&^fel^
Wb********!* mtsssA &*»*»*i:*is»
-------
EPA-454/R-92-026
PROCEDURES FOR THE
PREPARATION OF
EMISSION INVENTORIES FOR
CARBON MONOXIDE AND
PRECURSORS OF OZONE
VOLUME II: EMISSION INVENTORY
REQUIREMENTS FOR PHOTOCHEMICAL
AIR QUALITY SIMULATION MODELS
(REVISED)
Office Of Air Quality Planning And Standards i < o p -
Office Of Air And Radiation RpainnTr0nfr0i t:? " ^
U. S. Environmental Protection Agency 77^?° ,b; L^ary (F[ . i :, ,)'
Research Triangle Park, NC 27711
December 1992
-------
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. Any mention of trade names or commercial
. products is not intended to constitute endorsement or recommendation for use.
EPA^54/R-92-026
11
-------
ACKNOWLEDGEMENTS
This report was prepared by LuAnn Gardner of Systems Application International of San Raphael,
California. The work was conducted under EPA Contract No. 68-DO-0102, Work Assignment
No. 3-3. The author 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,
Mild Bouley and Chet Wayland of the Office Of Air Quality Planning And Standards (OAQPS),
EPA; and Marianne Causley and Jeremy Heiken of Systems Application International.
Additionally, the author wishes to acknowledge the work of Tom Lahre and David Misenheimer
(OAQPS/EPA); Jay Haney and Ralph Morris (Systems Application International); and Lowell
Wayne (Pacific Environmental Services) and Keith Rosbury (Pedco Environmental) in preparing
the previous two versions of this document
m
-------
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 tor carbon monoxide
(CO) and precursors of ozone (O3). 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 NC^. These inventories are required of the more serious O3 and
CO nonattainment areas only.
This is the second revision of this volume, which updates the versions released in 1979 (EPA-450/4-
79-018) and 1991 (EPA-450/4-91-014) to include current information pertinent to gridding,
speciation, and temporal allocation of emission inventories of CO and precursors of O3. This edition
includes changes and additions from the previous versions as summarized below:
o Inclusion of an additional section containing a brief overview of the Urban Airshed Model
(UAM) and the UAM Emissions Preprocessor System, Version 2.0-(EPS 2.0).
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 emission estimates by emissions component (e.g., exhaust,
evaporative) 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 EPS 2.0.
o Discussion of considerations specific to modeling for CO nonattainment applications.
-------
EXECUTIVE SUMMARY
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 of Ozone, Volume I: General Guidance for Stationary Sources, which outlines procedures
for compiling basic annual and seasonal emission inventories at a spatial resolution of county,
township, or equivalent level. Volume II provides guidance for identifying and incorporating the
additional detail required by photochemical air quality simulation models into an existing inventory of
the type described above, with a> special emphasis on fulfilling the input requirements of the Urban
Airshed Model.
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 described in Volume I.
The primary .additional requirements of the photochemical modeling inventory are summarized below.
i
o Emission estimates of precursor pollutants must be provided for each individual cell of a
grid system within the modeling domain instead of at a county or regional level;
o Emissions must be specified as hourly rather than annual or daily rates. Additionally,
annual or seasonal average rates should be adjusted to reflect episodic or day-specific
conditions as accurately as possible.
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)
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.
VII
-------
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 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 collection effort, which 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
«*
i
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 or seasonal county-level
emission inventory, generated in accordance with the methodologies 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 me proportions of VOC ana NOX to be assigned to the chemicai 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^
viii
-------
EXECUTIVE SUMMARY
for point sources, mobile sources, and area sources; specific data handling considerations are also
addressed for each of these source types. 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, general 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 widi the emission inventory data and classification scheme are
addressed, and special data handling considerations are outlined.
ix
-------
Contents
Title page : i
Disclaimer ii
Acknowledgements iii
Preface v
Executive Summary vii
List of Tables xiv
List of Figures xvii
1 INTRODUCTION 1-1
1.1 Purpose 1-1
1.2 Background 1-2
1.3 Contents of Volume II 1-3
2 INVENTORY PLANNING AND DESIGN CONSIDERATIONS 2-1
2.1 Selection of the Modeling Region and Grid System 2-1
2.2 Data Collection 2-2
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 2-5
2.3 Preparation of the Modeling Inventory 2-5
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 2-9
2.4 Emission Projections 2-9
2.5 Data Handling 2-12
2.6 Resource Requirements 2-13
2.7 Overview of Emission Inventory Planning Procedures 2-13
3 OVERVIEW OF THE URBAN AIRSHED MODEL (UAM) AND THE UAM
EMISSION PREPROCESSOR SYSTEM 3-1
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
3.3.1 EPS 2.0 Core Modules1 3-4
3.3.2 EPS 2.0 Utilities -. . . . 3-12
xi
-------
3.3.3 EPS 2.0 Input Requirements 3-15
3.3.4 EPS 2.0 Interface and Emission Display System 3-17
4 DETERMINATION OF THE GRID SYSTEM 4-1
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 4-6
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 4-11
5 POINT SOURCE EMISSIONS 5-1
5.1 Data Collection 5-1
5.2 Rule Effectiveness 5-3
5.3 Spatial Resolution 5-4
5.4 Temporal Resolution 5-4
5.5 Point Source Projections 5-7
5.5.1 Individual Facility Projections * 5-7
5.5.2 Aggregate Point Source Projections 5-8
5.5.3 Accounting for Regulatory Controls in Baseline Projections .... 5-14
5.5.4 Control Strategy Projections 5-15
5.5.5 Point Source Projection Review and Documentation 5-16
5.6 Data Handling Considerations 5-17
6 AREA SOURCES 6-f
6.1 Introduction 6-1
6.2 General Methodology for Spatial Resolution 6-6
6.2.1 Direct Grid Cell Level Determination of Emissions 6-6.
6.2.2 Surrogate Indicator Approach 6-7
6.3 General Methodology for Temporal Resolution 6-28
6.4 Area Source Projection Procedures 6-31
6.5 Data Handling Considerations 6-39
7 MOBILE SOURCE EMISSIONS 7-1
7.1 Introduction 7-1
7.2 Characterization of On-Road Motor Vehicle Emissions 7-6
7.2.1 Vehicle Classes 7-6
7.2.2 Roadway Types 7-6
7.2.3 Emission Components 7-8
7.3 Mobile Source Emission Factors 7-10
7.4 Mobile Emission Inventory Procedures 7-10
xii
-------
7.5 Spatial Resolution of Mobile Source Emissions 7-14
7.5.1 Link Surrogates 7-14
7.5.2 Non-link Mobile Emission Spatial Surrogates 7-17
7.6 Temporal Resolution of Mobile Source Emissions 7-19
7.7 Mobile Source Projections 7-21
7.7.1 VMT Projections 7-21
7.7.2 Future-Year Emission Factors 7-22
8 BIOGENIC EMISSIONS 8-1
8.1 Introduction 8-1
8.2 Overview of the UAM-BEIS 8-1
8.2.1 Leaf Biomass Factors 8-2
8.2.2 Emission Factors 8-2
8.2.3 Environmental Correction Factors 8-6
8.2.4 Processing Methodology 8-6
8.3 UAM-BEIS Input Requirements 8-10
8.4 Projection of Biogenic Inventories 8-13
9 SPECIATION OF VOC AND NOX EMISSIONS INTO CHEMICAL CLASSES 9-1
9.1 Introduction 9-1
9.2 The Carbon Bond-IV Mechanism 9-1
9.3 Chemical Allocation of VOC 9-1
9.4 Specification of NOX as NO and NOj 9-6
9.5 Projection of VOC and NOX Split Factors 9-7
9.6 Compatibility with Inventory Data and Source Categories 9-7
9.7 Data Handling Considerations 9-9
Appendix A:
Appendix B:
Appendix C:
Acronyms and Glossary
EPS 2.0 Reporting Codes
EPS 2.0 Temporal Profiles
xiii
-------
Tables
3-1 Input and output files for EPS 2.0 core modules 3-6
3-2 Types of data included on each record in the AFS work file 3-16
3-3 Types of data included on each record hi the AMS work file
for Area and Mobile Sources 3-18
5-1 Types of emissions data contained in the AIRS Facility Subsystem
and the level of detail at which each is maintained 5-2
5-2 Industrial groupings for BEA economic projections 5-10
5-3 Employment by place of work, historical years 1973-1988 and
projected years 1995-2040, for California 5-12
5-4 Example temporal factor file for individual point sources 5-19
6-1 Area source categories required for consideration in State Implementation
Plan emission inventories 6-2
6-2 Example spatial allocation surrogates for selected area source categories 6-8
6-3 Additional sources of information for spatial resqlution of emissions for
selected area source categories 6-9
6-4 Land use classification system used in USGS land use data bases 6-11
6-5 Land use categories for Tampa Bay area land use map 6-16
6-6 Demographic parameters used in San Francisco Bay Area for making zonal
allocations of area sources 6-20
6-7 Excerpt from ABAG cross classification table used in San Francisco Bay Area
for subcounty allocation of area source activities 6-21
6-8 Illustrative excerpts from zone-to-grid-cell correspondence table for
determining apportioning factors 6-24
6-9 Ozone season adjustment factors for selected area source categories 6-29
xv
-------
6-10 Diurnal patterns for gasoline stations in Tampa Bay, in percent of
daily operation 6-32
6-11 Example temporal resolution methodologies for selected area source categories 6-33
6-12 Preferred growth indicators for projecting emissions for area source categories 6-36
6-13 Example file of grid cell apportioning factors for area sources 6-40
7-1 AIRS AMS codes for onroad mobile sources 7-2
7-2 AIRS AMS codes for offroad mobile sources 7-3
7-3 AIRS AMS codes for aircraft, vessels, and locomotives 7-5
7-4 Vehicle class definitions used by the MOBILE models 7-7
7-5 Commonly used road type designations 7-7
7-6 EPS 2.0 internal source category codes for onroad motor vehicles 7-9,
7-7 Required input parameters for EPA's MOBILE models 7-11
7-8 Optional input parameters for EPA's MOBILE models 7-11
8-1 Vegetation types employed by the UAM-BEIS for user-specified county-level or
gridded land use data 8-3
8-2 Biomass density factors (g/m2) by forest group for each emission category . . . 8-3
8-3 Carbon Bond IV speciation for UAM-BEIS biogenic species 8-4
8-4 Biogenic emission factors (/xg/g'/h) by forest emission category used by UAM-BEIS
for canopy vegetation types and urban trees . ." 8-4
8-5 Emission rates Qng/m2-hr) and chemical speciation employed by UAM-BEIS for non-
canopy land use types 8-5
9-1 CBM-IV chemical species recognized by the UAM system, with molecular weights
for unit conversion 9-2
9-2 Example VOC speciation profile from die Air Emissions Species Manual .... 9-4
?-3 Zxamoie 'iniit factor" rlie 9-iO
xvi
-------
Figures
3-1 Overview of the EPS 2.0 core system 3-5
3-2 EPS 2.0 input file preparation utilities 3-13
3-3 EPS 2.0 support and reporting utilities 3-14
4-1 Schematic illustration of the use of the grid in the
Urban Airshed Model 4-2
4-2% The St. Louis area with a 4 km by 4 km modeling grid superimposed 4-3
4-3 UAM modeling region for the California South Coast Air Basin 4-7
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 4-8
4-5 Rotated modeling region encompassing the southern San Joaquin Valley
and Sierra Nevada 4-10
i«
6-1 Conceptual representation of the grid cell identification process 6-12
6-2 County grid cell assignments for the Atlanta, Georgia modeling region 6-13
6-3 Segment of land use map for Tampa Bay, Florida 6-15
6-4 Location of block group enumeration centroids for the Atlanta, Georgia
modeling region 6-26
6-5 Sample gridded population data for the Atlanta, Georgia modeling region .... 6-27
7-1 Depiction of typical link and grid cells 7-15
7-2 Onroad mobile source link surrogates developed for a UAM application of
the Dallas/Fort Worth region 7-18
7-3 Gridded annual average onroad mobile source emissions for a UAM application
of the Dallas/Ft. Worth region 7-20
8-1 Biogenic emission factor sensitivity to leaf temperature 8-7
xvii
-------
8-2 Schematic representation of forest canopy types 8-8
8-3 Temperature and solar flux variations by canopy layer 8-9
8-4 UAM-BEIS input and output files 8-11
xviii
-------
INTRODUCTION
1 INTRODUCTION
1.1 PURPOSE
This document supplements Procedures for the Preparation of Emission Inventories for Carbon
Monoxide and Precursors of Ozone, Volume I: General Guidance for Stationary Sources. Volume I
outlines procedures for compiling annual and seasonal emission inventories, which provide the basis
for development of the emission inventories required for use with photochemical grid models such as
the Urban 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 county, township, or equivalent level.
Volume 11 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 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 aiready described in detail in Volume I will not be repeated here. Thus, die reader should be
familiar with the contents of Volume I in order to thoroughly understand the procedures described in
this document.
Since the EPA-recommended photochemical model for urban applications is the Urban Airshed Model
(UAM), this document emphasizes methods for preparing emission inventories that fulfill the input
1-1
-------
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 developed.
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 concentra-
tions 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 photo-
chemical 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.
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 comroi 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. Photochemical models are also useful in basic scientific research, such as
in validation studies of atmospheric photochemistry and dispersion mechanisms.
1-2
-------
INTRODUCTION
Grid models (also called Eulerian models) calculate pollutant concentrations at fixed locations in space
at specified times. The concentrations estimated at each location result from interaction among
emissions, chemical reactions, and transport and dilution introduced by prevailing meteorological
conditions. Pollutant concentrations are calculated for each cubicle of a three-dimensional framework
in the entire region of interest. A cubicle might have horizontal dimensions of 1 to 10 kilometers on
a side and be 50 to 500 meters deep. Some Eulerian models are designed to provide vertical (as well
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 described 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 1.
The primary requirements of the gridded 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 modeling domain instead of at a county or regional level;
o Emissions must be specified as hourly rather than annual or daily rates. Additionally, annual
or seasonal average rates should be adjusted to reflect episodic or day-specific conditions as
accurately as possible.
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
o 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 documenr 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 II and its relationship to Volume /; it also includes an
introductory description of photochemical air quality simulation grid models and their emission
inventory requirements. Chapter 2 describes various technical considerations that aid in the planning
and design of the detailed emission inventory process. Chapter 2 is intended to provide an overall
perspective of the detailed inventory requirements for those who will actually be utilizing the
remainder of the document. Chapter 3 provides a brief overview of the Urban Airshed Model
1-3
-------
(UAM) and the UAM Emissions Preprocessor System, Version 2.0 (EPS 2.0), and Chapter 4
addresses selection of an appropriate modeling region and grid system. Finally, Chapters 5 through 9
provide detailed "how to" procedures for supplying the additional inventory detail required by the
photochemical grid model for point, area, mobile, and biogenic sources, including chemical speciation
of VOC and NOX emissions.
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 EPS 2.0 will be'enclosed in a gray box, as
shown here.
1-4
-------
PLANNING
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 the Empirical Kinetic Modeling
Approach (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, 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 mc;ude ail major emission sources which .may affect ozone formation in the urban area;
o to encompass as many ozone and precursor pollutant monitoring stations as possible (which
facilitates model validation);
2-1
-------
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 tune 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 of
a grid spacing, however, will result in excessive manpower and computer resource requirements,
because data must be collected and compiled for every grid cell hi 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 modeling 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.
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 .
2-2
-------
PLANNING
construction of various inventories for evaluating control strategies and/or analyzing the sensitivity of
model-predicted air quality parameters to emissions.) It is assumed that a conventional annual or
seasonal county-level 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
distributions, and for estimating the proportions of VOC and NO, to be assigned to the chemical
species 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 inventory must be examined to 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 NOX emissions into chemical classes and .detailed hourly emissions
information. Likewise, most of the county-level area source activity levels in existing local
inventories can be used as the basis for the model ing 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 episode.
2-3
-------
The only important source category for which the existing inventory does not ordinarily represent a
good starting point is highway motor vehicles. In 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-Iink 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, die 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.
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 rninimize resource
requirements, the data collection effort should focus on supplementing the existing spatial or temporal
resolution data for these sources. Many sources emit such minor amounts of VOC and NOX that
little, if any, additional effort is warranted in gathering temporally and spatially resolved data
regarding them.
Finally, the agency preparing the modeling inventory must work closely with the local metropolitan
planning organization (MPO) or other planning agencies in the area to determine what transportation
and land use planning models are currently being employed and what data from these models can be
2-4
-------
PLANNING
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, NOZ, and CO
The ozone modeling inventory development effort should be directed primarily toward obtaining
accurate emission data for VOC, NO,, 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 (UAM),
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.
2.2.4 Elevated Point Source Requirements
Some photochemical models, including the UAM, 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 ^js 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 mentioned previously, the
existing inventory will usually contain this information, and should be examined and utilized to the
greatest extent possible in order to minimize additional costs 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.
2-5
-------
23.1 Spatial Resolution of Emissions
In order for photochemical models to provide spatially resolved predictions of ozone 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 hi 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 most accurate (and
resource-intensive) approach is to obtain area source activity level 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 spatial apportioning 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,
population, or in some cases, employment statistics at the subcounty level; this data can be used to
spatially apportion 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 could then be
assigned to the appropriate grid cells.
Planning, land use, and transportation models are already hi 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 die 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 all the necessary- information that should be available from the MPO requires
2-6
-------
PLANNING
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 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 tune 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-attainment 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).
--K
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.
2-7 .
-------
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
NO2. (Some models do not require a NOX breakdown because they assume all NOX emissions to be
NO.) Literally hundreds of 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 UAM 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
several methods that can be employed for allocating VOC emissions because the same VOC
distribution is assumed to apply to each facility or process within a given source category.
In some instances, source-specific VOC species data may be available for certain individual facilities
(perhaps through source tests or material composition considerations), and the emissions modeler may
prefer 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.
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 pan 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 aiinimize 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.
2-8
-------
PLANNING
2.3.4 Special Considerations for CO Nonattainment Inventories
The 1990 Clean Air Act Amendments (CAAA) require the development, for each CO nonattainment
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 the 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 ozone precursors and
CO; a few special considerations for CO inventories, however, should be mentioned.
Ozone modeling episodes generally encompass several consecutive days; by contrast, CO simulations
are usually applied for much shorter time periods (e.g., 8 to 16 hours). Consequently, accurate
hourly allocation of emissions becomes more critical for CO simulations, and all temporal data for the
inventory should be carefully evaluated for both accuracy and completeness. With regard to point
sources, accurate and complete stack parameter data (see Section 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 both CO and ozone nonattainment
area emission inventories, consult Emission Inventory Requirements for CO SIP Nonattainment Areas
and Emission Inventory Requirements for Ozone SIP Nonattainment Areas (EPA, 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 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 RFPs 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
2-9
-------
nonattainment 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 persons responsible for RFP tracking to ensure consistency of projections and projection
methodologies. Consult EPA guidance on emission inventory projection techniques and on RFP
preparation for further information.
In many respects, the baseline projection modeling inventory will be the same as the baseline
projection inventory of annual or seasonal county-level emissions compiled for ozone nonattainment
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 shouid 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.
o To the 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 tfie near-term), and because highway vehicle and area
source emission patterns will directly reflect changes in the land use, employment, and transportation
data supplied by local planning agencies. With the exception 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 year inventory either because no changes are expected or because no data will be available to
forecast such changes. These considerations are discussed in more detail in succeeding chapters.
2-10
-------
PLANNING
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 tr^eat these particular stations as point sources rather than
lumping them ip with a general*^ervice station area source category.
o 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.
o Projection inventories will always be subject to criticism because of their somewhat
speculative nature. The technical credibility of emissions projections will be a function of
their reasonableness, the amount of research and documentation of assumptions, and the
procedures and methodologies used to make the projections. Some degree of uncertainty will
always accompany emission projections; this fact should be acknowledged openly. When
developing projection inventories, the emissions modeler should focus on minimising instead
of eliminating uncertainty. Internal and external review of the projection inventories will
improve their technical quality and enhance their credibility.
2-11
-------
2.5 DATA HANDLING
The large amount of data that must be gathered 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 hi 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 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 are usually disaggregated into species classes through the use of
an appropriate species distributionjbr &ch source category. NOX emissions are either assumed to be
all NO or are split into NO and NC^. 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
accordance 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 activity 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 model 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,
2-12
-------
PLANNING
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.
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 die emissions modeler in making emission projections.
2.7 OVERVIEW OF EMISSION INVENTORY PLANNING PROCEDURES
The remaining chapters 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 for each 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:
2-13
-------
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?
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 (hi 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 NO, 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?
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.
2-14
-------
UAM/EPS OVERVIEW
3 OVERVIEW OF THE URBAN AIRSHED MODEL (UAM)
AND THE UAM EMISSION PREPROCESSOR SYSTEM
3.1 INTRODUCTION
Under the 1990 Clean Air Act Amendments, air quality analyses using photochemical grid models are
required for areas designated as serious, severe, and extreme ozone nonattainment areas and multistate
moderate ozone nonattainment areas. In 1984 the EPA's Office of Air Quality Planning and
Standards proposed that the Urban Airshed Model (UAM) be a "recommended" (i.e., preferred)
model for "photochemical pollutant modeling applications involving entire urban areas." EPA
finalized this recommendation in 1986, noting that the UAM "is the most widely applied and
evaluated photochemical model in existence." Currently, the UAM is the recommended air quality
simulation model for use in ozone air quality analyses in the preparation of State Implementation
Plans (SIPs) as required in the 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;
4
o chemical composition of the emitted NOX and VOC;
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, NOX, 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, immediately
upwind of, and above the study region.
In a UAM application, these processes are simulated for the pollutant of interest (this may be either
summertime ozone concentrations or wintertime carbon monoxide concentrations). The UAM solves
the species continuity equation for each time step, in each grid cell of the modeling domain; the
maximum time step is a function of grid size and the maximum wind velocity. Typical time steps for
urban-scale simulations are oh 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. In 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 was made
available to the public in 1990 as the UAM Emissions Preprocessor System (EPS), version 1.0.
Since the passage of the 1990 Clean Air Act Amendments, a growing emphasis on the use of the
UAM in regulatory applications has led EPA to fund a series of enhancements to the initial EPS,
aimed at improving flexibility, providing a more efficient and user-friendly processing environment,
and expanding the capabilities of the system, especially regarding the implementation of proposed
control strategies. The enhanced system is called EPS
3-2
-------
UAM/EPS OVERVIEW
EPS. 2.0 provides the user with expanded capabilities to support the CAAA requirements, conform to
EPA emission inventory requirements, and provide a method for evaluating proposed control
measures for meeting RFP regulations. Regulatory requirements call for hydrocarbon to be reported
as reactive volatile organic compounds, motor vehicle emissions to be adjusted hourly for temperature
effects, and all State Implementation Plan emission inventories to be submitted to a central data base
(AIRS). EPS. 2.0 has features to process emission inventories that meet these requirements. Further
capabilities include processing of a link-based emissions, point-specific speciation, and locale-specific
temporal profiles.
The flexibility of EPA 2.0 provides the users with many options for preprocessing their emissions
inventory. The design provides the users with (1) a "turn-the-crank11 system for generating modeling
inventories, and (2) a means for the discriminating user to implement detailed, locally available data
such as source-specific speciation, temporal information, and episode specific emissions.
EPS 2.0 consists of a series of FORTRAN modules that perform the intensive data manipulations
required to incorporate spatial, temporal, and chemical resolution into an emissions inventory used for
photochemical modeling. The modules can be classified into four major components of the system:
(1) core EPS modules, (2) input preparation utilities, (3) support utilities to manipulate the internal
record format, and (4) reporting utilities.
Before executing EPS 2.0, 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 which point sources will
receive elevated (i.e., vertically resolved) treatment by the model. For guidance on selection
of an appropriate plume height cutoff, consult Guideline for Regulatory Application of the
Urban Airshed Model?-
o Run the EPA mobile source emission factor model MOBILE (Version 4.1 or higher) 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:
3-3
-------
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 Manual for the Emissions
Preprocessor System 2.0, Pan A: Core FORTRAN System (EPA-450/4-90-007D (R), May
1992), and
o User's Guide to MOBILE4.1 (Mobile Source Emission Factor Model) (EPA-AA-TEB-91-01,
June 1991) or User's Guide to MOBILES.
3 J.I EPS 2.0 Core Modules
Figure 3-1 provides an overview of the EPS 2.0 core system. In this figure, the ten core modules
have been divided into three categories: data loading modules (PREPNT, PREAM, and LBASE),
preprocessing modules (CNTLEM, CHMSPL, TMPRL, RPRTEM, PSTPNT, and GRDEM), and
final preprocessing modules (MRGUAM). After data has been loaded into EPS 2.0, the remaining
modules (except for MRGUAM, which reads UAM-formatted emissions files) employ a common
internal file format, referred to as EMBR ("Emissions Model Binary Record") format. This allows
these modules to be executed in any order, with the exception that the PSTPNT and GRDEM
modules must be executed last since these two modules produce the UAM-formatted emissions files.
(Note that the UAM preprocessor PTSRCE must be executed subsequent to EPS 2.0 to prepare the
final UAM elevated point source input file needed for modeling.) Table 3-1 lists the input and output
files for each of the core modules.
The ten modules and their primary functions are briefly described below. For a detailed description
of EPS 2.0 and its input file formats, see the User's Guide for the Urban Airshed Model: Volume IV,
Pan A?
PREPNT. The PREPNT module is the entry point to EPS 2.0 for point source emissions data. The
primary functions of PREPNT are
o identification of sources within the modeling domain;
o initial screening of which sources are to be treated as elevated point sources; and
o reformatting of an emissions inventory generated by the AIRS AFS into EMBR format.
PREPNT reads point source data in AIRS Facility Subsystem (AFS) work file format, including stack
parameters, process and geographical identification codes, and emission rates for any or all of the
following pollutants: NOX, VOC (or THC), CO, SOX, TSP, and PM-10. PREPNT converts the
emissions to an average daily rate for the time period of the AFS work file. PREPNT also calculates
a plume rise for each stack based on the Briggs effective height calculation (using default
3-4
-------
VAM/EPS OVERVIEW
Inventory Inputs
/ ATS / / MM /
/Point T / AIM. meoue/
Inventory / / im>M<«ni/ / '""ton/
^^^^.^M^^ /
Dirts Loading ±
MIPNTI I MfAM 1 I LBASE
Input point NotonrMt input potielly aliocat*
rnrtlel eoreenlng tor eeleeted WM rtfonntt to
tor «*!« Mure** rogtent rnttrnti format
t
/ /fmiMion*
,_ t / Me«*i
i ~/_
: f
Emissions Preprocessing I
CNTLEM I I CHMSPL
Apply pro|«ction«. CncmiMlly ip*elat* Twnperally llee«t«
r»gul«tery controlt. tor o*t«mie>l tor hour* of mooting
nd di»er«(ion«ry conwoli m*chtn»m ceitoa*
*»"TEM HTPNT aUDEM
O»n«rit« r>rae*t« **!« Sp*ti*|ly nocit*
total mvvntory pemt toure* arw miMion*
'pon« miwient
Final Emissions
Preprocessing
ln»w«ory / UAM
repent / ol»»«t«« point / / / ,, '
/ ooure* mput / / / i 7 MRGUAM
/ / / °*1"1 lew / p M»r9» multiple
level emieewn / UAM low level
*e» / emiuion diet
togenle Emi«*on / M.,gw, UAM
FIGURE 3-1. Overview of the EPS 2.0 core system.
3-5
-------
TABLE 3-1. Input and output files for EPS 2.0 core modules.
Module Input files
Output flies
PREPNT
PREAM
USERIN: /PREPNT/, /UAMREGN/, /COUNTY/
AFS work file
SCC(ASC)/speciation profiles cross reference
PREPNT message file
EMBR point source file
stack report
error records file (AFS format)
USERIN: /PREAM/ (optional), /COUNTY/
AMS work file
SCC(ASC)/speciation profiles cross reference
MADJIN: /DISAG/
PREAM message file
EMBR area source file
EMBR motor vehicle file
error records file (AMS format)
LEASE USERIN: /LEASE/, /UAMREGN/, /COUNTY/ LEASE message file
link based emissions file EMBR motor vehicle file
SCC(ASC)/speciation profiles cross reference error records file (link based data format)
MADJIN: {DISAG/ __
CNTLEM USERIN: /CNTLEM/, /COUNTY/ CNTLEM message file
control factors file EMBR file
SCC(ASC)/speciation profiles cross reference EMAR error records file
SIC/SCC(ASC) reporting codes glossary
MADJIN: /MVCNTL/
EMBR file
CHMSPL USERIN: /CHMSPL/, /EPISODE/ CHMSPL message file
SCC(ASC)/speciation profiles cross reference EMBR file
carbon bond split factors EMAR error records file
EMBR file
TMPRL USERIN: /TMPRL/, /EPISODE/ TMPRL message file
SCC(ASC)/temporal profiles cross reference EMBR file
temporal profiles EMAR error records file
MADJIN: /MVTMPRL/ (optional)
EMBR file
PSTPNT USERIN: /PSTPNT/, /EPISODE/, /UAMREGN/ PSTPNT message file
EMBR point source file ASCII elevated point source emissions file
GRDEM USERIN: /GRDEM/, /EPISODE/, /UAMREGN/ GRDEM message file
SCC(ASC)/spatial surrogate cross reference EMBR or UAM low level emissions file
gridded spatial surrogates EMAR error records file
link data (optional)
EMBR file
MRGUAM USERIN: /MRGUAM/ (optional), /EPISODE/ MRGUAM message file
one to ten UAM low level emissions files merged UAM low level emissions file
RPRTEM RPRTEM user input file RPRTEM message file
SIC/SCC(ASC) reporting codes glossary summary report file
SCCfASCVspeciation profiles cross reference
reporting code descriptions
one to six EMBR files
source: Reference 3
3-6
-------
UAM/EPS OVERVIEW
meteorological parameters representing stable conditions), 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.
PREAM. The PREAM module serves as the entry point to EPS 2.0 for area and non-link based
mobile source emissions data. The primary functions of the PREAM module are
o separation of area and onroad motor vehicle emissions data into two files;
o disaggregation of onroad motor vehicle emissions into emissions components; and
o reformatting of an emission inventory generated by the AIRS Area and Mobile Subsystem
(AMS) into EMBR format.
PREAM reads the AMS work file, separates the data into onroad motor vehicle sources and area
sources ("which include offroad mobile sources), converts the emissions to an average daily rate for
the time period of the AMS work file, and writes the output records to the appropriate EMBR files.
For motor vehicle sources, PREAM also disaggregates total emissions into exhaust, evaporative,
running loss, and resting loss components. In the event that link-based emissions data are available
and have been processed by the LEASE module, PREAM will subtract the link-based emissions totals
(provided by the user) from the county totals before processing the AMS record.
LBASE. The LBASE module incorporates link-based emissions estimates into the modeling
inventory; like the PREPNT and PREAM modules, LBA*SE serves as an entry point for the EPS 2.0
system. The primary functions of the LBASE module are
o identification of links within the modeling domain;
o spatial allocation of link-based emissions to grid cells;
o disaggregation of motor vehicle emissions into exhaust, evaporative, running loss, and resting
loss components; and
o reformatting of link-based emissions estimates into the EMBR format.
LBASE reads the link-based emissions data, processes the data for each link, and writes the processed
data to an EMBR file. The data may be hourly, seasonal, or annual average emission rates, and may
include any or all of the following pollutants: NOX, VOC, CO, SOX, TSP, and PM-10. LBASE
spatially allocates the link-based emissions by computing the fraction of the total link length contained
in each cell traversed by the link. Onroad motor vehicle emissions are also separated into exhaust,
evaporative, and running loss components; this disaggregation is required for correct Carbon Bond
Mechanism speciation of the emissions by the CHMSPL module.
3-7
-------
CNTLEM. The CNTLEM module allows the user to simulate the effects of various control
strategies on the emissions contained in the inventory. This module provides the user with methods
for
o projecting (or backcasting) a base year emission inventory to represent emissions levels for
another year, based on user-specified projection factors for changes in activity levels by
source category;
o assessing the effects of mandated regulatory controls (e.g., Maximum or Reasonably
Available Control Technologies, Control Technique Guideline controls, or allowable
emissions limits) on projection year inventories;
o assessing and comparing the effects on emissions levels (and consequently air quality) of
different control strategies under consideration.
CNTLEM may be executed at any stage of processing, but it is recommended as the first step after
the,input inventory has been loaded into EMBR format (i.e., immediately after running PREPNT,
PREAM, or LEASE). If the input EMBR file has been chemically speciated (using the CHMSPL
module), CNTLEM will apply control and projection factors to the appropriate carbon bond species
based on the criteria pollutant species from which each carbon bond species was derived.
Each record of the AFS and AMS work files from AIRS contains primary control equipment code,
control efficiency factor, rule effectiveness factor, and rule penetration factor fields; this information
is maintained in the EMBR files. CNTLEM assumes that the effects of existing control equipment
have been included in the emissions estimates in th&inpu\ AFS and AMS work files (in other words,
that the emissions represent controlled levels if control data is present for that source). All new
controls are assumed to represent replacement technologies.
CNTLEM will apply the following types of projection and control factors, which must be specified by
the user in the control factors input file:
o Control Technique Guideline (CTG) controls;
o Maximum Achievable Control Technology (MACT) and non-CTG Reasonably Available
Control Technology (RACT) controls;
o projection factors for changes in activity levels;
o onroad motor vehicle controls (and fleet turnover effects for projection inventories);
o other source- or source category-specific controls;
o allowable emissions limits; and
3-8
-------
UAM/EPS OVERVIEW
o discretionary control strategy controls (by source category, state or county FIPS, subregion
code, and user-specified subgrid, and by control strategy code type: activity, control, pod,
and speciation profile)
CHMSPL. The CHMSPL module allows the user to introduce the degree of chemical resolution
required by the UAM into the modeling inventory. The primary functions of CHMSPL are
o to disaggregate criteria pollutant emissions into the chemical species used in the Carbon Bond
IV (CB-IV) mechanism employed by the UAM; and
o to create a chemically speciated EMBR file for further processing by any of the other EPS 2.0
modules (excluding PREPNT, PREAM, and LEASE).
CHMSPL determines an appropriate speciation profile code based on the SCC code on the EMBR
record; the speciation profile code determines the "split factors" for that particular profile. The split
factors are multiplicative factors for converting grams of criteria pollutant emissions into moles of the
Carbon Bond Mechanism Version IV (CBM-IV) chemical species employed by the UAM. The user
may specify which carbon bond species will be generated for the UAM modeling application.
CHMSPL then applies the split factors to the emissions for each data record in the input EMBR file
and writes the speciated emissions for each of the selected species to an output emissions file, also in
EMBR format.
TMPRL. The TMPRL module allows the user to introduce the degree of temporal resolution (i.e.,
hourly) required for the modeling inventory by the UAM. The primary functions of TMPRL are to
o adjust annual or seasonal average emissions to episodic levels;
o apply hourly temperature adjustments to onroad motor vehicle emissions;
o allocate emissions to the hours of the modeling episode; and
o generate an EMBR file containing hourly emissions data.
If the input EMBR file contains average daily emissions derived from annual data or average daily
emissions for a specified period, these are first adjusted by applying a yearly profile to determine
emissions levels for the episode month. The emissions are then adjusted for the day of week
(Monday-Sunday) of the modeling episode based on weekly variations in activity levels. Finally, the
episode-adjusted daily emissions are temporally allocated to each hour using a diurnal variation
profile.
The Source Classification Code (SCC) for point sources or Area Source Category (ASC) code code
for area and mobile sources from each input emissions record is cross-referenced to a monthly, day of
week, and diurnal profile code which determines the temporal profiles applied to the emissions; the
codes currently defined in EPS 2.0 system input and glossary files are shown in Appendix C. To
3-9
-------
facilitate implementation of source- or geographical region-specific temporal profiles, the first three
fields of each record in the cross-reference file allow the user to also specify FIPS state and county
codes, plant identification codes, and AFS stack identification codes for which the temporal profile
should be applied. For each input EMBR record, TMPRL searches the cross-reference file to
determine the assigned temporal profiles for the FIPS state/county codes, facility identification codes,
and/or SCC(ASC) code combination most closely matching the input EMBR record.
PSTPNT. In addition to a low-level emissions file, the UAM requires a separate file containing
elevated source emissions data. Elevated source data must be processed through the UAM
preprocessor PTSRCE to create the binary file used by the UAM (the user will find a complete
description of the PTSRCE preprocessor in Volume n of EPA's User's Guide for the Urban Airshed
Model). Accordingly, the EPS 2.0 core system includes the PSTPNT module, whose primary
functions are
o to create the ASCII input file for the UAM PTSRCE preprocessor; and
o to provide the user with tabular summaries of elevated source emissions to assist in quality
control tracking.
Since the file of elevated source emissions data created by PSTPNT is intended for input directly to
the UAM PTSRCE preprocessor, PSTPNT must be executed after all other processing of the point
source data (in other words, PSTPNT should be the last module run for the point source emissions
data). PSTPNT reads the EMBR file to obtain the individual stack parameters and hourly emissions
data for each,,source which must be included in the PTSRCE input file. Only those EMBR records
which were assigned a record type of "E" (for "elevated") by the PREPNT module are processed and
written to the PTSRCE input file.
GRDEM. The GRDEM module allows the user to spatially allocate emissions to the grid cells of the
modeling domain. The GRDEM module will
o spatially allocate area and mobile sources based on a combination of gridded spatial surrogates
and (optional) link data;
o assign low-level point source emissions to grid cells based on source location; and
- o create either a gridded EMBR or UAM-format low-level emissions file.
For area sources, GRDEM spatially allocates regionally-aggregated emissions totals (e.g., county-
level or subregion-level emissions) by source category based on the spatial allocation surrogate for
that source- category, as specified in the SCC(ASC)/gridded surrogate cross-reference file. If.digitized
link data are available for some categories (e.g., light-duty automobiles on limited access roadways,
railroad locomotives, shipping vessels), GRDEM will determine the cells traversed by each link and
allocate emissions for that source category accordingly. Alternatively, the user may specify that
emissions from these sources be allocated based on a gridded spatial allocation surrogate. (Note the
3-10
-------
UAM/EPS OVERVIEW
distinction between link-based emissions, which are processed by the LEASE module and include the
effects of varying activity levels for each link, and spatial allocation of emissions using link data, in
which county-level emissions are allocated evenly along all links of that type within the county based
solely on the length of the link). For low-level point sources, GRDEM assigns emissions to grid cells
based on the UTM coordinates of the source.
GRDEM will produce a low-level gridded emissions file in either EMBR format or the format used
by the UAM. If the UAM format is chosen, the input EMBR file must have already been processed
through the TMPRL and CHMSPL modules to provide the hourly and chemical resolution required
by the UAM.
MRGUAM. Different types of emissions data (e.g., point, area, mobile, biogenic) can be processed
separately through EPS 2.0 to facilitate both control strategy analysis and quality control tracking.
Consequently, the EPS 2.0 core system includes a module for merging multiple emissions files into
one file for modeling. The MRGUAM module will
o combine the low-level emissions data from up to 10 UAM-format files into a single file; and
o apply domain-wide, across-the-board multiplicative factors by Carbon Bond Mechanism
(CBM) species for any (or all) of the input emissions files.
This second function permits the user to easily create emissions inventories for use in "control
sensitivity" applications of the UAM. As discussed in Section 2.4; this type of UAM application is
designed to estimate the overall amount of emissions reductions required to produce a desired change
in ozone concentrations. Control sensitivity applications should not be confused with control strategy
applications, which attempt to simulate as accurately as possible the effects of a specific set of control
measures applied at the individual source level.
Although different emissions files used in the same UAM application must all be appropriate to the
region and time period being modeled, they may contain different CBM species. Most UAM low-
level emissions files will have been created by the GRDEM module, with the notable exception of the
biogenic emissions file (which may be created using EPA's UAM Biogenic Emissions Inventory
System; refer to Chapter 8). MRGUAM will accept as input any UAM-ready low level emissions
file, regardless of its origin. It sums the emissions from all of the input emissions files, by hour, grid
cell, and CBM species, and writes the sum to a single file.
RPRTEM. The RPRTEM module summarizes emissions information hi a form that facilitates
comparative analysis of various control strategies, as well as information useful for quality control
tracking. RPRTEM allows the user to create
o reports of total emissions from up to six EMBR files; and
o summaries of the types of data (file type, included species, etc.) included in each input EMBR
file.
3-11
-------
RPRTEM will produce tabular reports of emission totals by species for different groupings of
emission sources. Sources can be grouped according to activity, process, control strategy, pod, or
speciation profile codes; these codes are defined in Appendix C. The user also has the option to
generate output tables by criteria pollutant, CBM species and selected counties or subregion. The
user should select the most applicable reporting category for the purposes for which the summary
output will be used.
Since RPRTEM reads EMBR files, it can be run at any stage of inventory processing after the entry
modules (PREPNT, PREAM, and optionally LEASE) have been executed. This flexibility allows the
user to immediately assess the effects of a particular step hi inventory processing on total emissions.
RPRTEM reads each record of the EMBR file and then looks up the appropriate reporting codes from
the SIC/SCC (ASC) reporting code glossary and cross-reference files based on the SIC and
SCC(ASC) codes specified in the EMBR record. If the SIC/SCC code pair for a particular EMBR
record is not included in the glossary, RPRTEM writes an appropriate message to a message file and
the reporting code is set to the default "miscellaneous" category.
3.3.2 EPS 2.0 Utilities
EPS 2.0 includes various support utilities for (1) generating input files required by EPS 2.0 core
modules, (2) manipulating the internal EMBR file structure, and (3) generating reports for quality
control. Figures 3-2 and 3-3 show the relationship of the input file preparation utilities and other
support utilities to the core EPS 2.0 system. The principle functions of the EPS 2.0 utility modules
are listed below:
BEAFAC Produce projection factors for point and area sources (used by CNTLEM).
EMSCVT Create or update the SCC(ASC)/speciation profiles cross reference file (required by
PREPNT, PREAM, LBASE, CNTLEM, CHMSPL, and RPRTEM) and the carbon
bond split factors file (read by CHMSPL).
MKGLOS Create the direct access reporting code glossary file used by CNTLEM and RPRTEM.
MVADJ Prepare the MADJIN motor vehicle adjustment factors file used by PREAM, LBASE,
CNTLEM, and TMPRL
TMPFAC Create or update the SCC(ASQ/temporal profiles cross reference file and temporal
profiles definition file required by TMPRL
QCEMBR Provide tabular summaries of the emissions data to augment the reports written by
each EPS 2.0 module. (The RPRTEM module also serves as a report utility, creating
summary reports of emissions totals for specified reporting categories.)
3-12
-------
UAM/EPS OVERVIEW
FIGURE 3-2. EPS 2.0 input file preparation utilities.
3-13
-------
/ UAM /
/PTSRCE/
j INPUT/
/Merged/
/ UAM/
FIGURE 3-3. EPS 2.0 support and reporting utilities.
3-14
-------
VAM/EPS OVERVIEW
EMBRET Convert the machine-dependent binary EMBR files to ASCII format for transfer to
other systems or modification using an ASCII test editor; will also extract selected
records based on user-specified criteria. The ASCII file produced by EMBRET is in
EMAR ("Emissions Modeling ASCII Record") format.
ATOBR Convert an ASCII EMAR file to binary EMBR format.
MRGEMB Merge two EMBR files into a single file.
3.3.3 EPS 2.0 Input Requirements
The EPS 2.0 system requires a variety of input data, which can be classified into five categories:
emissions data, system default parameters, region-specific data, episode-specific data, and optional
data.
Emissions Data. The user must supply the input emission inventory data that the EPS 2.0 requires to
process the modeling inventory. Before processing the emission data, the user should confirm that
the original inventory data that will be input to EPS 2.0 is appropriate for the purposes of the
modeling application. For example, a modeling inventory to be used for model performance
evaluation should reflect actual emission rates for the year of the episode. Accordingly, it would be
inconsistent to prepare the validation inventory by projecting (or backcasting) an inventory containing
allowable emission rates.
EPS 2.0 requires emissions data to be input as two distinct files: point sources and area sources. The
point source data must include physical stack parameters and data describing operational schedules as
well as total emissions. The area source emissions file should include small stationary sources not
included in the point source data file, and both offroad and onroad motor .vehicles. In order to
properly process the onroad motor vehicle portion of the inventory, the user must-also be able to
recreate the MOBILE 4.1 mobile source emission factor model runs (i.e., identify the values of the
MOBILE 4.1 input parameters) that produced the emissions factors used for estimating onroad motor
vehicle emissions in the input data file.
As stated above, the user must supply the emission inventory data for the EPS 2.0 system. These
data may either be retrieved from the AIRS system or developed independently by the user; in either
case the emissions data files must be formatted according to the AIRS Facility Subsystem (AFS) work
file format (for point source data) and the AIRS Area and Mobile Subsystem (AMS) work file format
(for area and mobile source data).
The AFS work file contains the point source data required by EPS 2.0 to process this portion of the
inventory. The information hi the AFS work file can be divided into seven types: inventory
description, geographical, source identification, stack characterization, operating schedule
information, control technology description, and emissions. Table 3-2 lists the individual data items
contained in the AFS work file for each of these seven categories. Certain data fields must contain
3-15
-------
TABLE 3-2. Types of data included on each record in the AFS work file. Arrows indicate
fields that must contain valid data for use in EPS 2.0. Missing data in other fields will be
replaced with default values or ignored.
Inventory Description Data
Inventory type as retrieved from AIRS:
Adjusted
Base
RFP
Modeling
Projection year
Base year
> Emissions type:
Actual
Allowable based on activity level limit
Allowable based on emission factor
limit
Allowable based on both activity level
and emission factor
limits
Allowable based on total emissions
Projected with base year controls
Projected with new controls
Allowable with base year controls
Allowable with new controls
> Inventory period:
Annual
Typical peak ozone season day
Typical peak CO season day
Other specified interval
> Period starting and ending dates and times
Geographical Data
> FIPS state and county codes
Subregion code (FIPS city code) (optional)
> Source location (geodesic or UTMs)
> UTM zone (only required if source location
is specified in UTM coordinates)
Source Identification Data
> Plant identification code
> Stack number
Point identification code
Segment number
> Standard Industrial Classification Code
> Source Classification Code
Stack Characterization Data
Height
Diameter
Exit gas temperature
Exit gas velocity
Operating Schedule Information
Seasonal percentages of annual throughput
Hours per day in operation
Days per week in operation
Hours/year in operation
Start hour (optional)
Control Technology Description
Primary control equipment code
Combined control effectiveness
Rule effectiveness
Rule penetration (not used for point
sources)
Emissions Data
t> AIRS pollutant code (SAROAD)
> Emissions for specified pollutant
source: Reference 3
3-16
-------
VAM/EPS OVERVIEW
valid data for use in EPS 2.0 (denoted in the table with an arrow, >). Although missing data in other
fields will be replaced with default values, the resultant modeling emission inventory will be less
accurately resolved, which will in turn affect the quality of the UAM modeling results.
Table 3-3 lists the data items contained in the AMS work file for area and mobile sources. The data
in this file can be divided into six categories: inventory description, geographical, source
identification, control technology description, and emissions. Data fields that must contain valid data
for use in EPS 2.0 are denoted with an arrow (>).
System Default Parameters. EPS 2.0 comes with a set of files containing default inputs for certain
data, which are intended to provide the user with an initial EPS setup. These files include:
o Speciation profiles for chemically allocating the emissions to the carbon bond species used by
the UAM (derived from data in EPA's Air Emissions Speciation Manual};
o Default speciation profile code assignments by Source Classification Code (SCC) and Area
Source Category (ASC) code;
o Default reporting code assignments (process, activity, control, process, and pod) by
SIC/SCC(ASC);
o Default reporting code descriptions;
o National average temporal allocation profiles;
I"
o Default temporal profile code assignments by SCC(ASC); and
o Economic and demographic projection data for developing projection factors (from the Bureau
of Economic Analysis).
In the absence of source- or region-specific data, the default speciation profiles and speciation profile
assignments and BEA projection data represent EPA-preferred data sources for speciation and
projection data, respectively. Likewise, the default reporting code assignments will prove adequate
for most EPS 2.0 applications. The user should review all default inputs for appropriateness for each
application.
3.3.4 EPS 2.0 Interface and Emission Display System
EPS 2.0 includes an optional interface and display system to assist the user in setting up an EPS 2.0
application and analyzing the modeling emissions data base created with EPS 2.0. The EPS Interface
is composed of two modules: the Setup module and the Graphics module. The Setup module
provides a user-friendly environment for creating or modifying several of the input files required to
3-17
-------
TABLE 3-3. Types of data included on each record in the AMS work file for Area and
Mobile Sources. Arrows indicate fields that must contain valid data for use in EPS 2.0.
Missing data in other fields will be replaced with default values or ignored.
Inventory Description Data
Inventory type as retrieved from AIRS:
Adjusted
Base
RFP
Modeling
Projection year
Base year
> Emissions type:
Actual
Allowable based on activity level limit
Allowable based on emission factor
limit
Allowable based on both activity level
and emission factor limits
Allowable based on'total emissions
Projected with base year controls
Projected with new controls
Allowable with base year controls
Allowable with new controls
>. Inventory period:
Annual
Typical peak ozone season day
Typical peak CO season day
Other specified interval
Period starting and ending dates and times
Geographical Data
> FIPS state and county codes
Subregion code (FIPS city code) (optional)
Source Identification Data
> Area Source Category code
Operating. Schedule Information
Period percentage of annual throughput
Days per week in operation
Weeks per year in operation
Hourly percentages of daily throughput
Weekday adjustment factor
Saturday adjustment factor
Sunday adjustment factor
Control Technology Description
Primary control equipment code
* Combined control effectiveness
Rule effectiveness
Rule penetration
Emissions Data
> AIRS pollutant code (SAROAD)
> Emissions for specified pollutant
source: Reference 3
3-18
-------
UAM/EPS OVERVIEW
run EPS 2.0. A series of interactive screens guides the user through the steps necessary to set up
user inputs and control factors for an EPS 2.0 application.
The Graphics module provides the user with tools for statistically analyzing the data emission
inventory data and graphically displaying the results. The display program supports a variety of
statistical, temporal, and spatial graphics (including bar charts, summary tables, time series plots, and
shaded tile maps). The user may perform quantitative analysis of emissions using the statistical
graphics, which allow the user to compare information for selected chemical species and/or source
types for either the entire modeling domain or selected counties within the domain. The temporal
graphics allow the user to portray variations in emissions over time; as for the statistical graphics, the
user may select chemical species, source types, and counties for which to display hourly variations in
emissions. The spatial graphics, while ignoring temporal variations in the data, provide valuable
information regarding the spatial distribution and variation hi emissions. The user may display either
total emissions for all source types for the modeling domain, or domain-wide emissions for a selected
source type. The Graphics module supports both EMBR and UAM emissions files formats, although
some options (e.g., display emissions for a selected source type) may not be available for UAM low
level emissions files, since the UAM format for low level emissions files only specifies emissions by
hour, chemical species, and grid cell.
Refer to the User's Guide for the Urban Airshed Model, Volume IV: User's Manual for the Emissions
Preprocessor System 2.0, Pan B: Interface and Emission Display System for information regarding
computing requirements, available options, and using the menu-driven screens for the interface and
emission display system.
90098* 03 3-19
-------
References for Chapter 3:
1. User's Guide for the Urban Airshed Model, Volume I: User's Manual for UAM (CB-IV), EPA-
450/4-90-007A, U.S. Environmental Protection Agency (OAQPS), Research Triangle Park, NC,
June 1990.
2. Guideline for Regulatory Application of the Urban Airshed Model, EPA^50/4-91-013, U.S.
Environmental Protection Agency (OAQPS), Research Triangle Park, NC, June 1991.
3. User's Guide for the Urban Airshed Model, Volume IV: User's Guide for the Emissions
Preprocessor System 2.0, Part A: Core FORTRAN System, EPA-450/4-90-007D (R), U.S.
Environmental Protection Agency (OAQPS), Research Triangle Park, NC, May 1992.
4. User's Guide for the Urban Airshed Model, Volume IV: User's Guide for the Emissions
Preprocessor System 2.0, Pan B: Interface and Emission Display System, EPA-450/4-90-007D
(R), U.S. Environmental Protection Agency (OAQPS), Research Triangle Park, NC, May 1992.
3-20
-------
GRID SYSTEM
4 DETERMINATION OF THE GRID SYSTEM
4.1 SELECTING AN APPROPRIATE GRID SYSTEM
Identification of the grid system which will be used to spatially reference emissions in the modeling
inventory influences all 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 guidance concerning establishment of the grid system, however, consult
the Guideline for Regulatory Application of the Urban Airshed Model.l
The first step in defining the grid system is selection of a grid boundary outlining the area to be
modeled. Once the grid boundary has been chosen, the region enclosed by the grid boundary
(subsequently referred to as the "modeling region" in this text) must be subdivided into grid cells.
Figure 4-1 illustrates the concepts of grid boundary and grid cells. The UAM 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 UAM 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 aiid
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 ceils 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.
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
4-1
-------
(a) The Area to Be Modeled
(b) Specification of the Grid
FIGURE 4-1. Schematic illustration of the use of the grid in the Urban Airshed Model.
4-2
-------
GRID SYSTEM
o
MMXWIU.I
}.
I
I
fUM
.
W i
. :
1
:
V.
1
^s 1
A\
, i n w
i i
i
! '
1 i
L
i i
t
',
!//
^
?r*"
's
S
f
'e
111
1
Vs i
COLWMIU 1
1
A
-------
cells that compose the grid. Defining the grid system before beginning the modeling inventory
development process will help minimize redundancy of effort.
In EPS 2.0, the modeling grid is defined in the /UAMREGN/ packet of the USERIN global user
input file. The user must supply the following information in this packet:
o Reference origin (UTM Easting and Northing) and UTM zone
o Grid origin in X and Y directions (in meters, with respect to reference origin)
o Number and size of cells in X and Y directions
o Number of vertical cells in lower and upper layers
o Minimum vertical cell height for lower and upper layers
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 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. Thus, before a final grid system is chosen, a
photochemical modeling specialist should be consulted regarding the effect of emission resolution on
modeling predictions.
4.1.1 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 hi 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.
-------
GRID SYSTEM \
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 areas may preclude the possibility of a clean
background along any boundaries that may be 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 spatially 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,
but may not exist for the outlying area. Technical problems may also be encountered if various
jurisdictions within the modeling region maintain information in different formats. For example, one
area may maintain records for townships and use EPA's Aerometric Information Retrieval System
(AIRS), whereas another area may maintain records for census tracts and use a locally developed data
handling system that is incompatible with AIRS.
If the exact area for which the photochemical model will be applied is not initially known, perhaps
because of uncertainties about future land use or the effect 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.
4-5
-------
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 Ceil 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.
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 hi 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 cells''(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 hi grid spacing between 2 to 5 km. Since
ozone formation occurs over an appreciable amount of tune 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.
4-6
-------
Glendora
San Bernardino
West Los
Anoetes
Palm Springs
Thousand
Q*s
Downtown
Los Angeles
Ventura
Anaheim
FIGURE 4-3. UAM modeling region for the California South Coast Air Basin.
1
0-J
1
-------
100km
(a) 2 km grid spacing (2.500 grid cells)
100km
.100 kin.
tbl 5 km grid spacing (4OO grid caHs)
FIGURE 4-4. Comparison of number of grid cells required for a 100 km by 100 km modeling
region for 2 km and 5 km grid sparings.
4-8
-------
GRID SYSTEM
Resource considerations, however, 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 hi 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.
4.2 MAP GRIDDING PROCEDURES
4.2.1 UTM 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 hi 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 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.
4-9
-------
550 600 650 700 750 BOO 850 900 950
4300
4250
4200 ~J
4150
4100 -^m
4050 -
4000
3950
3900
3850 -
3800 -
3850
-^ seco
550 600 650 700 750 BOO 850 900 950
UTM E«sting (Zone 10)
FIGURE 4-5. Rotated modeling region encompassing the southern San Joaquin Valley and Sierra
Nevada.
4-10
-------
GRID SYSTEM
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 cell 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 modeler 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 data base of computerized locations of
streets and other features may also prove useful for apportioning emissions from some area sources).
4-11
-------
References for Chapter 4:
1. Guideline for Regulatory Application of the Urban Airshed Model, EPA-450/4-91-013, U.S.
Environmental Protection Agency (OAQPS), Research Triangle Park, NC, June 1991.
4-12
-------
POINT SOURCES
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 or AIRS, and generally
contain the following types of information:
Source identification: county, facility, and source codes; Standard Industrial Classification (SIC)
of the facility; and location (latitude and longitude or JJTM coordinates) of each source.
Process information: Source Classification Code (SCC) or basic equipment codes for each
process; stack parameters (height, diameter, gas temperature, and gas exit velocity or flowrate);
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 3-2 in Chapter 3 lists the
data items included in the AIRS AFS work file format, which is the format required for input of point
source emissions data into EPS 2.0.
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 NC^, respectively. Techniques for speciation of
VOC and NOX emissions are discussed separately in Chapter 9.
In general, the point source data collection methodologies described in Procedures for the Preparation
of Emission Inventories for Precursors of Ozone, Volume f can also 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 detailed discussion of these techniques. The remaining
sections of this chapter focus on the additional data requirements of the modeling inventory and
specific data handling techniques.
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 data
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 (SCC)
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 1
5-2
-------
POINT SOURCES
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 compliance.
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 option of deriving local
category-specific rule effectiveness factors (within.tightly prescribed guidelines) as EPA deems
appropriate.
Rule effectiveness should be incorporated into all baseline and projected inventories with the
following exceptions: (1) sources not subject to the 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 the estimated control efficiency as shown in the following example.
> If uncontrolled emissions from a given source are 50 Ibs/day, and the estimated control
efficiency of a proposed measure is 90%, the actual controlled emissions accounting for a rule
effectiveness factor of 30% are calculated to be [50 Ibs/day] x [ 1 - $.90) x (0.30)], or 14
Ibs/day. The total emissions reduction is thus 72 percent.
5-3
-------
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 model-
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.
In the AIRS AFS work file-formatted point source emissions data file input to the EPS 2.0 module
PREPNT^ source location data may be reported in either UTM coordinates or as latitude and
longitude in decimal degrees. A user input flag specifies which coordinate system has been used
in the input emissions data file. If location 'data have been given as latitude and longitude, the
PREPNT module will convert the locations to UTM coordinates, since these are the units used by
the other EPS 2.0 modules. The GRDEM module assigns emissions from each point source to the
appropriate grid cell, based on the UTM location for each source.
5.4 TEMPORAL RESOLUTION
The modeling inventory should represent as accurately as possible day-specific emission estimates for
each hour of the modeling episode. By contrast, the existing point source inventory will more likely
contain annual or typical ozone season day estimates of emissions and a general description of the
operating schedule (seasonal fractions of annual throughput, and operating schedule in terms of
weeks/year, days/week, and hours/day in operation). This information may need to be augmented for
the modeling inventory. Several approaches for this augmentation are available, including contacting
the plant or local agencies, extrapolating from the information contained in the existing inventory, and
using engineering judgement to develop typical temporal profiles for the source types in question.
5-4
-------
POINT SOURCES
Ideally, each facility would be contacted to obtain hourly operating records for the modeling episode,
or, if this information is unavailable, representative operating schedules for a typical ozone season
day. Certain local agencies may also have this type of temporal information. Resource limitations,
however, generally make determination of source- or episode-specific operating schedules impractical
except for the largest emitters in the area. Some sources for which this type of data may be available
include the following: power plants (which generally keep detailed, hourly records of fuel firing rates
and power output for each day of operation), major industrial facilities such as automotive assembly
plants and refineries, and tank farms.
For many smaller point sources, reasonable temporal resolution can be obtained from the operating
data that are typically coded on each basic point source record.
> Consider an operation with annual emissions of 20 tons of VOC, with 40 percent of annual
throughput occurring in the summer. This source normally operates 12 hours per day and
seven days each week. Assuming uniform hourly emissions over a 13-week summer, the
% emissions rate is estimated to 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.
Above-ground fixed-roof petroleum product storage tanks present a unique situation in that breathing
loss emissions appear to be a function of time of day rather than operation.4 These tanks begin
expelling vapors when heated by 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. Daily emissions
from storage tanks can be estimated using procedures given in Volume I of EPA's AP-42 document.5
The EPS 2.0 input file preparation utility, TMPFAC, will create and/or update'the temporal
profiles file and the source/temporal profiles cross reference file. These files are required inputs
for the TMPRL module, which performs the actual temporal allocation. The TMPFAC utility also
allows the user to create source-specific temporal allocation profiles based on the throughput and
operating schedule information contained m the AIRS AFS (and AMS) workfiles. The monthly,
weekly, and diurnal profiles defined in the default temporal profiles file provided with EPS 2.0 are
shown in Appendix C.
TMPFAC examines the AIRS AFS workfile-formatted input emissions data files to retrieve the
operating schedule data for each record. TMPFAC then compares this information with the
profiles in the existing temporal profiles file. If none of the existing profiles match the operating
schedule information contained in the input emissions file, TMPFAC will create new profiles and
add a source-specific record to the source/temporal profile cross reference file to reflect the new
assignments. TMPFAC assigns monthly profiles based on the data hi the seasonal throughput
fields. The seasonal throughput for a particular season is divided evenly among the months for
that season, where the seasons are defined as followsf
5-5
-------
Winter: January, February, and March
Spring: April, May, and June
Summer: July, August, and September
Autumn: October, November, and December
TMPFAC assigns weekly profiles for the AFS work file based on the data in the days/week in
operation field; diurnal profiles are assigned based on the hours/day is operation. If an AFS
record contains valid data in the (optional) "start hour" field, TMPFAC will use the start hour and
the number of hours/day in operation to calculate the diurnal profile. Note that the diurnal profile
calculated by TMPFAC will result in equal distribution of emissions over each hour the source is
in operation. If actual hourly emissions data (perhaps reflecting hourly variations in activity
levels) are available for some sources in the modeling region, the user may define new source-
specific temporal profiles and incorporate the new profiles into the EPS 2.0 system input files
using an ASCII text editior. The new profiles must be added to the appropriate packets
(/MONTHLY/, /WEEKLY/, /DIURNAL WEEKDAY/, or /DIURNAL WEEKEND/) of the
temporal profiles file, and the source/temporal profiles cross-reference file must be modified to
reflect the new profile assignments.
TMPFAC assumes the following assignments regarding the operation schedule information
retrieved from the AIRS work files:
Days per week Hours per day
2 = Saturday and Sunday 8 = hours 0900 through 1600 (9 a.m. to 5 p.m.)
5 = Monday through Friday 12 = hours 0700 through 1800 (7 a.m. to 7 p.m.)
6 = Monday through Saturday 16 = hours 0800 through 2300 (8 a.m. to 12 a.m.)
7 = every day 24 = every hour
If a record specifies a number of days per week or hours per day other than those listed above
(e.g., 1, 3, or 4 days per week or 6 hours per day), TMPFAC will assign temporal profiles by
SCC code based on the default assignments listed in the source/temporal profiles cross reference
file. If any of the operating schedule information is missing, TMPFAC will assume a flat
operating profile for that source (i.e., 52 weeks/year, 7 days/week, and 24 hours/day, with no
seasonal variation).
Although the formats of the default profiles and cross-reference files provided with EPS 2.0 allow
these files to be input directly into die TMPRL core module, the user should use the TMPFAC
utility to generate source-specific temporal distributions based on the information in the input
emissions data file.
5-6
-------
POINT SOURCES
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 summarizes various methods for projecting point
source emissions; consult the EPA document Procedures for Preparing Emissions Projections6 for
definitive guidance on the 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 for 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. -
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
5-7
-------
additional controls would be expected, the current emission level could be multiplied by the
ratio 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
w *
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 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.
In 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 have been made with only eight of rnese. If production was expected to expand by 6
percent per year, on average, for the eight plants, then this rate could 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 ra"te 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.
5-8
-------
POINT SOURCES
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).7 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 5-2 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 MS As).
Note that MSA-level projections are not available for most two-digit SIC designations. Table 5-3
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.
In EPS 2.0, projection factors are applied by two-digit SIC designation. These projection factors
are expressed as ratios of future year to base year activity levels indicators (e.g., number of
employees); accordingly, a separate set of growth factors will be required for each projection year.
The projection factors are included in the control factors input file used by the CNTLEM module.
The BEAFAC utility provided with EPS 2.0 will produce these factors for the user based on data
contained in the Bureau of Economic Analysis' Regional Projections to 2040 data base file, which
is EPA's preferred data source for projecting future year activity for most stationary source
categories. This file contains the following data for each state: population for three age groups,
personal income (classified by major income component), and employment and earnings for 57
industrial groupings. For each of these categories, the BEA database contains historical data for
1973, 1979, 1983, and 1988, and projected data for 1995, 2000, 2005, 2010, 2020, and 2040.
For each state specified, BEAFAC calculates state-level projection factors for population .and
earnings by industry by dividing the projection year data by the base year data. If the specified
year is one of the years provided in the BEA data files, these data will be used directly. If the
year is not in the BEA data base, the data values required for calculating factors are determined by
interpolating between the two closest years in the file. The base and projection years must be
within the range of the BEA data base (i.e., between 1973 and 2040 inclusive). If the BEA data
base does not contain enough data to calculate a projection factor, BEAFAC assigns a factor of
1.0.
The user can also specify a list of projection factors for individual SIC codes (or ASC code for
area sources) that will override the data hi the BEA data base. The SIC codes used for the user-
supplied projection factors may be specified as either a 2-digit or 4-digit code, allowing projections
for individual as well as grouped SIC classifications.
5-9
-------
TABLE 5-2. Industrial groupings for BEA economk projections.
Industries projected for MSAs Industries projected for States and the Nation
1972 SIC code1
Farm
Agricultural services, forestry,
fisheries, and other
Mining
Construction
Manufacturing
Nondurable goods
Durable goods
Transportation and public
utilities
Farm 01,02
Agricultural servkes, forestry,
fisheries, and other 07, 08, 09
Agricultural services, forestry, and fisheries
Other2
Mining 11, 12
Coal mining 13
Oil and gas extraction 10
Metal mining 14
NonmetaUic minerals, except fuels 15, 16, 17
Construction
Manufacturing
Nondurable goods 20
Food and kindred products 21
Tobacco manufacturers 22
Textile mill products 23
Apparel and other finished textile products 26
Paper and allied products 27
Printing and publishing 28
Chemicals and allied products 29
Petroleum and coal products 30
Rubber and miscellaneous plastic products 31
Leather and leather products
Durable goods
Lumber and wood products, except
furniture and fixtures 24
Furniture and fixtures 25
Stone, clay, and glass products 32
Primary metal industries 33
Fabricated metal products 34
Machinery, except electrical 35
Electric and electronic equipment 36
Transportation equipment, except motor
vehicles 37 except 371
Motor vehicles and equipment 371
Ordnance3
Instruments and related products 38
Miscellaneous manufacturing 39
Transportation and public
utilities
Railroad transportation 40
Trucking and warehousing 42
Local, suburban, and tughway passenger
transportation 41
Air transportation 45
Pipeline transportation 46
continued
5-10
-------
POINT SOURCES
TABLE 5-2. Concluded.
Industries projected for MSAs Industries projected for States and the Nation
1972 SIC code
Transportation and public
utilities
Wholesale trade
Retail trade
Finance, insurance, and real
estate
Services
Government and government
enterprises
Federal, civilian
Federal, military
State and local
Transportation and public
utilities
Transportation services
Water transportation
Communication
Electric, gas and sanitary services
Wholesale trade
Retail trade
Finance, insurance, and real
estate
Banking
Other credit and securities agencies
Insurance
Real estate and combination offices
Services
Hotels and other lodging places
Personal, business, and miscellaneous repair
services
Automotive repair, services, and garages
Amusement and recreation services
Motion pictures
Private households
Health services
Private educational services
Nonprofit organizations
Miscellaneous professional services
Government and government
enterprises
Federal, civilian
Federal, military
State and local
47
44
48
49
50,51
52-59
60
61,62,67
64,64
65,66
70
72,73,76
75
79
78
88
80
82
83, 84, 86
81, 89
1 Historical data through 1974 are classified according to the 1967 SIC definitions; subsequent histoncal ^ma and
projections are classified according to the 1972 SIC definitions.
Refers to United States residents employed by international organizations.
3 The ordnance classification was discontinued in the 1972 SIC definitions. Earnings and employment
previously included in ordnance are now included in one or more of the following classes: fabricated metal
products (SIC 34); electric and electronic equipment (SIC 36); transportation equipment, except motor vehicles
(SIC 37, except 371); and instruments and related products (SIC 38).
source: Reference 7
concluded
5-11
-------
Ui
I
f-«
K)
TABLE 5-3. Employment by place of work (thousands of jobs), historical years 1973-1988 and projected years 1995-2040, for California (excerpt).
Manufacturing
Nondurable goods
Food and kindred products
Chemicals and allied products
Petroleum and coal products
Rubber and miscellaneous plastic
Durable goods
Primary metal industries
Electric and electronic equipment ....
Motor vehicles and equipment
Stone, clay, and glass products
Transportation and public utilities
Railroad transportation
Local and interurban passenger transit ....
Electric, gas, and sanitary services
Wholesale trade
Retail trade
1973
1686.5
557.7
170.8
55.6
24.5-
52.5
1128.8
59.7
261.3
42.2
57.1
499.6
38.1
24.7
67.7
473.4
1504.8
1979
2067.9
667.7
190.7
65.6
26.7
69.5
1400.2
59.8
321.8
29.9
62.8
579.5
32.8
29.2
70.5
606.4
1957.8
1983
2012.8
651.6
179.9
64.3
30.9
62.4
1361.1
43.4
368.0
30.4
54.1
589.5
24.8
28.5
76.3
636.0
2064.7
1988
2237.4
741.6
179.4
76.3
28.1
73.9
1495.8
44.0
396.8
35.6
63.8
662.2
18.8
36.8
89.3
777.1
2486.7
1995
2332.8
793.5
182.3
79.0
27.9
81.7
1539.3
43.1
397.1
33.8
66.9
736.0
15.3
41.4
98.4
868.0
2833.0
2000
2394.6
826.7
183.6
80.8
28.0
86.6
1568.0
43.0
397.5
32.6
68.5
779.8
13.9
44.3
105.3
920.0
3071.6
2005
2424.9
842.3
182.1
81.5
27.6
89.8
1582.5
42.5
398.0
31.5
69.7
807.5
12.8
45.9
109.3
950.9
3234.4
2010
2435.0
848.0
179.3
81.3
27.1
91.6
1587.1
41.7
" 398.3
30.6
70.7
825.2
12.1
46.8
111.7
992.3
3331.3
2020
2352.7
821.7
168.6
77.9
25.4
90.5
1531.1
39.1
383.0
28.2
69.1
816.6
10.7
46.1
110.5
995.4
3330.0
2040
2222.6
778.7
153.2
72.6
22.9
88.0
1443.9
35.2
359.6
25.0
66.4
797.2
8.8
44.8
107.7
989.0
3293.1
source: Reference 7
-------
POINT SOURCES
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-
3, 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.
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.
5-13
-------
5.5 J Accounting for Regulatory Controls in Baseline Projections
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. 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, SO2, 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 stringent local standard). Similarly, in control strategy projections, effects of any
alternative standards would have to be evaluated.
EPA has published a series of Control Technique Guideline (CTG) documents that specify control
requirements for a variety of sources and industries; these documents are summarized in Appendix C
of Volume I.3 The 1990 CAAA require that these controls be applied to all sources that meet the%
criteria specified in each document and which are located in moderate, serious, severe, and extreme
ozone nonattainment areas. For sources affected by more than one title of the 1990 CAAA, the
amendments mandate application of the more stringent control requirements. In particular, certain
processes that emit toxic compounds are also subject to Maximum Available Control Technology
(MACT) requirements. Under Section 112(d)(2) of Tide ffl, the 1990 CAAA require facilities that
emit 10 tons per year of any single air toxic or 25 tons per year of any combination of air toxics to
meet standards based on MACT.
Non-CTG Reasonably Available Control Technology (RACT) controls apply only to individual
processes not subject to CTG requirements. (Note that some CTGs specify the application of RACT
in the control guidelines; for these sources, an appropriate cutoff should be determined based on the
CTG requirements). Non-CTG RACT controls are applied to individual sources within a facility
when the total of emissions from all non-CTG regulated processes for that facility exceeds the cutoff
level for RACT applicability; the cutoff level is determined based on the ozone design value for the
area.
The /COUNTY/ packet of the EPS 2.0 global USERIN file includes the ozone design value for
each county in the modeling domain. CNTLEM uses this information to determine which of the
regulatory controls specified in the control factors input file are applicable for a given source. For
each CTG control to be applied, the user must specify the SCC codes for the processes subject to
the regulation, a control factor for NOX and VOC, and a cutoff limit for each pollutant which is
used to determine CTG applicability. Although the individual CTG documents determine
applicability based on a variety of parameters (throughput, capacity, etc,), CNTLEM requires the
cutoff limit to be specified in terms of emissions rates. For MACT and RACT controls, the user
need only specify control factors, since the cutoff limits for control applicability are determined
based on the ozone design value for the county.
5-14
-------
POINT SOURCES
5.5.4 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 a 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.
j In addition to the CTG, MACT, and RACT regulatory controls described above, the user may
! include additional controls to be applied for the sources in the modeling region in the CNTLEM
I control factors input file. These controls are referred to as "discretionary" controls to differentiate
I them from regulatory controls. The types of discretionary controls that may be specified in the
j control factors file are summarized briefly below.
\ Source- and Source Category-Specific Controls. The user may specify control parameters to be
| applied for individual sources or source categories using the /CONTROL EFFICIENCY/, /RULE
| EFFECTIVENESS/, and /RULE PENETRATION/ packets in the control factors files (rule
| penetration is a measure of the relative fraction of total sources for that source category which are
I subject to the regulation). Note that EPS 2.0 assumes that all new controls represent replacement-
I technologies. Accordingly, if the input EMBR emissions data record contains data in the control
j efficiency (CE), rule effectiveness (RE), and rule penetration (RP) data fields, CNTLEM will use
|that data to calculate an uncontrolled emissions rate before applying the new controls, as shown in
i Equation 5-1:
I In this equation, E is the emission rate, the subscript 0 indicates the values from the input EMBR
i record, and the subscript S indicates the control scenario-specific values. If any of the factors
i (control efficiency, rule effectiveness, or rule penetration) are not present, CNTLEM assumes a
I default value for the missing factor of zero for control efficiency (representing no control) and one
I for rule effectiveness and rule penetration (representing 100 percent effectiveness and penetration).
Allowable Emissions Controls. CNTLEM will also adjust emissions based on user-specified
5-15
-------
need to convert allowable emission limits expressed in terms of activity or emission factors.
Allowable limits may be specified at the state, county, source category, facility, and individual
source levels. A separate limit must be provided for each pollutant. Each control specified in the
/ALLOWABLE/ packet of the control factors file is identified as either a replacement limit or an
emissions cap. For replacement limits, the emissions in the input EMBR file are replaced with the
allowable emission rates specified in the control factors file. For limits identified as caps,
CNTLEM compares the emissions in the input EMBR file to the specified limit and substitutes the
limit if the input emissions exceed the specified value.
Control Strategy Code Controls. Hie CNTLEM module allows me user to specify additional
controls by control strategy code classification. CNTLEM supports five types of control strategy
code classifications: activity, control, pod, process, and speciation profile. Appendix B lists the
code definitions for each of die five types of control strategy codes. These code definitions
correspond to the reporting category codes employed by the RPRTEM module.
Other Discretionary Controls. CNTLEM also supports the application of controls by state,
county, and subregion code. In addition, the user may apply controls for a specified rectangular
subgrid of the modeling region.
5.5.5 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 presented to as many other
groups as possible for comment before being finalized. All assumptions, procedures, and data
sources must be carefully documented. Thorough review and documentation helps ensure that the
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 all assumptions, etc., are clearly stated for public review, and (4) defensible, because of all
the above characteristics.
Three ^key aspects of point source projections will invite criticism:
o the choice of indicators for projecting activity level growth;
o when and where this growth will occur, and whether it will be accommodated by expansion of
existing facilities or new construction; and
o what emissions will be associated with this growth, either in the baseline 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.
5-16
-------
POINT SOURCES
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.
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 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.
Some photochemical models do not require that elevated point sources be assigned to grid cells (in the
UAM, elevated point source locations are identified'by UTM coordinates). In these models,
preprocessor programs are available that make this assignment based on the point source coordinates
available from the annual, county-level inventory. If this is the case, point-tQrgrid-cell
correspondences need not be determined for these particular sources. Generally, however, since the
modelers may not know in advance which sources will be considered as elevated, and since
computerized assignments will be practiced in most instances, little extra effort will be expended in
simply making this assignment for all point sources. Thus, this information will always be available
in case it is needed at a later 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 space is available) 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 space is required. One 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 level. Note, however, 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 hi order to maintain the pollutant split
5-17
-------
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 5-4. The
entries for such a file, which are used to estimate hourly emission rates from annual emissions, are
determined using the procedures outlined hi 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.
Data handling requirements for the speciation data used to allocate VOC and NOX emissions into
chemical classes are discussed in Chapter 9.
5-18
-------
POINT SOURCES
TABLE 5-4. Example temporal factor file for individual point sources (excerpt).
Source
SCC" Ptb
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
5.23
5.23
3.76
2.99
5.00
5.00
2.99
Coded
S
D
1H
2H
3H '
4H
S
D
1H
2H
3H
4H
a 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.
5-19
-------
References for Chapter 5:
1. Love, R. A., and Mann, C. O., 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 1990.
2. User's Guide for the Urban Airshed Model, Volume IV: User's Manual for the Emissions
Preprocessor System 2.0, Pan A: Core FORTRAN System, EPA-450/4-90-007D (R), U.S.
Environmental Protection Agency (OAQPS), Research Triangle Park, NC, May 1992.
3. Procedures for the Preparation of Emission Inventories for Precursors of Ozone, Volume 1:
General Guidance for Stationary Sources, EPA-450/4-91-016, U.S. Environmental Protection
Agency (OAQPS), May 1991.
4. Breathing Loss Emissions from Fixed-Roof Petrochemical Storage Tanks (Draft), EPA Contract
» No. 68-02-2815, Work Assignment No. 6, Engineering-Science, Inc., July 1978.
5. Compilation of Air Pollutant Emission Factors, Volume I: Stationary Point and Area Sources,
Fourth Edition and Supplements, AP-42, U.S. Environmental Protection Agency, September
1985.
6. Procedures for Preparing Emissions Projections, EPA-450/4-91-019, U.S. Environmental
Protection Agency (OAQPS), July 1991.
7. BEA Regional Projections to 2040, Volume 1: States, U.S. Department of Commerce, Bureau of
Economic Analysis, June 1990.
5-20
-------
AREA SOURCES
6 AREA SOURCES
6.1 INTRODUCTION
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 the county-level inventory was prepared in
accordance with the guidelines given in Procedures 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 be handled individually in the point source inventory. In addition to small stationary
sources, the county-level area source inventory often includes emissions from offhighway mobile
sources, such as construction and agricultural equipment, aircraft, locomotives, and marine vessels.
As an example of a source classification scheme used to identify types of sources in the inventory,
Table 6-1 lists the area source categories that must be addressed in emission inventories developed for
State Implementation Plans.
> As a general rule, the maximum degree of source resolution should be maintained in the
modeling inventory. For examplejf sejterate emissions estimates have been prepared for dry
cleaners using perchloroethylene and cleaners using petroleum-based solvents, this distinction
should be maintained in the modeling inventory since it will permit more accurate speciation
of the VOC emissions associated with 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 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.
6-1
-------
TABLE 6-1. Area source categories required for consideration in State Implementation Plan
emission inventories.
Stationary Source Fuel Combustion:
Electric utility
Industrial
Commercial /Institutional
Residential
Industrial Processes:
Chemical manufacturing: SIC 28
Food & kindred products: SIC 20
Meat products
Grain mill products
Bakery products
Fermentation/beverages
Misc. food/kindred products
Primary metal: SIC 33
Secondary metal: SIC 33
Mineral processes: SIC 32
Concrete, gypsum, plaster products
Cut stone & stone products
Petroleum refining: SIC 29 « v
Asphalt paving/roofing materials
Wood products: SIC 24
Logging operations
Sawmills/planing mills
Millwork, plywood, & structural
rnembers
Miscellaneous wood products
Rubber/plastics: SIC 30
Fabricated metals: SIC 34
Coating, engraving, & allied services
Oil & gas production: SIC 13
Crude petroleum
Natural gas
Natural gas liquids
Construction: SIC 15-17
General building construction
Heavy construction
Road construction
Special trade construction
Industrial Processes (continued):
Machinery: SIC 35
Metalworking machinery: tool & die
makers
Mining & quarrying: SIC 14
Dimension stone
Crushed & broken stone
Sand & gravel
Gay, ceramic, & refractory
Chemical & fertilizer materials
In-process fuel use
All industrial processes
Solvent Utilization:
Surface coating (all solvent types)
Architectural coatings
Auto refinishing
Textile products
Flatwood products
Wood furniture
Metal furniture
Paper
Plastic products
Cans
Metal coils
Misc. finished metals
Electrical
Large appliances
Magnet wire
Motor vehicles
Aircraft
Marine
Railroad
Miscellaneous manufacturing
Degreasing (all solvent types)
Open top degreasing
Conveyorized degreasing
Cold cleaning degreasing
continued
6-2
-------
TABLE 6-1. Continued.
Solvent Utilization (continued):
Dry cleaning
Perchloroethylene
Petroleum
Graphic arts (all solvent types)
Lithography
Letterpress
Rotogravure
Flexography
Rubber/plastics
Miscellaneous industrial
Miscellaneous non-industrial
Film roofing: all solvent types
Adhesive application
Commercial/consumer
Pesticide application
Asphalt application (all solvent types)
Cutback asphalt
Emulsified asphalt
Asphalt roofing
Asphalt pipe coating
Solvent reclamation
' Tank/drum cleaning
All solvent use categories: all solvent types
Storage & Transport:
Gasoline marketing (service stations &
outlets)
Stage I (underground tank filling):
Splash fill
Submerged fill
Balanced submerged fill
Stage II (vehicle refueling):
Displacement losses
Spillage
Underground tank breathing &
emptying
Petroleum & petroleum product storage
Commercial/industrial (all products)
Bulk stations/terminals (all products)
Storage & Transport (continued):
Petroleum & petroleum product transport
Rail tank car (all products)
Marine vessel (all products)
Truck (all products)
Pipeline (all products)
Organic chemical storage (all products)
Commercial/industrial
Bulk stations/terminals
Organic chemical transport (all products)
Rail tank car
Marine vessel
Truck
Pipeline
Inorganic chemical storage (all products)
Commercial/industrial
Bulk stations/terminals
Inorganic chemical transport (all products)
Rail tank car
Marine vessel
Truck
Pipeline
Bulk materials storage (all products)
Commercial/industrial
Bulk stations/terminals
Bulk materials transport (all products)
Rail tank car
Marine vessel
Truck
Waste Disposal, Treatment, & Recovery:
On-site incineration
Industrial
Commercial/institutional
Residential
Open burning
Industrial
Commercial/institutional
Residential
continued
6-3
-------
TABLE 6-1. Concluded.
Waste Disposal, Treatment, & Recovery
(continued):
Landfills
Industrial
Commercial/institutional
Municipal
Wastewater treatment
Industrial treatment works (TWs)
Publicly owned TWs
Residential/subdivision owned TWs
Hazardous waste treatment, storage, &
disposal facilities (TSDFs)
Industrial TSDFs
Commercial/institutional TSDFs
Scrap & waste materials
Natural Sources:
Biogenic
Forests
Vegetarian
Soil
Geogenic
Volcanos
Geysers/geothermal
Wind erosion
Miscellaneous natural sources
lightning
Fresh water
Salt water
Miscellaneous Ana Sources:
Agricultural production - crops
Agricultural field burning
Orchard heaters
Country grain elevators
Agricultural production - livestock
Beefcattle feedlots
Poultry operations
Dairy operations
Hog operation
Other combustion
Forest wildfires
Managed (slash/prescribed) burning
Charcoal grilling
Structural fires
Firefighting training
Aircraft/rocket engine firing & testing
Cooling towers
Cooling towers
Process cooling towers
Comfort cooling towers
Catastrophic/accidental releases
Industrial accidents
Transportation accidents
Automotive repair shops
Auto top & body repair shops
Automotive exhuast repair shops
Tire retreading & repair shops
Miscellaneous repair shops
Welding repair shops
Health services
Hospitals
source: Reference 7
concluded
6-4
-------
AREA SOURCES
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 system input and glossary files provided with EPS 2.0 use the AIRS AMS source category
codes for area and mobile sources; this coding system classifies each source type using a ten-digit
numerical code. The AIRS AMS classification system is bierarchial, with four levels of detail
(specified by digits 1-2, digits 3-4, digits 5-7, and digits 8-10, respectively) as shown in the
example below:
AIRS AMS Code Description
24 xx xxx xxx Solvent utilization
24 01 xxx xxx Surface coating operations
24 01 001 xxx Architectural coatings
24 01 001 030 Architectural coatings: Acetone
To obtain a complete list of the AIRS AMS source category codes currently supported by the
AIRS system, consult the GEOCOMMON file which may be accessed using the online AIRS
system. ___ __ __
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
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.
6-5
-------
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 these 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 ceils.
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 additional 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
6-6
-------
AREA SOURCES
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 levels and emissions for
each grid cell is infeasible, the emissions modeler must implement some other apportioning scheme to
spatially allocate the emissions in the county-level area source inventory. The most straightforward
approach would be to distribute the total emissions for each county evenly over all of the grid cells in
the county; this approach, however, defeats the purpose of using a sophisticated grid model like the
UAM. Instead, the usual method employed to spatially distribute emissions to subcounty 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. Use engineering judgment to select appropriate
indicators for apportioning area source emission totals, and consult local authorities to verify the
applicability of the source category/spatial surrogate indicator pairings for a 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 6-2 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 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 6-3 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 by
traffic zone or census tract.
Often, the most representative way to spatially distribute emissions from some off-highway mobile
sources that are commonly included in the area source inventory, such as railroad locomotives,
6-7
-------
TABLE 6-2. Example spatial allocation surrogates for selected area source categories.
Emissions Category
Surrogate Indicator
Residential Fuel Combustion
Commercial/Institutional Fuel Combustion
Industrial Fuel Combustion
Onroad Vehicles - Limited Access Roadways
Onroad Vehicles - Rural Roadways
Onroad Vehicles - Urban Roadways
Off-Highway Vehicles
Railroad Locomotives
Aircraft - Commercial
Vessels
Gasoline Marketed
Unpaved Roads
Unpaved Airstrips
Forest Wild Fires
Managed Burning - Prescribed
Agricultural Operations
Structural Fires
Degreasing
Drycleaning
Graphic Arts/Printing
Rubber and Plastic Manufacturing
Architectural Coating
Auto 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
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
Housing
Urban Landuse
Urban Landuse
Link Location
Rural Landuse
Urban Landuse
County Area
Link Location
Airport Location
Water
Population
County Area
County Area
Composite Forest
Composite Forest
Agricultural Landuse
Housing
Population
Population
Population
Population
Population
Population
Population
Population
Population
Population
Population
Population
Population
Population
Water
Population
Population
Population
Population
County Area
Population
Population
Population
Population
Population
Population
Population
6-8
-------
AREA SOURCES
TABLE 6-3. 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
Administtation, Washington, D.C.
Aircraft, general Census of U.S. Civil Aircraft (Annual), U.S. Department of
Transportation, Federal Aviation AdmJnistratipn, Washington, D.C.
Aircraft, military Military Air Traffic Report (Annual), U.S. Department of
Jr^.P.o.rtatipn, Federal Aviation AdministTation, Washington, D.C.
Agricultural equipment Census of Agriculture, 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, Waterborne Commerce of the United States, (Annual), U.S. Army
river cargo, and small Corps of Engineers, Washington, D.C.
pleasure craft)
Gasoline handling Census of Business Selected Services Area Statistics, U.S.
^^f^^^o^Coimm&ccQ, ?yrea.H .9£!£e ^S?1:!?'. Washington, D.C.
Fuel combustion, Sales of Fuel Oil and Kerosene, Mineral Industry Surveys.
commercial/institutional
6.9
-------
aircraft, and vessels, is to treat these sources as "line" sources. Emissions from these sources can
be assumed to occur only in those grid cells that contain railroad track mileage, airports, or
waterways. The EPS 2.0 module GRDEMS will distribute emissions to grid cells using user-
specified link data. Specifically, GRDEMS allocates the emissions associated with each type of
link (e.g., railroad track mileage) to cacft grid cell based on die fraction of the total county link
distance for the link type occurring in that grid cell. Refer to Section 7.6, regarding spatial
distribution of onroad mobile source emissions, for additional information on the specification of
link data for use with EPS 2.0.
Developing Apportioning Factors from Land Use Patterns. For most urban areas, land use data will
be available for the present and several projection years; the emissions modeler can use this data to
develop apportioning factors for those area sources whose emissions will be distributed based on
various land use classifications. Although spatial apportioning factors can be developed manually
from maps, computerizing as many steps of this process as possible generally minimizes the required
effort. Unfortunately, computerized land use data may be unavailable for projection years, an
obvious drawback. In this case, the computerized land use data base can be used to develop
apportioning factors for the base year emissions inventory, and projected changes hi 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 6-4.
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 6-4), political unit code, USGS
hydrologic 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 boundaries of the modeling region represent grid ceil
(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
. 6-10
-------
AREA SOURCES
TABLE 6-4. 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
16 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
5. WATER
51 Streams and canals
52 Lakes evergreen
53 Reservoirs
54 Bays and estuaries
6. WETLAND
61 Forested wetland
62 Nonforested wetland
7. 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
8. TUNDRA
81 Shrub and brush tundra
82 Herbaceous tundra
83 Bare ground
84 Wet tundra
85 Mixed tundra
9. PERENNIAL SNOW OR ICE
91 Perennial snow fields
92 Glaciers
source: Reference 4
6-11
-------
FIGURE 6-1. Conceptual representation of the grid cell identiflcation process.
6-12
-------
\AREASOURCES
FIGURE 6-2. County grid ceil assignments for the Atlanta, Georgia modeling region.
6-13
-------
contribute to the total land use for any given cell; similarly, more than one county can contribute to
the total area within a grid cell, as shown hi 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,
computerized data bases and data base systems such as geographical information systems (GIS) can
be used to develop spatial apportioning factors, allowing complete automation of the spatial allocation
process.
* Assume that the existing inventory contains an estimate of total emissions from dry cleaning
for the entire study area and that no specific survey or other information is available for
individual dry cleaning establishments. The emissions modeler must select a spatial surrogate
indicator that will 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 6-5 shows the
coding system used in 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 hi this
computation.
The emissions for each grid cell are then estimated as a simple fraction of the total, as
follows:
E, = ET (Si / ST) (6-1)
where E denotes emissions, S indicates surrogate indicator, i indicates the value hi grid cell i,
and T indicates the total for the county or region.
6-14
-------
AREA SOURCES
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
FIGURE 6-3. Segment of land use map for Tampa Bay, Florida.
6-15
-------
TABLE 6-5. Land use categories for Tampa Bay area land use map (Figure 6-3).
1. URBAN OR BUILT-UP LAND
10 Multi-family residential
H Single family residential
12 Commercial and service
13 Industrial
14 Transportation, communication and
utilities
15 Industrial and commercial combined
16 Mixed urban, or built-up land
17 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 r
29 Citrus groves
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
5. WATER
51 Streams and canals
52 Lakes
53 Reservoirs
54 Bays and estuaries
6. WETLAND
63 Freshwater forested wetland
64 Freshwater marsh
65 Saltwater forested wetland
66 Saltwater marsh
7. BARREN LAND
72 Beaches
73 Sandy areas other than beaches
75 Extractive
76 Transitional areas
source: Reference 5
6-16
-------
AREA SOURCES
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 (Si /
ST) for grid cell (15,15) will be 0.2/26.3, 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 total dry cleaning emissions from the entire region. Mathematically,
this can also be expressed by Equation 6-2,
fi.k = Si,k/S?,1(Si)k) (6-2)
where f; k 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.
One disadvantage of developing apportioning factors from maps other than land use 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.
The foregoing discussion dealt only with the allocation of area source emissions based on a single
surrogate indicator. In some cases, no one parameter may accurately describe the subcounty
distribution of emissions from a particular area source category. In this situation, apportioning
factors can be based on two or more surrogate 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 6-5).
The emissions inodeier 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).
6-17
-------
> One third of miscellaneous solvent emissions may be assigned to multifamily residences (land
use 10), one third to single family residences Qand use 11), and one third to commercial and
service use (land use \2). Hence, if county-level emissions from miscellaneous solvent use
are 12 tons per day, 4 tons 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 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
r / 3 \i
(6-3)
where Wjk is the weighting factor selected for land use type j in relation to source category k,
and Sj: is the value of the surrogate indicator (i.e., the area) of land use type j hi cell i.
The summation term appearing in the numerator above is essentially a composite surrogate
indicator for the entire category. Thus, if solvent emissions are 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 cell is (0.6
x 1) + (0.2 x 3) + (0.2 x 5), or 2.2. The entire category is then apportioned as usual, based
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.
6-18
-------
AREA SOURCES
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 hi 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.
* In the San Francisco Bay Area of California, emissions from 58 area sources are apportioned
using combinations of the 19 demographic parameters shown hi Table 6-6, all of which are
compiled at the subcounty level by the local MPO as part of transportation planning studies.
For some area source categories, a single parameter from Table 6-6 is used 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 6-6, which
includes agriculture production and services. Similarly, the source category "printing" is
distributed with the variable "MFGI," 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 6-7 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 6-7.
Assume that area source degreasing emissions in a given county are 42 tons/day of VOC.
According to Table 6-7, 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**1 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)+4.2(0.01)+4.2(0)
= 0*0714 ton/day
- 6-19
-------
TABLE 6-6. Demographic parameters used in San Francisco Bay Area for making
zonal allocations of area sources.
Variable2
Name
DWELL
AGRI
MIN
MFG1
MFG2
MFG3
MFG4
MFG5
MFG6
TRAN
WHOL
FIN
SERV 1
SERV2
GOV
RET
BUS. SERV.
RET. SERV
OTHER
SERV.
SICb Classification
(not applicable)
1,7-9
10, 13, 14
27
26, 28, 29, 32, 33
20
19, 36, 38
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
a 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
Column 3).
b Standard Industrial Classification Code
source: Reference 6
6-20
-------
AREA SOURCES
TABLE 6-7. Excerpt from ABAC cross classification table used in San Francisco Bay Area for
subcounty allocation of area source activities.
Source Classification
CHEMICAL
Misc. Chem. Proc
OTHER IND./COM.
Metallurgical
Mineral - Concrete
Mineral - Stone/Sand/Gravel
Mineral - Sand Blast
Misc. Mineral Proc.
Farming Operations
Food/ Agricultural Proc.
Paint Spray Mist
Wood Products Mfg.
Misc. Inc. /Com. Proc.
GASOLINE DISTRIBUTION
Vehicle Fill Station -
- Spillage
- Storage Tanks
- Vehicle Tanks
OTHER ORG. COMP. EVAP.
Storage Tanks -
- Solvent
- Misc. Org. Comp.
Ind. Coat. - Solv. Base
Ind. Coat. - Water Base
Com. & Dom. Coat.
- Solv. Base
- Water Base
De greasers
Drycleaners - PERC
Drycleaners - Misc. solvents
Rubber Fabrication
Plastic Fabrication
Printing
Misc. Org. Evap.
1 7 8 9 10 11 12 13 14
Dwell Agri Min Mfgl MTg2 Mfg3 Mfg4 MfgS Mfg6
90 10
100
10 40 50
10 5
10 90
100
100
5 10 70
10 80
30 30 30
i'»
9
20 20 20 20
5 10 20 15 10 10 10
10 10 10 50 10
90 5
64 3 3
64 3 3
10 60
100
90
100
10 5 5 50 555
15
Tran
-10
5
5.
5
10
10
3
3
continued
6-21
-------
TABLE 6-7. Concluded.
Source Classification
16 17
18
19
Whol Fin Servl Serv2
20 21 22 23 24
Bus Ret Serv
GOT Ret Serv Serv Other
CHEMICAL
Misc. Chem. Proc
OTHER IND./COM.
Metallurgical
Mineral - Concrete
Mineral - Stone/Sand/Gravel
Mineral - Sand Blast
M.isc. Mineral Proc.
Fanning Operations
Food/Agricultural Proc.
Paint Spray Mist
Wood Products Mfg.
Misc. Inc./Com. Proc.
GASOLINE DISTRIBUTION
Vehicle Fill Station -
Spillage
- Storage Tanks
- Vehicle Tanks
OTHER ORG. COMP. EVAP.
Storage Tanks -
- Solvent
- Misc. Org. Comp.
Ind. Coat. - Solv. Base
Ind. Coat. - Water Base
Com. & Dom. Coat.
- Solv. Base
- Water Base
De greasers
Drycleaners - PERC
Drycleaners - Misc. solvents
Rubber Fabrication
Plastic Fabrication
Printing
Misc. Org. Evap.
10
10
10
10
90
90
90
3
3
10
15
10
10
10
5
3
3
20
100
100
10
10
10
5
5
5
3
3
10
source: Reference 6
concluded
6-22
-------
AREA SOURCES
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 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, these amounts will be allocated in the appropriate subcategories, using the
corresponding demographic parameter as the surrogate indicator hi each case.
The preceding apportioning calculation assumes that apportioning factors are compiled at the grid cell
level. In actuality, as mentioned at the outset of this section, the spatial surrogate indicators (such as
the demographic parameters shown in Table 6-6) used for apportioning are initially compiled at the
zonal level for transportation and other planning purposes rather than at the grid cell level. For
example, the San Francisco Bay area local MPO develops its population, land use, and employment
data for 440 zones, each of which comprises one to seven census tracts. By contrast, there are some
5,000 grid cells to which area source emissions are 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-cell 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 6-8 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.
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:
*ik = sj Sijbjk (6-4)
6-23
-------
N)
TABLE 6-8. Illustrative excerpts from zone-to-grid-cell correspondence table for determining apportioning factors.
Zone
1
2
17
18
23
GC
A
g
GC
A
g
GC
'A
g
GC
A
g
GC
A
g
(01,01)
1.0
.19
(01,03)
0.3
.08
(05,14)
0.6
1.0
(05,14)
0.1
0.33
(06,24)
0.3
0.33
(01,02)
1.0
.19
(01,04)
1.0
.26
(06,14)
0.2
0.67
(06,25)
0.6
0.67
(01,03)
0.7
.13
(02,02)
0.5
.13
t
(02,01)
1.0
.19
(02,03)
1.0
.26
(02,02)
0.5
.10
(02,04)
0.6
.15
(03,01)
0.7
.13
(03,01)
0.2
.05
(04,01)
0.3
.06
(03,03)
0.2
.05
(03,04)
0.1
.03
Total, £A
5.2
3.9
0.6
0.3
0.9
Legend:
GC - Grid Cell
A - Area of intersection of zone with grid cell
g - Apportioning factor from zonal level to grid cell (all activities assumed uniform throughout zone)
-------
AREA SOURCES
where a^ is the value of the kth demographic parameter, aggregated to grid cell i; bjk is the
value of the kth demographic parameter, as compiled for zone j; and g^ is the fractional area
of zone j in cell i. Note that the value of gj is given by
gij - AS / Ej (Aij) (6-5)
where A^ 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 f- k) 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.,
fi,k = *ik / ^ a* (6-6)
The same normalizing factor can, of course, be obtained by totaling the zonal values; that is,
Ei a* = Ej bjk (6-7)
except for necessary corrections for any zone which falls partly outside the county. The
apportioning factors (f; k) are applied in the same way as the (Sj / Sp) factors (determined
from land use or other maps) in Equation 6-1 that were determined from land use or other
maps.
r *
The most difficult part of the zone-to-grid-cell conversion process as described above is determining
the g^: fractions. 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 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 UAM application for Atlanta, Georgia; note that
some grid cells, particularly in the urban area of Atlanta (located at the center of the modeling
region) 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 cells based on the BGED centroid locations. Although the population data available from
the Census Bureau will usually be somewhat outdated because of the infrequency of data
compilation, the emissions modeler can still use this data to develop apportioning factors for
6-25
-------
10
20
30
FIGURE 6-4. Location of block group enumeration centroids for the Atlanta, Georgia modeling
region.
6-26
-------
AREA SOURCES
//
r?§825
- S765
- 3745
3705
665
FIGURE 6-5. Sample gridded population data for the Atlanta, Georgia modeling region.
6-27
-------
the base year modeling inventory, provided that no significant changes in population density
distributions have occurred.
When estimating the gld fractions, which represent the area! fractions by grid cell for each
demographic 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 sometimes 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 6-9
lists recommended seasonal adjustment factors for selected 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 EPS 2.0 core module TMPRL temporally allocates area (and mobile) source emissions based
on the temporal profiles assigned for each area source category (ASQ in the source/temporal
profiles cross reference file. The temporal profiles are defined in a separate file, which may
contain up to four packets of profile definitions: /MONTHLY/, /WEEKLY/, /DIURNAL
WEEKDAY/, and /DIURNAL WEEKEND/. Appendix C shows the profiles which are currently
included in the default temporal profiles definition file provided with EPS 2.0.
Temporal profiles in the /MONTHLY/ packet are defined by assigning a relative weight factor for
each month which indicates the monthly contribution to total annual activity. The TMPRL module
divides the weight factor for each month by the sum of all 12 weight factors to determine the
fraction of annual activity occurring hi each month. The seasonal adjustment factors hi Table 6-9
can be converted to monthly weight factors for use in this file as described below.
6-28
-------
AREA SOURCES
TABLE 6-9. Ozone season adjustment factors for selected area source categories.
Category
Seasonal Adjustment Factors
Gasoline Service Stations
Tank trucks in transit
Tank truck unloading (Stage I)
Vehicle refueling (Stage U)
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
Commerc i al /consumer
Uniform
Uniform
1.3
Uniform
Uniform
Uniform
0
1.3
Uniform
Waste Management Practices
POTWs
Hazardous waste TSDFs
Municipal landfills
1.4
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
6-29
-------
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 6-9 for architectural surface coating is 1.3;
accordingly, the monthly fraction will be (1.3) / (12), or 0.108. For the other nine 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
[l-(3x0.108)]/9, or 0.075,
In the temporal profiles definition file, the weight factors for each month must be expressed as
integer values. Accordingly, the values calculated above may be multiplied by 1000 to construct
the weight factors to be entered into this file, as shown below:
Season Weight factor for each month in season Total weight for season
Winter 75 75 75 225
Spring 75 75 75 225
Summer 108 108 108 324
Fall 75 75 75 225
999
If the inventory already contains peak season or episodic emission estimates (i.e., if the inventory
type designation in the input emissions data file is "PO", "PC", or "S", indicating peak ozone
typical day, peak CO typical day, or specified interval, respectively), TMPRL will not apply a
seasonal adjustment.
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.
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.
6-30
-------
AREA SOURCES
Table 6-10 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.
In EPS 2.0, this type of hourly operating information can be incorporated into the modeling
inventory by defining new temporal profiles for the appropriate area source categories. An ASCI!
text editor can be used to add the new profiles to die temporal profiles definition file and modify
the corresponding assignments in the source/temporal profiles cross reference file accordingly.
The hourly operating information in Table 6-10 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 6-11 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 6-11. 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 to project future year levels 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. The emissions modeler should consult current EPA guidance on projection of future
year emission inventories when identifying appropriate growth indicators for the various source
categories.
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.
6-31
-------
TABLE 6-10. Diurnal patterns for gasoline stations in Tampa Bay, in percent of daily
operation.
Hour
6 - 7 am
7-8
8-9.
9- 10
10-11
11-12 noon
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 Station3
Small
5
6
6
5
6
6
5
5
7
7.
9
9
6
6
5
5
1
1
Medium
4
4
6
5
7
7
7
7
6
7
8
8
8
7
3
3
2
1
Large
8
8
8
7
2
2
8
9
5
6
13
13"
4
4
2
1
a 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
6-32
-------
I
CO
UJ
TABLE 6-11. Example temporal resolution methodologies for selected area source categories.
Source Category
Gasoline handling
Dry clean ing
Degreasing
Nonindustrial surface
coating
Cutback asphalt
Pesticide application
Miscellaneous solvent use
Aircraft, general
Aircraft, commercial and
military
Agricultural equipment
Construction equipment
Industrial equipment
i
Lawn and garden
equipment
Off-highway motoi cycles
Snowmobiles
Seasonal
varies from area to area. Use average
temperatures for a summer day where
appropriate.
uniform
uniform
uniform
uniform spring through fall
coincides with growing season
uniform
uniform
estimate on an individual basis; contact
local airport authorities, Federal Aviation
Administration, and appropriate military
agencies
uniform throughout the agricultural season
20% December-February
25% March-May
30% June-August
25% September-November
20% December-February
25% March-May
30% June-August
25% September-November
uniform through months which have an
average temperature of 38°F or higher
base on monthly off-highway fuel use
base on monthly off-highway fuel use
Daily
Monday-Saturday
uniform Monday-Saturday
uniform Monday-Saturday
uniform Monday-Saturday
Monday-Friday
uniform
uniform
60% Mon-Fri; 40% Sat-Sun
uniform
Monday-Saturday
uniform Monday-Saturday
50% Mon-Fri
50% Sat-Sun
30% Mon-Fri; 70% Sat-Sun
30% Mon-Fri; 70% Sat-Sun
Hourly
uniform 0600 to 2000, otherwise zero
uniform 0700 to 1900, otherwise zero
80% 0700 to 1900; 20% 1900 to 2400
uniform 0700 to 1900, otherwise zero
uniform 0700 to 1900, otherwise zero
daylight hours (0700 to 1900)
80% from 0700 to 1900
20% from 2000 to 2400
uniform 0700 to 2100, otherwise zero
uniform 0700 to 2100, otherwise zero
uniform 0700 to 1900 Monday-Friday
uniform 0700 to 1200 Saturday
80% from 0700 to 1900
20% from 1900 to 2400
uniform 0900 to 1900, otherwise zero .
continued
-------
I
U>
TABLE 6-11. Concluded.
Source Category
Small pleasure ciaft
Railroad locomotives
Ocean-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
Waste management
practices (POTWs,
TSDFs)
Seasonal
uniform through months which have an
average temperature of 45 °F or greater
uniform
uniform
10% of emissions uniform throughout the
year
90% of emissions uniform during months
having an average temperature less than
68°F
25% of emissions unifonn throughout the
year
75 % of emissions uniform during months
having an average temperature less than
68"F
uniform
uniform
10% winter
70% spring
0% summer
20% fall
uniform
uniform
Daily
30% Mon-Fri
70% Sat-Sun
uniform
uniform
uniform
95% Mon-Sat
5% Sunday
unifonn Mon-Sat
91% Mon-Fri
9% Saturday
uniform
unifonn
unifonn
Hourly
unifonn from 0700 to 1800, otherwise
zero
70% from 0700 to 1800
30% from 1800 to 0700
75% from 0700 to 1900
25% from 1900 to 0700
uniform
90% from 0600 to 2400
10% from 2400 to 0600
80% from 0700 to 1800
20% from 1800 to 2400
otherwise zero
unifonn from 0600 to 1700
uniform from 0500 to 2100
unifonn
unifonn
source: Reference ± I, 2
concluded
-------
AREA SOURCES
* If a survey of dry cleaners has been performed and the average estimated growth in the
modeling area is 5 percent per year, then in 5 years, dry cleaning activity would be projected
to increase by a factor of [1.05]5, or 1.28 (a 28% increase). As another example, a local
asphalt trade association may be able to project cutback asphalt usage.
When considering such estimates, the inventorying agency must recognize the possibility of deliberate
or inadvertent biases due to wishful thinking or self-serving motives, and should strive to obtain
opinions which are as objective as possible. The agency should also be careful to determine whether
or not such estimates of future activity levels reflect the effects of anticipated control measures, an
important consideration since some such 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 the absence of local projections, the BEA
economic indicators described in Section 5.5 can be used to develop growth indicators for area
sources.
In EPS 2.0, area source emissions are projected based on either the first four digits of the area
source category (ASC) code or the entire ten-digit code. The projection factors for each ASC,
which are applied using the CNTLEM module, are included in the control factors input file. The
BEAFAC utility, described in Section 5.5.2, will calculate projection factors for four-digit ASCs
using the BEA projections of population and industrial earnings.
Table 6-12 lists growth indicators and potential information sources for selected area source
categories, as suggested in the EPA guidance document Procedures for Preparing Emissions
Projections* The following example illustrates use of 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 hi 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 weil.
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 hi the base year is
multiplied by the area source activity level in the base year to yield the projection year activity level.
6-35
-------
TABLE 6-12. Preferred growth indicators for projecting emissions for area source categories.
Source Category
Gasoline marketing
Dry cleaning
Degreasing (cold cleaning)
Architectural surface coating
Automobile refinishing
Small industrial surface coating
Graphic arts
Asphalt use - paving
Asphalt use - roofing
Pesticide application
Commercial/consumer solvent use
Publicly Owned Treatment Works
(POTWs)
Hazardous Waste Treatment,
Storage, and Disposal Facilities
(TSDFs)
Municipal solid waste landfills
Commercial/institutional fuel
combustion
Industrial fuel combustion
Growth Indicators j Information Sources
projected gasoline consumption { MOBILE fuel consumption model
population; retail service I solvent suppliers; trade
employment ( associations
<
industrial employment ! trade associations
population or residential dwelling I local MPO
units j
industrial employment
industrial employment
population
consult industry
industrial employment;
construction employment
historical trends in agricultural
operations
population
site-specific information
state planning forecasts
state waste disposal plan
residential housing units or
population
commercial/institutional
impioyment; population
BEA
BEA
state planning agencies; local
MPO
consult industry
local industry representatives
state department of agriculture;
local MPO
local MPO; state "planning
agencies
state planning agencies
state planning agencies; local
MPO
local MPO; state planning
agencies
local MPO
local MPO; land use projections:
state planning agencies
continued
6-36
-------
AREA SOURCES
TABLE 6-12. Concluded.
Source Category
Aircraft (commercial and general)
Aircraft (military)
Railroads
Ocean-going and river cargo
vessels
Vessels, small pleasure craft
Off-higfiway motorcycles
Agricultural equipment
Construction equipment
Industrial equipment
Lawn and garden equipment
On-site incineration
Open burning
Fires: managed burning,
agricultural field burning, frost
control (orchard heaters)
Forest wildfires
Structural fires
Growth Indicators
site-specific forecasts
site-specific forecasts
revenue ton-miles
cargo tonnage
population
population
agricultural land use; agricultural
employment
industry growth (SIC code 16)
Industrial employment (SIC codes
10-14, 20-39, 50-51) or industrial
land use area ^
single-unit housing
based on information gathered
from local regulatory agencies
based on information gathered
from local regulatory agencies
areas where these activities occur
historical average
population
i Information Sources
j local airport authorities and
i commercial carriers
1 local airport authorities;
j appropriate military agencies
! American Association of
; Railroads and local carriers
i local port authorities; US
i Maritime Administration; US
i Army Corps of Engineers
| local MPO
1 local MPO
1 local MPO; Census of
i Agriculture
! local MPO
! local MPO
local MPO
local regulating agencies and
MPO; state planning agencies
local agencies; state planning
agencies; local MPO
US Forest Service; state
agriculture extension office
local, state, and federal forest
management officials
local MPO; state planning
agencies '
source: Reference 8
concluded
6-37
-------
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 cell
level. 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 used directly, as described above, to determine changes in spatial
emission patterns.
If the surrogate indicators used for apportioning certain area source emissions are not projected at a
subcounty level, engineering judgment must be used to decide whether spatial distributions of various
activities will change enough to warrant the effort of identifying new patterns. Changes may be
warranted in rapidly 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 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 wirnin each grid.
6-38
-------
AREA SOURCES
This type of methodological inconsistency will likely lead to changes in the emissions inventory that
are not due to growth or control measures but, rather, to changes in the inventory procedures
themselves.
As a test to determine whether different base and projection year methodologies are mutually
consistent, apply the projection year methodology to the base year case and see if the results are
identical. If important discrepancies exist, then one methodology should be chosen for use for both
years. Generally, any methodology which applies growth factors to base year estimates 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 6-13 shows a sample excerpt
from such a file. This file basically consists of a matrix of apportioning factor values by grid cell. In
Table 6-13, 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.
The EPS 2.0 module GRDEMS performs the calculations described above. This module uses a
cross reference input file of area source categories and spatial surrogate indicators, which specifies
which indicators will be used to allocate emissions from which source categories. The default
cross-reference file provided with EPS 2.0 may be modified to include county- or source category-
specific spatial surrogate assignments, or to redefine the spatial surrogate indicator codes. As an
example, the user might redefine the^surrogate code designations for a rarely used land use
surrogate (such as barren, or rocky with lichens) to incorporate special spatial apportioning
information for one or more counties, such as detailed location data for gasoline service stations.
Consult the User's Guide for the Urban Airshed Model, Volume IV: User's Guide for the
Emissions Preprocessor System3 for additional guidance and specific format requirements for the
spatial apportioning file.
6-39
-------
TABLE 6-13. Example Hie of grid cell apportioning factors for area sources (excerpt).
Grid Cell
Coordinates8
272,784
274,784
274,784
274,784
280,784
252,786
254,786
256,786
258,786
260,786
Apportioning Factors for
SIlb
.001
.001
.001
.001
.001
.001
.011
.013
.001
.001
SI2b
.001
.001
.001
.001
.001
.002
.011
.014
.001
.001
SI3b
.001
.001
.001
.001
.001
.002
.012
.015
.001
.001
SI4b
.0
.0
.0
.0
.0
.0
.0
.0
.0
.100
SI5b
.004
.004
.004
.004
.003
.004
.0
.0
.004
.004
SI6b
.0
.0
.0
.0
.0
.0
.045
.270
.009
.0
a 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.
6-40
-------
AREA SOURCES
The sequence of steps described above applies in cases where each area source category is
apportioned using only one surrogate indicator. If more than one surrogate indicator must be used 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 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 ar.eal 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 fractions for each zone, (2) summing over
all 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 involved.
Estimating hourly area source emissions requires essentially the same data handling procedures as are
described 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. Typically, one set of temporal operating
factors will be assigned for each area source category, which are applicable for the entire modeling
area. Determining appropriate temporal factors for the area source temporal factor file is a manual
procedure, as described in Section 6.3.
6-41
-------
References for Chapter 6:
1. Procedures for the Preparation of Emission Inventories for Precursors of Ozone, Volume I:
General Guidance for Stationary Sources, EPA-450/4-91-016, U.S. Environmental Protection
Agency (OAQPS), May 1991.
2. NAPAP (National Add Precipitation Assessment Program) Emissions Inventory: Overview of
Allocation Factors, 1985, EPA-600/7-89-010a, Alliance Technologies Corporation, October 1989.
3. User's Guide for the Urban Airshed Model, Volume IV: User's Manual for the Emissions
Preprocessor System 2.0, Part A: Core FORTRAN System, EPA-450/4-90-007D (R), U.S.
Environmental Protection Agency (OAQPS), Research Triangle Park, NC, May 1992.
4. Procedures for Estimating and Allocating Area Source Emissions of Air Toxics, Working Draft:
Appendix A, EPA Contract No. 68-02-4254, Work Assignment No. 105, Versar, Inc.,
Springfield, Virginia, March 1989.
*
5. L.G. Wayne and P.C. Kochis, Tampa Bay Area Photochemical Oxidant Study: Assessment of the
Anthropogenic Hydrocarbon and Nitrogen Dioxide Emissions in the Tampa Bay Area, EPA-904/9-
77-016, U.S. Environmental Protection Agency, Region IV, Air and Hazardous Materials
Division, Atlanta, GA, September 1978.
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.
7. Emission Inventory Requirements for Ozone State Implementation Plans, EPA-450/4-91-010, U.S.
Environmental Protection Agency (OAQPS), March 1991.
8. Procedures for Preparing Emissions Projections, EPA-450/4-91-019, U.S. Environmental
Protection Agency (OAQPS), July 1991.
6-42
-------
MOBILE SOURCES
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. For inventory purposes, mobile sources are typically
categorized as
o Onroad vehicles;
»
o Offroad vehicles;
o Aircraft;
o Railroad locomotives; and
jfc
o Vessels. *
Onroad 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. Offroad
vehicles include all recreational vehicles and machinery used in off-road situations, such as farm
equipment, construction equipment, snow mobiles, off-road motorcycles, etc. Together, offroad
sources and aircraft, railroad locomotives, and vessels (which represent all vehicles used in air, rail,
and water transportation, respectively) are sometimes referred to as "other mobile" sources; these
sources, however, should be maintained as separate source categories in the modeling inventory to
facilitate spatial, temporal, and chemical resolution of emissions. To provide an example of a source
classification scheme for mobile sources, Tables 7-1, 7-2, and 7-3 list the source category codes
employed by the AIRS Area and Mobile Subsystem for onroad, offroad, and other mobile sources.
As mentioned in Chapter 6, an existing annual or seasonal area source emission inventory generally
contains adequate estimates of emissions for all sources except onroad motor vehicles. Since the
Urban Airshed Model will be applied for a particular episode, the mobile source emissions must be
either calculated 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 onroad vehicles (for offroad mobile sources, the methods described in Chapter
6 may be employed).
7-1
-------
TABLE 7-1. AIRS AMS codes for onroad mobile sources (each code is ten characters long; characters 1
and 2 are always "22", indicating Mobile Sources).
Characters 3 and 4: Fuel type
Characters 5, 6, and 7: Vehicle Type
01 Highway Vehicles - Gasoline
30 Highway Vehicles - Diesel
001 Light Duty Gas Vehicles (LDGV)
020 Light Duty Gas Trucks 1 (LDGT1)
040 Light Duty Gas Trucks 2 (LDGT2)
060 Light Duty Gas Trucks 1 & 2 (LDGT)
070 Heavy Duty Gas Vehicles (HDGV)
080 Motorcycles (MC)
001 Light Duty Diesel Vehicles (LDDV)
060 Light Duty Diesel Trucks (LDDT)
070 Heavy Duty Diesel Trucks (HDDT)
Characters 8, 9, and 10: Roadway Classification
000 Total: All Road Types
110 Interstate: Rural Total
111 Interstate: Rural Time 1
112 Interstate: Rural Time 2
113 Interstate: Rural Time 3%
114 Interstate: Rural Time 4
130 Other Principal Arterial: Rural Total
131 Other Principal Arterial: Rural Time 1
132 Other Principal Arterial: Rural Time 2
133 Other Principal Arterial: Rural Time 3
134 Other Principal Arterial: Rural Time 4
150 Minor Arterial: Rural Total
151 Minor Arterial: Rural Time 1
152 Minor Arterial: Rural Time 2
153 Minor Arterial: Rural Time 3
154 Minor Arterial: Rural Time 4
170 Major Collector: Rural Total
171 Major Collector: Rural Time 1
172 Major Collector: Rural Time 2
173 Major Collector: Rural Time 3
174 Major Collector: Rural Time 4
190 Minor Collector: Rural Total
191 Minor Collector: Rural Time 1
192 Minor Collector: Rural Time 2
193 Minor Collector: Rural Time 3
194 Minor Collector: Rural Time 4
210 Local: Rural Total
211 Local: Rural Time 1
212 Local: Rural Time 2
213 Local: Rural Time 3
214 Local: Rural Time 4
230 Interstate: Urban Total
231 Interstate: Urban Time 1
232 Interstate: Urban Time 2
233 Interstate: Urban Time 3
234 Interstate: Urban Time 4
250 Other Freeways <& Expressways: Urban Total
251 Other Freeways & Expressways: Urban Time 1
252 Other Freeways & Expressways: Urban Time 2
253 Other Freeways & Expressways: Urban Time 3
254 Other Freeways & Expressways: Urban Time 4
270 Other Principal Arterial: Urban Total
271 Other Principal Arterial: Urban Time 1
272 Other Principal Arterial: Urban Time 2
273 Other Principal Arterial: Urban Time 3
274 Other Principal Arterial: Urban Time 4
290 Minor Arterial: Urban Total
291 Minor Arterial: Urban Time 1
292 Minor Arterial: Urban Time 2
293 Minor Arterial: Urban Time 3
294 Minor Arterial: Urban Time 4
310 Collector: Urban Total
311 Collector: Urban Time 1
312 Collector: Urban Time 2
313 Collector: Urban Time 3
314 Collector: Urban Time 4
330 Local: Urban Total
331 Local: Urban Time 1
332 Local: Urban Time 2
333 Local: Urban Time 3
334 Local: Urban Time 4
Note:
In addition to classification by vefaicie and roadway type, the AIRS AMS structure allows the user to speedy
emissions for up to four time periods per day (e.g., morning peak, noon, afternoon peak, and rest of day) in
order to incorporate temporal variations due to the effects of parameters, such as average speed and traffic
volumes, which may vary widely depending on the time of day.
7-2
-------
TABLE 7-2. AIRS AMS codes for offroad mobile sources (each code is 10 characters long; characters
Characters 3 and 4
60 Off-Highway Vehicle
Gasbline, 2-Stroke
65 Off-Highway Vehicle
Gasoline, 4-Stroke
70 Off-Highway Vehicle
Diesel
60 Off-Highway Vehicle
Gasoline, 2 -Stroke
65 Off-Highway Vehicle
Gasoline, 4-Stroke
70 Off-Highway Vehicle
Diesel
Characters 5, 6 and 7
000 All Off-Highway Vehicle:
Gasoline, 2 -Stroke
000 All Off-Highway Vehicle:
Gasoline, 4-Stroke
000 All Off-Highway Vehicle:
Diesel
001 Recreational Vehicles
002 Construction Equipment
003 Industrial Equipment
1-2 are always "22", indicating Mobile Sources).
Characters 8, 9, and 10
000 Total
000 Total
000 Total
000 Total
010 Motorcycles: Off-Road
020 Snowmobiles
030 All Terrain
000 Total
003 Asphalt Pavers
006 Tampers/Rammers
009 Plate Compactors
012 Concrete Pavers
015 Rollers
018 Scrapers
021 Paving Equipment
024 Surfacing Equipment
027 Signal Boards .
030 Trenchers
033 Bore/Drill Rigs
$)36 Excavators
039 Concrete/Industrial Saws
>
000 Total
010 Aerial Lifts
020 ForkJifts
040 Minibikes
050 Golf Carts
060 Speciality Vehicle Carts
042 Cement A Mortar Mixers
045 Cranes
048 Graders
051 Off-highway Trucks
054 Crushing/Proc. Equipment
057 Rough Terrain Forklifts
060 Rubber Tire Loaders
063 Rubber Tire Dozers
066 Tractors/Loaders/Backhoes
069 Crawler Tractors
072 Skid Steer Loaders
075 Off-Highway Tractors
078 Dumpers/Tenders
081 Other Construction Equipment
030 Sweepers/Scrubbers
040 Other General Industrial Equipment
050 Other Material Handling Equipment
continued
-------
TABLE 7-2. Concluded.
Characters 3 and 4
i
60 Off-Highway Vehicle
Gasoline, 2-Stroke
65 Off-Highway Vehicle
Gasoliiie, 4-Stroke
70 Off-Highway Vehicle
Diesel
Characters 5, 6 and 7
004 Lawn & Garden
Equipment
005 Farm Equipment
006 Light Commercial
*
007 Logging Equipment ^
008 Airport Service Equipment
Characters 8, 9, and 10
000 Total
010 Lawn mowers
015 Rotary Tillers < 5 HP
020 Chain Saws < 4 HP
025 Triromers/Edgers/Brush Cutters
030 Leafblowers/Vacuums
035 Snowblowers
040 Rear Engine Riding Mowers
000 Total
010 2 -Wheel Tractors
0 1 5 Agricultural Tractors
020 Combines
025 Balers
030 Agricultural Mowers
000 Total
005 Generator Sets < 50 HP
010 Pumps < 50 HP
015 Air Compressors < 50 HP
000 Total
005 Chain Saws > 4 HP
010 Shredders > 5 HP
000 Total
005 Airport Support Equipment
045 Front Mowers
050 Shredders < 5 HP
055 Lawn & Garden Tractors
060 Wood Splitters
065 Chippers/Stump Grinders
070 Commercial Turf Equipment
075 Other Lawn & Garden Equipment
035 Sprayers
040 Tillers > 5 HP
045 Swathers
050 Hydro Power Units
055 Other Agricultural Equipment
020 Gas Compressors < 50 HP
025 Welders < 50 HP
030 Pressure Washers < 50 HP
015 Stridden
020 Fellers/Bouchers
010 Terminal Tractors
concluded
-------
MOBILE SOURCES
TABLE 7-3. AIRS AMS codes for aircraft, vessels, and locomotives (each code is 10 characters long;
characters 1-2 are always "22", indicating Mobile Sources).
Characters 3 and 4
Characters 5, 6 and 7
Characters 8, 9, and 10
75 Aircraft
000 All Aircraft Types &
Operations
001 Military Aircraft
020 Commercial Aircraft
050 General Aviation
060 Air Taxi
070 Aircraft Aux. Power Units
085 Unpaved Airstrips
000 Total
900 Refueling: All Fuels
000 All Processes
101 Displacement Loss/Uncontrolled
102 Displacement Loss/Controlled
103 Spillage
201 Underground Tank: Total
202 Underground Tank: Breathing &
Emptying
80 Marine Vessels,
Commercial
001 Coal
002 Diesel
003 Residual Oil
004 Gasoline
000 Total, All Vessel Types
010 Ocean-Going Vessels
020 Harbor Vessels
030 Fishing Vessels
040 Military Vessels
82 Marine Vessels,
Recreational
005 Pleasure Craft, Gasoline 2-
Stroke
010 Pleasure Craft, Gasoline 4-
Stroke
020 Pleasure Craft, Diesel
000 Total
005 Inboards
010 Outboards
015 Stemdrive
020 Sailboat Aux. Inboard
025 Sailboat Aux. Outboard
85 Railroads
002 Diesel
000 Total
005 Line Haul Locomotives
010 Yard Locomotives
7-5
-------
MOBILE SOURCES
7.2 CHARACTERIZATION OF ONROAD MOTOR VEHICLE EMISSIONS
The emission factors used to estimate emissions from onroad 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 onroad 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 onroad 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 and road type (see Table 7-1 for an example).
In annual or seasonal inventories, emissions for each category will typically be reported as county-
level totals. To facilitate accurate spatial allocation and chemical speciation of mobile source
emissions, and to simplify the implementation and analysis of detailed control strategies, emissions
from onroad mobile sources should be reported by both vehicle type (e.g., light-duty gasoline
automobiles) and roadway classification. In addition, onroad mobile source emissions should be
disaggregated into the different emissions components, such as exhaust, evaporative, running losses,
etc. These three types of categorization are discussed below.
7.2.1 Vehicle Classes
The registered vehicle fleet can be divided into subgroups, or classes, such as passenger automobiles,
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 7-4, 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 two light duty
gasoline truck categories (LDGT1 and LDGT2) are often combined into a single LDGT category.
7.2.2 Roadway Types
Onroad mobile source emissions should also be distinguished by road type in the inventory. Road
types for which the Federal-Highway Administration (FHWA) maintains statistics are listed in Table
7-5; these road types are commonly used in mobile emission inventories. Emission factors will vary
by road type because of the variance in parameters such as speed and fleet distributions associated
with each different road type.
7-6
-------
MOBILE SOURCES
TABLE 7-4. Vehicle class deflnitions
Vehicle Class (Abbreviation)
Light-duty Gasoline Vehicles (LDGV)
Light-duty Gasoline Trucksl (LDGT1)
Light-duty Gasoline Trucks2 (LDGT2)
Heavy-duty Gasoline Vehicles (HDGV)
Light-duty Diesel Vehicles (LDDV)
Light-duty Diesel Trucks (LDDT)
Heavy-duty Diesel Vehicles' (HDDV)
Motorcycles (MC)
used by the MOBILE models.
GVWf Specification
not applicable
less than 6500 Ibs.
6500 to 8500 Ibs.
more than 8500 Ibs.
not applicable
less than 8500 Ibs.
more than 8500 Ibs.
not applicable
t Gross Vehicle Weight
source: Reference 2
TABLE 7-5. Commonly used road type designations.
Rural and Urban Interstates
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
7-7
-------
7.2.3 Emission Components
In addition to the categorization of onroad 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 onroad vehicle emissions- are defined below:
o Exhaust emissions: emissions resulting from the combustion processes associated with the
operation of motor vehicles. In EPA's MOBILE models, exhaust emissions are composed of
three components, representing three different operating modes: cold start, hot start, and hot
stabilized.
o Evaporative emissions: emissions resulting from the volatilization of gasoline and solvents
due to rising ambient temperatures or engine heat after motor vehicle shutdown. EPA's
MOBILE models (versions 4.1 and higher) recognize five components of evaporative
emissions. Diurnal emissions (resulting from ambient temperature changes which occur when
the vehicle is not in use), hot soak emissions (emissions that occur when fuel in the engine is
vaporized by residual engine heat following vehicle shutdown), and crankcase emissions are
lumped together under "evaporative." Resting loss and running loss emissions are reported
separately.
o Running loss emissions: evaporative VOC emissions which occur during the operation of the
vehicle, typically at warm temperatures and low speeds.
o Resting loss emissions: evaporative VOC emissions from nonoperating motor vehicles that
result from vapors permeating parts of the evaporative emissions control system, migrating
out of the carbon canister, or evaporating liquid fuel leaks. Resting losses are distinct from
diurnal emissions in that they do not result from rising ambient temperatures.
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 five groups, necessitating separation of the four components.
The AIRS AMS source category classifications for onroad motor vehicles (Table 7-1) do not
support specification of emissions by component (e.g., exhaust, evaporative). Sines this distinction
must be maintained in the modeling inventory in order to adequately characterize the emissions,
EPS 2.0 employs an expanded source category code classification for onroad mobile sources. The
EPS 2.0 internal motor vehicle source category codes are shown in Table 7-6.
7-8
-------
MOBILE SOURCES
TABLE 7-6. EPS 2.0 internal source category
code is 10 characters long; characters 1 and 2
codes for onroad motor vehicles (each source category
are always equal to "MV").
Characters 3-4: Vehicle Class
01 Light Duty Gas Vehicles (LDGV)
02 Light Duty Gas Trucks 1 (LDGT1)
03 Light Duty Gas Trucks 2 (LDGT2)
04 Heavy Duty Gas Vehicles (HDGV)
Characters 5-7: Roadway Classification
000 Total: All Road Types
001 Total: Urban
Total: Suburban
Total: Rural
Total: Limited Access
Interstate: Rural Total
Interstate: Rural Time 1
Interstate: Rural Time 2
Interstate: Rural Time 3
Interstate: Rural Time 4
Other Principal Arterial: Rural Total
Other Principal Arterial: Rural Time 1
Other Principal Arterial: Rural Time 2
Other Principal Arterial: Rural Time 3
Other Principal Arterial: Rural Time 4
Minor Arterial:. Rural Total
Rural Time 1
Rural Time 2
Rural Time 3
Rural Time 4
002
003
004
110
111
112
113
114
130
131
132
133
134
150
151
152
153
154
170
171
172
173
174
190
191
192
193
194
210
211
212
Minor Arterial:
Minor Arterial:
Minor Arterial:
Minor Arterial:
Major Collector: Rural Total
Major Collector: Rural Time 1
Major Collector: Rural Time 2
Major Collector: Rural Time 3
Major Collector: Rural Time 4
Minor Collector: Rural Total
Minor Collector: Rural Time 1
Minor Collector: Rural Time 2
Minor Collector: Rural Time 3
Minor Collector: Rural Time 4
Local: Rural Total
Local: Rural Time 1
Local: Rural Time 2
Characters 7-10: Emissions Mode
EXH Exhaust emissions
EVP Evaporative emissions
RNL Running loss emissions
RST Resting loss emissions
HOT Hot soak emissions
05 Light Duty Diesel Vehicles (LDDV)
06 Light Duty Diesel Trucks (LDDT)
07 Heavy Duty Diesel Trucks (HDDT)
08 Motorcycles (MC)
213 Local: Rural Time 3
214 Local: Rural Time 4
230 Interstate: Urban Total
231 Interstate: Urban Time 1
232 Interstate: Urban Tune 2
233 Interstate: Urban Time 3
234 Interstate: Urban Time 4
250 Other Freeways & Expressways
251 Other Freeways & Expressways
252 Other Freeways & Expressways
253 Other Freeways & Expressways
254 Other Freeways & Expressways
270 Other Principal Arterial: Urban
271 Other Principal Arterial: Urban
272 Other Principal Arterial: Urban
273 Other Principal Arterial: Urban
274 Other Principal Arterial: Urban
290 Minor Arterial: Urban Total
291 Minor Arterial: Urban Time 1
292 Minor Arterial: Urban Time 2
293 Minor Arterial: Urban Time 3
294 Minor Arterial: Urban Time 4
310 Collector: Urban Total
311 Collector: Urban Time 1
312 Collector: Urban Time 2
313 Collector: Urban Time 3
314 Collector: Urban Time 4
330 Local: Urban Total
331 Local: Urban Time 1
332 Local: Urban Time 2
333 Local: Urban Time 3
334 Local: Urban Time 4
Urban Total
Urban Time 1
Urban Time 2
Urban Time 3
Urban Time 4
Total
Time 1
Time 2
Time 3
Time 4
DNL Diurnal emissions
BG1 Exhaust emissions: Bag 1 (cold start)
BG2 Exhaust emissions: Bag 2 (hot stabilized)
BG3 Exhaust emissions: Bag 3 (hot start)
7-9
-------
7.3 MOBILE SOURCE EMISSION FACTORS
The EPA's MOBILE series of models calculate VOC, NOX, and CO emission factors for the onroad
vehicle fleet. These models incorporate the data from the EPA Surveillance Program designed to
quantify and characterize the emission factors encountered in the onroad vehicle fleet. The MOBILE
models are run with a single input file containing both model parameters and user supplied
information. Table 7-7 contains the required input parameters, and Table 7-8 lists additional optional
parameters. The complete definitions of the input parameters and the required format of the input
file should be obtained from the user's guide for the model, available from EPA. Consult Volume IV
for guidance on determining appropriate values for these inputs.
In July 1991, EPA released MOBILE4.1, which updated the previous version of the model
(MOBILE4). This is the version required for use by states in determining motor vehicle emission
factors for all base year emission inventories, adjusted base year emission inventories, and CO
projection inventories required by the 1990 CAAA (MOBILE4.1 does not apply to California
vehicles). MOBILE4.1, which was intended as an interim version of the model which would allow
states to meet the regulatory deadlines for base year SIP inventory submittal prior to the official
release of MOBILES, does not incorporate VOC and NO, emission reductions from CAAA-mandated
motor vehicle control measures, and accordingly cannot be used to project future-year emission
factors without modification. MOBILES will include the effects of these reductions in addition to the
benefits of the current Federal Motor Vehicle Control Program. A draft version of MOBILES has
already been released, and a final version should soon be available.
7.4 MOBILE EMISSION INVENTORY PROCEDURES
In general, the emissions modeler may employ either one of two methods to develop the on-road
vehicle portion of the modeling inventory:
o compiling an episode-specific onroad vehicle emission inventory using the methods
described in Procedures for Emission Inventory Preparation, Volume IV: Mobile Sources3
(hereafter referred to as Volume TV); or
o adjusting an existing annual or seasonal inventory (e.g., AIRS or SAMS) to reflect
episodic conditions.
The remainder of this section addresses how to determine which method is appropriate for a particular
modeling application, and describes the general methodology for adjusting an existing inventory for
episodic conditions.
The time requirements for developing an original inventory are not considerably greater than for
developing the modeling inventory from an existing inventory, so this option is less infeasible than it
7-10
-------
MOBILE SOURCES
TABLE 7-7. Required input parameters for EPA's MOBILE models.
Calendar year In-use RVP*
ASTM volatility class In-use RVP start year*
Ambient daily temperature* . Region altitude
Minimum and maximum daily temperature Speeds
Base RVP Operating modes*
* MOBILE default values are recommended.
* Not always used by MOBILE, but input record is required.
TABLE 7-8. Optional input parameters for EPA's MOBILE models.
Alternate basic emission rates'
Alternate vehicle tampering rates*
Fleet Characterization Data:
Fleet registration distribution* Diesel penetration rate*
Fleet mileage accumulation* Vehicle class distribution*
Inspection & Maintenance Programs:
start year stringency %
model years inspected waiver rates
compliance rate program type
frequency of inspection vehicle classes inspected
test type alternate credits
special mechanic training
Anti-Tampering Programs:
start year model years inspected
vehicle classes inspected program type
frequency of inspection compliance rate
list of equipment inspected alternate credits
Refueling Programs (Stage II):
start year phase-in period
LDGV % system efficiency HDGV % system efficiency
* MOBILE default values are highly recommended.
* MOBILE default values available (national averages) but local or regional data recommended.
7-11
-------
might first appear. Additionally, the emissions modeler may only need to generate original emissions
estimates for some subcategory of the mobile fleet. Some of the primary reasons for -developing an
original inventory instead of adjusting an existing one include
o The motor vehicle source category classifications employed in the existing inventory are
inadequate for accurate spatial, temporal, or chemical resolution of the modeling inventory, or
for evaluating particular control strategies under consideration;
o The procedures used to develop the existing emission inventory contain uncertainties which
would produce questionable model results.
Detailed procedures for development of mobile source emission estimates are presented in Volume TV;
this document should be referred to as the definitive guidance for constructing a mobile source
inventory.
For some areas, detailed transportation modeling results may be available for some portion of the
modeling domain which allow onroad mobile source emissions to be estimated individually for each
link of the transportation modeling network. Note the distinction between link-based emission
estimation and the spatial allocation of county-level emissions using link surrogates (addressed in
Section 7.3). Link-based emissions estimates incorporate actual activity data and other parameters
(e.g., average vehicle speed) for each link in the transportation network; accordingly, the emissions
associated with each link will vary depending on the activity for that link. When county-level
emissions are spatially allocated using link surrogates, the emissions are distributed evenly over all the
links of that type located in the county, based solely on the length of each link.
In EPS 2.0, the PREAM and LBASE modules serve as the entry points for county-level and link-
based mobile source emissions data, respectively. The input emissions data requirements for
county-level mobile source emissions data are identical to those for area source data, shown in
Table 3-3. For link-based emissions data, the user must provide UTM coordinates for the
beginning and end nodes of each link, the FTPS code for the county in which the link is located,
the AIRS AMS source category code, and the emissions associated with each link.
Alternatively, the mobile source modeling emission inventory may be developed by adjusting an
existing annual or seasonal inventory for episodic conditions. To adjust an existing inventory, the
emissions modeler must duplicate the emission factors used to generate the existing mobile source
inventory. Accordingly, accurate and complete information regarding the data used to generate the
existing inventory is required; if necessary, contact the developer of the existing inventory to obtain
this information. Some instances where the emissions modeler might choose this approach include
o Emissions from mobile sources do not contribute significantly enough to the total inventory to
warrant development of an inventory from original data (note that this is usually not the case);
7-12
-------
MOBILE SOURCES
o Time constraints prevent development of an original inventory; or
o Required data such as locale-specific VMT data is not available.
The change in emissions associated with the adjustment of annual or seasonal estimates to reflect
episodic conditions can be summarized by the following equation:
MEE = MEB (EFE/EFe) (VMTE/VMTB) (7-1)
In Equation 7-1, ME refers to mobile emissions, EF represents the emission factor, the subscript "B"
signifies the variables associated with the existing annual or seasonal inventory, and the subscript "E"
indicates the episode-specific variables. In this equation, 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 (EPE/EFg) (VMT Factor) (7-2)
Equation 7-2 can then be incorporated into the following steps which are required to adjust the onroad
mobile inventory for episodic conditions. 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 onroad 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.
o Apply equation 7-2 to generate annual average onroad mobile emissions for each county
within the modeling region.
o Spatially allocate emissions (Section 7.5) to produce a gridded inventory.
o Temporally adjust the mobile emissions (Section 7.6) to reflect seasonal, daily, and hourly
diurnal variations.
If the existing inventory reports total VOC emissions instead of VOC emissions by component (e.g.,
running losses), the total VOC may be disaggregated into the emission components by applying the
following fraction to the total VOC mobile source emissions:.
7-13
-------
VOC, - VOCT (EF, / EFT) (7-4)
where the subscript T refers to each emission component, and the subscript "T" refers to the total
VOC emissions and emission factors. From the component totals in Equation 7-4, Equation 7-2 can
then be used to determine the scenario emission totals.
' EPS 2.0 includes a utility module, MVADJ, to prepare the motor vehicle adjustment factors input
file ("MADJIN") required by the PREAM, LBASE, CNTLEM, and TMPRL modules. The
adjustment factors contained in the MADJIN file are used to disaggregate total motor vehicle
emissions into the emission components, estimate die effects of proposed vehicle emission control
strategies, and perform hourly temperature corrections. The MVADJ utility reads a series of
output files from EPA's MOBILE model (versions 4.1 or 5.0) and calculates ratios of emission
factors which are used to simulate each of the effects luted above. Refer to the User's Guide for
the Urban Airshed Model, Volume IV, Part A for information regarding the MVADJ utility.
7.5 SPATIAL RESOLUTION OF MOBILE SOURCE EMISSIONS
Mobile sources differ from stationary source categories in that their-spatial variation is more
accurately described using a link-based rather than a surrogate-based gridding procedure. In a link-
based spatial allocation approach, emissions are distributed only to those grid cells that contain
transportation pathways (e.g., roadways, railways, airports, shipping channels, etc.). This approach
is usually used in conjunction with a surrogate-based procedure to complete the spatial resolution of
the mobile source inventory. The following section describes the methodology for spatial allocation
of emissions using link data; spatial allocation of mobile emissions using gridded surrogates is
discussed in Section 7.5.2.
7.5.1. Link Surrogates
Emissions from onroad vehicles on limited access roadways (interstates and expressways), railroad
locomotives, aircraft, and vessels are often spatially allocated with a link-based procedure, since the
transport routes used by these vehicles are both easily identifiable and readily modeled as a series of
lines or links. This results in more accurate allocation of emissions from these sources than could be
achieved using surrogates such as population or land use.
Figure 7-1 illustrates a typical link with respect to a portion of a modeling grid. In this 8xample, the
link, which has start and end point coordinates of (Xj, Yj) and (X2, ₯2), crosses three grid cells.
The fraction of the total link length in each of the three grid cells, designated as llt £2, and ^3 in the
figure, is easily calculated using basic trigonometric relationships. In order to spatially allocate
county-level emissions using link surrogates, all of the links of a particular type (e.g., interstates or
7-14
-------
MOBILE SOURCES
S, + 2As
Sy + As
S, S, + As S, + 2As A
Grid size - As
FIGURE 7-1. Depiction of typical link and grid cells.
7-15
-------
airport runways) located within the county must be identified. For counties only partially located
within the photochemical modeling domain, the links located outside of the domain must also be
identified, since this information will be required to determine what fraction of the county-level
emissions should be allocated to the portion of that county located within the modeling domain. After
all links for a particular county have been identified, the length of each link located within each grid
cell must be calculated, and the lengths summed for all links crossing that grid cell. The total length
of all links of that type for the county must also be calculated, so that emissions may be allocated to
grid cells as shown in Equation 7-5 below:
EOJ) - BT ' 0-tU) 7 ^ (7'5)
where E represents emissions, L is link length, the subscript (i,j) indicates each grid cell in the
county, and T refers to the total values for the county.
The EPS 2.0 core module GRDEM automatically performs the calculations described above, given
user-specified link definition data. GRDEM calculates the length of each link located within each
grid cell and the total link length distance by link type for each county. The county-level
emissions for the selected sources (e.g., motor vehicle travel on interstates) are then distributed to
individual grid cells based on the fraction of the total link length for each county located within
that grid cell.
For each link located within the counties covered by the modeling domain, the user must
specify the following parameters in the optional link definition input file for the GRDEM
module.
o Type of link: The link types currently supported by EPS 2.0 are defined as follows:
101 Limited access roadways (interstates, expressways)
102 Railroads
103 Airports (runways)
104 Shipping channels (rivers, sea lanes, ports)
105 Disposal sites
In order to redefine the existing types, or add new types, the source
category/spatial allocation surrogate cross-reference glossary file must be modified
to correspond to the new link type assignments.
o County designation: Links should be identified by county so that the total link length
associated with each county can be determined. Digitized link segments should end at
county borders (i.e,, each link should be located entirely within a single county).
o Beginning and end point coordinates: The UTM Easting and UTM Northing
coordinates of the endpoints of each link must be specified in meters.
7-16
-------
I MOBILE SOURCES
The first step in determining the parameters required for specifying link surrogates is to obtain a map
identifying the appropriate transportation pathways. The U.S. Geologic Survey (USGS) maps, which
are readily available in a variety of scales, clearly indicate railroads, airport runways, rivers, and
ports. In addition, the USGS maps reference locations using both UTM coordinates and latitude and
longitude in degrees, simplifying the conversion of link locations to the modeling coordinate system.
More detailed street maps, however, may be required to identify different motor vehicle roadway
types. Since no standard coordinate system is usually identified on street maps, several reference
points on the street map must be identified whose coordinates are known or can be determined from a
USGS map in order to convert the street map locations to the modeling coordinate system. County
line intersections make particularly good reference points, since they are clearly marked on most
street maps.
To facilitate the determination of the link coordinates, it is highly recommended that an electronic
digitizer be used to map out the start and end points of each link. A digitizer is an electronic sensor
that can translate any position on the 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 the digitizer board in a fixed position and moving the sensor to each link start
and end point location. In addition, two reference points must be digitized to provide the information
required to convert the digitizer coordinate system to the coordinate system of the modeling region.
7.5.2 Non-link Mobile Emission Spatial Surrogates
Non-link surrogates are commonly used to spatially allocate mqbile emissions in the following
situations:
o Links are too numerous to define and process, as is typically the case for onroad rural and
urban vehicles and for offroad vehicles.
o Emission totals are too insignificant when compared with emissions, from other sources in the
modeling domain to warrant the development of link data.
o Use of gridded spatial surrogates based on landuse or population data provides a more
accurate allocation of vehicle emissions. For example, recreational boating activities may be
distributed approximately equally over the surface of a large lake.
The procedure for the allocation of emissions using landuse data is identical to that outlined in Section
6.2.2.
In most modeling applications, a combination of link and landuse surrogates is used for the spatial
allocation of mobile source emissions. As an example. Figure 7-2 shows the links identified for a
7-17
-------
640
680
720
760
40
FIGURE 7-2. Onroad mobile source link surrogates developed for a UAM application of the
Dallas/Fort Worth region.
7-18
-------
MOBILE SOURCES
UAM application for the Dallas/Fort Worth area. In this figure, 131 links were defined for the
allocation of emissions from limited access roadways and commercial airports. Figure 7-3 shows the
gridded onroad mobile source inventory for this area, which was prepared using urban and rural
landuse surrogates for those source categories for which links were not developed.
The default source category/spatial allocation surrogate cross-reference glossary file provided with
the EPS 2.0 contains the following assignments for onroad mobile sources:
AIRS roadway classification Spatial allocation surrogate
Interstates (both urban and rural) Limited access roadways (link type 101)
All other urban road classifications Urban landuse
All other rural road classifications Rural landuse
Unless the user modifies this file, EPS 2.0 will use these surrogates to spatially allocate emissions
from onroad mobile sources.
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); see Figures B-x
and B-y in Appendix B, which show the default weekday and weekend diurnal variation profiles
assumed by EPS 2.0 for onroad mobile sources, for an example. Accordingly, the emissions modeler
should be careful to select a diurnal variation pattern for onroad motor vehicle emissions which is
appropriate for the modeling episode. If hourly vehicular speeds and VMT distributions are available
from the local Metropolitan Planning Organizations (MPOs), these should be utilized in estimating
hourly mobile source emissions.
As discussed in Section 6.3, the TMPFAC utility of EPS 2.0 will construct source category-
specific temporal variation profiles for each county based on the information contained in the AMS
workfile-formatted input emissions data. In the absence of such data, EPS 2.0 will use the
temporal profiles assigned to the onroad mobile source categories in the source category/temporal
profiles cross-reference system input file. The default profile assignments in this file (refer to
Appendix C for profile definitions) for all onroad mobile source categories are
Monthly variation: Profile 1 (no variation by month)
Weekly variation: Profile 13 (weekend activity equal to 75% of weekday levels)
Diurnal variation: Profile 50
7-19
-------
640
680
720
760
.!.'..'. I .«..«. I .'. Jl ( i. I «.«.».« I « «. ' ' .". f .«..'. I .1. .»::!.'..». I '. .'. I .'. X::I.:;.:T.::I..
3720
-3680
.= 3640
- 3600
30
40
iJsseo
DAL1AS UTEEKDAY MOTOR VEHJC1Z EMISSIONS
NOx (kg/day)
Total - 320702.60
HGURE 7-3. Gfidded annual average onroad mobile source emissions for a UAM application
of the Dallas/Fort Worth region.
7-20
-------
MOBILE SOURCES
For each region modeled, these defaults should be reviewed to assess their applicability for that
region. Note that the assumption of no monthly variation of onroad mobDe emissions is not
characteristic of any particular region, but is an indication of the wide variance of monthly
variation depending on location. Data describing monthly variations in VMT may be directly
incorporated into the calculations described in Section 7.3, or may be used to develop region-
specific monthly temporal profiles for the onroad mobile source categories.
7.7 MOBILE SOURCE PROJECTIONS
Future-year emissions estimates for onroad mobile sources must incorporate anticipated changes in
vehicle activity (i.e., VMT) as well as in the vehicle fleet itself. As more and more older, higher
emitting vehicles are phased out of service and replaced with new vehicles, the average emission rate
for the fleet decreases, although this effect is partially offset in most areas by increased activity. In
addition, changing regulations (federal, state and local) affecting the onroad vehicle fleet, such as new
vehicle emission standards and various control measures required under the 1990 Clean Air Act
Amendments, must also be addressed. Consult Sections ffl.E and IV.D of Procedures for Preparing
Emissions Projections (EPA, 1991) for additional guidance.
7.7.1 VMT Projections
EPA has published several documents which provide guidance for the estimation of future year motor
vehicle activity. In addition to the document cited above, the Section 187 VMT Forecasting and
Tracking Guidance document, developed to provide guidance for those areas required to submit ozone
or carbon monoxide State Implementation Plans, describes EPA's recommended methods for
developing future year VMT estimates for these areas. The preferred method involves the use of a
zonal-based travel demand model which has been validated using 1985 or more recent ground counts.
Alternatively, short term VMT projections can be based on Federal Highway Administration's
Highway Performance Monitoring System (HPMS) (U.S. DOT, 1987). For areas located outside of
the travel demand modeling domain and/or the HPMS reporting area, EPA permits use of any
reasonable method for forecasting VMT.
Transportation Control Measures (TCMs). TCMs consist of a wide array of control measures which
are designed for the reduction of traffic activity and/or congestion. TCMs are being studied and
implemented in many CMS As. incorporation of TCMs into the modeling emission inventory requires
that either a VMT reduction or speed adjustment factor be estimated for each TCM. If a speed
adjustment factor is used to characterize the effectiveness of the TCM, an associated emissions
reduction may be estimated using the MOBILE emission factor model.
7-21
-------
7.7.2 Future-Year Emission Factors
As discussed in Section 7.3, MOBELE4.1 does not incorporate VOC and NOX emission reductions
from CAAA-mandated motor vehicle control measures and thus cannot be used to project future-year
emission factors. Once the final version of MOBILES becomes available, it should be used to
prepare future year onroad mobile source emission inventories instead of earlier versions, since
MOBILES addresses the new control requirements. The following paragraphs summarize some
future-year conditions that may need to be modeled, and recommend methods for addressing these
conditions in the emission factor calculations.
Fuel Oxygenate Additives. In the past, oxygenates such as methyl tertiary butyl ether (MTBE) and
ethanol have been added to fuel in winter months to reduce CO emissions; the use of these additives,
however, has become sufficiently widespread that fuel oxygenates are beginning to be used all year.
Since oxygenates generally lower exhaust VOC emissions and raise evaporative VOC emissions,
determining fuel oxygen content for both the base and future modeling years is important. CO
reductions due to oxygenates can be modeled using a specialized version of the MOBILE model called
OXY4. OXY4 is available from EPA QMS, but only adjusts CO emissions. To determine VOC
adjustments, the emissions modeler should consult the EPA guidance document on fuel blends.4
Clean Air Act Amendments. The Clean Air Act Amendments of 1990 require that ozone
nonattainment areas implement a variety of regulations addressing onroad motor vehicles, including
the following:
o Vehicle inspection and maintenance programs;
o Stage II refueling vapor recovery programs;
o Fuel reformulation (including Reid Vapor Pressure (RVP) limits, fuel additives, and fuel
composition); and
o New vehicle emission certification standards (will be incorporated into MOBILE 5.0).
The first two of these requirements can be modeled directly with the MOBILE models. For fuel
reformulation effects, the MOBILE models address the effect of RVP limits; consult Reference 4 for
guidance regarding fuel additive effects. In general, changes hi fuel composition changes do not
affect emission factors, but will affect the chemical speciation of VOC emissions (Chapter 9).
7-22
-------
MOBILE SOURCES
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 for Inventory Preparation, 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 1988.
5. User's Guide for the Urban Airshed Model, Volume IV: User's Manual for the Emissions
Preprocessor System 2.0, Part A: Core FORTRAN System, EPA-450/4-90-007D (R), U.S.
Environmental Protection Agency (OAQPS), Research Triangle Park, NC, May 1992.
6. Procedures for Preparing Emissions Projections, EPA^50/4-91-019, U.S. Environmental
Protection Agency (OAQPS), July 1991.
7-23
-------
BIOGENIC SOURCES
8 BIOGENIC EMISSIONS
8.1 INTRODUCTION
In recent years, air quality modelers have begun to recognize that biogenic emissions (naturally
occurring emissions from vegetation) can contribute significantly to the total emission inventory, even
in predominantly urban regions. Some of these naturally occurring organic species are quite
photochemically reactive (e.g., isoprene). Accordingly, the modeling inventory must include some
estimate of biogenic emissions for completeness.
In a collaborative, effort, researchers at Washington State University and EPA developed a
computerized system to estimate hourly gridded biogenic emissions, called the Biogenic Emissions
Inventory System (BEIS). To support different inventory and modeling input requirements, several
versions of the BEIS model are available, including PC-BEIS (which runs on a personal computer and
estimates county-level biogenic emissions), ROM-BEIS (for Regional Oxidant Model applications),
and UAM-BEIS (for Urban Airshed Model applications). Although the remainder of this section
focuses on the specifics of the UAM-BEIS, similar methodologies are employed in each version.
Several changes to the original version of the UAM-BEIS have recently been implemented, including
(1) a more accurate algorithm for calculating solar energy; (2) a new algorithm to calculate soil NOX
emissions and a conversion factor correction; and (3) additional capabilities for processing user-
supplied county-level or gridded land use data. The current version, UAM-BEIS Version 1.1, is
available to the public from EPA. The following overview of the BEIS is taken from information
contained in the paper "Development of a Biogenic Emissions Inventory System for Regional Scale
Air Pollution Models"1 and in Section 8 of the User's Guide for the Urban Airshed Model, Volume
IV: User's Manual for the Emissions Preprocessor System 2.0, Part A: Core FORTRAN System.2
8.2 OVERVIEW OF THE UAM-BEIS
The UAM-BEIS estimates biogenic emissions based on various biomass, emission, and environmental
factors. In general, the basic equation for these calculations can be expressed as
ERi-EjDBFj EFy F(S,T)]. (8-1)
where ER is the emission rate, i indicates each chemical species (e.g., isoprene or monoterpene), j
indicates vegetation type, BF is a biomass density factor, EF is the biogenic emission factor, and
F(S,T) is an environmental factor accounting for solar radiation (S) and leaf temperature (T). Each of
8-1
-------
the variables in Equation 8-1 are discussed below, followed by a brief description of the processing
methodology employed by the UAM-BEIS.
The UAM-BEIS produces one output file, a binary UAM-fonnat low-level emissions file. This file
contains hourly gridded biogenic emission rates (which have been corrected for episode-specific
environmental conditions) for olefins, paraffins, isoprene, aldehydes, NO, and NO2. This file may be
used directly as input to UAM or merged with the UAM low level anthropogenic emissions file using
EPS 2.0.
8.2.1 Leaf Biomass Factors
The default leaf biomass data base provided with the UAM-BEIS was derived from land use data in
the Oak Ridge Laboratory's Geoecology Data Base. The land use data base is resolved at the county
level and includes areal coverage for different types of forests, agricultural crops, and other areas
such as grasslands and water. Alternatively, UAM-BEIS will estimate emissions using user-specified
county-level or gridded land use data for the 25 vegetation types listed hi Table 8-1. Note that
although water area is included in the default data base provided with UAM-BEIS, and is a required
field in the optional user-specified county-level or gridded land use data files, UAM-BEIS does not
currently estimate aquatic natural emissions. It is expected that future versions of the UAM-BEIS
model will include this calculation, so the water area data field has been maintained hi the input land
use data file structures.
Each of the forest types in the land use data base is assigned to one of three forest groups: oak forest,
other deciduous forest, and coniferous forest. The leaf biomass for each forest group is partitioned
into four emission categories: high isoprene deciduous, low isoprene deciduous, non-isoprene
deciduous, and coniferous. Table 8-2 shows the biomass density factors assumed by UAM-BEIS for
each emission category for the three forest groups. The UAM-BEIS seasonally adjusts biomass based
on the frost dates for each county using a simple step function. For each month, deciduous vegetation
within a county is assumed to have either full biomass or no biomass. Since most high ozone
episodes occur during the summer months, this is not usually a critical assumption.
8.2.2 Emission Factors
The emission factors used in UAM-BEIS are based largely on Zimmerman's study of biogenic
emission rates in the Tampa/St. Petersburg Florida area.3 Emissions are calculated for four
hydrocarbon species: isoprene, a-pinene, other monoterpenes, and other nonmethane hydrocarbons.
The Carbon Bond IV speciations for these four species are shown in Table 3-3 (consult Chapter 9 for
a discussion of the Carbon Bond FV mechanism). The UAM-BEIS processes canopy vegetation types
and non-canopy vegetation types. For forest types (i.e., oak, other deciduous, and coniferous
forests), UAM-BEIS estimates emissions for each forest emission category using the factors shown in
Table 8-4. Emissions for non-forest vegetation types are estimated from areal coverage by land use
type using the emission rates given in Table 8-5.
8-2
-------
BIOGENIC SOURCES
TABLE 8-1. Vegetation types employed by the UAM-BEIS for user-specified county-level or
gridded land use data. ^
1 Oak forest
4 Urban oak
5 Urban deciduous
6 Urban coniferous
7 Alfalfa
8 Sorghum
9 Hay
10 Soybean
11 Corn
Canopy Land Use Types
2 Deciduous forest
Non-Canopy Land Use Types
12 Potato
13 Tobacco
14 Wheat
15 Cotton
16 Rye
17 Rice
18 Peanut
19
20
21
22
23
24
25
3 Coniferous forest
Barley
Oats
Scrub
Grass
Urban grass
Miscellaneous crops
Water
source: Reference 2
TABLE 8-2. Biomass density factors (g/m2) by forest £roup for each emission category.
Forest Group
Oak
Other deciduous
Coniferous
High isoprene
deciduous
185
60
39
Low isoprene
deciduous
60
185
26
Non- isoprene
deciduous
60
90
26
Non-isoprene
coniferous
70 j
135
559 1
source: Reference 1
8-3
-------
TABLE 8-3. Carbon Bond IV speciation for UAM-BEIS biogenic species (moles CB-FV
species/mole chemical species).
Chemical Species
Isoprene
ocPinene
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
» *
TABLE 8-4. Biogenic emission factors (/tg/g/h) by forest emission category used by UAM-
BEIS for canopy vegetation types and urban trees (standardized for full sunlight and 30° C).
Chemical Species
Isoprene
a-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
8-4
-------
BIOGENIC SOURCES
TABLE 8-5. Emission rates (ng/nf-hr) and chemical speciation employed by UAM-BEIS for
non-canopy land use types (except urban trees).
Non-Canopy Land Use Type
7 Alfalfa
8 Sorghum
9 Hay
10 Soybean
1 1 Corn
12 Potato
*
13 Tobacco
14 Wheat
15 Cotton
16 Rye
17 Rice
1 8 Peanut
19 Barley
20 Oats
21 Scrub
22 Grass
23 Urban grass
24 Miscellaneous crops
Emission
Rate
Otg/n^-hr)
37.9
39.4
189.0
22.2
0.5
48.1
294.0
30.0
37.9
37.9
510.0
510.0
37.9
37.9
189.0
281.0
281.0
37.9
Percent of Emission Rate by Chemical Species
Monoterpenes
10
25
25
0
10
25
10
10
25
25
25
'25
25
25
25
25
25
25
o-Pinene
10
25
25
0
10
25
10
10
25
25
25
~25
25
25
25
25
25
25
Isoprene
50
20
20
100
0
0
0
50
20
20
20
20
20
20
20
20
20
20
Other
30
30
30
0
80
50
80
36
30
30
30
30
30
30
30
30
30
30 !
8-5
-------
Some natural sources also emit quantities of 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. The previous version on UAM-BEIS estimated soil
NOX emissions from grasslands only. The current version, UAM-BEIS 1.1, has been modified to
estimate NO emissions for forest, agricultural crops, and urban trees as well as grasslands; additional
changes include an increase to the NO emission factor for grasslands, the removal of NO2 emissions,
and the correction of a conversion factor. UAM-BEIS calculates soil NO emissions using the
following equation:
FNO = A exp (0.071 T.) (8-2)
In Equation 8-2, FNO is the NO flux in units of ng Nitrogen/n^-s, T8 is the soil temperature in
Celsius, and A is a constant equal to 0.9 for grasslands and pasture, 0.07 for forests and urban trees,
and 0.2 for agricultural croplands.
8.2.3 Environmental Correction Factors
w *
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 UAM-BEIS are
standardized for full sunlight and 30° Celsius. The UAM-BEIS adjusts these emission factors to
account for the effects of variations in ambient conditions using relationships derived by Tingey 4>5>6
The emission factor sensitivities to leaf temperature for isoprene and monoterpene are shown
graphically in Figure 8-1.
%
BEIS also simulates the vertical variation of leaf temperature and sunlight within the forest canopy.
The canopy model employed by UAM-BEIS assumes that sunlight decreases exponentially through the
hypothetical forest canopy; the rate of attenuation depends on the assumed biomass distribution.
Figure 8-2 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 level; Figure 8-3 presents the assumed temperature
and solar flux variation by canopy layer for deciduous and coniferous forests.
8.2.4 Processing Methodology
UAM-BEIS calculates the emission rates of ot-pinenes, other monoterpenes, isoprene, and other
nonmethane hydrocarbons by multiplying biomass for each vegetation type by the appropriate
emission factors. The calculated emission rates are then adjusted for specific environmental
conditions of tne modeling episode using data from the user-supplied meteorology file and the output
files of the temperature and winds UAM'meteorology preprocessors. Corrections are performed for
8-6
-------
BIOGENIC SOURCES
2.S
2.O -
1 .3 -
1.0 H
o.o
2.S
O.O
ISOPRENE
BOO uE/rn"/ii
-*OO
200
1OO
S 1O IS 2O 2S
Leaf temperature (C)
MONOTERPENE
4-0
pin«n«
«rp«
unia«ntif i
'L_eof temperature
*o
FIGURE 8-1. Biogenic emission factor sensitivity to leaf temperature.
8-7
-------
M
DECIDUOUS
CONIFEROUS
15m
v
u
20m
FIGUUE 8-2. Schematic representation of forest canopy types.
-------
CANOPY MODEL
10_ T-29.7. RH=56«. U=1.5 M/S. SOLAR«=1.15 LY/MIN
EC
LU
6-
4-
DECIDUOUS
CONIFEROUS
20 22 24
26 26 30 32 34 36
TEMP (C)
10
EC
LJ
Ambient flux * 2548 uE
DECIDUOUS
CONIFEROUS
SCO 1000 1500 2000
SOLAR FLUX (UE)
2500
FIGURE 8-3. Temperature and solar flux variations by canopy layer.
-------
each grid cell of the modeling domain using the hourly, gridded temperature and surface wind speed
fields data contained in these two UAM input files.
For forested areas, land use area data for each county are multiplied by biomass density factors for
each forest type. EPA provides a default land use/biomass data file for use with UAM-BEIS which
contains county-level land use area for each of the 25 vegetation types listed in Table 8-1. The
calculated county-level biomass is gridded using a county allocation file, which indicates the
percentage of each grid cell occupied by a particular county. If more current land use data are
available, the user may include these data by using the optional gridded land use data file or the
optional county-level land use data file. The next step in processing forested areas involves adjusting
the emission rates for temperature and solar energy changes within the forest canopy. The adjusted
emission rates are then grouped into categories for five Carbon Bond Mechanism FV (Gery et al.,
1989) species (olefins, paraffins, isoprene, aldehydes, and nonreactives) and nitric oxide (NO).
Although UAM-BEIS calculates emission rates for nonreactive hydrocarbons, it does not output this
information because it is not used by UAM.
Nonforested areas, or noncanopy areas (including agricultural areas and urban trees), are handled a
bit differently. The land use data found in the land use/biomass file is multiplied by an on/off flag
(also found in this file) to indicate whether the vegetation is growing or not for the episode month.
Next the data are gridded using the county allocation file. Land use is then multiplied directly by an
emission factor to produce emission rates. Speciation and environmental adjustments are handled the
same as for forested areas, except that the nonforest emission rates are not adjusted for canopy
effects.
8.3 UAM-BEIS INPUT REQUIREMENTS
As indicated above, the UAM-BEIS uses three types of data files: UAM preprocessor data, user-
supplied data, and data supplied to the user by EPA. Figure $-4 shows a flow chart of the UAM-
BEIS-flow of information. Each of the input data files is briefly described below. Section 8 of the
User's Guide for the Urban Airshed Model, Volume IV: User's Manual for the Emissions
Preprocessor System 2.0, Part A: Core FORTRAN System2 contains detailed format descriptions of the
various input files.
UAM Preprocessor Data. Two of the UAM preprocessor files are also used by UAM-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. In addition, Version 1.1 of the UAM-BEIS allows the user to specify land use data, either
8-10
-------
BIOGENIC SOURCES
PREPROCESSOR
DATA
DATA SUPPLIED
TO THE USER
BY EPA
OPTIONAL USER-
SUPPUED DATA
USER-SUPPLIED
METEOROLOGY, COUNTY
ALLOCATION DATA
RAWMET CNTYALO
UAM-BEIS
FIGURE 8-4. UAM-BEIS input and output flies.
8-11
-------
county-level or for each grid cell of the modeling domain. Each of these files are described briefly
below:
RAWMET The meteorology file RAWMET contains hourly surface meteorological information
on relative humidity, cloud coverage, and cloud height for one station in the user's
UAM domain. The file is created by the user with data which may be obtained from
the National Weather Service.
CNTYALO The county allocation file contains the percentage of each grid cell that a given county
occupies. To speed processing, this file should be sorted by FEPS code. The
CNTYALO file may be created by the user; however, the EPA has a Gridded
Surrogate/County Allocation Utility which will create the file given the UAM domain
coordinates and grid cell size. To request this data, contact the nearest EPA Regional
Office.
GRBIO If the user wishes to use more up-to-date land use data which are already gridded, the
optional gridded land use file, GRBIO, may be used. For each grid cell, the user
must input values in hectares for the 25 land use types used by UAM-BEIS.
UCBIO If the user has more up-to-date land use data on a county level, the optional county
land use file, UCBIO, may be used. For each FIPS code, the user must input values
in hectares for the 25 land use types used by UAM-BEIS.
In addition to the user-supplied data files listed above, the user must supply certain input control data
required by UAM-BEIS, consisting of domain and scenario-specific information along with control
option flags. The user must input start date and hour, end date and hour, month of scenario, number
of columns and rows, number of hours from Greenwich Mean Tune (GMT) to local time, number of
counties, and a list of counties. Three flags which control several options in UAM-BEIS must also be
set by the user. The first flag, called GIFLAG, must be set to TRUE if the user wishes to input his
own gridded land use data using the optional GRBIO file. The second flag, called CFLAG, must be
set to TRUE if the optional UCBIO file is being used to input the user's own county land use data.
The third flag, called GROWFLAG, must be set if the user inputs gridded land-use data. This flag is
set to TRUE to indicate that vegetation is growing and FALSE if only coniferous vegetation is
growing, which might be the case in a cold season like late fall or winter.
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, 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.
8-12
-------
BIOGENIC SOURCES
8.4 PROJECTION OF BIOGENIC INVENTORIES
In general, the same emissions factors will be used to estimate biogenic emissions for both base 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 cell.
» 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 this 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.
8-13
-------
References for Chapter 8:
1. T. E. Pierce, B. K. Lamb, and A. R. Van Meter, "Development of a Biogenic Emissions
Inventory System for Regional Scale Air Pollution Models", Paper No. 90-94.3, presented at
the 83rd Air and Waste Management Association Annual Meeting at Pittsburgh, Pennsylvania,
June 1990.
2. User's Guide for the Urban Airshed Model, Volume IV: User's Manual for the Emissions
Preprocessor System 2.0, Part A: Core FORTRAN System, EPA-450/4-90-007D (R), U.S.
Environmental Protection Agency (OAQPS), Research Triangle Park, NC, May 1992.
3. P. Zimmennan, 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.
* *
4. D. Tingey, Atmospheric Biogenic Hydrocarbons, J. Bufalini and R. Arnts, eds., Ann Arbor
Science Publication, Ann Arbor, MI, 1981, pp. 53-79.
5. D. Tingey, R. Evans, and M. Gumpertz, "Effects of Environmental Conditions on Isoprene
Emissions from Live Oak," Planta (-152): 565 (1981).
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).
8-14
-------
SPEC1ATION
SPECIATION 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, NOX may have to be
specified as NO and NC^. Each model's classification requirements differ somewhat; this chapter
focuses on speciation of emissions according to the Carbon Bond IV Mechanism employed by the
UAM.
9.2 THE CARBON BOND-IV MECHANISM
The currently available version of the UAM uses version IV of the Carbon Bond Mechanism (CBM-
IV). Since every reaction of all of the organic species found in an urban atmosphere cannot be
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,
the CBM-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 7.1
The differential equations that describe the CBM-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 of the state species. Table 9-1 lists the
carbon bond species used in the CBM-FV version of the UAM.
In EPS 2.0, each carbon atom of total VOC emissions is assigned to one of the following ten
species listed in Table 9-1: olefim'c carbon bond (OLE), paraffinic carbon bonds (PAR), toluene
(TOL), xylene (XYL), formaldehyde (FORM), high molecular weight aldehydes (ALD2), emene
(ETH), methanol (MEOH), ethanol (ETOH), and isoprene (ISOP). NOX emissions are partitioned
into NO and NO2. Emissions of carbon monoxide (CO) should also be included in the UAM
modeling inventory, since CO is a photochemically reactive species.
9.3 CHEMICAL ALLOCATION OF VOC
Generally, the basic annual inventory will contaia estimates of either total VOC or non-methane
VOC, depending on what emission factor information is used for computing emissions. The basic
9-1
-------
TABLE 9-1. CBM-IV chemical species recognized by the UAM system, with molecular
weights (grams per mole) for unit conversion.
UAM Species Species Name
Default species supported by EPS 2.0:
NO Nitric oxide
NO2 Nitrogen dioxide
CO Carbon monoxide
SO2 Sulfur dioxide
AERO Aerosols (particulates)
PM10 Paniculate matter, diameter <, 10 microns
OLE1" Olefinic carbon bond (C=C)
PARf Paraffmic carbon bond (C C)
TOL+ Toluene (C6H5-CH3)
XYL* Xylene (C6H6-(CH3)2)
FORM* Formaldehyde (CH2=O)
ALD21" High molecular weight aldehydes CRCHO, R*H)
ETHf Ethene (CH2=CH2)
MEOHf Methanol (optional)
ETOH* Ethanol (optional)
ISOPf Isoprene
Other species:
O3 Ozone
O2 Oxygen
H2O Water
CO2 Carbon dioxide
CH4 Methane
TOTAL HC Total hydrocarbons
OXIDANT Photochemical oxidant
SO4 Sulfate
' CRES Cresol and higher molecular weight phenols
MGLY Methyl glyoxal (CH3C(O)C(O)H)
OPEN Aromatic ring fragment acid
PNA Peroxynitric acid (HO2NO2)
NOX Total nitrogen oxides (NO + NO2 + N2O5 + NO3)
SOX Total sulfur oxides
PAN Peroxyacyl nitrate (CH3C(O)O2NO2)
MONO Nitrous acid
H2O2 Hydrogen peroxide
HNO3 Nitric acid
T Default molecular weight for carbon bond species derived from hydrocarbons
number multiplied by the molecular weight of methane.
Molecular
Weight
30
46
28
64
1
1
32
16
112
128
16
32
32
16
32
80
48
32
18
44
16
16
48
96
128
72
86
79
46
64
121
47
34
63
is the carbon
source: Reference I
9-2
-------
SPECIATION
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 split factors for this particular source are
0.2 tons OLE/ton VOC, 0.5 tons PAR/ton VOC, and 0.3 tons ALD2/ton VOC. Simple
multiplication of each factor by the total tonnage of VOC yields the quantity of VOC in each
carbon bond class, in this case 2, 5, and 3 tons per day, respectively.
This allocation step, would, of course, have to be performed for each emission total developed in the
inventory using different split factors appropriate for each source or source category. Please note that
the example above is simplified; the UAM requires that split factors be provided in terms of moles
CBM-IV species per gram total VOC. Calculation of split factors in these units will be discussed
further below.
As can be seen from the above example, the VOC allocation step is not difficult once the split factors
are available for each source. The major difficulty in this process is determining which split factors
are most appropriate. Two basic approaches can be followed for determining split factors. Ideally,
VOC split factors should be source-specific, reflecting the actual composition of VOC emissions from
each individual source.
» 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, chemical manufacturers, etc.), and
solvent composition data could be solicited from smaller commercial and industrial
establishments (dry cleaners, degreasers, etc).
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 develop CBM-IV split factors from generalized
speciation data, the individual chemical compounds typically present in the emissions from each
source type (and their molecular weights and weight fractions of the emissions mixture) must first be
identified. Then, each of the chemical compounds present in the modeling inventory must be
classified according to the CBM-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 taken from EPA's VOC/PM Speciation Data Base Management System
(SPEC1ATE).2 The SPECIATE data base contains over 250 VOC "emission profiles" like the
example in Table 9-2 for various point and area source categories; the data base also contains profiles
for emissions from motor vehicles and aircraft. In each profile, individual chemical compounds are
9-3
-------
TABLE 9-2. Example VOC speciation profile from the SPECIATE data base
Profile Name: Fixed Roof Tank 0 Crude Oil Production
Profile Number: 0296
Control Device: Uncontrolled
Data Source: Engineering evaluation of test data and literature data
SAROAD
Number
43115
43116
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
111-16-9
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
Mol.
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
Percent
Weight
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
9-4
-------
SPECIATION
listed with their corresponding molecular weights and weight percentage of the mixture. (The EPA
speciation profile codes and descriptions are listed in Appendix Q.
The type of information contained in Table 9-2 can be used with the CBM-IV species assignments for
individual chemical compounds from Guidelines for Using O2JPM-4 with CBM-IV or Optional
Mechanisms3 to calculate composite split factors by speciation 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 VOQ 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
split factor for each CBM species i = ^ [(WFj IMWJ) (moles of/ / moley)] (9-1)
where / is the CBM-IV species, j is each chemical compound in the mixture (e.g., carbon
tetrachloride), WFj is the weight fraction ofy in the mixture, and MWj is the molecular weight of
chemical compound j.
If source-specific speciation data are unavailable, the emissions modeler can use the speciation
profiles contained in the SPECIATE database (with the CBM-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.
EPS 2.0 includes an input file preparation utility, called EMSCVT, which will create the split
factors and SCC(ASQ/speciation profile cross reference files required by the CHMSPL module
(which converts criteria pollutant emissions into CBM-IV species). The split factors file consists
of multiplicative factors for converting grams of criteria pollutant emissions into moles of CBM-IV
species for each speciation profile. The SCQASQ/speciation profiles cross-reference file contains
the speciation profile code assignments by SCC code for point sources and by Area Source
Category (ASC) code for area and mobile sources; this file may also contain source-specific
speciation profile assignments. The user may either generate.completely new versions of these
files or update the existing versions. EMSCVT calculates the split factors for each speciation
profile based on the weight fractions of the individual chemical compounds associated with that
profile and the carbon bond speciation for each compound, as described above. The EMSCVT
utility has been designed to allow the user to easily incorporate source-specific speciated emissions
data, if such data are available.
The EMSCVT utility will produce split factors to disaggregate either total or reactive VOC,
depending on the value specified for a user input flag. The split factors used to speciate emissions
must be compatible with the way emissions have been reported in the inventory (e.g., if emissions
have been reported as total VOC, the split factors must be calculated for total VOC also). See
Section 9.6 for more on this topic.
9-5
-------
Whether or not the agency intends to employ a model that incorporates the carbon bond mechanism, a
photochemical modeling specialist should be consulted to review all 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 case
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 NOX AS NO AND NO2
Some photochemical models do not require that nitrogen oxides be distinguished as either nitric oxide
or nitrogen dioxide. Instead, these models assume that all NO, 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 NC^. In this
sense, allocating NOX into NO and NO2 is analogous to utilizing a 2-class 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 NO2", which means that a molecular weight of 46 is attributed
to NO as well as NC^, 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 NOX "as
NO2", 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 value of NO "as NO^." 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 NOj per hour. Given split
factors of 95 and 5 percent by weight "as NO2H, then NO and NO2 emissions would be
equivalent to 950 and 50 kg "as NC^" 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 NO-,.
Two sources of such data are References 4 and 5. As a rough average, 97 percent (by weight as
NO2) of the NOX emitted from most boilers will be NO.
9-6
-------
Unless the user has explicitly included source- or source category-specific split factors for
speciating NOX emissions in the split factors file, the CHMSPL module will use default split
factors for all sources, which assume 90% and 10% by weight of NO (as NGg) and N<>2,
respectively.
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 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 perchloroethylene). Likewise, since no evidence suggests that NO/NO2
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. For example, if split factors for total VOC were to be applied
(erroneously) to RVOC emissions estimates, the resulting emission estimates in each VOC class
9-7
-------
would be potentially underestimated, since a specified nonreactive fraction of the emissions would be
subtracted before allocation of the reactive portion to carbon bond classes.
Converting the split factors from a total VOC to an RVOC basis is a relatively simple procedure that
involves recalculating the average molecular weight of the mixture without the nonreactive
components (in most cases, primarily methane); the methane-free molecular weight may then be used
with the CBM-IV species assignments by chemical compound to generate non-methane VOC split
factors using the method described in Section 9.3.
If the existing inventory is not in terms of VOC or non-methane VOC, and instead utilizes some sort
of species classification scheme that is incompatible with the chemistry 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 specialist if this situation exists.
Another important consideration is compatibility of the split factors with the source classification
scheme. The source categories and subcategories chosen for the basic inventory may fail to
distinguish between sources having substantially different emission compositions, requiring different
sets of split factors. There are several possible solutions to this problem, as shown in the example
below.
* Area source degreasing may be considered as a single category in the emission inventory, but
different degreasing solvents are used in different 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.
«t
Second, if there are many degreasing operations, each of which emits only 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
cells containing degreasing emissions. In this case, degreasing can be treated as a single area
source category, with a composite set of split factors reflecting 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 die 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.
A similar situation may be encountered in dealing with motor vehicle emissions; different speciation
profiles are available for exhaust and evaporative VOC emissions, whereas the basic inventory may
9-8
-------
SPECIATION
only provide a single lumped estimate combining these emission components. Since mobile sources
comprise a significant portion of the anthropogenic inventory in most urban areas, these emissions
should be recalculated in terms of the 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 model to evaluate control strategies that may affect exhaust and evaporative emissions differently.
If EPS 2.0 is being used to develop the modeling inventory, emissions from onroad motor vehicles
will already have been disaggregated into exhaust and evaporative (consisting of diurnal, refueling,
and running loss components) by the PREAM or LBASE modules.
As a special consideration when speciating emissions from on-road motor vehicles, the effects of
RVP, oxygenated fuel blends, and alternative fuel use on VOC speciation must also be quantified.
Specific guidance on the evaluation of these effects is provided in the EPA Technical Memorandum
Motor Vehicle VOC Speciation for SIP Development.6
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 cliss totals. A single set of
split factors is, therefore, adequate for all external combustion sources operating on a given type of
fuel, and no subcategories would be 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.
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
9-9
-------
TABLE 9-3. Example "split factor" file (excerpt).
Source
Category*
sec
30600801f
30600802
30600803
40300106
40300107
40300152
40300205
30000606
30000608
Pollutant
Codeb
HC
HC
HC
HC
HC
HC
HC
NX
NX
Class lc
SF1 MWe
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 3C
SF MW
7.9 71
3.9 72
6.6 72
11.2 72
11.2 72
37.0 31
37.0 31
Class 4°
SF MW
.7 96
7.3 92
4.4 96
1.4 92
1.4 92
4.6 101
4.6 101
a Source category by SCC code (eight digits) *
b Code: HC = VOC, NX = NOX
0 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 - NO2
d Split factor, percent of total, by weight
e Average molecular weight
f Each line constitutes a record, either for VOC or NOX, for one source category
9-10
-------
SPECIATION
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 stored in 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 NO, emission totals.
The method for allocating area and mobile source VOC and NO, 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.
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.
9-11
-------
References for Chapter 9:
1. User's Guide for the Urban Airshed Model, Volume I: User's Manual for UAM (CBM-IV), EPA-
450/4-90-007A, U.S. Environmental Protection Agency (OAQPS), Research Triangle Park, NC,
June 1990.
2. VOC/PM Speciation Data Base Management System (SPECIA7E), Version 4.1, U.S.
Environmental Protection Agency, (OAQPS), Research Triangle Park, NC, 1991.
3. H. Hogo and M. W. Gery, Guidelines for Using OZIPM-4 with CBM-IV or Optional
Mechanisms, Volume 1, EPA Contract No. 68-02-4136, Systems Applications Incorporated, 1986.
4. R. J. Milligan et al., Review ofNOx Emission Factors for Stationary Combustion Sources, EPA-
450/4-79-021, U.S. Environmental Protection Agency, September 1979.
5. T. W. Tesche and W. R. Oliver, Technical Basis for Refinement ofSAI 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 Speciation for SIP Development, Technical Memorandum to
John Calcagni and William Laxton, U.S. EPA Office of Air and Radiation, Ann Arbor, MI,
March 23, 1989.
9-12
-------
APPENDIX A:
ACRONYMS AND GLOSSARY
A-l
-------
ACRONYMS
ACRONYMS AND ABBREVIATIONS
AEROS Aerometric and Emissions Reporting System
AIRS Aerometric Information Retrieval System
AFS AIRS Facility Subsystem
AMS Area and Mobile Subsystem (of AIRS)
ASC AIRS Source Category (used to identify area and mobile source categories)
ATP Anti-tampering program
BACT Best Available Control Technology
BEA Bureau of Economic Analysis
'BEIS EPA's Biogenic Emissions Inventory System
CBM Carbon Bond Mechanism
CMSA Consolidated Metropolitan Statistical Area
CO Carbon monoxide
CTG Control Technique Guideline
EMAR Emissions Model ASCII Record
EMBR Emissions Model Binary Record
EPA Environmental Protection Agency
EPS Urban Airshed Model Emissions Preprocessor System
FTP Federal Implementation Plan
FTPS Federal Information Processing Standards
FMVCP Federal Motor Vehicle Control Program
FTP Federal Test Procedure (for motor vehicles)
HDGV Heavy duty gasoline vehicles
I/M Inspection and maintenance program
JCL Job Command Language
LDGV Light duty gasoline vehicles
A-3
-------
LDGT Light duty gasoline trucks
MACT Maximum Available Control Technology
MDGV Medium duty gasoline vehicles
MSA Metropolitan Statistical Area
NAAQS National Ambient Air Quality Standard
NAPAP National Acid Precipitation Assessment Program
NEDS National Emission Data System (superseded by AIRS)
NOX Nitrogen oxides
O3 Ozone
OAQPS EPA's Office of Air Quality Planning and Standards
RACT Reasonably Available Control Technology
RHC Reactive hydrocarbons
ROM Regional Oxidant Model
RVOC Reactive Volatile Organic Compounds
RVP Reid vapor pressure
SAMS SIP Air Management System
SAROAD Storage and Retrieval Air Quality Data code (superseded by AIRS)
SCC Source Classification Code
SIC Standard Industrial Classification
SIP State Implementation Plan
SMSA Standard Metropolitan Statistical Area
SOX Sulfur oxides
THC Total hydrocarbons
TNMHC TotaJ nonmethane hydrocarbons
TSP Total suspended paniculate matter
UAM Urban Airshed Model
USGS United States Geological Survey
VMT Vehicle miles traveled
VOC Volatile organic compounds
A-4
-------
GLOSSARY
GLOSSARY
Activity level
Any variable parameter associated with the operation of a source of emissions which is
proportional to the quantity of pollutant emitted.
Actual emission inventory
An emission inventory which represents the actual emission rates over the specified time interval
for each source.
Adjusted base year inventory
The base year inventory minus (1) biogenic emissions, (2) emission reductions from Federal
Motor Vehicle Control Program (FMVCP) regulations promulgated prior to 1 January 1990, (3)
emissions reductions from RVP rules promulgated prior to 1990 Clean Air Act Amendment
(CAAA) enactment or required under CAAA Section 211(h), and (4) any sources outside of the
nonattainment area. Required for ozone nonattainment areas ranked moderate, serious, severe,
and extreme.
Allowable emission inventory
An emission inventory which represents maximum allowable emission rates for each source.
Allowable emission rates may be based on emission factor limits, activity level limits, or both.
!
Annual average daily emissions
The average daily emission rate from a particular source or source category, calculated by
dividing the annual emissions for that source by 365 days per year.
Anthropogenic emissions
Emissions from man-made sources; commonly subdivided into point, area, and mobile sources for
inventory purposes.
Area source emissions
Emissions which are assumed to occur over a given area rather than at a specified point; often
includes emissions from sources considered too small or numerous to be handled individually in
the point source inventory.
Base year inventory
A comprehensive and accurate inventory of actual emissions of VOC, NOX, and CO from all
stationary point and area sources (including biogenic sources), on-road mobile sources, and non-
road mobile sources in an area. Required for all nonattainment areas by the 1990 Clean Air Act
Amendments (CAAA). If a nonattainment area is,required to perform photochemical modeling
A-5
-------
(UAM) by the 1990 CAAA, the base year inventory must also include emissions from attainment
areas located within the modeling domain.
Biogenic Emissions
Naturally occurring emissions from vegetation.
Carbon bond mechanism
The chemical kinetics mechanism employed by the Urban Airshed Model, in which various
hydrocarbons 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 a molecular lumping approach would group reactions of entire
molecules.
Crankcase evaporative emissions
Evaporative emissions emitted from motor vehicle crankcases.
Design value
A numerical value indicating the air quality for an area. For ozone, the design value is defined as
the fourth highest monitored ozone value for a single monitor with three complete years of data
(i.e., the greatest of the fourth-highest ozone values measured at each monitor within the area).
Diurnal evaporative emissions
Evaporative emissions from motor vehicles that occur when fuel vapor is emitted from partially
filled fuel tanks of non-operating vehicles during periods of rising ambient temperatures.
Effective stack height
The sum of the actual physical stack height and the plume rise. Effective stack height is defined
as the height at which a plume becomes passive and subsequently follows ambient air motion.
Emission factor
A factor, usually expressed as mass pollutant per throughput or activity level, used to estimate
emissions for a given source.
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 codes and stack parameters.
Evaporative emissions
Emissions resulting from the volatilization of gasoline and solvents due to rising ambient
temperatures, or engine heat after motor vehicle shutdown. EPA's MOBILE4.1 emission factor
mode! recognizes five components of motor vehicle evaportive emissions: hot soak, diurnal, and
crankcase, which are lumped together under "evaporative," and resting and running losses, which
are reported separately.
A-6
-------
Exhaust emissions
Emissions resulting from the combustion processes associated with the operation of motor
vehicles. In EPA's MOBILE4.1 model, exhaust emissions are composed of three components,
representing three different operating modes: cold start, hot start, and hot stabilized.
Grid cell
The three-dimensional box-like cell of a grid system; also commonly used to refer specifically to
the groundlevel horizontal layer of grid cells over which emissions are allocated for modeling.
Grid layer
See vertical resolution.
Grid model
An air quality simulation model that provides estimates of pollutant concentrations for a gridded
network of receptors, using assumptions regarding the exchange of air between hypothetical box-
like cells in the atmosphere above an emission grid system. Mathematically, this is known as an
"Eulerian" model; the Urban Airshed Model (UAM) is an example of a photochemical grid
model.
Growth surrogate
See projection surrogate.
Hot soak evaporative emissions
Evaporative emissions from motor vehicles that occur when fuel in the engine is vaporized by the
residual heat of the engine after the vehicle is shut off.
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
DDD is the day; for example, May 3, 1990 is written as 90123 in Julian notation.
Land use
A description of the major natural or man-made features contained in an area of land, or a
description of the way the land is being used. Examples of land use categories include forest,
desert, cropland or agricultural, urban, grasslands, and wetlands.
Line source
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; see also mobile
source emissions.
A-7
-------
Link
A surrogate generated to model allocation of line source emissions. It takes the form of a line
(designated by start and endpoint coordinates), or a group of lines; spatial allocation is performed
on the basis of link length per grid cell.
Lumping
In chemical mechanisms, the strategem of representing certain compounds by surrogate or
hypothetical species in order to reduce the assumed number of elementary reactions to a
manageable number.
Mobile source emissions
Emissions from non-stationary sources. Also commonly used to designate emissions from on-
road motor vehicles only (as opposed to "other mobile" sources). 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
Abbreviated as NOX. With respect to air pollutants, nitric oxide (NO) and nitrogen dioxide
together comprise nitrogen oxides (NOX).
PCBEIS
A version of EPA's Biogenic Emission Inventory System, designed to run on a personal
computer, which produces county-level biogenic emissions estimates.
Periodic inventory
An emission inventory based on actual emission rates and addressing both VOC and NOX
emission sources, primarily used for tracking emissions reductions, particularly relating to
Reasonable Further Progress (RFP) requirements. Required for all ozone nonattainment
classifications by the 1990 Clean Air Act Amendments.
Photochemical model
An air quality simulation model which simulates the photochemical reactions that occur over an
area during each hour of the day or days for which the model is being applied.
Point source emissions
Emissions which are inventories as occur ing 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
emisisons (due either to gas temperatures higher than the ambient air or the momentum of the
emissions as they leave the stack).
PM-10
Paniculate matter with particle diameters of 10 microns and smaller.
A-8
-------
GLOSSARY
POD
A grouping of sources based on similar SCC(ASC) codes.
Projection inventory (for future year)
An inventory for a future year, derived from a base year inventory, in which emissions have been
adjusted to reflect differences in activity levels and/or implemented controls between the base and
future years.
Projection surrogate
A quantity for which official projections are known and whose growth may be assumed similar to
mat of activity for a particular source category or type.
Reactivity
Measure of the tendency of a chemical species to react with other species.
Receptor
A hypothetical sensor or monitoring instrument, usually a unit of a network overlaid on a map of
the area being modeled. Eulerian models usually assume one receptor at the center of each grid
cell.
Resting losses
Evaporative emissions from nonoperating motor vehicles that result from vapors permieating parts
of the evaporative emission control syste, migrating out of the carbon canister, or evaporating
liquid fuel leaks. Resting losses are distinct from diurnal evaporative emissions in that they do
not result from rising ambient temperatures.
RFP projection inventory
A projected inventory based on allowable rather than actual emissions and used for tracking
Reasonable Further Progress (RFP) for State Implementation Plans. Required for nonattainment
areas ranked moderate and above by the 1990 Clean Air Act Amendments.
Running losses
Evaporative emissions from motor vehicles that occur while the vehicle is in operation.
Seasonal adjustment
Adjustment of emissons from an annual to a seasonal level, usually based on seasonal variations
in activity levels or temperature.
Standard Industrial Classification code
Abbreviated as SIC; a 4-digit integer code designating the primary type of business for a facility
(e.g., petroleum refinery, electric utility, gasoline service station, dry cleaner, etc.).
Source
A process or activity resulting in the release of pollutants to the atmosphere.
A-9
-------
Source Classification Code
Abbreviated as SCC; an 8-digit integer code used to characterize a process. For example, SCC
code 40200902 corresponds to the use of acetone as a thinning solvent in surface coating
operations.
Source/receptor relationship
A model that predicts ambient pollutant concentrations based on precursor emission levels.
Photochemical models are one type of source/receptor relationship.
Sulfur oxides
Abbreviated as SOX; comprised primarily of sulfur dioxide (SC^) and sulfate (SO4).
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 spatial 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
the chemical mechanism (e.g., the carbon bond mechanism) employed in a photochemical air
quality simulation model.
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 air
quality simulation model.
Stack parameters
Characteristic parameters of a stack and its associated plume, as required for input into some
photochemical simulation models. Typical stack parameters include stack height, inner diameter,
volumetric flow rate, and gas exit velocity and temperature; stack parameters are required to
calculate plume rise.
A-10
-------
Temporal resolution
Disaggregation of annual, seasonal, or daily emission rates into hourly rates. (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 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 emissions grid
system. Mathematically, this is known as a "Lagrangian" model.
Vertical resolution
(1) Allocation of emissions to vertical layers of grid cells based on plume calculations. (2\In
regard to meteorological parameters and concentrations of pollutants in ambient air, tfie provision
(in a model) of a means of 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 (CO^, carbonic acid, carbonates, and
metallic carbides.
A-ll
-------
APPENDIX B:
EPS 2.0 REPORTING CODES
TABLE B-l. Activity codes.
TABLE B-2. Control codes.
TABLE B-3. Process codes.
TABLE B-4. POD codes.
TABLE B-5. Speciation profile codes.
v
WS
B-l
-------
EPS REPORTING CODES
TABLE B-l. Activity codes.
000 UNSPECIFIED ACTIVITIES
100 RESOURCE DEVELOPMENT & AGRICULTURE
110 AGRICULTURAL PRODUCTION
111 AGRICULTURAL CROPS
112 AGRICULTURAL LIVESTOCK
113 AGRICULTURAL SERVICES
120 FORESTRY
130 MINING '
131 METAL MINING
132 COAL MINING
133 STONE & CLAY (MINING)
134 CHEMICALS & FERTILIZER MINERAL
140 OIL & GAS EXTRACTION
141 LIQUID GAS PRODUCTION
200 MANUFACTURING & INDUSTRIAL
210 FOOD & KINDRED
211 FRUIT/VEG PRESERVATION
212 GRAIN MILL PRODUCTS
213 BAKERY PRODUCTS
214 VEGETABLE OIL
215 SUGAR MFG/REFINING
216 MALT BEVERAGES
217 WINES & BRANDY
220 LUMBER & WOOD PRODUCTS
230 PAPER & ALLIED
231 PULP & PAPER MILLS
240 . CHEMICAL & ALLIED
241 RUBBER & PLASTICS MFG
242 DRUGS
243 CLEANING/TOILET PREP
244 PAINT MFG
245 AGRI CHEMICALS
260 PETROLEUM REFINING/RELATED
261 PETROLEUM REFINING
262 PAVING & ROOFING MATERIALS
263 PET COKE/BRIQUETTE
270 MINERAL PRODUCTS
271 GLASS/GLASS PRODUCTS
280 METALLURGICAL
281 - IRON/STEEL PRODUCTION
282 IRON/STEEL FOUNDRY
B-3
-------
TABLE B-l. Continued.
283 NONFERROUS METALS
290 MISC. MANUFACTURING
291 TEXTILES & APPAREL
292 FURNITURE & FIXTURES
293 FABRICATED METAL
294 MACHINERY
295 TRANSPORTATION EQUIPMENT
296 RUBBER & PLASTICS FAB.
297 TOBACCO MANUFACTURING
298 INSTRUMENTS
300 SERVICES & COMMERCE
310 ELECTRIC UTILITIES
320 PETROLEUM & GAS MARKETING
321 BULK PLANTS
322 SERVICE STATIONS
323 PIPE LINES
330 MISC. SERVICES
331 STEAM SUPPLY
332 PRINTING & PUBLISHING
333 LAUNDRY & DRYCLEANERS
334 SANITARY & WATER
335 HEALTH SERVICES
336 EDUCATIONAL SERVICES
400 TRANSPORTATION
410 ON-ROAD TRAVEL
420 RAIL TRANSPORT
430 WATER BORNE
440, AIR TRANSPORTATION
450 LIGHT DUTY AUTOMOBILE
451 LDA - COLD START
452 LDA - HOT STABILIZED
453 LDA - HOT START
454 LDA - HOT SOAK
455 LDA - DIURNAL LOSSES
456 LDA - RUNNING LOSSES
460 LIGHT DUTY TRUCK
461 LDT - COLD START
462 LDT - HOT STABILIZED
463 LDT - HOT START
464 - LDT - HOT SOAK
465 LDT - DIURNAL LOSSES
B-4
-------
EPS REPORTING CODES \
TABLE B-l. Concluded.
466 LOT - RUNNING LOSSES
470 MEDIUM DUTY GAS VEHICLES
471 MDGV - COLD START
472 MDGV - HOT STABILIZED
473 MDGV - HOT START
474 MDGV - HOT SOAK
475 MDGV - DIURNAL LOSSES
476 MDGV - RUNNING LOSSES
480 HEAVY DUTY GAS VEHICLES
481 HDGV - COLD TART
482 HDGV - HOT STABILIZED
483 HDGV - HOT START
484 HDGV - HOT SOAK
485 HDGV - DIURNAL LOSSES
486 HDGV - RUNNING LOSSES
500 DOMESTIC
510 RESIDENTIAL
520 RECREATIONAL
600 MISC. ACTIVITIES
610 CONSTRUCTION
611 BUILDING CONSTRUCTION
612 ROAD CONSTRUCTION
620 NATURAL SOURCES
630 GOVERNMENT
631 NATIONAL SECURITY
801 SEEPS/BIOGENIC
802 CHANNEL SHIPPING
803 OCS AND RELATED SOURCES
804 TIDELAND PLATFORMS
B-5
-------
TABLE B-2. Control codes.
000 UNSPECIFIED
101 UTILITY BOILERS - LIQUID FUELS
102 UTILITY BOILERS - GASEOUS FUELS
103 REFINERY BOILERS & HEATERS - LIQ. FUEL
104 RESIDENTIAL SPACE HEATERS - NATURAL GAS
105 RESIDENTIAL WATER HEATERS - NATURAL GAS
107 NON-UTILITY I. C. ENGINES - GASEOUS
108 UTILITY RECIPROCAL - LIQUIDS
109 INDUSTRIAL BOILERS
110 CEMENT KILNS
111 GLASS MELTING FURN CONTNR/SDLM
112 MARINE DIESEL ENGINES
113 NON-FARM EQUIPMENT (DIESEL)
114 SULFUR IN FUEL
116 UnLITY TURBINES - LIQUIDS
117 REFINERY BOILERS & HEATERS - GAS. FUEL
118 STEAM GENERATORS - LIQUIDS
121 PIPELINE HEATERS
122 MARINE VESSELS - COMBUSTION
124 UTILITY TURBINES - GASEOUS
125 COGENERATION
126 TEOR STEAM GENERATORS - GASEOUS
127 NON-UTIL I.C. ENGINES - LIQUID
128 RESOURCE RECOVERY
129 BOILERS-SPACE HEATERS-LIQ 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 & COIL - SURFACE COATING
307 CAN & COIL - SOLVENT USE
308 METAL PARTS & PROD. - SURFACE COATING
309 METAL PARTS & PROD. - SOLVENT USE
310 PAPER - SURFACE COATING
311 PAPER - SOLVENT USE
312 FABRIC - SURFACE COATING
313 FABRIC - SOLVENT USE
B-6
-------
EPS REPORTING CODES
TABLE B-2. Continued.
314 DEGREASING-NON-SYNTH&MISC-(IND)
315 DEGREASING-NON-SYNTH&MISC-(COMM)
316 CUTBACK ASPHALT PAVING MATERIALS
317 DRY CLEANING (NON-SYNTHETIC)
318 DRY CLEANING (SYNTHETIC&NGSQ
319 GRAPHIC ARTS-EXCPT LITHO/L PTRS
320 WOOD FURNITURE - SURFACE COATINGS
321 WOOD FURNITURE - SOLVENT USE
323 AUTO REFINISHING - SURFACE COATINGS
325 SHIPS - SURFACE COATING
326 SHIPS - SOLVENT USE
327 AEROSPACE - SURFACE COATING
328 AEROSPACE - SOLVENT USE
331 DECREASING SYNTHETIC (INDUS) . .
332 DECREASING SYNTHETIC (COMM)
333 FLATWOOD PRODUCTS
334 GRAPHIC ARTS - LlTHO/LTTR PRESS
398 OTHER INDUSTRIAL SURFACE COATING
399 UNSPECIFIED IND. SOLVENT USE
401 GASOLINE WORKING LOSS - BULK STORAGE
402 .. GASOLINE WORKING LOSS - TANK TRUCKS
403 GASOLINE WORKING LOSSES - UNDGRND TANK
404 GASOLINE WORKING LOSSES - VEHICLE TANK
405 FDCED ROOF TANKS AT REFINERIES
406 FLOATING ROOF TANKS AT REFINERIES
407 MARINE VESSEL OPERATION - EVAP.
410 OIL PRODUCTION FIELDS STORAGE TANKS
411 MARINE LIGHTERING
412 GASOLINE BREATHING LOSS - UNDG
413 GASOLINE BREATHING LOSS - ABOVEG
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 PROD.-VALVES,FLANGES,CONNECTORS
507 SMALL RELIEF VALVES
508 NON-REFINERY VALVES
510 VEGETABLE OIL PROCESSING
511 PAINT MANUFACTURING
512 RUBBER PRODUCTS FABRICATION '
B-7
-------
TABLE B-2. Continued.
513 CHEMICAL MANUFACTURING
514 PHARMACEUTICAL MANUFACTURING
515 RUBBER PRODUCTS MANUFACTURING
518 OIL PROD. STEAM DRIVE WELL
519 WINERIES
520 CARBON BLACK MANUFACTURING
522 PUMPS & COMPRESSORS
523 REFINERY SEWERS & DRAINS
524 REFINERY PUMPS & COMPRESSORS
526 REFINERY VACUUM SYSTEM
530 OIL PROD. PUMP AND COMPRESSORS
531 OIL PROD. HEAVY OIL TEST STATION
532 OIL PROD. CYCLIC WELL VENTS
533 OIL PROD. PSEUDO-CYCLIC WELL
534 OIL PRODUCTION SUMPS AND PITS
535 NATURAL GAS PLANT FUGITIVES
601 CONSTRUCTION & DEMOLITION
602 WASTE SOLVENT DISPOSAL
603 PESTICIDES (SYNTHETIC)
604 ROOFING TAR POTS
605 PESTICIDES (NON-SYNTHETIC)
606 AEROSOL PROPELLANT SYNTHETIC
607 AEROSOL PROPELLANT NON-SYNTHETIC
608 WASTE DISPOSAL LANDFILL
609 DOMESTIC SOLVENT USE
610 AEROSOL CONSUM PROD PROPELLANT
611 AEROSOL CONSUM PROD SOLVENT
612 NON-AEROSOL CONSUM PROD SOLVNT
620 AGRICULTURAL PESTIC - SYNTHETIC
621 AGRICULTURAL PESTIC - NON-SYNTH
622 OTHER PESTIC - SYNTHETIC
623 OTHER PESTIC - NON-SYNTH
651 UNPAVED CITY/COUNTY ROAD DUST
711 LDGV-EXHAUST
712 LDA - HOT START
713 LDA - HOT STABILIZED
714 LDGV - EVAPORATIVE
715 LDGV - RUNNING LOSSES
716 LDA - CRANKCASE BLOWSY
717 LDA - TIRE WEAR
718 LDGV - REFUELING
B-8
-------
TABLE B-2. Continued.
719 GASV - OFF-HIGHWAY EXHAUST
720 GASV - OFF-HIGHWAY EVAPORATIVE
721 LDT - COLD START
722 LDT - HOT START
723 LDT - HOT STABILIZED
724 LDT - HOT SOAK EVAP.
725 LDT - DIURNAL EVAP.
726 LDT - CRANKCASE BLOWBY
727 LDT - TIRE WEAR
731 MDGV-EXHAUST
732 MDT - HOT START
733 MDT - HOT STABILIZED
734 MDGV - EVAPORATIVE
735 MDGV - RUNNING LOSSES
736 MDT - CRANKCASE BLOWBY
737 MDT-TIRE WEAR ' '
738 MDGV - REFUELING
741 HDGV - EXHAUST
742 HDG - HOT START
743 HDG - HOT STABILIZED
744 HDGV - EVAPORATIVE
745 HDGV - RUNNING LOSSES
746 HDG - CRANKCASE BLOWBY
747 HDG - TIRE WEAR
748 HDGV - REFUELING
751 HDD - EXHAUST
753 HDD - HOT STABILIZED
757 HDD - TIRE WEAR
759 HDD - OFF-HIGHWAY EXHAUST
761 MCY - COLD START
762 MCY - HOT START
763 MCY - HOT STABILIZED
764 MCY - HOT SOAK EVAP.
765 MCY - DIURNAL EVAP.
766 MCY - CRANKCASE BLOWBY
767 MCY - TIRE WEAR
801 NON-FARM EQUIPMENT (GASOLINE)
802 FARM EQUIPMENT (DIESEL)
803 LAWN & GARDEN EQUIP (UTILITY)
804 OFF-ROAD MOTORCYCLES
805 PLEASURE CRAFT i BOATS)
B-9
-------
TABLE B-2. Continued.
806 RAILROAD LINE HAUL OPERATIONS
807 COMM./CIVIL PISTON AIRCRAFT
808 COMM. JET AIRCRAFT
809 FARM EQUIPMENT (GASOLINE)
811 LDA - NCAT - COLD START
812 LDA - NCAT - HOT START
813 LDA - NCAT - HOT STABILIZED
814 LDA - NCAT - HOT SOAK
815 LDA - NCAT - DIURNAL
816 LDA - NCAT - CRANKCASE
817 LDA - NCAT - TIREWEAR
821 LDA - CAT - COLD START
822 LDA - CAT - HOT START
823 LDA - CAT - HOT STABILIZED
824 LDA - CAT - HOT SOAK
825 LDA - CAT - DIURNAL
826 LDA - CAT - RUNNJNG LOSSES
827 LDA - CAT - TIREWEAR
831 LDA - DSL - COLD START
832 LDA - DSL - HOT START
833 LDA - DSL - HOT STABILIZED
837 LDA - DSL - TIREWEAR
841 LDT - NCAT - COLD START
842 LDT - NCAT - HOT START
843 LDT - NCAT - HOT STABILIZED
844 LDT - NCAT - HOT SOAK
845 LDT - NCAT - DIURNAL
846 LDT - NCAT - RUNNING LOSSES
847 LDT - NCAT - TIREWEAR
851 LMDT - CAT - COLD START
852. LMDT - CAT - HOT START
853 LMDT - CAT - HOT STABILIZED
854 LMDT - CAT - HOT SOAK
855 LMDT - CAT - DIURNAL
856 LMDT - CAT - RUNNING LOSSES
857 LMDT - CAT - TIREWEAR
861 LMDT - DSL - COLD START
862 LMDT - DSL - HOT START
863 LMDT - DSL - HOT STABILIZED
867 LMDT - DSL - TIREWEAR
873 HOT - NCAT - HOT STABILIZED
B-10
-------
I EPS REPORTING CODES
TABLE B-2. Concluded.
874 HOT - NCAT - HOT SOAK
875 HOT - NCAT - DIURNAL
876 HDT - NCAT - CRANKCASE
877 HDT - NCAT - TEREWEAR
881 HDT - CAT - COLD START
882 HDT - CAT - HOT START
883 HDT - CAT - HOT STABILIZED
884 HDT - CAT - HOT SOAK
885 HDT - CAT - DIURNAL
886 HDT - CAT - RUNNING LOSSES
887 HDT - CAT - TIREWEAR
891 SEEPS/BIOGENIC
892 CHANNEL SHIPPING
893 OCS AND RELATED SOURCES
894 TIDELAND PLATFORMS
901 FOREST MANAGEMENT CONTROL BURNING
902 WILD FIRES CONTROL BURNING
903 LIVESTOCK WASTE
999 MISC. CONTROL TACTICS
B-ll
-------
TABLE B-3. Process codes.
000 UNSPECIFIED PROCESSES
100 FUEL COMBUSTION
110 BOILERS & HEATERS
111 BOILERS
112 SPACE HEATERS
113 ORCHARD HEATERS
114 PROCESS HEATERS
120 IN-PROCESS FUEL
130 STATIONARY I.C. ENGINES
131 TURBINE - CMBSTN GASES
140 EQUIPMENT
141 UnLITY EQUIPMENT
142 MOBILE EQUIPMENT
200 WASTE BURNING
210 INCINERATION
211 CONICAL BURNER
220 OPEN BURNING
221 AGRI DEBRIS
222 RANGE IMPROVEMENT
223 FOREST MANAGEMENT
300 SOLVENT USE
310 DRY CLEANING
320 DECREASING
330 SURFACE COATING
340 ASPHALT PAVING
350 PRINTING
400 LIQUID STORAGE & TRANSFER
410 TANKS
420 TANK CARS & TRUCKS
430' MARINE VESSELS
440 VEHICLE REFUELING
500 INDUSTRIAL PROCESSES
510 CHEMICAL PROCESSES
520 FOOD & AGRICULTURAL
530 PETROLEUM & RELATED
540 MINERAL PROCESSES
550 METAL PROCESSES
551 PRIMARY METAL
552 SECONDARY METAL
553 METAL FABRICATION
560 1 WOOD & PAPER PROCESSES
B-12
-------
EPS REPORTING CODES
TABLE B-3. Concluded.
570 RUBBER & PLASTICS
600 MISC PROCESSES
610 PESTICIDE APPLICATION
620 SOLID WASTE LAND FILL
621 WASTE DISPOSAL
630 FARMING OPERATIONS
640 CONSTRUCTION & DEMOLITION
650 ROAD TRAVEL
651 UNPAVED ROAD
652 PAVED ROAD
660 UNPLANNED FIRES
661 WILD FIRES
662 STRUCTURAL FIRES
700 VEHICULAR SOURCES
710 ON-ROAD MOTOR VEHICLES
720 OFF-ROAD MOTOR VEHICLES
730 TRAINS
740 SHIPS
750 AIRCRAFT
801 SEEPS/BIOGENIC
802 CHANNEL SHIPPING
803 OCS AND RELATED SOURCES
804 TIDELAND PLATFORMS
B-13
-------
TABLE B-4. POD codes.
0 Combustion point and area sources
1 Solvent metal cleaning
2 Printing/publishing
3 Dry cleaning
4 Fixed roof tanks - crude
5 Fixed roof tanks - gasoline
6 EFR tanks - crude
7 EFR tanks - gasoline
8 Bulk terminals - splash fill
9 Bulk terminals - subm. bal.
10 Bulk terminals - not bal.
11 Stage I refueling controls
15 Ethylene oxide manufacturing
16 Phenol manufacturing
17 Terephthaiic acid manufacturing
18 Acrylonitrile manufacturing
19 SOCMI fugitives
20 Refinery fugitives
21 Cellulose acetate manufacturing
22 Styrene butadiene manufacturing
23 Propylene manufacturing
24 Polyethylene manufacturing
25 Ethylene manufacturing
26 Refinery wastewater treatment
27 Refinery vacuum distillation
28 Vegetable oil processing
29 Paint and varnish manufacture
30 Rubber tire manufacturing
31 Green tire spray
32 Carbon black
33 Automobile coating
34 Can coating
35 General surface coating
36 Paper coating
37 Miscellaneous surface coating
301 Surface coating - coils
302 Surface coating - large appliances
303 Surface coating - fabrics
304 _ Surface coating - magnet wire
305 Surface coating - misc. metal parts
38 Food/agricultural starch mfg.
B-14
-------
EPS REPORTING CODES
TABLE B-4. Continued.
39 Coke byproduct plants
43 Marine vessel loading
46 Charcoal manufacturing
47 Whiskey production
48 Plastic parts coating
49 Wood furniture coating
50 Utility - pulv coal
661 Utility - coal stokers
51 Utility boilers - oil
52 Utility boilers - gas
53 Utility boilers - other
54 Industrial inprocess fuel combustion
55 Industrial process heaters
56 Industrial ext comb - space heaters
57 Industrial ext comb - other
58 Comm/inst - pulv coal
663 Comm/inst - coal stokers
59 Comm/inst - oil
60 Comm/inst - gas
61 Comm/inst ext comb - other
63 Internal combustion - other
64 Solid waste disposal
70 Indl oil turbine
71 Indl oil recip
72 Indl gas turbine
73 Indl gas recip
74 Utility oil turbine
75 Utility oil recip
76 Utility gas tubine
77 Utility gas recip
78 Indl 1C - dist oil turbine, cogen
79 Indl 1C - dist oil engine, cogen
80 Indl 1C - nat gas turbine, cogen
81 Indl 1C - nat gas engine, cogen
33 Indl ext comb - coal, cogen
84 Indl coal boilers
85 Indl boilers - resid oil
86 Indl boilers - resid oil
87 Indl boilers - resid oil
662 Indl boilers - dist oil
88 Indl boilers - nat gas
B-15
-------
TABLE B-4. Concluded.
89 Indl boilers - nat gas
90 Indl boilers - nat gas
95 Aircraft coating
96 SOCMI reactors
97 SOCMI distillation
99 Unspecified
1040 Paper coating
1041 Degreasing
1042 Dry cleaning
1043 Printing
1044 Rubber and Plastics manufacture
1045 Misc. surface coating
1046 Automobile Refinishing
1047 Architectural surface coating
1048 Misc. industrial solvents
1049 Consumer solvents
1060 Light Duty Gasoline Vehicles
1061 Light Duty Gasoline Trucks
1062 Heavy Duty Gasoline Vehicles
1063 Heavy Duty Diesel Vehicles
1064 Off highway vehicles
1065 Railroads
1066 Open burning, forest fires, prescribed burns
1067 Area source incineration
1068 Aircraft and Marine Vessels
1070 Treatment, storage and disposal facilities
1071 Bakeries
1072 Cutback Asphalt
1073 Public Treatment works
1074 SOCMI fugitives
1075 Gasoline bulk terminals and plants
1076 Petroleum refinery fugitives
1077 Pharmaceutical manufacture
1078 Synthetic fiber manufacture
1079 Oil and natural gas production fields
1080 Service stations - Stage ! - truck unloading
1081 Service stations - Stage II vehicle refueling
1082 Gasoline refueling - spillage
1083 Gas serv. station underground tank breathing loss
B-16
-------
REPORTING CODES \
TABLE B-5. Speciation profile codes.
0000 Overall Average
0001 External Combustion Boiler - Residual Oil
0002 External Combustion Boiler - Distillate Oil
0003 External Combustion Boiler - Natural Gas
0004 External Combustion Boiler - Refinery Gas
0005 External Combustion Boiler - Coke Oven Gas
0007 Natural Gas Turbine
0008 Reciprocating Diesel Fuel Engine
0009 Reciprocating Distillate Oil Engine
0011 By-Product Coke Oven 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 Fug. Ems. - Covered Drainage/Sep. Pits
0035 - Refinery Fugitive Emissions - Cooling Towers
0039 Refinery Fug. Ems. - Compressor Seals Refinery Gas
0047 Refinery Fug. Ems. - Relief Valves, L.P.G.
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
0166 Printing Press - Letterpress Inking Process
0182 Printing Press - Gravure General Solvent
B-17
-------
TABLE B-5. Continued.
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 Solvent - 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 - Trichloroethylene
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
0321 Pump Seals - Composite
0332 Printing Press - Lithography Inking and Drying
0333 Lithography - Inking and Drying-Direct Fired Dryer
1001 Internal Combustion Engine - Natural Gas
B-18
-------
EPS REPORTING CODES
TABLE B-5. Continued.
1002 Chemical Menufacturing - Carbon Black Production
1003 Surface Coating Application - Solvent-base Paint
1004 Plastics Production - Polystyrene
1005 Plastics Production - Polyester Resins
1006 Phthalic Anhyd. - o-Xyl. Oxidation- Process Stream
1007 Mineral Products - Aspbaltic Concrete
1008 Rubber and Misc. Plastics Prod.- Styrene/Butadiene
1009 Plastics Prod.-Acrylontrl.-Butadiene-Styerne Resin
1010 Oil & Gas Production - Fugitives - Unclassified
1011 Oil & Gas Prod. - Fug. - Valves & Fit. -Liq. Serv.
1012 Oil & Gas Prod. - Fug. - Valves & Fit. -Gas Serv.
1013 Surface Coating Application - Water-base Paint
1014 Gasoline - Summer Blend
1015 Gasoline - Winter Blend
1016 Surface Coating -Thinning Solvents - Composite
1017 Surface Coating Application - Lacquer
1018 Surface Coating Application - Enamel
1019 Surface Coating Application - Primer
1020 Surface Coating Application - Adhesives
1021 Degreasing - Open Top - Chlorosolve
1022 Print./Publishing-Thin. Solv.-Met. Isobutyl Ketone
1023 Terepht.Acid/Dimet.Terepht.-Crys., Sep.,Drying Vat
1024 Terepht. Acid/Dimet.Terepht. - Distil. & Rec. Vent
1025 Terepht. Acid/Dimet.Terepht. - Prod. Transfer Vent
1026 Surface Coating - Thinning Solv. - Hexylene Glycol
1027 Ketone Production - Methyl Ethyl Ketone (MEK)
1028 Acetone - Light Ends Distillation Vent
1029 Acetone - Acetone Finishing Column
1030 Aldehydes Prod. - Formaldehyde - Absorber Vent
1031 Surface Coating - Thinning Solv. - Ethylene Oxide
1032 Aldehydes Prod. - Acrolein - Distillation System
1033 Aldehydes Prod. - Acrolein - Reactor Blowoff Gas
1034 Chloroprene -Butadiene Dryer
1035 Chloroprene - Clprene Stripper & Brine Stripper
1036 Secondary Aluminum - Pouring and Casting
1037 OrganohJgns.-Ethyiene DiCJ-Dir. Chlorin.-Dist. Ven
1038 Organohlgns. Prod.-Ethylene DiCl-Oxychlorination
1039 Organohlgns. Prod.-Ethylene DiCl-Caustic Scrubber
1040 Flourocarbons/Chlorofluorocarbons - General
1041 Flourocarbons/Qilorofluorocarbons - Dist. Column
1042 Flourocarbons/Chlorofluorocarbons- Fug. Ems.- Gen.
B-19
-------
TABLE B-5. Continued.
1043 Acrylic Acid - Quench Absorber
1044 Organic Acids Production - Formic Acid
1045 Organic Acids Prod.-Acetic Anhyd.-Dist. Column Yen
1046 Esters Production - Acrylates - Ethyl Acrylate
1047 Esters Production - Butyl Acrylate
1048 Cumene Prod. - Cumene Distillation System Vent.
1049 Cyclobexane - General
1050 Cyclohxnone/Cyclohxnol- Phenol Hydrogen.- Dist.Ven
1051 Vinyl Acetate - Inert Gas Purge Vent
1052 Vinyl Acetate - CO2 Purge Vent
1053 Vinyl Acetate - Inhibitor Mix Tank Discharge
1054 Vinyl Acetate - Refining Column Vent
1055 Organic Chemical Storage - Methylamyl Ketone
1056 Ethylene Oxide- O2 Oxidation Proc.-CO2 Purge Vent
1057 Ethylene Oxide- O2 Oxidation Proc.-Argon Purge Ven
1058 Ethylene Oxide - Stripper Purge Vent
1059 Met. Methacrylate - Hydrolysis, Light Ends, Dist.
1060 Met. Methacrylate (MMA) - Acid Dist. & MMA Purif.
1061 Nitrobenz. -Reactor & Sep. Vent- Wash. & Ntrl. Ven
1062 Benzene
1064 Olefins Prod. - Ethylene - Compressor Lube Oil Ven
1065 Propylene Oxide - Chlorohydronation Process - Gen.
1066 Styrene - General i
1067 Styrene - Benzene Recycle
1068 Styrene - Styrene Purification
1069 Organic Chemical Storage - N-Propyl Acetate
1070 Alcohols Production - Methanol - Purge Gas Vent
1071 Alcohols Production - Methanol - Distillation Vent
1072 Chlorobenzene - Tail Gas Scrubber
1073 Chlorobenzene - Benzene Drying Distillation
1074 Monochlorobenzene
1075 Chlorobenzene - Vacuum System Vent
1076 Chlorobenzene - Dichlorobenzene Crystallization
1077 Clbenzene - DiClbenzene Crystal Handling/Loading
1078 Railcar Clean.- Low Pres.High Vise (Ethyl. Glycol)
1079 Railcar Clean.- Low Pres,Med Vise (o-DiClbenz.)
1080 Railcar Clean.- Low Pres,High- Vise (Cresote)
1081 Tank Truck Clean.- Med Pres.Med Vise (MMA)
1082 Tank Truck Clean.- Low Pres,Low Vise (Phenol)
1083 Tank Truck Clean.-Low Pres,High Visc(Propyl. Glyc.
1084 Residential Wood Combustion (C1-C6)
B-20
-------
EPS REPORTING CODES
TABLE B-5. Continued.
1085 External Combustion Boiler - Coal-Slurry Fired
1086 Printing/Flexographic
1087 Organic Chemical Storage/i-Butyl i-Butyrate
1088 Surface Coating Operations - Adhesive Application
1089 Sec. Metal Prod. -Gray Iron Fndr. -Pouring/Casting
1090 Fluorocarbon Manufacturing - CF 12/11
1091 Plastics Prod. - Polyvinyl Chlorides & Copolymers
1092 Synthetic Organic Fiber Prod. - Nylon Batch Prod.
1093 Fluorocarbon Manufacturing - CF 23/22
1094 Paint Manufacture - Blending Kettle
1095 Textile Prod. - Gen. Fabric Oper. -Dyeing & Curing
1096 Textile Prod. - Gen. Fabric Oper. - Tenter Frame
1097 Aircraft Landing/Takeoff (LTO) - Military
1098 Aircraft Landing/Takeoff (LTO) - Commercial
1099 Aircraft Landing/Takeoff (LTO) - General Aviation
1100 Gasoline Refueling
1101 Light Duty Gasol ine Vehicles
1103 1-Pentene
1104 Acetaldehyde
1105 Acetic Acid
f
1106 Acetic Anhydride
1107 Acrolein
1108 Acrylic Acid
1-109 Acrylonitrile
1110 Adipic Acid
1111 Aniline
1112 Benzyl Chloride
1114 Butyl Aery late
1115 Butyl Carbitol
1116 Butyl Cellosolve
1118 Carbitol
1119 Carbon Tetrachloride
1120 Acetylene
1121 Chloroform
1122 Cresol
1123 Camene
1124 Cyclohexanol
1125 Cyclohexanone
1126 Cyclopentene
1127 Diethylene Glycol
1128 Diisopropyl Benzene
B-21
-------
TABLE B-5. Continued.
1129 Dipropylene Glycol
1130 Dodecene
1131 EpichJorohydrin
1132 Etbanolamines
1134 Ethyl Acrylate
1135 Ethyl Benzene
1136 Ethyl Ether
1137 Ethyl Mercaptan
1138 Ethyl Dibromide
1139 Ethyleneamines
1140 Formaldehyde
1141 Formic Acid
1142 Furfural
1144 Heptenes
1145 Isobutyraldehyde
1146 Isobutyl Acrylate
1147 Isobutyl Alcohol
1148 Isoprene
1149 Methanoi
1150 Methyl Acetate
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-Butyraldehyde
1163 N-Decane
1164 N-Dodecane
1165 o-Xylene
1166 Pentadecane
1167 Residential Wood Combustion
1168 Piperylene
1171 Propionaldehyde
1172 Propionic Acid
1173 Propylene Oxide
1174 p-Xylene
1175 Tert-Butyl Alcohol
1176 Toluene Diisocyanate
B-22
-------
TABLE B-5. Continued.
1178 Coal-Fired Boiler - Electric Generation
1185 Coal-Fired Boiler - Industrial
1186 Heavy-Duty Gasoline Trucks
1187 Citrus Coating
1188 Fermentation Processes
1189 Pulp and Paper Industry - Plywood veneer Dryer
1190 Gasoline Marketed
1191 Graphic Arts - Printing
1192 Degreasing
1193 Drycleaning
1194 Auto Body Repair
1195 Degreasing Composite
11% Drycleaning Composite
1197 Isooctane
1198 Pentane
1199 Isopentane
1200 Cyclopentane
1201 Light-Duty Diesel Vehicles
1202 Primary Aluminum Production
1203 Light-Duty Gasoline Vehicles - Exhaust Emissions
1204 Light-Duty Gasoline Vehicles - Evaporative Ems.
5001 Light-Duty Automobiles - Cold Start Emissions
5002 Light-Duty Automobiles - Hot Stabilized Emissions
5003 Light-Duty Automobiles - Hot Start Emissions
5004 Light-Duty Automobiles - Hot Soak Emissions
5005 Light-Duty Automobiles - Diurnal Losses
5006 Light-Duty Automobiles - Running Losses
5011 Light-Duty Truck - Cold Start Emissions
5012 Light-Duty Truck - Hot Stabilized Emissions
5013 Light-Duty Truck - Hot Start Emissions
5014 Light-Duty Truck - Hot Soak Emissions
5015 Light-Duty Truck - Diurnal Losses
5016 Light-Duty Truck - Running Losses
9001 External Combustion Boilers - Industrial - Average
9002 Internal Combustion - Average
9003 Industrial Processes - Average
9004 Chemical Manufacturing - Average
9005 Plastics Production - Average
9006 Synthetic Organic Fiber Production - Average
9007 Alcohols Production - Average
9008 Food and Agriculture - Average
B-23
-------
TABLE B-5. Concluded.
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 Prod. - 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 & Marketing of Petroleum Prod. -Avg
9028 Organic Chemical Storage - Average
9029 Org Chem Strg - Fixed Roof Tank - Alcohols -Avg
9030 Org Chem Strg - Fixed Roof Tank - Alkanes -Average
9031 Org Chem Strg - Fixed Roof Tank - Alkenes -Average
9032 Org Chem Strg - Fixed Roof Tank - Amines - Average
9033 Org Chem Strg - Fixed Roof Tank - Aromatics - Avg
9034 Org Chem Strg - Fixed Roof Tank - Carb. Acids -Avg
9035 Org Chem Strg"- Fixed Roof Tank - Esters - Average
9036 Org Chem Strg -Fixed Roof Tank -Glycol Ethers -Avg
9037 Org Chem Strg - Fixed Roof Tank - Glycols - Avg
9038 Org Chem Strg - Fixed Rf Tk -Halogenated Org - Avg
9039 Org Chem Strg - Fixed Roof Tank - Isocyanates -Avg
9040 Org Chem Strg - Fixed Roof Tank - Ketones - Avg
9041 Org Chem Strg - Float. Roof Tank - Aldehydes - Avg
9042 Org Chem Strg - Float. Roof Tank - Alkanes - Avg
9043 Org Chem Strg - Float. Roof Tank - Ethers - Avg
9044 Org Chem Strg - Float. Rf Tk -Halogenated Org- Avg
9046 Org Chem Strg - Pres Tanks - Alkenes - Average
9047 Org Solvent Evaporation - Miscellaneous - Average
B-24
-------
APPENDIX C:
EPS 2.0 TEMPORAL PROFILES
C-l
-------
Fraction
Fraction
Fraction
Fraction
ooooo
o
c
3
C
lO
n
T3
0
o
O
o
(ti
o
o
S"
O
£°
C
c
o
c
10
n
T>
CD
O
10
O
~r
u -^
o o
~\r
n
=fc
o
>(
3
c
3
c
c
tr
o
o
c
3
t_
C_
C
ID
n
o
O
0
o
<
n»
3 0 O O O C
D '-» k> 04 -£> b
D ,0 0 0 0 C
- 3
3
- 3
- 3
- 3
- 3
3
- 3
3
5
Mill
*
E
Jill
p p p p p p
b '-*
"o 9
-i -*
o
ST
;T* C
o-
(0
2
o
*
2
o
c
3
C_
C
lO
l/J
^3
O
o
Nov
Maximum
o
2 o
n>
n
j
- 3
- 3
_ 3
3
- 3
- 3
- 3
p o o o c
1 1 1 1
E
i
Jill
D
-o
-t
o
ST
*
P
^
.
-------
TEMPORAL PROFILE
/MONTHLY/
CM
01
O)
CN
CN
V
Profile #: 5
Maximum: 0.15
0.40
c
.2 0.30
S 0.20
U- 3
0.10
0.40
| 0.30'
0
2 0.2Q
0.10
o.og
Jc
0.50
0.40
.2 0.30
o
2 0.20
. 0.10
0.00
Jc
0.50
0.40
.2 0.30
u
£ 0.20
L_
0.10
°-°a
_
L ^~~~~"^
k J^ J
-
-
S* -
m Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Profile #: 6
,
r "^ *^~^-*
- , , , \
1
Maximum: 0,24
\^~ ?f
f
m Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Profile #: 7
.._
N.
N.
Maximum: 0.33
/
/
/
in .Feb Mar Apr May Jun Jul Aug. Sep Oct Nov Dec
Profile #: 8
-
1
~ ₯ ' * '
~ 1 1 1
sn Feb Mar Apr M
^~ .
I
av Ji
* .-,
jn J
ul A
Maximum' 0 1 2
_
_
-
* -__
^ ^
'
u.
O
C-4
-------
N
M
c
o
o
o
L.
u.
c
^o
u
0
l_
u.
c
_o
u
D
L.
U-
k-
_o
u
o
u
U_
itLivirurcAL rKUriLQ
/MONTHLY/
0.50
0.40
0.30
0.20
0.10
0.09]
0.50
0.40
0.30
0.20
0.10
o.oo3
Jc
0.50
0.40
0.30
0.20
0.10
0.00
J<
0.50
0.40
0.30
0.20
o.io
o.oo*
Profile #: 9
_
_
EPS Temporal Profiles
Maximum: 0.10
__
- -
* xn» ^^*"^^
k Jk
I I I I I I I
*- * ^
I I
£
Jn Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Profile #: 10
_^
y ^^-^ "
i "f I
I I I I
Maximum: 0.1 1
_
y
i ^
* -,
I I
e
Jn Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Profile #: 11
^_
-
, - * *- ^
I T >
!
.^* « *-
till
Maximum: 0.13
_
_
I |
*
«
=n - Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Profile #: 12
-
u
f ₯ T I
I I 1 1
Maximum' 0. 1 3
,
1 1^^
Jan
Feb
Mar
Apr May Jun Jul Aug Sep Oct
Nov Dec
C-5
-------
TEMPORAL PROFILE
/MONTHLY/
CN
CN
a
t>
Profile #: 13
Maximum: 0.15
c
o
o
o
V.
u_
c
^o
o
D
i_
U_
C
_o
o
o
L.
u_
_5
o
D
i_
U_
U.
0.
0.
0.
0.
0.
0.
0.
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
ou
40
30
20
10
°9!
JO
50
40
30
20
10
°9J
50
.40
.30
.20
.10
°9
.50
.40
.30
.20
.10
°9
_
^*^ \\_
c Jc ^r^"^ i i i i i i i > .)
n Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Profile #: 14
/* \
/ \
X ' ' ' \ i
Maximum: 0.33
-
.s
I X X J/
t
in Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Profile #: 16
^
_
( * * x x *
I I I I I
!
Maximum: 0.10
»
* * * *
I I ' I
3n Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Profile #: 17
-
-
£ * -K * 3* *-
I 1 I ' 1 1
Maximum- O.OS
1
-
K X "^ ₯ ~
I I I I
\
on Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
O
K
Q.
C-6
-------
p
O.SOr
0.40-
.1 0.30-
0
S 0.20-
0.10*
0.00-
w Q
f
0.50
0.40
I 0.30
S 0.20
u_
0.10
0.00
JC
0.50
0.40
.1 0.30
'S 0.20
0.10
0.09
0.5C
0.4C
I 0.3C
o
S 0.2C
u_
0.1C
n nr
TEMPORAL HKUhILL
/MONTHLY/
rofile f. 18
* * * * *~~ * * X "
1 | 1 1 1 1 ! 1
n Feb Mar Apr May Jun Jul Aug Sep
=rofile #: 19
EPS Temporal Profiles
Maximum: 0.10
I I
Oct Nov Dec
Maximum: 0.10
-
, I I I t I I _J 1
jn Feb Mar Apr May Jun Jul Aug Sep
Profile #: 20
.
-
I I I I I I I - -
i 1 1 > -
an Feb Mar . Apr May Jun Jul Aug Sep
Profile #: 21
.
-
)-
)-
Jt- "X 5K X X 5K
> I I I I 1 1 1 1
Oct Nov Dec
Maximum: 0.10
I
Oct Nov Dec
Maximum' 0. ' '
I I
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug Sep
Oct Nov Dec
C-7
-------
01
O
CN
CM
u
O
o
V
LL.
O
EC
Q.
TEMPORAL PROFILE
/MONTHLY/
Profile #: 22
0.50
0.40-
: 0.25
.1 0.30
"*"* i
S 0.20
u_
0.10
-*-
*-
an Feb Mar
Profile #: 23
Apr May
Jun
Jul
Aug Sep
Oct Nov Dec
Maximum: 0.10
0.50
0.40-
.2 0.30
S 0.20
u_
0.1
0 00' ' ' ' '
Jan Feb Mar Apr May Jun
Profile #: 24
Jul
Aug Sep
Oct Nov Dec
Maximum: 0 1 0
0.50
0.40h
.1 0.30
0.20
0.10
o.oo1 ^-..
Jan Feb Mar
Profile #: 25
Apr May
Jun
Jul
Aug Sep
Oct Nov Dec
Maximum1 0 1 C
0.50
0.40-
J 0.30-
o
P 0.20
0.10
O.OO1
Jon
Feb Mar Apr May Jun Jul Aug Sep
C-8
Oct
Nov Dec
-------
TEMPORAL PROFILE
/MONTHLY/
EPS Temporal Profiles
Profile #: 26
Maximum: 0.27
c
o
o
o
u
U.
U.DU
0.40
0.30
7
0.20
0.10
n nn
-
\ ~
\
" , , \ , * , ~
Jon Feb Mar Apr May Jun
Profile #: 27
Jul
Aug Sep
Oct Nov Dec
Maximum: 0.12
o
0
u
o
u.ou
0.40
0.30
0.20
0.1C)
0.00
Jc
-
- -
- -
*
' . x ₯ **"
1 1 I I I I I I I I
:n Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Profile #: 28 f Maximum: 0.10
u.ou
0.40
0.30
0.20
0 10
0.00
Jc
.
-
_ _
_ _
_£ XX X X ^
I I I I I I I I I !
in Feb Mar Apr May Jun Jul Aug Sep Oct Nov D£
C-9
-------
CM
a>
01
a>
O
CN
rsi
0
o
0
ui
c
_o
u
o
u.
|j
C
o
o
o
l_
u_
3
o
o
i_
u_
0.60
0.48
0.36
0.24
0.12
0.00
M<
0.60
0.48
0.36
0.24
0.12
°-°9
M
0.60
0.48
0.36
'0.24
0.12
°-°s
M
0.50
0.48
0.56
0.24
0.12
o.og
M<
ILIVII UKML ri\UI IL.L- r»c T ' '» 'n_ ^
'WrrKI " / EPS Temport Profiles
/ WLLrsL I /
Profile #: 1 Maximum: 0.20
« ' - -i-
_ .
1 1 I ^^^S: )
Jn T"« Wed Thu Fri Sat Sun
Profile #: 2 Maximum: 0.50
/
" * ./
c. -k -k k x
r
5n Tue Wed Thu Fri Sat Sun
Profile #: 3 Maximum: 0.20
V
»-*
_
i ' i i i ^^S-
Tue Wed Thu ' Fri Sat Sun
Profile #: 4 Maximum: 0.20
_
"~ .-
1 1 1 ^^N-
3n Tue W«d Thu Fri Sat J
C-ll
-------
PROFILE: Wed Ocl 7 22:09 -42 1992
Fraction , Fraction Fraction Fraction
oooooo oooooo oooooo oppopp
c
a
o
c
m
1
a
3-
C
?
I/)
a
in
3
D » K> O4 -£». C
D Kt, ^ 01 CO C
I
.
-
-
^
- 5
^
1
)
3
'III
(
[
__L I__L_.
J> C
D a
"0 °
-^ 3
§,
n
^%=
OJ
c
m
(D
a
H
C
?
M
S °
O ^
X
3
c
3
o
c
3
D '-» to OJ -^ t
D M 4^ 01 CO C
1
5
- 3
l;
k 1 1 1
< *
1 I J
Jl C
D a
"D °
-^ D
O
5"
=%:
vj
C
a
-H
3"
C
5
s: °
Q ^
X
3
c
3
o
^ 2 5
D-»KJCJ-f!>.cn o-'tooi^cn
D K) > Ol CO O 2O NJ ^-^ O) CO O
1
- )
- J
;
1 II
*
/
/
/
/
2,
r»
=?fc
Ol
c
n
1
a
3-
c
3
Q **
X
3
c
3
0
^ c 3
3
J "
3
- '
- }
-1 1 1
I
1 1 1 1
o
S"
=*:
m
P
r~ i
r
n
Q
X
3
c
3
o
k>
o
-------
C
O
'f
u
o
o
o
u
D
0.60
0.48
0.36
0.24
0 12
I L-IVII UUML I I\UI ILL. _.. _ ' -- _,
'Wrri'l V / EPS Temporal ProJUes
/WLLlXLY/
Profile #: 21 Maximum: 0.22
,
II II I
*
' ^°n Tue Wed Thu Fri Sat Sun
0.60
0.48
0.36
0.24
0.12
M
0.60
0.48
0.36
0.24
0.12'
n on
Profile #: 22 Maximum: 0.16
- _
- _
I I I I 1 ^
d
Tue Wed Thu Fri Sat Sun
Profile #: 23 Maximum: 0.15
1
' -
- -
"" * ~" . i i
1 1 1 1 I
don
TUS
Wed
Sat
Sun
C-13
-------
tN
01
01
D
u
S
ii
/DIURNAL WEEKDAY/
x> J
JTt'UJUd*
Profile #: 6
Maximum: 0.17
0.40
0.32-
.2 0.24
-r->
u
S 0.16
0.08
-*-
-**-
0.00* * * * jc
I
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Maximum: 0.14
Profile #: 7
0.40
I
*->
0.32-
0.24-
0.16-
0.08-
o oo«*****«~^
w ^ 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Profilers Maximum: 012
0.40
c
_o
lw
u
ul
0.32-
0.24-
0.16-
0.08-
Q.00^_t.4
Profile #: 9
15 16 17 18 19 20 21 22 23 2-
Maximum: 0.1 1
C
o
"^
u
0.40
0.32
0.24
0.08
o.oa
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
C-15
-------
/DIURNAL WEEKDAY/
Profile #: 10
Maximum: 0.10
0.40-
0.32-
.2 0.24-
o
2 0.16-
U-
0.08-
0.00*
F
0.40
0.32
.2 0.24
E 0.16
U-
0.08
O.OQ
0.40
0.32
.2 0.24
u
S 0.16
0.08
o.oo
0.4C
3 0.32
c
£ .2 0.24
* £ o.ie
(N u_
0.0£
° O.OC
u
lii
234567
'rofile #: 11
/
i i i i I i I i L_Na. >L
8 9 10 11 12 13 14 15 16 17 18 19 20 21
-
* *
22 23 24
Maximum: 0.09
/
234567
Profile #: 12
.
-
1 23456
Profile #: 13
i
[-
5-
T 2 3" 4 5 6
iZ
O
or
0.
/: : : ; : : \
-
-
8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Maximum: 0.08
*
1
7
, t i I i I t I i I i ^
& *
8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 2-
Maximum: 0.08
/ A J- -u J- a ' ' T
. ./I i 1 i 1 i i 1 1
7 8 9 10 11 12 13 14 15 '16 17 18
C-16
I i
-i
i
\ * '
-H
1
19 20 21 22 23 24
-------
TEMPORAL PROFILE
WEEKDAY/
Profile #: 14
PBV »_ ^=»
EPS Temporal ProJHes
Maximum: 0.07
o
D
a_
u.tu
0.32
0.24
0.16
0.08
n or*
-
-
i i i i i i i i i i i i i \j
-
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Profile #: 15 Maximum: 0.07
c
^o
o
0
i_
u_
u.tw
0.32
0.24
0.16
0.08
o.or>
- _
-
_
.
^ x x- x- x- i: x x I i I i I i I i I i I i I i \
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Profile #: 16 ' Maximum: 0.06
.2
u
u.tu
0.32
0.24
0.16
0.08
o.oo
I.. . I I I I I ill ill!
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Profile #: 17 Maximum- 0.06
w.tw
0.32
.2 0.24
u
E 0.16
0.08
O.OO
_
Jf. %. 5K 5K 3K- XXXXXXXXXXX~
^ ^ 4^ ^j ± ^ J^-J . _l I . 1 i 1 i 1 i I i i !
1 2 3 4 5 6 7 8 9 10 11 12 13
14 15 16"~17 18 19 20 .21 22 23 24
C-17
-------
04
cn
01
01
O
(N
r>.
.TEMPORAL PROFILE
/DIURNAL WEEKDAY/
Profile #: 18
Maximum: 0.06
c
.2
u
0
h
c
0
o
ll
c
.2
o
D
u_
.2
u
o
l_
ll
0.32
0.24
0.16
0.08
0 OO
0 40
0.32
0.24
0.16
0.08
0 OO
0 40
0.32
0.24
0.16
.0.08
0 OO
0.32
0.24
0.16
0.08
O.OO
.
-
CX X X X K I i 1 i I i 1 i 1 i I i I I I i
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 2
Profile #: 19 Maximum: (
^
-
'--t Y ' t i i i t i i i i i i i \j
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 2
Profile #: 20 Maximum: C
_
-
-
^ ^ ^ -1 1 1 1 1 1 1 1 1 1 1 1 1 ! 1 1
2 3 4 5 5 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 2
Profile #: 21 Maximum: C
',
< ^± i i i i i i i i T- i i . i i i , i ,
3 2
105
_
_
3 2
105
V,
3 2'
105
_
_
X
7 8
9 10 11 12 13 14 15 16 17 .18 19 20 21 22 23 24
o
ce
a.
C-18
-------
TEMPORAL PROFILE
/DIURNAL WEEKDAY/
Profile #: 22
i «_ -,
EPS Temporal Profiles
Maximum: 0.05
o
o
u.t-u
0.32
0.24
0.16
0.08
n on
-
-
-
/i 1 i 1 i I i 1 i 1 i 1 i 1 i 1 i 1 i 1 t l\
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Profile #: 23 Maximum: 0.04
o
o
u.tu
0.32
0.24
0.16
0.08
n nn
- _
- -
- _
Xl 1 1 1 1 1 1 1 1 1 I 1 I 1 1 1 I 1 I 1 I 1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Profile #: 24 Maximum: .04
u
0
u.tv
0.32
0.24
0.16
0.08
o.oo'
-
-
- - _
I i L . L . .. j iTiTi?i?i'"i~l7'T
Profile #: 26
11 12 13 14 15 16 17 18 19 20 21 22 23 24
Maximum' 0 1 3
u
D
-------
q
(M
3
UJ
.
|
TEMPORAL PROFILE
/DIURNAL WEEKDAY/
Profile #: 27
Maximum: 0.05
0.40
0.32
c
.2 0.24
o
2 0.16
0.08?
0.00
, i i I i n * n * JC T X * 'C -i i ^ i 1 ' 1 1
Profile #: 31 ' Maximum: 0.09
0.40
0.32k
I 0.24
*s
o
S 0.16
0.08
o.oo5
_L
1 - 1
'1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Profile #: 33 Maximum: 0.08
0.40
0.32
I 0.24
+~i
U
2 0.16
u_
0.0.8
o.ocy
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Profile #: 34 Maximum: 0.14
0.40
0.32
2 o.24
0
2 0.16
0.08
n nrfc
.
"I
~~j
^, w w v^*^* ^~ ^~ \
. ^
/ i- i i I , i i I V .1 J. J, J. i * ! -I. 1 * J- 1 » )
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
C-20
-------
F
0.40-
0.32-
.1 0.24-
S 0.16-
h-
0.08
H
0.00.
I
0.40
0.32
| 0.24
E 0.16
u.
0.08
O.OQ
0.40
0.32
° 0.24
u
S 0.16
0.08
0.00
0.40
0.32
.2 0.24
o
S 0.16
0.08
n nr
ILMHUKAL hKUMLL EPS Tfatporal Profiles
/DIURNAL WEEKDAi/
>rofile #: 35 Maximum: 0.05
,
\, i ^ i t ^-^ 3"^ i 1 i 1 i 1 i .I .__ i 1 1 1
2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 21 22 23 24
Profile 0: 37 Maximum: 0.08
_ -
_ -
*
^ j. j. j( _^ y^T i . i i i i i > i i i^ 4 >
2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 21 22 23 24
Profile #: 38 Maximum: 0.15
| ~~
. ~
A*-»-_ "
". 1^X^V-^-*^r^MS-^^^l ^*^\J
*
e
1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17f 18 19 20 21 22 23 24
Profile #: 40 Maximum: 0. 1C
-
_r jltf~"^ ^^W IK
w -* -^.,.^-L i i i i i i i i i i 1^~^-4 * & * &
1
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
C-21
-------
PROFILE: Wed Oct 7 22 09 -to 1992
o
to
Fraction
o o o o
a
x
3
C
3
o
00
Fraction
p p p p
-» M Lrl -f^
0) -&. Ki O
Fraction
Fraction
ro
tn
_*
O)
NJ
K
NJ
J__J L
T)
(^
ST
=te
a
x
3
c
3
p
o
o
U
0>Sf-
ooje-
CJ
NJ
(0
p p
b '-*
00 O)
pop
fo GJ ^
^ ro o
1
1
I I L
K>
o o ->
CD O)
0135-
0)
K) .,
f. s
C M j|
3 N
p K
xl K» ^
O O
01 i>.
N) O
35-
3f-
I I I
a
x
3'
C
3
p
Ui
(O
"0
o
>
m TJ
m TO
o -r\
-------
TEMPORAL PROFILE
/DIURNAL WEEKDAY/
JEPS Temporal Profiles
Profile jjf: 48
Maximum: 0.10
0.40
0.32
.2 0.24
o
£ 0.16
u_
0.08
o.od
^>«-
-* *
f .1. v t .Jy y fanfgjf _ ^r^ i I 1 I - - _ ' ' 1..I i I 1.
T 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
_ ... _- Maximum: 0.10
Profile #: 50
0.40
o
o
0.32-
0.24-
0.16-
0.08-
o.od
Ti ir^- -J*T-
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Profile #: 51 Maximum: 0.17
(J
o
0.40
0.32
0.24
0.16
0.08
O.OO
-
: .r^\
^ j. .,. .j. /i i i ' H- t -4 * ^ ^ ± ± ± ± * »
I 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 2
Dr«f;io U- ^ Maximum: 0.12
0.40
0.32-
I
0.24-
0.16-
0.08-
0.00*-
1 i I i I 1 1 L
_L
J_
2 3. 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
C-23
-------
TEMPORAL PROFILE
/DIURNAL WEEKDAY/
Profile #: 53
Maximum: 0.18
0.40-
0.32-
.2 0.24-
"o
S 0.16-
L.
0.08-
O.Ooi
f
0.40
0.32
1 0.24
S 0.16
U-
0.08
O.OQ
0.40
0.32
.2 0.24
u
2 0.16
_ 0.08
0.00
0.4C
Si 0.32
01
v. ,2 0.2-
en U
° 2 O.U
CM U_
" O.OS
° o.oc
o
UJ
:
xb X J. X . .g. ^
5* * 5* 5s 5* 5ft
23456789
:>rofile #: 54
.L .L ^ X a.
i 1
10 11
\
\
A,
* ^
12 13 14 15 16 17 18 19
,
20 21
-
-5k 5k 5
22 23 24
Maximum: 0.07
.
23456789
Profile |: 55
_
-
1 23456
Profile #: 56
I
.f-
k
5- ~ ^
>, j, X si/ I i
)fe 5fe * 5(f ' '
T 2 3 4 5 6
(T
o
a:
O.
X
i !
10 1
L Jx ^
7 8 9 10 1
i 1
1
- J
I
i 1
1
1 12 13 14 15 16 17 18 19 20 21
t
-
~
,\ ,
22 23 24
Maximum: 0.10
1 12 13 14 15 16
L j, J
.
I ± X
/, I
17 18 19 20 21 22 23 24
Maximum' 0 06
J
. y Mx^^^MXN^V^-^ -SI/ -^ W Xt
7 8
I i _
I i
I i
1 i
9 10 11 12 13 14 15 16
I i I i\l
17 18 1
k * -Jc
9 20 21 22 23 24
C-24
-------
Profile #: 50
0.40
0.32
| 0.24
o
2. 0.16
0.08
TEMPORAL PROFILE
/DIURNAL WEEKEND/
EPS Temporal Profiles
Maximum: 0.10
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
C-25
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
454/R-92-026
4. TITLE AND SUBTITLE procedures for the Preparation of
Emission Inventories for Carbon Monoxide and Precursors
of Ozone, Volume II: Emission Inventory Requirements
for Photochemical Air Quality Simulation Models
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
December 1992
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORGANIZATION NAME AND ADDRESS
U.'S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Technical Support Division (MD-14)
Research Triangle Park, NC 27711
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68D00102
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
EIB/TSD/OAQPS
IS. SUPPLEMENTARY NOTES
Project Officer - Keith A. Baugues
16. ABSTRACT ' ~~~ ' 1
This is a companion document to Volume I, which describes procedures for compiling
the annual countywide inventory of volatile organic compounds (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.
*
This is the second revision of this volume, which updates the versions released
in 1979 (EPA-450/4-79-018) and 1991 (EPA-450/4-91-014).
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS [c~ COSATI Field/Grouo
Emissions inventory
Gridding
Hydrocarbons
Nitrogen Oxides
Photochemical Models
Spatial Resolution
Temporal Resolution
Species Resolution
Volatile Organics
18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (This Report)
21. NO. Of PAGES
242
20. SECURITY CLASS (This page)
22. PRICE
EPA Form 2220-1 (R«». 4-77) PREVIOUS EDITION is OBSOLETE
------- |