oy venicle type
adjust r.caile emissions for episoaic conciticns
( temoerature, RVP, fleet turnover effects, etc.',
RDEMS
CENTEMS
• spatially allocate area and r.ooile source emissions using
spatial surrogates ana link cata
• assign temnorai distribution coaes oy source category
adjust annual-average daily emissions oy month of year ana
day of week to episodic levels
• allocate emissions by hour
remove non-reactive fraction from total r.ydrocaroon
emissions
assign emissions to carcon cons classes
create emissions inout file for the 'JAM elevatea coir.t
source prearocessor program. PTSRCE
create a UAM-reaay low-level emissions file
merge UD to six low-level anthropogenic -AM emissions files
into one file
create a summary report describing the mergea inventory
MRGEMS • merge two low-level 'JAM emissions files into one file
(generally used to merge anthropogenic ana oicgenic files;
3-;
-------
.ABLE 2. Ir.f crsaticr. rscuirsc zv EPS fcr sacr. rscorc of tr.e ar.nua... aver3.se
inventory.
Point source inventory
state and county identification coaes
facility and point identification codes
• process identification codes (SIC, SCO
location
stack parameters (exhaust gas flow rate, stacic
height and diameter, temperature)
• operating schedule (seasonal thrcugnputs ana
weetcs per year, days per weeK, ana hours cer :ay
in operation
annual average emissions
range of points sharing a common stack v octicr.a.
Area and sooile source .r.ventory
state and county identification coaes
• source category code
annual average emissions ay county
3-8
-------
TABLE 3. NAPA? area sourcs catezcrv coces.
V!A?A,? Area Sourcs Cateeorv
I Residential Fuei - Antr.racite Ccal
2 Residential Fuel - Bituminous Coal
3 Residential Fuel - Distillate Oil
1 Residential Fuel - Residual Oil
5 Residential Fuel - Natural Gas
6 Residential Fuel - Woca
7 Commercial/Institutional Fuel - Anthracite Coal
8 Commercial/Institutional Fuel - Bituminous Coal
9 Commercial/Institutional Fuel - Distillate Oil
10 Commercial/ Institutional Fuel - Residual Oil
'I Commercial/ Institutional Fuel - Natural Gas
'2 Commercial/ Institutional Fuel - Wood
'3 Industrial Fuel - -.ntnracite Coal
'U Industrial Fuel - Eitur.ir.ous Coal
15 Industrial Fuel - Coke
'6 Industrial Fuel - Distillate Oil
17 Industrial Fuel - Resiouai Oil,,
18 Industrial Fuel - Natural Gas
19 Industrial Fuel - Wocc
20 Industrial Fuel - Process Gas
21 On-Site Incineration - Resiaential
22 On-3ite Incineration - Industrial
23 Cn-Site Incineration - Commercial/Institutional
24 Ooen Burning - Residential
25 Coen Burning - Industrial
25 Cpen Burring - CciKsercial/Ir.stituticr.ai
27 Light Duty Gasoline Venicles - Limited Access Hoaas
28 Light Duty Gasoline Vehicles - Rural Roads
29 Lignt Duty Gasoline Vehicles - Suburban Roaas
30 Lignt Duty Gaso^ina itui •'• -r r'rcan Roads
31 Medium Duty Gasoline Vehicles - Limited Access Roaas
32 Medium Duty Gasoline Vehicles - Rural Roads
33 Medium Duty Gasoline Vehicles - Suburban Roaas
34 Medium Duty Gasoline Vehicles - Urban Roads
35 Heavy Duty Gasoline Vehicles - Limited Access Rcaas
36 Heavy Duty Gasoline Vehicles - Rural Roads
37 Heavy Duty Gasoline Venicles - Suburban Roaas
38 Heavy Duty Gasoiir.e /enioies - Urcan Roaas
39 Off Highway Gasonr.2 Venicles
-10 Heavy Duty Diesel Ver.ioles - Limited Access* Roaas
U1 Heavy Dutv Diesel Vehicles - Rural Roaas
3-9
-------
sntir.uea.
Area Source Cateeorv
^2 Heavy Duty Diesel Vehicles - Suburcar. Roacs
^3 Heavy Duty Diesel Vehicles - Urban Roaas
U4 Off Highway Dieses Vehicles
U5 Railroad Locomotives
U6 Aircraft LTOs - Military
U7 Aircraft LTOs - Civil
US Aircraft LTOs - Commercial
U9 Vessels - Coal
50 Vessels - Diesel Oil
51 Vessels - Hesiauai Oil
52 Vessels - Gasoiir.e
53 (MOT USED)
5^ Gasoline Martcetea
55 Unpaved Roaa Travel
56 'Jnpaved Airstnz _7Ds
57 (MOT USED)
53 (MOT USED)
59 (MOT USED)
60 Forest Wild Fires
61 Managed Burr.ir.g - Prescrnea
62 Agricultural Fieic Burr.ir.g
63 Frost Control - Crcr.ars Heaters
64 Structural Fires
65 I MOT USED)
66 Aamonia Er.issicr.s - '..gr.t I-ty Gasoline Vehicles
67 Ammonia E.T.issicr.z - -'eavy I-ty Gasoline Vehicles
68 Ammonia Ecissicr.s - Heavy Zuty Diesel Vehicles
69 Livestock Waste Management - Turkceys
70 Livestock Waste Management - Sheep
"1 -ivestock Waste Management - Beef Cattle
72 Livestock Waste Management - Dairy Cattle
73 Livestock Waste Management - Swine
7^ Livestock Waste Management - Broilers
75 Livestock Waste Management - Other Chickens
76 Anhyarous Ammonia Fertilizer Application
'7 Beef Cattle Feea Lots
"3 Decreasing
"9 Dry Cleanir.z
80 Graonic Arts/Pnr.tir.g
81 Ruboer and Plastics manufacture
32 Arcnitectural Ccatir.es
3-10
-------
TABLE 3. Ccr.tir.uea.
*JAPAP Area Source Catezorv
83 Auto 3ody Repair
84 Motor Vehicle Manufacture
85 Pacer Coating
86 Fabricated Metals
87 Macninery Manufacture
88 Furniture Manufacture
89 Flatwood Products
90 Other Transportation Equipment Manufacture
91 Electrical Equipment Manufacture
92 Shicouilding and Repairing
93 Miscellaneous Industrial Manufacture
9* MOT USED;
95 Miscellaneous Solvent Use
96 Mir.or Point Sources - Ccai Boilers
97 Mir.or Point Sources - Oil Boilers
98 Mir.or Point Sources - Natural Gas Boilers
99 Minor Point Sources - Other
100 Publicly Owned Treatment Worxs iPOTWs)
101 CutsacK Asonalt Paving Coeration
102 Fugitives from Synthetic Organic Chemical Manufacture
103 Bulk Terminals and Bulk Plants
10U Fugitives from Petroleum Refinery Operations
'05 Process Emissions from Bakeries
106 Process Emissions from Pharmaceutical Manufacture
'07 "recess Emissions from Syncnetic risers Manufacture
'C8 Iruce Oil and Natural Gas Proauction Fields
109 Hazardous Waste Treatment, Storage and Disposal Facilities
'S-expafiCec MAPftP source category tocl^s for -T~- vehicle emissions:
Exhaust Emissions:
27 LDGV - Limited Access Roaas
28 LDGV - Rural Roaas
29 LDGV - Suburban Roacs
30 LDGV - Urban Roaas
3* MDGV - Limited Access Roacs
32 MDGV - Rural Roaas
33 MDGV - Suburoan Roacs
3a MDGV - Urban Roaas
35 HDGV - Limitea Access Roaas
-------
:ABLE
Exhaust Emissions uor.t,,:
36 HDGV - Rural Roaas
37 HDGV - Suburban Roacs
38 HDGV - Urban Roaas
39 Off Highway Gasoline veniciss
UO HDDV - Limited Access Roaas
U1 HDDV - Rural Roaas
42 HDDV - Suburaan Roads
US HDDV - Urban Roads
uu Off Highway Diesel Vehicles
Evaporative
227
&te »
223
229
230
23 *
w ^
232
233
234
f *> 1
£ J J
236
237
233
239
240
~ »t *
2-2
243
244
Refueling
327
328
329
330
331
332
333
334
335
336
Emissions:
LDGV - Rural Roaas
LDGV - Suburban Roa>s
LDGV - Urban Roaas
MDGV - Limited Access Roaas
MDGV - Rural Roaas
MDGV - Suburban Roaas /.
MDGV - 'Jrcar. Roaas
HDGV - Rural Roaas
HDGV - Suburban Roacs
HDGV - Urban Roaas
Off Hignway Gasoiir.e venicies
HDDV - Limited Access Roaas
HDDV - Rural Roaas
HDDV - Suburban Roaas
HDDV - Urban Roads
Off Hignway Diesel Vehicles
Loss Emissions:
LDGV - Limited Access Roads
LDGV - Rural Roaas
LDGV - Subursan Roads
LDGV - Urnan Roaas
MDGV - Limited Access Roaas
MDGV - Rural Roads
MDGV - Suburcan Roaas
MDGV - Urban Roaas
HDGV - Limited Access Roaas
HDGV - Rural Roaas
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
( SffT NSC )
-------
TABLE 2. Conciudea.
Refueling
337
338
339
340
341
342
313
344
Loss Emissions (.cone.;:
HDGV - Suburoan Roaas
HDGV - Urban Roads
Off Highway Gasoline vehicles
HDDV - Limited Access Roads
HDDV - Rural. Roads
HDDV - Suburoan Roaas
HDDV - Urban Roads
Off Highway Diesel Vehicles
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
Running Loss Emissions:
U27
U28
U29
430
431
U32
433
434
435
436
437
438
-39
UliO
UU1
W2
-;i:3
auu
LDGV - Limited Access Roaas
LDGV - Rural Roaas
LDGV - Subursan Roaas
LDGV - Urban Roaas
MDGV - Limited Access Roaas
MDGV - Rural Roads
MDGV - Suburoan Roads
MDGV - Urban Roaas
HDGV - Limited Access Roaas
HDGV - Rural Roaas
HDGV - Suburban Roads
HDGV - Urban Roads
Off Highway Gasoline venicles
HDDV - Limited Access Roaas
HDDV - Rural Roads
HDDV - Suburnan Roaas
HDDV - Urban Roads
Off Highway Diesel Vehicles
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
;SA: "so
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
(SAI NSC)
-------
employee in tnis cisaegregation. Cammomy usea indicators mciuce population
ana/or noustng, ianause categories (e.g., urban, agricultural, or rangeiana), anc ::r.x
information ^locations of major roadways, shipping cr.anneis, airports, raiiways,
etc.). Distribution of the surrogates used to spatially allocate emissions must :e
determined for tne entire moaeiing region by grid ceil. Landuse by grid ceil can oe
estimated using maps available from local planning agencies; USGS maps are often
useful for locating such sources as raiiways and airports.
Spatial surrogate indicator distributions should also be estimated for future year
inventories which will be developed from the baseline inventory. Future year spatiai
surrogate distributions should reflect anticipated changes in landuse, requiring
coordination with local planning agencies.
For compatibility with EPS, spatial surrogate distributions should be tabulated as
fractions of the county total for that surrogate for each grid ceil associated with the
county. Grid ceils can be identified by either UTMs or (1,3) coordinates, aithougn tne
final gridded surrogate field used as input to the EPS module GRDEMS must be
referenced by (I,J) coordinates. If the exact extent of the modeling domain nas yei:
to oe determined when the spatial surrogate distribution-data is collected, identify-
ing grid ceils by LTM coordinates provides a more versatile data base. Spatial surro-
gate distribution can then be determined over a larger region than is likeiy to ce
modeled; the final gridded surrogate field can easily be extracted from this cata
base, and UTM coordinates converted to (l»J) modeling coordinates with minimal
effort. The modeling domain, however, should ideally be identified prior to collect-
ing any data in order to minimize resource requirements. All spatial surrogate dis-
tribution data except for link data should be included in a single file; separate file:.
will be maintained for base year and future year distributions.
Link data is generally digitized from maps; features that can be included in the link
file include maior roadways, raiiways, airports, and shipping channels. To be com-
patible with EPS, individual link segments must end at county borders. Currently,
EPS does not support spatial allocation based on weighted travel by link; ir.steac.
-------
In particular, iink aata generates for attribution of on-roaa motor venicie emissions
should iceauy incorporate cifferent travel weightings by iink wmcn are representa-
tive of the aay of weex being moaeiea. Separate iink files wouid then need to oe
maintamea for each aay of muitiday moaeiing episodes.
IV. TEMPORAL ADJUSTMENT AND DISTRIBUTION OF EMISSIONS
Annual emission estimates must be adjusted to reflect seasonal, weekday/weekend,
and diurnal variations in either activity levels or emission factors. The default tem-
poral distribution profiles and factors by source category provided with EPS may not
be representative of actual patterns for a given region. For example, shorter grow-
ing seasons in some portions of the country may result in a different monthly profile
for emissions from agricultural operations than wouid be appropriate for a warmer
region. Similarly, emission factors used to estimate evaporative emissions from sol-
vent use, gasoline marketing, and other activities can be strongly temperature-
depenaent. The effect of regional amoient meteorological conditions on emission
leveis may accordingly result in seasonal distribution profiles whicn differ signifi-
cantly from the EPS defaults.
Ideally, locale-specific temporal variation data snouid be collected for ail sources
which contribute significantly to the inventory. Monthly fractions of annual leveis
are preferanie to seasonal fractions, but either can be used to construct the monthly
activity profile required by CRDEMS. In addition to seasonal profiles, regional
weekday/weekend activity levels and diurnal variation by source category can be
determined through special surveys or estimated using engineering judgment.
Locale-specific temporal data can be incorporated into the emission inventory in
several ways, either oy modifying EPS input files or by developing supplemental
software. Locale-specific temporal distribution factors by area source category can
usually be incorporated into the moaeiing inventory without aeveioping supplemental
software. Monthly fractions of annual activity, weeidv profile codes wnicn identify
typical activity profiles By day of weeK, and diurnal profile cooes are assignee by
source category in an input to the GRDEMS module. This file should be eaitea so
that temporal profile assignments reflect locally applicable rather than default
variations in activity. Diurnal and weekday codes and profiles currently defied !.-.
EPS inputs are shown in Tables 4 and 5. If none of the existing profiles mater, tne
desired temporal distribution for a given source category, additional profiles can be
created. All new profiles must be added to the files defining weeKday ana aiurnai
distribution profiles used as inputs to the EPS module CENTEMS.
Alternatively, diurnal variations in emissions can be incorporated directly into the
modeling inventory. The day-specific modeling emission record format (merf), shown
in Table 6, allows the user to specify the hourly fractions used to diurnaily allocate
total daily emissions. Unusual diurnal orofiles not currently defined in the temporal
30127 2
3-1:
-------
TABLE 4. Diurr.ai variation coaes usea :r. the emissions
preprocessor system.
Coae
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Emissions Contr
000000001 1
000000001 1
000000001 1
000000001 1
000000001 1
000000001 1
000000001 1
000000001 1
000000001 1
OOOOOQQQ1 '
00000000 i
00000000 1
00000000'.'
000000001 '
00000000 1
00000000 1
00000000 1
00000000 1
000000001
1 1 1 1 ' 1 1 1 1
111111111
i i : i • : : • :
II. 1
1 1 I 1 II
311111133
222222 21010
0883 310101010
10 1 1 1 ' 1 1 8 8
isution ay Hour (0-23) Daily
11111100000000
1 1 1 1 1 100000000
11111100000000
11111100000000
11111100000000
11111100000000
11111100000000
11111100000000
1 1 1 1 1 100000000
'1111100000000
! ' 1 1 1 1 '00000000
' ' 1 1 1 1 '00000000
: : i ' ' 1 i i i i i i ' i •
1 ' 1 1 1 1 -V. 1 1 1 1 1 1 1 1
111-1111111111:
111111111111111
111111111111111
111111111111111
111111111111:1:
111111111111111
1 • 1 1 1 1 1 1 1 1 1 1 < 1 •
: • : • 1 l : i i i 1 ' : : •
i i : l , • ! i
i i ' : i 1 i i 1 1 1 1 i 1 :
5555555 510101010 773
665555555 510101010 2
000000000000000
101010101010101010101010101010
3
3
3
3
8
a
3
3
8
3
3
3
16
16
16
16
16
16
16
24
24
24
24
,6
128
72
•32
continued
3-iO
-------
TABLE
Coae
36
37
38
39
40
41
42
43
U4
45
47
48
49
50
51
52
53
54
55
56
57
53
60
6.
1
62
53
f \ .
64
65
66
67
f O
68
69
4. Conduces.
Emissions Concncution ay Hour iu-23) Daily
00000 1 36 91010101010101010 9631000
0000026622124421 1 310 87610
0000234UU44U33221 1000000
00000 0185923 000000000000000
00000000000000 0283735 000000
1411 32 41927 05078889^999399 3 3 3987050443314
3821 1 U12364S547C72-16875768U.786556423525 8
1 1 1 1 1 1 610 6 5 5 5 5 5 5 6JO 8 6 4 i 1 : 1
oooooo • • : : i 'oooooooooooo
00000 1 61010101010 633334UOOOOO
000000022222010000000000
0 1 1 1 1 ' ' ' ' i : " ' 1 1 1 1 1 1 ' ' 1 1 0
o ' : i : ' ' ' ' : : • i 1 1 1 1 i i i i 1 i :
3 o i '••'•• i i •• i 1 i 1 i i •• i • o
300 1 1 1 1 1 1 1 ' ' 3 0
300 ' 1 1 •'•'•' 0
300000''. '11' 111111000000
0 0 0 0 0 0 ' ' 1 1 ' 1 1 1 1 1 1 1 1 1 1 1 1 1
0000000' ' 1 ' ' 1 1 1 1 10000000
3000000 ' : ' ' ' 1 1 1 1 ' 0 0 0 0 '.' 0
0 0 0 0 0 0 0 ' 1 : ' • 1 1 1 1 1 1 1 1 1 1 1 '
OOOOOOOO: ' ' ' 1 1 1 1 10000000
000000001 ' ' ' 1 1 1 1 1 1 1 1 1000
0 0 0 0 0 0 0 0 • 1 ' • 1 1 1 1 1 1 1 1 1 1 0 0
0 0 0 0 0 0 3 0 1 •"'• 1 1 1 1 1 1 1 1 1 i 0
000000000' ' ' 1 1 1 100000000
118
68
41
100
100
999
999
96
33
11
"
23
2 1
' ?
' 2
, o
18
'0
1 7
9
'3
'4
'5
7
3-1
i /
-------
TABLE 5. Wee.Kcay vana'icr. ccces used :r.
the emissions orecrocessor system.
Code
1
2
3
u
5
6
7
3
Q
:o
* i
^2
13
1U
15
16
17
'8
• n
20
Emissions contribution
oy (Mon-Sun)
1111100
000001 1
' 1 1 1 1 0 0
' 1 1 1 1 0 0
1111100
• 1 1 1 1 1 0
* * 1 « 1 1 «
I I I
Total
Days
5
2
5
5
5
6
7
22 '010101010 7 U 61
23 5 5 5 = 5 -i U 33
3-13
-------
rABL£ 6. lay-specific .Toaeiir.g emissions reccra fcrmat (merf).
Lir.e Vanaaie
1* ISRG
IF!?
SIC
PS
ICL
JCL
IYR
IDYCCD
IHKCCD
FID
FST
FCNTY
VMNTH
CO
CNO
SOX
THC
DM
SLBL
Either -'.
secono 12
Columns
1-3
4-8
9-^2
13-20
21-23
24-26
27-28
29-30
31-32
33- 1
42—6
47-52
53-56
57-H6
117-126
127-136
137-146
147-156
157-166
167-168
169-^5
Type •
-
T
i
R
A
*
I
I
I
i
A
A
I
-
R
R
R
R
R
0
-
A
DescriDtion
Griddea surrogate code (not usea.
skipped)
FIPS state/county code (not used,
skipped)
Either Source Industrial Classification
or NAPAP Source Category code
Either Standard Classification Code or
NAPAP Source Category code
I coordinate of grid cell
J coordinate of grid cell
Year, two digits (e.g.. 89 for 1989)
Diurnal variation code
Weekday variation code
Facility ID (0 or blank for area
sources )
Stack ID (0 or blank for area sources)
AEROS state/ county code
(Not usea, skipped)
Array of 12 hourly factors^
CO episode emissions (kg/day)
NOX episode emissions (kg/day)
SOX episode emissions (kg/day)
TOC episode emissions (kg/day)
?M episode emissions (kg/day)
(Not used, skipped)
Scenario label (not used. skiDDea)
or -2, corresponaing to the first 12 hours of the day cr me
hours of the day,
Always equal to 0 in
-1 Factors us
ied to allot
respectively.
day-specific merf.
*ar.e da
ilv pmi«!sinn«? tn hour?; of riav : f!orr?«nand
to first 12 hours of day if IDYCOD = -1 and second 12 hours of :ay :f
IDYCCD = -2.
3-19
-------
input files :o £?S can tr.us easuy oe accommoaatea; this format also supports :.-.s
inclusion of houriy episoce-specific emissions cata into tne moaeiing inventory.
Table 7 shows tne stanaara meri for comparison. Note that in the aay-speciiic .-r.er:.
emissions reflect episoaic levels; the stanaara merf contains annual average caiiy
emissions.
The day-specific merf is especially useful for incorporating source- or stack-soec::ic:
operating information or episoaic emissions into the modeling inventory. Day-
specific meri records must currently be generated outside of EPS; the CENTEMS
module, however, accepts both standard and day-specific merf input files.
V. MOBILE SOURCE PARAMETERS
Since EPS was originally designed for use with the NAPAP inventory, mobile source
emissions are assumed to be annual, county-level composite emissions for the follow-
ing vehicle types and road classes: light duty gasoline vehicles, light duty gasoline
trucks, heavy duty gasoline vehicles, and heavy duty gasoline trucks for limitea
access roaaways, rural roadways, suburban, and urban roadways. The EPS moduie
PRECRD disaggregates composite mobile source emissions into exhaust, evaporative,
running loss, and refueling components and adjusts emissions from annual average to
episodic levels based on regional episodic conditions (e.g., temperature and fuel
RVP). The factors used for disaggregation and adjustment are calculated from ratios
of MOBILES emission factors generated fdr episodic conditions to MOBILES emission
factors generated using the annual average inputs used to construct the NAPAP
inventory, assuming average speeds of 55 mph for limited access, 45 mph for rural,
and 19.6 mph for suburban and urban roadways.
Mobile sources often constitute a significant portion of the urban anthropogenic
inventory. Consequently, to minimize uncertainties in the modeling inventory, a
mobile source inventory generated specifically for the modeling episode snouid be
used insteaa of an annual-average inventory whenever feasible. Detailed guidance on
construction of mobile source inventories is provided in Procedures for Emission
Inventory Preparation. Volume IV; Mobile Sources (EPA, 1989). In short, a traffic
model is employee in conjuction witn a mootie emission factor model (the recom-
mended emission factor model is EPA'-> MOSILE'O to estirr.-uo mooiie source emis-
sions oy link. Locale- and episode-specific parameters such as link VMT, average
speed by link, and VMT mix ana fleet mix by vehicle type should be used to ensure
that the mobile source inventory reflects actual episodic conaitions as accurately as
possible.
MOBILES calculates exhaust, evaporative, running loss, and refueling emission
factors for eight vehicle types (light duty gasoline automobiles, Light duty ciesei
automooiies, two classes of light duty gasoline trucks, light duty diesel trucks, r.eavy
duty gasoline vehicles, heavy duty diesei vehicles, and motorcycles). Accorc:ngiy,
90127 2
3-20
-------
TABLE 7. Stanaars moaeiir.z emissions recarc format (-erf).
Line Vanaole
U ISRG
IFIP
SIC
PS
T ^**
w WM
• vo
I3YCCD
IWKCCD
IT " "*,
r^r
•CJ1TY
VHNTH
CO
CNO
SOX
THC
:u
SLEL
Columns
1-3
tt-8
9-12
13-20
21-23
24-26
27-28
29-30
31-32
33-ai
42-46
47-52
53-56
57-116
117-126
127-136
137-146
147-156
157-166
167-168
169- '75
Type
T
r
R
A
I
T
r
i
.
A
A
I
-
R
R
R
9
R
R
-
f\
Description
Gridded surrogate coae (not usea,
skipped)
FIPS state/county code (not usea.
skipped)
Either Source Industrial Classification
or NAPAP Source Category code
Either Standard Classification Code or
NAPAP Source Category code
I coordinate of grid cell
J coorainate of grid ceil
Year, two digits (e.g., 89 for '989)
Diurnal variation coae
Weekday variation coae
Facility ID (0 or blankc for area.
sources)
Stack ID (0 or blank for area sources;
AERflS state/county coae
(Not used, skipped)
Array of 12 monthly factors
CO annually averaged emissions (kg/day;
NOX annually averaged emissions (kg/day;
SOX annually averaged emissions (kz/day,
TOC annually averagea emissions (kg/day,
PM annually averaged emissions (kg/day;
(Not used, sicippea)
Scenario label (not usea. sKicsec:
Li.
-------
•jse of an eoisoae-soeciiic rr.ooi:e source inver.:ory mav reau:re exoanair.e tr.e area
source category cesignations to induce new cooes for tr.ese aaaitionai sources isee
Section VII of this memorandum, aacressmg aadition of new source categories. :cr
specific guidance).
If resource limitations prevent construction of an episode-specific mobile source
inventory, emission factors from MOBILES can be used to generate the ratios usea by
EPS to aajust an available annual average inventory to episodic conditions. At tr.e
minimum, this adjustment will require the following data: maximum and minimum
ambient temperatures on the episode day o
3- £.2
-------
in tne profile, :~e [(weigr.; fraction of compouna j for tr.at profiie)/(moiecuiar v-eigr.:
of compouna )) • (rr.otes of carson-oona species i for eacn moie of compouna ;)], as
shown in tcuation 1:
for eacn
caroon-bonc
species i,
fc
each
chemical
compound j
we frac of '
moi we of j
moie J
Modified carbon-bond splits corresponding to redistribution of the weight fraction
contributions of the individuai chemicai compounds can thus be constructed fairly
easiiy. However, difficulties may arise if chemical compounds not present in the file
designating carbon-bond class assignments by compound are added to the profile,
since carbon-bond distributions for the new chemicals may be unavailable from
published literature. An appropriate carbon-bond split must then be determined by
someone experienced in both photochemistry and the carbon-bond mechanism.
Locale-specific speciation data are more often available for individuai sources man
for categories of sources. As these data are collected, each chemicai compound in
the data base should be identified by a universally recognized standardized code
rather than by name alone; the SAROAD designation is especially appropriate, since
this is the coding system used to identify chemicals in both EPA publications ana EPS
input files. Additionally, all other data included in a point source inventory (facility,
point, and process identification codes, location, operating schedule information,
etc.) should be linked with each emission record to facilitate further processing. EPS
does not currently support direct incorporation of pre-speciated emissions by stacx
into the modeling inventory. Consequently, additional software must be deveiooed to
assemble emissions from these sources into a UAM-compatible emissions file. This
software must accomplish the following tasks:
Disaggregate emissions of individual cnemicais into carbon-bona species:
Apply appropriate temporal factors to adjust for seasonal and day-of-weex
variations:
Disaggregate emissions by hour of day basea on diurnal profiles:
Distinguish between elevated and low-ievei sources:
Create a UAM-compatible emission file containing low-ievel emissions:
and
30127
-------
Create an eievatec Domt source emission file for inout to tne LAM eieva-
:ea aomt source creDrocessor, FTSRCE.*
VH. PROJECTION OF FUTURE YEAR INVENTORIES
Often, future year inventories must be generatea for planning purposes or contrc:
strategy evaluations. EPS allows the baseline inventory to be "grown" using ratios c:
future-to-base-year expectea activity levels or activity levei indicators for general
categories of sources. Upper bounds on point source emissions based on maximum
future year permittable emission levels may also need to be incorporated into tne
projected inventory. Bounded emissions projections might also result from process
capacity limits. The current EPS will not impose an upper-bound limit on emissions:
such processing must occur outside of EPS.
Common indicators used to estimate future year activity levels for source categories
include employment by industrial category (the Bureau of Economic Analysis
publishes projections for the nation, states, and MSAs), demographic characteristics
such as population and housing, and vehicle miles traveled by vehicle type and roao
type. Appropriate pairing of growth indicators with source categories should be
evaluated using engineering judgment.
Alternatively, estimated future year activity levels for individual sources may be
obtained through surveying the facilities in»question or screening permit applications;
to determine expected expansions or new construction. This is a resource-intensive
approach, so collection of projected future year activity data for individual sources
should be limited to the major sources in the region.
EPS projects emissions based on 2-digit SIC code for point sources and NAPAP
Source Category Code for area and mobile sources. Projection factors are expressed
as a ratio of future-to-base year activity. Source-specific projections (incorporating
such information as anticioatea shut-aowns or construction of aaditionai facilities)
must be implemented outside of EPS and then incorporated into the remainder of the
inventory. For point sources, it may be desirable to assign growth based on proiectsc
fractional increases of industrially zoned land by grid cell instead of assuming tnai
all growxn will occur at existing plants. This approach, however, will probabiv
* If source-specific speciated emissions are not available for ail elevated stacxs
within the modeling region, THC emissions from some elevated stacks must se
speciatea through EPS by source classification code using either EPA or locaie-
specific VOC speciation profiles. Supplemental software will need to be deveiooec
to merge the source-specific and EPS-generated elevated point source emission
files before the eievatea oomt source orocessor can be.run.
90127 :
3-24
-------
require the creation or hypothetical point source records; ail information associated
with a point source recora (inducing process identification codes ana stack
parameters) must be estimated for- eacn hypothetical record. Additionally, supple-
mental software or multiple runs of EPS modules may be required if different growtn
rates are used for sub-regions of the modeling domain.
VHI. INCORPORATION OF ADDITIONAL SOURCE CATEGORIES
For some regions, emission estimates may be available for source types not included
in Table 3. Incorporation of additional source categories into the modeling inventory
may require minor computer code changes to EPS as well as modifications to input
files.
Area source categories may be added to the inventory in one of two ways. If some of
the NAPAP categories listed in Table 3 are not applicable to the region in question,
their NAPAP source category codes may be reassigned to new types of sources. All
temporal profiles ana spatial allocation surrogate pairings (usea as input to GRDEMS)
associated with the reassigned codes must be reviewed and revised if necessary.
Additionally, the speciation profiles and reporting category code assignments used by
the CE.NTEMS module (activity code, old inventory category code, process code, and
control code) must be reassigned to ensure compatibility with the reassigned NAPAP
source category designations.
Alternatively, the user can create new code designations in addition to the cate-
gories in Table 3. In this case, the maximum parameters set in each of the EPS
modules used to process the area source portion of the inventory (i.e., PREGRD,
GRDEMS, and CENTEMS) must be assessed and modified where necessary to ensure
that array dimension bounds will not be exceeded because of the additional cate-
gories. New spatial allocation surrogate and temporal profile assignments must be
made for ail new categories; approriate growth indicators must also be selected and
all data incorporated into the appropriate input files. Likewise, appropriate specia-
tion profiles and reporting category code assignments must be identified and included
Ln the CENTEMS and POSTEMS inputs.
Disaggregation of mobile source emissions into source categories besides exhaust,
evaporative, running, and refueling losses (e.g., hot soak, hot stabilized, cold start,
etc.) or inclusion of additional sources (such as motorcycles) requires all of the modi-
fications discussed above for new source category code designations. Additionally,
extensive code modifications to the PREGRD module will be necessary, or the func-
tions provided by PREGRD for adjustment of mooile source emissions (discussed in
the previous section on mobile source parameters) must be duplicated outside of EPS,
perhaps in a supplemental module either parallel or subsequent to PREGRD and prior
to GRDEMS.
30127 2
3-25
-------
"The aaaition of point source SIC.'SCC comoinations not currently inciuaea in tr.e
.nput files to CENTEMS requires :ess effort than inclusion of additional area or
Tiooiie source categories. For eacn.new SIC/SCC comnination, appropriate reporting
category codes must be assignee: the SCC/speciation profile pairings should also ne
checked if the SCC in the new pair does not appear in any existing SIC/SCC com-
binations, since the new SCC may not be present in the CENTEMS input file pairing
SCCs with speciation profiles.
IX. MULTISTATE AND MULTINATION MODELING DOMAINS
For some areas* the inclusion of ail major sources that may affect the uroan region
requires extending the modeling domain across state or sometimes national borders
(namely, Canada and Mexico). In these cases, it is often necessary to obtain
emissions data for portions of the domain from other agencies. These data may have
to receive special treatment, such as conversion to an EPS-compatible format,
before it can be incorporates into the modeling inventory.
A province-level NAPAP inventory has been compiled for Canada: the technical
memorandum "Recommended Software and System Upgrades: the UAM Emission
Preprocessor System (EPS)" discusses some of the problems encountered in combining
emissions estimates from the Canadian and U.S. NAPAP inventories into a single
photochemical modeling inventory (SAI, 1990). Specifically, the Canadian inventory
employs a different SIC, SCC, and source category coding system that must be cross--
referenced with the U.S. designations. Additional problems include the lack of
readily available data bases for construction of gridded spatial surrogate fields ana
speciation of emissions.
Emissions estimates for Mexican sources are likely to be less readily available than
Canadian information. Mobile source emission factors for Mexican vehicles may
need to be estimated by modifying the fleet distribution and anti-tamper ing rates
used as input to MOBILES; it may also be desirable to modify the zero mile emission
factors and deterioration rates internal to the MOBILES code (any modifications to
MOBILES code should be aiscussed with, and approved by, the EPA Office of Mobile
Sources). The industrial source inventory may need to be generated by hand from
available information describing source characterization, location, and activity
levels. Special attention should be given to identifying types of control equipment
widely employed in the United States but not required in Mexico, which might make
U.S. emission factors by source classification inapplicable. Gridded spatial
allocation surrogate fields and link files must also be developed to distribute area
and mob tie source emissions.
Developing the mobile source portion of the photochemical modeling inventory for a
muitistate modeling domain requires a slightly different procedure than is used for a
single-state region. If episoaic mobile source emissions are estimated by adjusting
an annual average inventory as aescr±ed in Section V, the. MOBILES inputs usea to
3-26
-------
estimate eacn state's mooiie source emissions in the annual inventory must ae
identified and a separate set of momie source adjustment factors defined for eacn
state.
Similarly, the growth factors used to project future year inventories will probaoiy
differ by state. For these reasons, it may be necessary to process each state's
contribution to the area and the point source inventory separately through the
PREGRD and PREPNT modules. The output merf emission files for each state from
these modules can then be merged before further processing. For point sources,
however, a separate PREPNT run using dummy growth factors and a combined annual
inventory of all point sources within the modeling domain should be made so that a
single stack control file (used as input to the CENTEMS module) containing all stacks
in the region is generated.
X. BIOGENIC EMISSIONS
In recent years, air quality modelers have begun to recognize that biogemc emissions
(naturally occurring emissions from vegetation) can contribute significantly to the
total hydrocarbon inventory, even in predominantly urban regions. Some of the
chemical species commonly found in biogemc emissions are also quite photocnemi-
cally reactive (e.g., isoprene). Accordingly, the photochemical modeling inventory
should include an estimate of biogenic emissions for completeness.
j*
EPS Version 1.00 does not contain a module for generating biogenic emission inven-
tories; however, EPA's recently developed Biogenic Emission Inventory System (BEIS)
is distributed with EPS. This system produces a UAM-formatted biogenic inventory
by using a biomass data base (derived from landuse data from the Oak Ridge National
Laboratory's Geoecoiogical Data Base) in conjunction with emission factors oy forest
type (developed by Zimmerman). BEIS provides algorithms for calculating the
effects of temperature and light intensity on biogenic emission rates and can
accordingly be used to aeveiop a biogenic emission inventory which includes the
effects of region-specific hourly temperatures and light intensity. The canopy moaei
utilized in BEIS also requires as input the gndded hourly wind fields used by CAM;
these are used to account for varying leaf surface temperatures in different strata of
the canopy. Incorporation of additional locale-specific information (such as a more
highly resolved landuse data base or emission factors for plant species indigenous to
the region) requires modification of the BEIS code.
XI. SUMMARY
This memorandum identifies methods by which locale-specific information can be
incorporated into a photochemical modeling emission inventory. Specific topics
addressed included the following:
30127 2
-------
Spatially allocating emissions using cistricution surrogates ana/or ::r.x
Temporally Distributing ana aajusting emissions;
Generating or adjusting mooiie source inventories using episode-specific
parameters;
Speciating VOC emissions;
Projecting future year inventories;
Adding source categories; and
• Incorporating biogenic emissions.
A brief review of the emission inventory input requirements of the Urban Airshea
Model and the UAM Emission Preprocessor System Version 1.00 developed for EPA is
provided. Additionally, for each of the categories listed above, modification of EPS
Input files to accommodate locale-specific data and supplemental software require-
ments are discussed in detail.
-------
REFERENCES
User's Guide for the Urban Airshed Model. Volume IV" User's Guide for the Emissions Preprocessor
System. EPA-450/4-90-007D. U.S. Environmental Protection Agency (OAQPS), Researcn
Triangle Park. NC, June 1990.
User's Guide to MOBILE4. EPA-AA-TEB-89-01, U.S. Environmental Protection Agency, February
1989.
Procedures jor Inventory Preparation. Volume IV: Mobile Sources. EPA-450/4-81-026d (Revised),
U.S. Environmental Protection Agency, July 1989.
Air Emissions Species Manual. Volume I: Volatile Organic Compound Species Profiles. EPA-450/2-
88-003a. U.S. Environmental Protection Agency (OAQPS), Research Triangle Park, NC.
April 1988.
H. Hogo and M. W. Gery. Guidelines for Using OZIPM-4 with CBMJV or Optional Mechanisms.
Volume i. EPA Contract No. 68-02-4136, Systems Applications Incorporated. 1986.
Development of a Biogenic Emissions Inventory System for Regional Scale Air Pollution Models. T.
E. Pierce, B. K. Lamb, and A. R. Van Meter, Paper No. 90-94.3, presented at the 83rd Air
and Waste Management Association Annual Meeting at Pittsburgh, Pennsylvania. June 1990.
• ft
P. Zimmerman. Determination of Emission Rates of Hydrocarbons from Indigenous Species of
Vegetation in the Tampa Bay/Petersburg. Florida Area. EPA-904/9-77-028, U.S.
Environmental Protection Agency, Atlanta, GA, 1979.
»J27 il B-29
-------
-------
TECHNICAL REPORT DATA
•00/4-91-014
l- CESSION NO
TfbdidWe!u¥oPElhe Preparation of Emission Inventories !
for Carbon monoxide and Precursors of Ozone, Vol. II: I
[mission Inventory Requirements for Photochemical Air I
Ouaiitv Si'rnn tAt-i -m Mnrlnlc
7 AUTHOHIS) '
9. PERFORMING ORGANIZATION NAME AND ADDRESS
US Environmental Protection Agency
Office of Air Quality Plannina and Standards
Technical Support Division (MD-14)
Research Triangle Park, NC 27711
12. SPONSORING AGENCY NAME ANO ADDRESS
3 HCPORT DATF I
June 1991 1
5. PERFORMING ORGANIZATION CODE I
3. PERFORMING ORGANIZATION REPORT NO 1
IO PROGRAM ELEMENT NO.
11 CONTRACT/GRANT NO.
13. TYPE OF REPORT ANO PERIOD COVERED
14. SPONSORING AGENCY CODE
IS. SUPPLEMENTARY NOTES
Project Officer - Keith A. Baugues
Fhis is a companion document to Volume I, which describes procedures for
:omoiling the annual countywide inventory of volatile organic compound (VOC)
emissions. Volume II describes procedures for converting the annual countywide
emission inventory to the detailed inventory needed for photochemical models.
The detailed inventory contains hourly gridded emissions (by species class for
VOC and NOx) and CO for the days to be simulated in the photochemical model.
his document is an update to the original, (450/4-79-018), published is 1979.
KEY WORDS ANO DOCUMENT ANALYSIS
DESCRIPTORS
b.lOENTIFICRS/OPEN ENDED TERMS [c. COSATI Fkld/Gioup
Emissions Inventory Spatial Resolution
Gridding Temporal Resolution
Hydrocarbons Species Resolution
Nitrogen Oxides Volatile Organics
Photochemical Models
118. DISTRIBUTION STATEMENT
11. SECURITY CLASS /TTiii Rtponi
M. SECURITY CLASS iTtia pf**l
21. NO. OF PAGES
239
22. PRICE
EPA Firm 1220-1 (R»». 4-77) »«cviout COITION <• OMOUCTC
------- |