United States
Environmental Protection
Agency
Air Pollution Training Institute (APTI)
Mail Drop E14301
Research Triangle Park, NC 27711
September 2004
/a Š Preparation of Fine
f | 1
Particulate Emission
Inventories
Student Manual
APTI Course 419B
Developed by
ICES Ltd.
EPA Contract No. 68D99022
-------
Notice
This is not an official policy and standards document. The opinions
and selections are those of the authors and not necessarily those of the
Environmental Protection Agency. Every attempt has been made to
represent the present state of the art as well as subject areas still under
evaluation. Any mention of products or organizations does not constitute
endorsement or recommendation by the United States Environmental
Protection Agency.
This project has been funded wholly or in part by the United States
Environmental Protection Agency under Contract No. 68D99022 to ICES, Ltd.
Availability
This document is issued by the Air Pollution Training Institute, Education
and Outreach Group, Office of Air Quality Planning and Standards, USEPA.
This workbook was developed for use in training courses presented by the
U.S. EPA Air Pollution Training Institute and others receiving contractual or
grant support from the Institute. Other organizations are welcome to use
the document.
This publication is available, free of charge, to schools or governmental
air pollution control agencies intending to conduct a training course on the
subject matter. Submit a written request to the Air Pollution Training
Institute, USEPA, Mail Drop E14301, Research Triangle Park, NC 27711.
Sets of slides designed for use in the training course of which this
publication is a part may be borrowed from the Air Pollution Training
Institute upon written request. The slides may be freely copied. Some films
may be copied; others must be purchased from the commercial distributor.
This consent does not extend to copying or transmission by any means
- graphic or electronic - for any other purpose, such as for advertising or
promotional purposes, for creating a new collective work, for resale, or for
information storage and retrieval systems.
State and local air pollution control agencies, U.S. EPA offices, and
federal offices designated by U.S. EPA are authorized to make copies of
this document in connection with telecourses.
Printed on recycled paper in the United States of America.
-------
Course Description
APTI 419B: Preparation of Fine Particulate Emission Inventories is a two-
day, resident instructional course designed to present an advanced view of all
major, practical aspects of developing an emission inventory for fine particulate
matter. The course is intended primarily for employees that have a working
knowledge of emission inventory terminology and techniques. The course
focuses on the principal stationary nonpoint area and nonroad mobile source
categories emitting PM fine particles. For select categories, the course provides
a brief summary of how emissions are estimated for EPA's National Emissions
Inventory (NEI), and how state/local/tribal agencies can improve upon those
estimates. Case studies are used to provide real-world examples of how state or
lo9cal agencies collected their own data to prepare inventories that are
improvement to the NEI methods. The lessons include information on an
overview of fine PM, an overview of the NEI, onroad mobile inventory
development, onroad mobile inventory development, point source inventory
development, area sources, fugitive dust area sources, combustion area
sources, and other related topics.
The course is taught at an instructional level equivalent to that of an
advanced, undergraduate university course. The Air Pollution Training Institute
curriculum recommends APTI 419B: Preparation of Fine Particulate Emission
Inventories as an advanced course for all areas of study. The student should
have minimally completed a college-level education and APTI Course SL419A-
Introduction to Emission Inventories or have a minimum of six months of
applicable work experience.
How to Use This Manual
This manual is to be used during classroom instruction and telecourse
sessions. The workbook contains instructional objectives and materials for each
of the nine subject areas.
Each chapter provides a lesson goal, instructional objectives, subject
narrative, and reference materials that may guide your study. A separate
Student Workbook also contains a reproduction of lecture slides intended to
guide your notetaking. The slides are presented to generally follow the
course outline; however, the instructor may on occasion vary the order of
presentation or present material not included in the workbook. Each
student, therefore, should take thorough notes of the lecture content
throughout the course, but not rely solely upon graphic reproductions for the
course content.
-------
DISCLAIMER
This document does not constitute U.S. Environmental Protection Agency policy.
Mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
iv
-------
TABLE OF CONTENTS
Table of Contents v
List of Figures vi
List of Tablesvii
Chapter 1: PM2.5 Overview 1-1
Chapter 2: The National Emissions Inventory and Emissions Inventory Tools 2-1
Chapter 3: Onroad Mobile Inventory Development 3-1
Chapter 4: Nonroad Mobile Inventory Development 4-1
Chapter 5: Point Source Inventory Development 5-1
Chapter 6: Area Sources 6-1
Chapter 7: Fugitive Dust Area Sources 7-1
Chapter 8: Ammonia Emissions from Animal Husbandry 8-1
Chapter 9: Combustion Area Sources 9-1
V
-------
LIST OF FIGURES
Figure 1-1 PM2.5 in Ambient Air 1-2
Figure 1 -2 MSA to Non-MSA Comparison of PM Emissions 1-3
Figure 1-3 Example of "Urban Excess" 1-4
Figure 1-4 Comparison of Urban -Rural Ratios 1-5
Figure 1-5 NOx National Emissions 1-6
Figure 1-6 SO2 National Emissions 1-6
Figure 1-7 NH3 National Emissions 1-7
Figure 1-8 Primary Carbon in PM2.5 1-9
Figure 1-9 Characteristics of Primary Carbon 1-10
Figure 1-10 Primary Carbon Emission Density Ratios 1-11
Figure 1-11 Comparison of Emission Density Ratios 1-12
Figure 1 -12 Summary of Important PM2 5 Source Categories 1-13
Figure 1-13 Summary of PM2.5 Primary Emission Sources 1-14
Figure 2-1 Evolution of the NEI 2-2
Figure 6-1 Fugitive Dust Emissions in VISTAS States 6-5
Figure 6-2 NH3 National Emissions 6-5
Figure 6-3 Fugitive Dust Emissions in VISTAS States 6-8
Figure 6-4 Urban Annual Averages 6-8
Figure 6-5 Capture Fraction Conceptual Model 6-10
Figure 7-1 Example Crop Calendar for Corn 7-5
Figure 8-1 NH3 - Precursor to Ammonium Sulfate and Nitrate 8-2
Figure 8-2 NH3 - Example Manure Management Train 8-5
vi
-------
LIST OF TABLES
Table 3-1 Onroad Sources 3-5
Table 4-1 Engine Types Included in the NONROAD Model 4-3
Table 4-2 Equipment Categories Included in the NONROAD Model 4-3
Table 4-3 Aircraft SCC 4-6
Table 4-4 LTO-Based PM Emission Factors 4-7
Table 4-5 Commercial Marine Vessel SCC 4-9
Table 4-6 Small Commercial Marine Vessel PM10 Emission Factors 4-11
Table 4-7 Large Commercial Marine Vessel PM10 Emission Factors 4-11
Table 4-8 Locomotive SCC 4-12
Table 4-9 NEI Locomotive PM Emission Factors 4-12
Table 5-1 Reading List 5-7
Table 6-1 Web Address for References Cited in this Chapter 6-2
Table 6-2 Key Chapters of Volume III of the EIIP Area Source Guidance for
Sources of PM Emissions 6-3
Table 6-3 Typical Nonpoint Source Categories 6-4
Table 6-4 Capture Fraction Estimates 6-10
Table 6-5 Example Capture Fraction Calculations 6-11
Table 7-1 NEI Silt Content Values 7-3
Table 7-2 Number of Tillings in NEI 7-3
Table 7-3 Land Preparation Emission Factors 7-6
Table 7-4 Harvest Emission Factors 7-6
Table 7-5 NEI Default Emission Factor Input Values 7-11
Table 7-6 SCCs for Construction 7-12
Table 7-7 Relationship Between Housing Units and Residential Housing
Structures 7-13
Table 7-8 Assumed Values for Residential Construction 7-14
Table 7-9 NEI PM10 Residential Construction Emission Factors 7-14
Table 7-10 Road Construction Conversion Factors 7-18
Table 8-1 Overview of New Estimation Methodology 8-3
Table 8-2 NH3 - Comparison of'99 and'02 NEIs 8-7
Table 9-1 Sample Frame 9-4
Table 9-2 Sample Frame 9-5
Table 9-3 NEI SCCs for Residential Wood Combustion 9-6
Table 9-4 Climate Zones 9-8
Table 9-5 Urban/Rural Apportionment Data 9-8
Table 9-6 Apportionment for Woodstoves and Fireplaces with Inserts 9-9
Table 9-7 Apportionment for Woodstoves and Fireplaces with Inserts 9-10
Table 9-8 Residential Opening Burning SCCs and Pollutants 9-11
Table 9-9 Vegetation Adjustment Values 9-12
Table 9-10 Fuel Loading Factors 9-14
vii
-------
THIS PAGE INTENTIONALLY LEFT BLANK
viii
-------
Chapter 1: PM 2.s Overview
LESSON GOAL
Demonstrate, through successful completion of the chapter review exercises, a
general understanding of the composition of fine particulate matter in the atmosphere;
how the components of fine particulate matter are formed; and the types of sources
that contribute to the formation of fine particulate matter.
STUDENT OBJECTIVES
When you have mastered the material in this chapter, you should be able to:
1. Explain the difference between primary and secondary particles.
2. Explain the geographical differences in PM2.5 concentrations.
3. Define the term "urban excess."
4. Describe the sources that emit precursors to the formation of secondary particles.
5. Explain the difference between crustal and carbon emissions.
6. Describe possible control strategies for reducing fine particulate matter
concentrations.
Preparation of Fine Particulate Emissions Inventories
1-1
-------
Chapter 1: PM 2JS Overview
1.1 PM COMPOSITION
In learning about particulate matter it is important to understand the difference
between directly emitted or primary particles and secondary particles that are formed
in the atmosphere from precursor gases. The distinction is graphically depicted in
Figure 1-1. Primary particles consist mostly of elemental carbon (EC) and primary
organic aerosol (POA) but will also contain crustal matter and a few other materials.
Secondary particles consist of secondary organic aerosol (SOA) formed from volatile
organic compounds, ammonium sulfate formed from SO2 and ammonia gases, and
ammonium nitrate, formed from NOx and ammonia gases. The term total
carbonaceous matter is used to describe the combined mass of EC, POA and SOA.
Figure 1-1. PM2.5UI Ambient Air
Primary Particles
(Directly Emitted)
Secondary Particles
(From Precursor Gases)
1.1.1 Urban Sites
A review of data from EPA's urban speciation trends network shows that particulate
matter in the eastern half of the United States is very homogenous in terms of
composition. Another feature of Eastern sites is that the PM is comprised of mainly
carbonaceous aerosol and ammonium sulfate in roughly comparable amounts. It is
also important to note that the data shows the crustal component of PM2.5 is very
Preparation of Fine Particulate Emissions Inventories
1-2
-------
small in both Western and Eastern urban monitoring sites, with the exception of a few
places in the southwest and the central valley of California.
1.1.2 Urban and Rural Comparisons
Figure 1-2 represents the magnitude of the emissions of primary PM and the various
precursors throughout the 37 state eastern and central United States. As this figure
shows, about half of the PM primary is emitted in the Metropolitan Statistical Areas
(MSAs) and about half in the rural areas. This figure also shows that ammonia is the
only precursor with larger emissions in the rural areas than in the urban areas. This is
due to the large contribution of agriculture to ammonia emissions. However, it
should be noted that there is still some ammonia in the urban areas because of
agriculture in the urban areas, and mobile sources.
Figure 1-2. MSA to Non-MSA Comparison of PM Emissions
100%
80%
60%
40%
20%
0%
~ MSA Counties
~ Non-MSA Counties
PM25-PRI S02 NOX NH3 VOC
An examination of ambient monitoring data from both urban and rural sites in the
speciation trends network shows that there is more sulfate than carbon in the non-
urban sites. Sulfate concentrations are only slightly higher in the urban areas than in
the surrounding non-urban areas; however, carbon concentrations do increase
substantially in the urban areas. The conclusion from this monitoring data is that
sulfate is very much a regional problem. Carbon, on the other hand, does have a have
a regional component, there is a significant excess of carbon in the urban areas, as
evidenced by the marked increase in carbon from rural to urban areas. Urban air
quality data is often compared to rural air quality data by noting the amount of "urban
excess" for a particular component.
This concept is illustrated using data from the Atlanta area. As shown in Figure 1-3,
almost all of the sulfate is associated with the regional contribution. In other words
the sulfate that you find in Atlanta is only 10-15% higher in concentration than the
Preparation of Fine Particulate Emissions Inventories
1-3
-------
sulfate that you find in the surrounding rural sites. The ammonium concentrations
show the same pattern as the sulfate concentrations because most of the ammonium is
associated with sulfate. The data also show that the nitrate and carbon concentrations
are about twice as high in the urban areas as they are in the rural areas, a significant
"excess". The top part of the bars in Figure 1-3 shows this "urban excess." It is also
important to note that the concentration of total carbonaceous material is greater than
the sulfate concentrations in Atlanta and that the concentration of crustal material is
very small.
Figure 1-3. Example of "Urban Excess"
Atlanta, GA / Ring of 5 Rural Locations
Gray: Regional Contribution
Black: Urban Excess
H
ii
Sulfate Est. Ammonium Nitrate TCM Crustal
Figure 1-4 shows another comparison of urban and rural information by comparing
emission densities with ambient concentrations. The density of NOx emissions per
square mile are about four times higher in urban areas than they are in the rural areas and
the concentrations of nitrate are only about twice as high in the urban areas as in the rural
areas. This suggests that the higher concentration of ammonium nitrate in urban areas is
associated with the higher NOx emissions in the urban areas. Sulfate has a much higher
density of emissions in the urban areas, but this ratio is not reflected in the ambient data.
As seen in the Atlanta example, there is virtually no urban excess of sulfate there. This
lack of urban excess sulfate is found throughout the East.
Preparation of Fine Particulate Emissions Inventories
1-4
-------
Figure 1-4. Comparison of Urban -Rural Ratios
~ Emissions Density
~ Ambient Levels
NOx Nitrate S02 Sulfate
The reason for these differences between the urban excesses for nitrates and sulfates is
due at least in part to the following. The NOx to nitrate reaction occurs fairly quickly and
much of the transformation occurs before it gets transported very far. Also, nitrate is a
little less stable and may revert to other compounds during transport. Sulfate, on the
other hand, has a very long lifetime. Once it is converted from SO2 to sulfate it stays
around, sometimes for weeks, as a sulfate particle and can be transported long distances.
So, even though the emission density of SO2 is much higher in the urban areas than it is
in the rural areas, the concentrations are fairly uniform over broad geographic areas. As a
result, sulfate is considered a regional pollutant in terms of the impact on PM2.5.
1.2 PM2.5 SOURCE CATEGORIES
1.2.1 NOx Emissions
National data indicates that NOx emissions are about 23 million tons a year. Figure 1-5
shows that about 35% of those emissions are from highway vehicles, twenty five percent
are from electric utilities, eighteen percent are from mobile sources, and fifteen percent
are from industrial and commercial fuel combustion. All of these NOx emission sources
are associated with fuel combustion with the exception of the "Other" category, which is
mostly emissions from industrial processes.
re
1-
3
QŁ
E
rtJ
-G
o
o
re
DC
Preparation of Fine Particulate Emissions Inventories
1-5
-------
Figure 1-5. NOx National Emissions
Highway Vehicles
lilecirie Utilities
Off road Mobile
Ind. & Comm Fuel Comb.
Other
0% 5% 10% 15% 20% 25% 30% 35% 40%
1.2.2 S02 Emissions
Figure 1-6 shows the source categories that contribute to the national SO2 emissions.
Electric utilities are responsible for about 70-75% of the emissions of SO2. As stated
previously, even though emissions from sources such as electric utilities tend to be
concentrated more in the urban areas where the people live, the impacts of sulfate stretch
across large geographic areas, due to the long lifetime of sulfate particles and their ability
to transport long distances.
Figure 1-6. SO2 National Emissions
lilectric LJliililies
1
Other Fuel Comb.
Mobile Sources
Other []
0% 10% 20% 30% 40% 50% 60% 70% 80%
Preparation of Fine Particulate Emissions Inventories
1-6
-------
1.2.3 NH3 Emissions
Animal husbandry is by far the largest source of ammonia emissions, as indicated in
Figure 1-7. The ammonia from animal husbandry operations comes from animal waste
and depends on the manner in which the waste is processed. The largest single
contributor to the animal husbandry category is cattle, followed by hogs and poultry.
Fertilizer application is also a source of approximately 15-20% of the ammonia. There is
a small percentage from highway vehicles, which can be important in an urban area.
Figure 1-7. NH3 National Emissions
1
Animal Husbandry
-
Fertilizer Application
Highway Vehicles
Industrial Processes
Waste Disposal
Oilier
0'
Zo
1 1 1 1 1 1 1 1
10% 20% 30% 40% 50% 60% 70% 80%
Ammonia emissions are spread out across large parts of the east and the Midwest. This is
not surprising since this is the farm belt where a lot of the animals are raised. This is
consistent with the pattern of measured ammonium ion deposition from the National
Atmospheric Deposition Program (NADP). Specifically, there is a qualitatively good
agreement between where the ammonia is deposited and where the ammonia emissions
are estimated to occur.
1.2.4 Carbon and Crustal Emissions
1.2.4.1 Crustal Emissions
Crustal material mainly comes from fugitive dust. The main sources of fugitive dust are
unpaved roads, agricultural tilling, construction, and wind-blown dust which is found to
occur mostly in the arid areas of the west. A less significant source of crustal material is
fly ash. Fly ash that comes out of a coal- or oil-fired boiler is chemically similar to
crustal material.
There is a huge disparity between the crustal data in an emissions inventory and the
crustal material found in ambient air quality samples. The ambient data say that there is
less than a microgram per cubic meter of crustal material across most of the U.S., with
Preparation of Fine Particulate Emissions Inventories
1-7
-------
the exception of the southwest. On the other hand, the emissions data indicates that
PM2.5 emissions are about 2.5 million tons a year, which is comparable to the carbon
emissions.
This apparent anomaly can be explained by looking at what happens to the fugitive dust
after it is emitted. Fugitive dust emissions are not always transported very far because
they are emitted very close to the ground and get trapped in shrubbery, vegetation,
buildings, etc. In short, fugitive dust emissions may not all be transported very far from
where they are released. When fugitive emissions data are used in air quality dispersion
models to simulate the impact of that dust several miles away, these models fail to take
into account the fact that a lot of the fugitive dust is going to be deposited within a few
hundred yards to a few miles of the source. It is estimated that, on average, about half of
the fugitive dust emitted in eastern metropolitan areas are removed by surface features
near the source. This inventory adjustment only applies when the inventory is being used
in regional chemical transport modeling. Thus, as will be discussed later, this adjustment
is made in the emissions processor, not in the emissions inventory.
In summary, crustal materials are a relatively small part of PM2.5 in the ambient air.
Fugitive dust is released near the ground, and surface features often capture the dust near
its source.
1.2.4.2 Carbon Emissions
Carbon is a huge component of PM2.5 in the ambient air. You will recall from the
previous section that carbon particles are those that are primary (or directly emitted) and
secondary organic aerosol (SOA) particles, which are formed in the atmosphere primarily
from VOCs. Primary carbon particles are comprised of elemental (or black) carbon (EC
or BC), and those that have an organic structure, primary organic aerosol (POA). On
average, approximately 20% of the primary carbon emissions are EC and the other 80%
are POA.
Figure 1-8 shows that primary carbon nationwide comes from wildfires, mobile sources,
industrial and commercial combustion, residential heating and open burning, burning of
construction debris, industrial and commercial processes; agricultural burning; and
fugitive dust. Nationally, there are about 2.5 million tons per year of crustal materials
emitted as compared to about 2 million tons per year of primary carbon emissions.
However, the carbon emissions are found in a lot more abundance in the ambient air.
Preparation of Fine Particulate Emissions Inventories
1-8
-------
Figure 1-8. Primary Carbon in PM2.5
Wildland Fire
Mobile Sources
1
Intl. & Comm. Combustion
1
Res. Healing & Open Burning
1
Ind. & Comm. Processes
1
Agricultural Burning
1
Transportable Fugitive Dust
1
1 1 1 1 1 1 1
0% 5% 10% 15% 20% 25% 30% 35%
Figure 1-9 shows that the ratio of POA mass to EC mass for most sources is roughly 10
to 1. However, a major exception to this is diesel engines and diesel-powered vehicles,
ships, trains and planes, where elemental carbon is a larger fraction than organic carbon.
This higher elemental to organic ratio in diesels is due in part to the relatively higher
combustion temperatures in diesel-fueled engines, which tends to more completely
combust the organic carbon. Conversely, the lower temperature combustion processes
will emit more organic matter, as a result of less complete combustion.
It is important to be aware that the organic carbon reported from analysis of a source or
ambient sample does not include the oxygens, hydrogens and other elements that
comprise the organic carbonaceous matter (OCM). The organic carbon matter is often
called primary organic aerosol (POA). The OC to POA multiplier for "fresh" POA in the
emissions is usually estimated as
POA = OCx 1.2
to approximate the amount of oxygen and hydrogen that is found in POA emissions. In
the atmosphere, these particles "age" through oxidation. As such, a different "multiplier"
is often applied to the POA by (within) the chemical transport models to account for the
"aging" or further oxidation of the POA emissions:
POA = OC x 1.4 to 2.4
Atmospheric transport and transformation models contain this additional multiplier, but
only apply it to the POA, not the EC or SO A. It is important to note that the multiplier is
not related to the model's estimate of secondary organic aerosol formed in the
atmosphere from precursor gases. It is purely to account for further oxidation of primary
particle emissions as the aerosol "ages". Transport models contain a separate module to
simulate the amount of secondary organic carbon formed in the atmosphere from
precursor gases and the OCM of those particles is estimated directly by that module.
Preparation of Fine Particulate Emissions Inventories
1-9
-------
The derivation of a multiplier for ambient OC is much more complicated because the
sample usually contains both POA and SOA, but the relative proportions of each are not
known. Thus, a single multiplier is applied to ambient OC, to adjust both primary and
secondary OC that may be in the sample. The use of a single multiplier introduces error
since it is likely that the multipliers would not be the same for both fractions. A
multiplier of 1.4 to 2.4 is often used for ambient data.
As of this writing, there is no agreed upon standard adjustment that is consistently
applied in either monitoring and modeling studies.
Figure 1-9. Characteristics of Primary Carbon
Category
Ratio of Organic
Carbon Mass to
Elemental Carbon
Mass (Average)
Potential Range
of Ratios
Forest Fires
9.9
6-28
Managed Burning
12
6-28
Agricultural Burning
12
2.5-12
Open Burning - Debris
9.9
Non-road Diesel Engines & Vehicles
0.4
©
j-
l
Ui
On-road Diesel Vehicles
0.4
©
j-
1
Ui
Trains, Ships, Planes
0.4
0.4 - 25
Non-road Gas Engines & Vehicles
14
0.25-14
On-road Gas Vehicles
4.2
0.25 - 14
Fugitive Dust - Roads
22
3 - 65
Woodstoves
7.4
3 - 50
Fireplaces
7.4
3 - 50
Residential Heating - Other
26
Commercial Cooking
111
13 - 111
Figure 1-10 shows a comparison of the emission density ratios for urban carbon
emissions and rural carbon emissions. There is about three times as much primary
carbon emitted in urban areas as there is in the rural areas; about 80% of this is primary
organic aerosol and about 20% is elemental carbon.
Preparation of Fine Particulate Emissions Inventories
1-10
-------
Figure 1-10. Primary Carbon Emission Density
Ratios*
o
TO
GC
&
CTJ
-------
Figure 1-11. Comparison of Emission Density Ratios*
w
o
'ŠP
CO
DC
"(5
Ł =
43 o:
Q
.11
m i-
M 3
E
LU
3
2
1
0
Primary Carbon
80% POA
20% EC
Aromatics
70% Mobile
Terpenes
Biogenic
Eastern US
1.3 SUMMARY
Figure 1-12 presents a summary of the larger source categories of PM2.5 direct and
precursor emissions. These are presented in no particular order; however, the larger
categories are in boldfaced type.
Preparation of Fine Particulate Emissions Inventories
1-12
-------
Figure 1-12. Summary of Important PM2 5 Source Categories
Direct Emissions
Precursor Emissions
Combustion a.b
S02 c
NH3
Open Burning (all types)
Non-Road & On-Road
Mobile
Residential Wood Burning
Wildfires
Power Gen
Boilers (Oil, Gas, Coal)
Boilers (Wood)
Power Gen (Coal)
Boilers (Coal)
Boilers (Oil)
Industrial Processes
On-Road Mobile
Animal Husbandry
Fertilizer Application
Wastewater Treatment
Boilers
Crustal / Metals b
NOx
VOC d
Fugitive Dust
Mineral Prod Ind
Ferrous Metals
On-Road Mobile (Gas,
Diesel)
Power Gen (Coal)
Non-Road Mobile
(Diesel)
Boilers (Gas, Coal)
Residential (Gas, Oil)
Industrial Processes
Biogenics
Solvent Use
On-Road (Gas)
Storage and Transport
Residential Wood
Petrochemical Industry
Waste Disposal
a Includes primary organic particles, elemental carbon and condensable
organic particles; also some flyash
b Impact of carbonaceous emissions on ambient PM 5 to 10 times more
than crustal emissions impact
c Includes SO and SO and HSO condensable inorganics
d Contributes to formation of secondary organic aerosols
NOTE: Categories in BOLD
are most important
nationally. Their relative
importance varies among
and between urban and rural
areas.
Figure 1-13 illustrates several important features of PM2.5 emissions. First, it shows
that the majority of both elemental and organic carbon comes from combustion
sources. It also shows that almost all of the crustal materials are associated with
fugitive dust and very little of the total carbon is associated with fugitive dust.
Emissions of primary carbonaceous PM2.5 are about two million tons per year, and
about a fourth of that is elemental carbon. However, the emissions of crustal
materials is about 2.5M tons per year, roughly similar in magnitude. However, due
the further adjustments made to the EI for carbon and crustal materials previously
discussed, (carbon emissions increase and the crustal emissions are reduced), carbon
is usually found in much greater quantity on ambient PM2.5 samples than are crustal
materials.
Preparation of Fine Particulate Emissions Inventories
1-13
-------
Figure 1-13. Summary of PM2.5 Primary Emission Sources
'Industrial Processes -
Mobile
Source Fuel Š
Combustion
Stationary Source
Fuel Combustion, Š
Open / Biomass
and
Waste Burning
Fugitive Dust
~ (Incl Sand &
Mineral
Productsl)
Elemental
Carbon
(500,000 TPY)
r
Industrial Processes
Mobile
Source Fuel
Combustion
Stationary Source
Fuel Combustion.-
Open I Biomass
and
Waste Burning
Fugitive Dust _
(Incl Sand &
Mineral
Productsl)
Organic Total
Carbon Carbon
(1,500,000 TPY) (2,000,000 TPY)
Crustal Materials
& Misc.
Compounds
(2,500,000 TPY)
The formation of secondary organics from terpenes associated with VOC emissions
from vegetation occurs relatively fast. The formation of secondary organics from
aromatics associated with VOC emissions from mobile sources occurs slower than
the terpene reaction. From a control strategy standpoint it is important to recognize
that reducing aromatics would reduce secondary organic aerosol.
Ammonium sulfate is formed from SO2 that is emitted from the combustion of sulfur
containing fuels. Compared to ozone, the sulfate forms and deposits more slowly and
therefore may be transported much longer distances than either ozone or nitrate. If
there were insufficient ammonia the formation product would be partially neutralized
particles of ammonium bisulfate, or possibly even sulfuric acid. From a control
strategy standpoint, reducing emissions of S02 will lower ammonium sulfate
concentrations.
Ammonium nitrate is formed from NOx that is emitted from fuel combustion.
Nitrates are formed relatively quickly. If there is insufficient ammonia, the ammonia
will react to form ammonium sulfate before it forms ammonium nitrate. Higher
temperatures and a lower relative humidity will shift equilibrium so that less nitrate
and more nitric acid will be formed. Reducing NOx emissions may reduce nitrates,
sulfates and secondary organic aerosols, but the outcomes are complicated, involve
ozone chemistry and can't be generalized.
In general a reduction in VOC emissions would reduce ozone levels and that would
result in less secondary organic aerosols, sulfate and nitrate formation. However, this
is very complicated issue and must collectively consider ozone formation, ozone
precursors, sulfates, nitrates, and the secondary organics because the reactions are
interrelated.
Preparation of Fine Particulate Emissions Inventories
1-14
-------
Review Exercises
1. Which of the following is not a component of secondary particles that are formed in the
atmosphere?
a. ammonium sulfate
b. ammonium nitrate
c. crustal particles
d. secondary organics
2. Which of the following characteristics are common to particulate mater in the eastern half of
the United States?
a. It is composition is chemically similar across a number of urban areas
b. It is comprised of mostly carbon
c. The ammonium and sulfate components are comparable to the carbon component
d. All of the above
3. The density of NOx emissions per square mile are about times higher in urban
areas than they are in the rural areas, but the concentrations of nitrate are only about
times higher in the urban areas as in the rural areas.
a. four, two
b. two, four
c. ten, five
d. three, two
4. Twenty-five percent of total NOx emissions are associated with , the second
largest contributor.
a. Highway vehicles
b. Farm animals
c. Fugitive dust
d. Electric utilities
5. Which of the following is the least significant source of crustal material?
a. Unpaved roads
b. Agricultural tilling
c. Construction
d. Coal and oil-fired boilers
6. Approximately percent of primary carbon is elemental.
a.
10
b.
25
c.
50
d.
75
Preparation of Fine Particulate Emissions Inventories
1-15
-------
7. Which of the following sources of primary carbon has a low ratio of organic carbon mass
to elemental carbon mass?
a. Wildfires
b. Fugitive dust
c. Diesel engines
d. All of the above
8. Which of the following is not an aromatic VOC precursor to secondary organic formation?
a. Benzene
b. Toluene
c. Terpenes
d. Xylene
9. Secondary organic aerosol formation when the ambient temperature increases.
a. increases
b. decreases
c. remains constant
d. ceases
10. is (are) the only precursor with much larger emission density in rural areas than in
urban areas.
a. Terpenes
b. Benzene
c. Ammonia
d. Sulfates
Preparation of Fine Particulate Emissions Inventories 1-16
-------
Review Answers
1. c. carbonaceous particles
2. d. All of the above
3. a. four, two
4. d. Electric utilities
5. d. Coal and oil-fired boilers
6. b. 25
7. c. Diesel engines
8. c. Terpenes
9. a. increases
10. c. ammonia
Preparation of Fine Particulate Emissions Inventories
1-17
-------
THIS PAGE INTENTIONALLY LEFT BLANK
Preparation of Fine Particulate Emissions Inventories 1-18
-------
Chapter 2: The National Emissions
Inventory and Emission Inventory Tools
LESSON GOAL
Demonstrate, through successful completion of the chapter review exercises, a
general understanding of the National Emissions Inventory and the process by which
it was developed. Also, the student should be able to describe the emission inventory
preparation tools that are available as well as those that are under development.
STUDENT OBJECTIVES
When you have mastered the material in this chapter, you should be able to:
1. Describe the purpose of the National Emission Inventory (NEI).
2. Describe some of the data contained in the NEI.
3. Identify inventory preparation tools that currently exist.
4. Explain the purpose of process based emission models.
5. Identify source categories where better data for estimating emissions is needed.
Preparation of Fine Particulate Emissions Inventories
2-1
-------
Chapter 2: The National Emissions
Inventory and Emission
Inventory Tools
2.1 THE NATIONAL EMISSIONS INVENTORY
The information contained in the NEI includes data on 52,000-point sources by
latitude/longitude with over 4,500 types of processes represented. There are about
400 categories of highway and nonroad mobile sources and 300 categories of area
sources in the NEI. Emissions for area and mobile sources are allocated by county.
The NEI includes annual emissions, the dates that sources started or stopped
operations, and stack parameters. There are also HAPs emissions for over 6,000
types of processes.
Figure 2-1. Evolution of the NEI
1985 NAPAP
PM10
"4lh Priority"
Use "Latest" NEI for
Prelim. Planning
Bridge the Gap:
State/local Involvement
Stakeholder Involvement
Tools Development
NAPAP - National Acidic Precipitation Assessment Program
NPI National Particulate Inventory
NET National Emission Trends Inventory
NEI Merger of NET and Nat'l Toxics El
Figure 2-1 presents a timeline showing the evolution of the National Emissions
Inventory (NEI). The first PM inventory in the NEI was in the 1985 National Acidic
Precipitation Assessment Program (NAPAP). This inventory was for PMio and it was
developed without any input from the states. It has only been recently that the states
have become involved in the development of the NEI for PM.
During the early 1990s there was minimal activity on developing PM inventories,
although a National Particle Inventory was prepared in 1993. In 1996 it was called the
Preparation of Fine Particulate Emissions Inventories
2-2
-------
National Emissions Trends Inventory (NET). The NET was updated in 1999 and was
renamed the NEI. Integration of the National Air Toxics Assessment Inventory was
begun with the 1999 NEI and was completed with the 2002 NEI. Improving and
updating the NEI is a continuing process.
In the 2002 NEI, some areas of the US began to include large fires as point sources.
Data on when they started, when they ended, and where they were located are
essential to accurately model their impact on air quality. Smaller fires may continue
to be treated as area sources. These are allocated to counties using data on forested
land area and emissions are assigned to months of the year using temporal allocation
factors. Treating fires as point sources is important, for example, where a fire may
have a major impact on a Class 1 area it could relate to the 20% worst days at that
area. When fires are treated as area sources, it is impossible to know where the fire
occurred or when. As a result, it is impossible to determine if the particulate matter
emissions in the Class 1 area are attributable to the fire.
The process of developing the NEI begins with data on emission factors and models,
various databases for source activity levels, default values for emissions related
variables, existing point source data, and growth factors for source categories. This
data is combined to form what is called the preliminary NEI. The preliminary NEI is
provided to State and local air agencies for their refinements and improvements.
Working with stakeholders and using factor and model improvements, and local
activity levels and variables provided by State and local agencies, the preliminary
NEI is transformed into the improved NEI. This process is repeated yearly, but
emphasized every three years.
2.2 INVENTORY PREPARATION TOOLS
2.2.1 Introduction
One of the tools used to prepare the NEI is the Factor Information and Retrieval database
(FIRE), which is available at www.epa.gov/ttn/chief. There are about 20,000 emission
factors in this database that are used in developing the NEI. However, since industrial
processes vary over time and from facility to facility, there are representativeness issues
with many of these factors.
Another inventory preparation tool is process-based emission models. TANKS is an
example of a model that is used for estimating storage tank emissions of VOCs. The
NONROAD model is used to estimate emissions from non-road vehicles and BEIS is
used to estimate biogenic emissions.
Other tools include special characterization and locator aids such as GIS (Geographic
Information Systems) and global positioning systems (GPS). Satellites are beginning to
be used to locate fires, especially in Mexico and Canada. It should be noted that there are
some severe limitations in using satellites to locate certain fires, such as those that are
below a certain size or through cloud cover, for instance.
Preparation of Fine Particulate Emissions Inventories
2-3
-------
2.2.2 Emissions Processors
The purpose of an emissions processor is to provide an efficient tool for converting
emissions inventory data into the formatted emission files required by an air quality
dispersion model.
After the NEI emissions data is developed it goes into an emissions processor such as
SMOKE (Sparse Matrix Operator Kernel Emissions) modeling system. Speciation
factors are applied to the emissions data by the SMOKE emissions processor. These
steps happen to the inventory after the NEI, and it is something that is generally
performed when the inventory is to be used by air quality dispersion modelers. It is
important to recognize that emissions modeling depends on speciation factors,
temporalization factors, and species allocation factors. The data flow is from the NEI to
the emissions processor and then into the air quality model. The output of the emissions
processor is a gridded, hourly emissions file speciated into elemental carbon, organics,
primary sulfates, and primary nitrates. It should be noted that the speciated inventory
data is especially useful in modeling for regional haze. For example, carbon particles
absorb and scatter light with a different efficiency than other particles, such as sulfate
particles. As a result, it is necessary to consider different types of particles separately
when doing regional haze work.
Area source data is input to the emissions processor as an annual county level inventory
and point source data is input as annual data, located by latitude and longitude. CEM
data feeds into the emission processor separately through a CEM database. The
emissions processor contains default factors and profiles, including county to grid
allocation factors, temporal allocation profiles, and speciation profiles. Using this data,
the emissions processor turns the annual, (sometimes) county level inventory into a
gridded, hourly emission file speciated into EC, POA, Primary S04, Primary Nitrate and
Other, which contains crustal materials/fugitive dust and unidentified species. It is then
ready to be used as input to a dispersion model.
The emissions processor assigns all of the PM2.5 sources to one of several dozen
speciation profiles. Note that elemental carbon and the primary organic aerosols are
derived within the emission processor from PM2.5 data using speciation profiles. As
such, they are not part of the NEI inventory.
There are some issues with compiling a carbon inventory. The split between
elemental carbon and primary organic aerosols is subject to some analytical
uncertainties and there are a lot of questions about how to do that type of analysis. It
is an operational definition of what to call elemental carbon and what to call organic
carbon when doing those analyses. Also these analyses provide data for organic
carbon, not the organic carbonaceous matter that accounts for all the oxygens and
hydrogens. As mentioned in Section 1, it is necessary to use a multiplier or a
compound adjustment to go from organic carbon to primary organic aerosols.
Preparation of Fine Particulate Emissions Inventories
2-4
-------
2.2.3 Processed-based Emission Models
There is another set of tools becoming available called process-based emission
models. Process-based models consider spatially and temporally available activity
parameters as a part of the emissions estimation in an effort to reflect real world
conditions like wind, temperature, relative humidity, vegetation type, soil type,
moisture, etc. For example, these models will eventually include algorithms to
develop a wind-blown dust inventory by examining the wind fields for the whole
modeling domain and deciding when the wind is going to blow fast enough to create
dust emissions. Other examples are a model under development to estimate fire
emissions by taking into account such factors as relative humidity, moisture, and
wind speed and a fugitive dust, model.
These models would have to links to various databases such as MM5, the meteorological
data processor. They would also be linked to GIS coverages of soil and vegetation types,
and would contain emission algorithms responsive to these variables. Currently there are
several models containing some aspects of process-based emission models. These
include the MOBILE 6 model and the BEIS 3 model because they take temperature into
account. Consideration of temperature is critical for estimating biogenic emissions.
There are a number of other needs for process-based emission models. Some examples
include estimating emissions of ammonia and residential wood burning. In the future,
some of these process-based models will be integrated with the emissions processor and
some will be stand-alone. Currently, the biogenics model (BEIS) is always integrated
with the emissions processor. The onroad model MOBILE 6 can optionally be integrated
with the emissions processor. The development of process-based emissions models for
ammonia, fugitive dust and wildland fires are currently underway.
2.2.3.1 Wildland Fires Model
The inputs for the wildland fire model include fire locations, duration, and size. This
model will access meteorological data for wind speed and moisture and uses fuel-loading
defaults from a one-kilometer resolved national map of fuel loadings. Although a fuel
map currently exists (e.g., NFDRS) there is a project currently being funded to develop a
map that will provide better fuel-loading data. Fuel moistures are calculated using the
MM5 data. Fuel consumption will be done using CONSUME or the First Order Fire
Effects Model (FOFEM) for fuel consumption. The emissions projection model, together
with the Briggs' Plume Rise equation modified for fires calculates emissions, heat
release, and plume rise. When it is completed, this emissions module will provide
gridded, hourly emissions and plume characteristics that will take into account the real
world meteorological conditions that would effect fire behavior and emissions.
2.2.3.2 Fugitive Dust Model
Another model that is currently under development is a fugitive dust model. The
approach for developing a fugitive dust model is to establish a consistent database of
resource information such as soil, land use, vegetation, moisture, precipitation, and wind
speed that can be used to estimate emissions for use with grid models. Currently, a proof
Preparation of Fine Particulate Emissions Inventories
2-5
-------
of concept of this emission model is being demonstrated for wind-erosion, unpaved
roads, construction, and other dust sources.
2.2.4 Receptor Models
Receptor modeling is an important toolset that can be used in the development of an
emissions inventory. An estimate of the amount of fossil versus contemporary carbon
can be obtained by using radiocarbon analysis. Some receptor models can use specific
tracers for gas versus diesel particles by looking at the specific organic compounds that
make up those particles and identifying whether those carbon particles are emitted from
gas or diesel engines and also whether they are emitted from cold starts or smokers.
Other tracers (or groups of tracers, called source profiles) are available for a variety of
source types. One model commonly used is the Chemical Mass Balance, a dedicated
weighted least squares model to infer source contribution estimates from ambient
speciated data and source profiles. Multivariate models are the Positive Matrix
Factorization (PMF) model and UNMIX, which is somewhat similar to PMF.
2.3 OTHER NEEDS
There are a number of specific PM2.5 categories that generally need better emissions
models and emissions data. Some of these categories include wildland, forest and
rangeland burning, particularly private and state and tribal burners. For these
categories, data on acreage burned, fuel loadings for the largest fires, and the timing
of those fires is needed.
Other source categories that need better emissions data include:
Residential open burning, household waste, and yard waste. For these
categories, data on the volumes, burning practices, regulations and their
effectiveness, and local surveys of burn activities are needed.
Construction debris and logging slash. Data on the regulations and their
effectiveness, including local surveys of burn activities, is needed.
Agricultural field burning is another source category and data on acreages,
fuel loadings, and timing of the burn events is needed.
Residential wood combustion, fireplaces and wood stoves. Data from local
surveys of fuel burn is needed. This includes data on whether the wood is
being burned in a fireplace or a wood stoves.
Area-specific industrial process sources are another category for which better
data are needed. However, since these source constitute a small percentage
of the industrial process sources, it is important to pick those sources that have
the biggest errors associated with them.
Finally, data on local conditions contributing to fugitive dust are needed in some cases.
Preparation of Fine Particulate Emissions Inventories
2-6
-------
Review Exercises
1. Which of the following was the first particulate matter inventory in the NEI?
a. National Acidic Precipitation Assessment Program
b. National Particle Inventory
c. National Emissions Trends Inventory
d. None of the above
2. Which of the following best describes how large fires are currently handled in the NEI?
a. Large fires are treated as point sources.
b. Large fires are treated as area sources.
c. Large fires are not currently included in the NEI.
d. Large fire emissions are allocated seasonally across the entire state.
3. Which of the following is not an example of an existing emissions inventory preparation
tool?
a. FIRE
b. TANKS
c. BIES3
d. NET
4. models apply space and time sensitivities to emissions.
a. Receptor
b. Grid
c. Process-based emission
d. Photochemical dispersion
5. Which of the following best describes the emissions data that is the output of an emissions
processor?
a. Unspeciated annual averages
b. Speciated hourly averages
c. Speciated annual averages
d. Unspeciated hourly averages
6. Process-based emission models are needed for estimating emissions from .
a. ammonia sources
b. fugitive dust
c. wildland fires
d. All of the above
Preparation of Fine Particulate Emissions Inventories
2-7
-------
7. Which of the following is not an input that would be needed for a process-based emissions
model for wildland fires?
a. Fire location
b. Fire temperature
c. Size of the fire
d. Fuel loading
8. models can identify the source of an emission.
a. Receptor
b. Grid
c. Process-based emission
d. Photochemical
9. Efforts to develop data for sources should be limited to those sources that have the
biggest errors associated with them.
a. mobile
b. open burning
c. specific industrial processes
d. fugitive dust
Preparation of Fine Particulate Emissions Inventories
2-8
-------
Review Answers
1. a. National Acidic Precipitation Assessment Program
2. b. Large fires are treated as area sources.
3. d. NET (National Emissions Trends Inventory)
4. c. Process-based emission
5. b. Speciated hourly averages
6. d. All of the above
7. b. Fire temperature
8. a. Receptor
9. c. specific industrial processes
Preparation of Fine Particulate Emissions Inventories
2-9
-------
THIS PAGE INTENTIONALLY LEFT BLANK
Preparation of Fine Particulate Emissions Inventories 2-10
-------
Chapter 3: Onroad Mobile Inventory
Development
LESSON GOAL
Demonstrate, through successful completion of the chapter review exercises, a
general understanding of EPA's MOBILE 6 model and the National Mobile
Inventory Model (NMIM). Also, the student should be able to describe the concept
of vehicle miles traveled and how it is used to calculate emissions from onroad
vehicles.
STUDENT OBJECTIVES
When you have mastered the material in this chapter, you should be able to:
1. Describe the inputs required to run the MOBILE 6 model.
2. Explain the purpose of the NMIM.
3. Identify sources for obtaining VMT data.
4. Explain the approach for calculating onroad vehicular emissions from VMT data.
5. Identify additional sources of information for calculating onroad emissions.
Preparation of Fine Particulate Emissions Inventories
3-1
-------
Chapter 3: Onroad Mobile Inventory
Development
3.1 MOBILE 6
3.1.1 Overview
EPA's Office of Transportation and Air Quality (OTAQ) has developed MOBILE 6 to
estimate emissions from mobile sources. The MOBILE 6 model includes emission
factors for PM2.5, SO2, NH3, PM10, VOC and CO. The MOBILE6 model and the User
Guide can be downloaded from www.epa.gov/otaq/m6.htm. The PM2.5 and the PMi0
emission factors represent primary emissions. Data on vehicle miles traveled (VMT) is
matched to the corresponding MOBILE 6 emission factors to form the basis of emission
calculations.
PART 5 was EPA's prior model for modeling PM emissions and the data and algorithms
that were previously in PART 5, with some updates, have been integrated into the
MOBILE 6 model. However, the fugitive dust emission factors that were included in
PART 5 have been excluded from MOBILE 6. Consequently, the calculation of
emissions from re-entrained road dust is done separately outside the model. In addition,
MOBILE 6 also includes emission estimates for gaseous S02 as well as ammonia.
3.1.2 Modeling Inputs
In most cases MOBILE 6 uses the same type of inputs that were required for prior
versions. This includes registration, distribution, ambient conditions such as temperature
and humidity, speeds and speed distributions, and fuel parameters such as the Reid Vapor
Pressure of gasoline and oxygenated fuel. It also includes control programs such as Stage
II or Inspection and Maintenance programs, and data on VMT by vehicle type.
One additional data input required for MOBILE 6 modeling that was not required in the
past is the diesel sulfur content expressed in parts per million. Also, there are additional
commands needed for generating PM2.5 inventories in MOBILE 6. These are described
in the MOBILE user's guides that OTAQ has developed. One thing to note is when
developing a PM inventory you cannot do a PM2.5 and a PM10 inventory at the same time.
As a result, it is necessary to specify just one particle size per each run.
Preparation of Fine Particulate Emissions Inventories
3-2
-------
3.2 NATIONAL MOBILE INVENTORY MODEL
The National Mobile Inventory Model (NMIM) is a tool developed by EPA's Office
of Transportation and Air Quality (OTAG) to create national or sub-national emission
inventories for any calendar year using county-specific input parameters. It is a
consolidated emissions modeling system for EPA's MOBILE and NONROAD
models. It combines a graphical user interface, MOBILE, NONROAD, and a
database that contains modeling information for each county in the United States.
Currently this database contains the most recent information (e.g., fuel parameters,
registration data, temperatures, etc.) used by EPA to generate the default National
Emission Inventory (NEI) estimates for each county.
NMIM is capable of calculating both criteria (including ammonia) and HAPs for the
source categories included in the MOBILE6 and NONROAD models. The true
beauty of NMIM is that it consolidates all the model inputs into a single data base
such that all the estimates are based on the same input parameters in each county
(e.g., fuel programs, inspection/maintenance, humidity, temperatures).
EPA used a draft version of NMIM to generate the preliminary EPA default 2002
NEI inventories for nonroad engines. For states, NMIM is an optional tool that
should simplify estimating mobile source inventories by organizing and automating
emission inventory development for highway vehicles and NONROAD categories. It
is not a substantively different approach than directly using MOBILE6.2 and
NONROAD2002.
EPA expects to complete NMIM and release it for general use in 2004 but states will
not be required to use it to generate inventory estimates. This tool was developed to
make creating inventories easier and does not change the answers that are obtained
from running MOBILE or NONROAD individually. In the future, states may wish
to tailor all or part of their own inventory generation process to the NMIM model
approach to take advantage of its efficiency and transparency and to align the NEI
inventory results more closely with their own inventory estimates. State and local
agencies will be able to use the database to view the county-level default values and
to replace them with data that better represents their geographic areas.
3.3 VMT DATA
3.3.1 Sources
State departments of transportation typically provide VMT data. In addition,
metropolitan planning organizations (MPOs) track these data for certain areas.
However, VMT data should be used from whatever source it is available. As a case
in point, the 1999 NEI included VMT data that was provided by eight states and this
data was used in conjunction with MOBILE6 emission factors. VMT data for the
remaining states were obtained from the Federal Highway Administration's (FHWA)
data summaries. The FHWA data contains vehicle miles traveled by roadway type,
Preparation of Fine Particulate Emissions Inventories
3-3
-------
by state, as well as VMT by roadway type for specific urban areas. The 1999 NEI
relied upon a national distribution for the VMT mix by vehicle type. As a result, the
same mix of vehicles was assumed for all areas unless the state provided their own
data. Documentation for the 1999 NEI can be found at this web address:
www.epa.gov/ttn/chief/net/1999inventory.html.
3.3.2 Approach
In the case of the NEI, the VMT data was developed for use in conjunction with
MOBILE 6 by using the distributions of VMT by roadway type and vehicle type. In
some cases this activity data may be available by hour of the day. Regardless of the
format, these fractions can be applied directly to the total VMT, or they can be
included within the MOBILE 6 input files in order to generate a weighted emission
factor in MOBILE 6.
It should be noted that it is important to have speeds matched to the roadway types,
either as an average speed or as speed distributions by speed ranges. This latter
approach is the approach needed for link-based VMT development and some
transportation demand models.
3.3.3 Level of Detail
Ideally, the level of VMT data that should be used is by county and by the various
roadway types or link level if modeling at that level is planned. Using data by vehicle
type is important since emission rates can vary greatly among the different vehicle types.
Using vehicle type data will allow the adjustments to be made to the national defaults that
are typically used. Finally, it is important to match the VMT data (daily or hourly) to the
appropriate time period for modeling.
3.4 CALCULATING EMISSIONS
VMT data needs to be matched to a corresponding MOBILE 6 emission factor and
mapped according to speed, roadway type, vehicle type, and time period. Emissions
are calculated by multiplying the VMT data by an emissions factor as shown in the
following equation.
Emissions = VMT * EF * K
where: Emissions = emissions in tons by roadway type and vehicle type
VMT = vehicle miles traveled by roadway type and vehicle type
EF = emission factor in grams/mile by roadway type and vehicle type
K = conversion factor
Preparation of Fine Particulate Emissions Inventories
3-4
-------
3.5 ADDITIONAL RESOURCES
Since this has been cursory treatment of onroad sources, Table 3-1 provides a number
of online resources that should be consulted when developing an emissions inventory
for onroad sources. This includes EPA's online user's guide for using MOBILE 6.1
and 6.2 as well as technical documentation describing how all the defaults were
developed. There are also links to training materials that have been developed as
MOBILE 6 has been updated.
Table 3-1 ONROAD SOURCES
Additional Resources
Reference
Web Site
User's Guide to MOBILE
6.1 and MOBILE 6.2:
Mobile Source Emission
Factor Model, EPA420-R-
02-028, October 2002.
www.epa.gov/otaq/m6.htm
MOBILE 6.1 Particulate
Emission Factor Model
Technical
Description,Draft,EPA420
-R-02-012, March 2002
www.epa.gov/OMS/models/mobile6/r02012.pdf
Links to MOBILE 6
Training Materials
www.epa.g0v/0taq/m6.htm#m6train
Documentation for the
Onroad NEI for Base
Years 1970 -2002
ftp: //ftp. epa. gov/Emi slnventory/ finalnei 99ver3/hap s/documentati o
n
/ onroad/ neionroad j an04. pdf
Preparation of Fine Particulate Emissions Inventories
3-5
-------
Review Exercises
1. Which of the following pollutants does not have an emission factor included in MOBILE
6?
a. PM2.5
b. Ammonia
c. Carbon dioxide
d. Volatile Organic Compounds
2. Which of the following MOBILE inputs are required for MOBILE 6, but were not required for
MOBILE 6.0?
a. Reid Vapor Pressure
b. VMT data by vehicle type
c. Ambient humidity
d. Sulfur content of diesel fuel
3. The National Mobile Inventory Model is a graphical interface that uses
a. MOBILE 6.2
b. NONROAD 2002
c. a county level database
d. All of the above
4. The development of VMT data in the NEI for use in conjunction with MOBILE 6 is done
by using the distributions of VMT by .
a. vehicle type and speed ranges
b. roadway type and vehicle type
c. roadway type and speed ranges
d. roadway type and link level
5. VMT data needs to be matched to a corresponding MOBILE 6 emission factor and mapped
according to .
a. speed
b. roadway type
c. vehicle type
d. All of the above
6. Which of the following is needed to calculate emissions from onroad vehicles?
a. VMT data
b. Emission factor by roadway type
c. Emission factor by vehicle type
d. All of the above
Preparation of Fine Particulate Emissions Inventories 3-6
-------
Review Answers
1. c. Carbon dioxide
2. d. Sulfur content of diesel fuel
3. d. All of the above
4. b. roadway type and vehicle type
5. d. All of the above
6. d. All of the above
Preparation of Fine Particulate Emissions Inventories
3-7
-------
THIS PAGE INTENTIONALLY LEFT BLANK
Preparation of Fine Particulate Emissions Inventories 3-8
-------
Chapter 4: Nonroad Mobile Inventory
Development
LESSON GOAL
Demonstrate, through successful completion of the chapter review exercises, a
general understanding of EPA's NONROAD model. Also, the student should be able
to describe the approaches for estimating PM emissions from aircraft, commercial
marine vessels, and locomotives.
STUDENT OBJECTIVES
When you have mastered the material in this chapter, you should be able to:
1. Describe the source categories and pollutants that are included in the NONROAD
model.
2. Explain the methodology used by the NONROAD model to estimate emissions.
3. Define the source categories that comprise the aircraft category.
4. Explain the methodology used to estimate aircraft emissions in the NEI and how a
state or local agency can improve on those results.
5. Identify the source categories that comprise the commercial marine vessel category.
6. Explain the methodology used to estimate commercial marine vessel emissions in
the NEI and how a state or local agency can improve on those results.
7. Identify the source categories that comprise the locomotive category.
8. Explain the methodology used to estimate locomotive emissions in the NEI and how
a state or local agency can improve on those results.
Preparation of Fine Particulate Emissions Inventories
4-1
-------
Chapter 4: Nonroad Mobile Inventory
Development
The discussion of EPA's National Mobile Inventory Model (NMIM) that was
presented in Chapter 3 is also applicable to the nonroad category, but will not be
repeated here. It should be noted that aircraft, railroad, and commercial marine
vessel inventories are not included in the NONROAD model and are estimated
independently.
4.1 NONROAD Model
The latest version of EPA's NONROAD model can be accessed at this web address:
www. epa. gov/otaq/nonrdmdl. htm.
This web site contains documentation, a user's guide, as well as technical reports to
describe the sources and development of all the default input values (e.g., equipment
populations, geographic allocations, growth factors, and emission rates).
4.1.1 Sources
Table 4-1 lists the source categories that are included in the NONROAD model. The
four-digit source classification code (SCC) generally denotes the engine type, or fuel that
is used in the nonroad equipment. There are two exceptions where the four-digit SCC
denotes the equipment type instead of the engine type: the recreational marine and
railroad maintenance categories.
Preparation of Fine Particulate Emissions Inventories
4-2
-------
Table 4-1. Engine Types Included in the NONROAD Model
SCCs
Type
2260xxxxxx
2-Stroke Gasoline
2265xxxxxx
4-Stroke Gasoline
2267xxxxxx
Liquefied Petroleum Gasoline (LPG)
2268xxxxxx
Compressed Natural Gas (CNG)
2270xxxxxx
Diesel
2282xxxxxx
Recreational Marine
2285xxxxxx
Railroad Maintenance
Table 4-2 lists the 12 different equipment categories denoted by the seven-digit SCC that
are included in the NONROAD model. Within each of these categories there are
multiple applications that are specified at the 10-digit SCC level.
Table 4-2. Equipment Categories Included in the NONROAD Model
Equipment Category (7-digit SCC denotes equipment)
Airport Ground Support
Logging
Agricultural
Recreational Marine Vehicles
Construction
Recreational Equipment
Industrial
Oil Field
Commercial
Underground Mining
Residential/Commercial
Lawn and Garden
Railway Maintenance
4.1.2 Pollutants
The pollutants included in the NONROAD model are PMio and PM2.5 (representing
primary PM ), CO, NOx, VOC, SO2 and CO2. Ammonia is not a direct output of the
NONROAD model, but it can be estimated based on fuel consumption estimates that are
obtained from the model and EPA emission factors derived from light-duty onroad
Preparation of Fine Particulate Emissions Inventories
4-3
-------
vehicle emission measurements. In addition to exhaust pollutants, the NONROAD
model estimates evaporative VOC components from crankcase emissions, spillage, and
vapor displacement.
4.1.3 Emission Calculations
The NONROAD model calculates exhaust emissions by assuming that they are
dependent on equipment activity. The model takes into account a number of
measures of equipment activity, including how many hours per year the equipment is
used, the load factor at which the engine is operating, the average rate of horse power
of the engine, and the equipment population (i.e., how many pieces of equipment are
in use). These equipment activity measures are multiplied by emissions factors in
tons per horsepower-hour to obtain emission estimates as shown in Equation 4-1.
Equation 4-1. NONROAD Model Emission Equation
lexh = Egxh *A*L*P*N
where: Iexh = Exhaust emissions (tons/year)
Eexh = Exhaust emission factor (ton/hp-hr)
A = Equipment activity (hours/year)
L = Load factor (proportion of rated power used on average basis)
P = Average rated power for modeled engines (hp)
N = Equipment population
Values for the equipment activity measures are generally obtained from a market
engine research firm that conducted telephone surveys of equipment owners and
operators to generate default values for the different equipment categories. However,
there are some exceptions that are described in the documentation for the NONROAD
model.
The emission factors are dependent on the engine type as well as the engine size, or
horsepower. Future year emission controls or standards are reflected in revised
emission rates, so that as older engines are scrapped and new engines replace them,
revised emission rates are applied to the new engines to reflect the standards that they
need to meet.
SO2, CO2 and evaporative VOC emissions are based on fuel consumption. In the
NONROAD model, PM10 is assumed to be equivalent to total PM and for gasoline
and diesel engines, PM2.5 is assumed to be 0.92 times PMi0. For liquefied petroleum
gas (LPG) and compressed natural gas engines, all PM is assumed to be less than
PM2.5.
4.1.4 Geographic and Temporal Allocation
Because there are no estimates of county level populations, the NONROAD model
estimates those populations using surrogate indicators. The model starts with
Preparation of Fine Particulate Emissions Inventories
4-4
-------
national or state level equipment populations (either by equipment type or
horsepower range) and allocates them to the county level by using surrogate
indicators that correlate with nonroad activity for a specific equipment type.
The NONROAD model also accounts for temporal variations in activity. The
temporal profiles vary by month and depend on the equipment category and the
geographic region of the country. The model contains typical weekday and weekend
day activity profiles by equipment category, however, those do not vary by region.
4.1.5 Improving the NONROAD Results
One way to improve EPA's latest 2002 model results is to specify local fuel
characteristics and the ambient temperatures specific to the area being modeled.
Also, if possible, the NONROAD default activity inputs should be replaced with state
or local data. However, it can be resource intensive to obtain reasonable estimates to
replace the default values. In order to obtain this data it would be necessary to
perform a local survey of equipment owners and users.
Another way to improve the model results is to obtain local information to improve
the geographic allocation (i.e., going from state to county). Obtaining local data used
for the temporal profiles can also improve the model results.
Finally, another approach to improve the model results is to focus on priority
categories and obtain better data for those categories. For example, for fine PM,
priority categories would be diesel construction, diesel farm, diesel industrial,
gasoline lawn and garden, and gasoline recreational marine.
4.2 AIRCRAFT
The SCCs representing the aircraft categories that have been historically reported in
the NEI are listed with their definitions in Table 4-3. The activity data used for
aircraft are known as a landing and takeoff operations, or LTO. Emissions are
estimated by applying emission factors to the LTO data that are either specific to an
aircraft or engine type. If the make-up of the aircraft fleet is unknown, fleet averages
are available to be applied to the emission factors.
Preparation of Fine Particulate Emissions Inventories
4-5
-------
Table 4-3. Aircraft SCC
SCC
Aircraft Type
Definition
2275020000
Commercial Aircraft
Aircraft used for scheduled
service to transport
passengers, freight, or both
2275050000
General Aviation
Aircraft used on an
unscheduled basis for
recreational flying,
personal transportation,
and other activities,
including business travel
2275060000
Air Taxis
Smaller aircraft operating
on a more limited basis to
transport passengers and
freight
2275001000
Military Aircraft
Aircraft used to support
military operations
The LTO cycle consists of different modes including: the approach, taxi idle in, taxi
idle out, take off, and climb out. The operation time in each of these modes is
dependent on the aircraft category, meteorological conditions, as well as how the
airport is operating (e.g., the length of time waiting to take off). In addition, there can
be substantial variations in these modes from airport to airport. Because different
emission rates result when the aircraft are operating in each of these modes, it is
important to consider all of these factors in estimating emissions from aircraft.
4.2.1 NEI Method
The NEI estimated emissions for commercial aircraft by using national-level FAA
LTO data by aircraft type and emission rates from the Emissions and Dispersion
Modeling System (EDMS) version 4.0. The NEI estimated emissions from the
general aviation, air taxi, and military aircraft categories by also using national LTO
data, however, data was not available for specific aircraft types within each of the
aircraft categories. Consequently, emissions for these three categories were estimated
by multiplying total LTO by an emission factor as shown in Equation 4-2.
Equation 4-2. NEI Method - General Aviation, Air Taxi, and Military Aircraft
National Emissionsc p = National LTOc * EFC p
where: LTO = Landing and take-off operations
EF = Emission factor
c = Aircraft category
p = criteria pollutant
Preparation of Fine Particulate Emissions Inventories
4-6
-------
Using PM as an example, the emission factors are LTO-based and represent a fleet
average emission factor for the general aviation, air taxi, and military aircraft
categories. Table 4-4 presents these PM emission factors. The PM2.5 primary
emissions are estimated, as they are for many combustion sources, by applying a
particle-sized multiplier of 0.92 to the PMi0.
Table 4-4. LTO-Basec
PM Emission Factors
Category
PM Emission Factor
(lbs/LTO)
General Aviation
0.2367
Air Taxi
0.60333
Military Aircraft
0.60333
Once national emissions are calculated for the four aircraft categories, the NEI
allocates them to the county level based on airport level LTO data. This is shown in
Equation 4-3. Using La Guardia airport as an example, the NEI assumes that a
fraction of the total LTO is assigned to La Guardia, and the emissions calculated
from this allocation are assigned to the corresponding county.
Equation 4-3. Emissions Allocation for Aircraft Categories
Airport Emissionsc p x = National Emissionsc p * AFC p x
where: AF = allocation factor
x = airport (e.g., La Guardia)
c = Aircraft category
p = criteria pollutant
More information on the NEI methodology for estimating emissions from aircraft
categories can be found at the following web address:
ftp://ftp.epa.gov/EmisInventory/finalnei99ver3/criteria/documentation/nonroad/99no
nroad_voli_oct2003 .pdf.
4.2.2 General Approach
Although it may be acceptable to rely upon the NEI data for smaller airports in an
area, a bottom up inventory should be developed for the larger airports. There are
seven steps for developing an aircraft inventory for a specific airport.
Step 1 - Determine the mixing height to be used to define the LTO cycle. The
mixing height is important because above the mixing height, emissions are not
expected to contribute much to ground level pollutant concentrations.
Step 2 - Define the fleet make-up for the airport.
Preparation of Fine Particulate Emissions Inventories
4-7
-------
Step 3 - Determine airport activity in terms of the number of LTO by aircraft
and their associated engine-type.
Step 4 - Select emission factors for each engine model that is associated with
the aircraft fleet at the airport being inventoried (Instead of using defaults that
EDMS may apply for a specific aircraft type).
Step 5 - Estimate the time-in-mode for the aircraft fleet at the airport.
Step 6 - Calculate the emissions (based on the aircraft LTO data, the emission
rates for each aircraft engine model, and the time-in- mode data).
Step 7 - Aggregate the emissions across aircraft to obtain a total for the
airport.
4.2.3 NEI Improvements
Developing an emissions inventory for a local airport involves determining the engine
types associated with the local aircraft types. This data is an improvement over the
assumptions used in the NEI for the commercial aircraft category. In addition,
developing information on climb-out, take-off, approach time, and taxi idle times will
be an improvement over the defaults used in the NEI.
Because the current version of EDMS does not include PM emission rates, EPA
recommends that the few PM emission factors that are available in the 1992 version
of the Mobile Sources Procedures document be matched to the aircraft engines in the
local fleet as best as possible. EPA is aware of this limitation and work is underway
to try to get better data on PM emission factors for commercial aircraft. Some
regional inventories have looked at using emission factor ratios to develop the PM
emission rates for commercial aircraft. Specifically, the ratio based on the PMio and
NOx emission factor ratios for air toxics was applied to the commercial aircraft NOx
emissions.
For the other categories (general aviation, air taxis, and military aircraft) the NEI can
be improved by obtaining local LTO estimates (i.e., the LTO not covered by the FAA
data). Obtaining this data from smaller airports that may not be reporting to the FAA
would be an improvement. The same is true for military bases, although the
heightened security over the last couple years has made it harder to obtain data from
military operations.
Another improvement is to obtain information on the aircraft/engine types that
comprise the fleet for these other categories. If data on the mix of aircraft types in a
fleet are available, engine specific emission factors or EDMS could be used to
estimate emissions. Finally, the NEI can be improved by maintaining the
latitude/longitude of the airport so the emissions are not "smeared" across the entire
county.
Preparation of Fine Particulate Emissions Inventories
4-8
-------
4.3
COMMERCIAL MARINE VESSELS
The SCCs representing the commercial marine vessel categories that are currently
used in the NEI are listed in Table 4-5. This includes diesel activity for ships in port
and underway, as well as residual or steamships for those two categories.
Table 4-5. Commercial Marine Vessel SCC
SCC
Type
2280002100
Diesel, In Port
2280002200
Diesel, Underway
2280003100
Residual, In Port
2280003200
Residual, Underway
4.3.1 NEI Method
The NEI methodology for commercial marine vessels is a top down method that splits
national diesel and residual emissions into port and underway components. The
methodology makes assumptions about what portion of the activity for both diesel
and residual ships takes place in ports and what portion takes place underway (i.e.,
away from ports or on their way between ports). These are allocated separately since
port activity surrounds a port area, while underway covers a larger area such as along
a river system. Both port and underway emissions are assigned to counties, however,
port emissions are assigned to a single county in a port area.
More information on the NEI methodology for estimating emissions from the
commercial marine vessel categories can be found at the following web address:
ftp://ftp.epa.gov/EmisInventory/finalnei99ver3/criteria/documentation/nonroad/99non
road_voli_oct2003 .pdf.
4.3.2 NEI Improvements
One approach to improving the NEI emission estimates for the commercial marine
vessel category is to review the spatial allocation of commercial marine emissions
that is included in the NEI. The NEI method looks at port traffic for the 150 largest
ports in the United States and only allocates those emissions. However, there are
additional ports that are not accounted for in the allocation method. Identifying
smaller ports that are not accounted for in the NEI would be an improvement.
Another approach to improving the NEI method is to allocate port emissions to the
appropriate counties. Port emissions in the NEI are being assigned to a single county
Preparation of Fine Particulate Emissions Inventories
4-9
-------
in the port area. While that may hold for some ports (e.g., deep sea ports or coastal
ports) where the port activity is centered on one county, there are other ports along
the Mississippi and the Ohio Rivers that span multiple counties and even state
boundaries. Assigning these port emissions to the appropriate counties and states is
another way to improve the NEI results.
Another approach to improving the NEI results is to conduct a bottom-up inventory
by obtaining activity estimates at the state or local level from the DOT or Port
Authority. This can include data on fuel consumption, as well as data to define the
actual categories and characteristics of the vessels in terms of the number, size and
horsepower in each category. Similar to aircraft, there are different emission rates
depending on the operating mode of the vessels (i.e., cruising or reduced speed zone,
maneuvering or hotelling), so data on the fraction of the time engines are spent in
those modes would also be an improvement.
Finally, underway emissions can be improved by using available GIS data to monitor
vessel movement.
Equation 4-4 shows the methodology for calculating emissions from commercial
marine vessels. It requires data on vessel populations, horsepower, load factor, and
the time-in-mode operation. Applying this emission equation with this data will
produce a better inventory.
Equation 4-4. Emission Methodology for Commercial Marine Vessels
Emissions = Pop * HP * LF * ACT * EF
where: Pop = Vessel Population or Ship Calls
HP = Average Power (hp)
LF = Load Factor (fraction of available power)
ACT = Activity (hours)
EF = Emission Factor (g/hp-hr)
4.3.3 Activity Profiles
In 1999 EPA completed two studies that provide commercial marine activity profiles
for select ports, and present a method for an inventory preparer to allocate detailed
time-in-mode activity data from a typical port to another similar port. These studies
are Commercial Marine Activity for Deep Sea Ports in the United States and
Commercial Marine Activity for Great Lake and Inland River Ports in the United
States. The specific variables that are collected for the typical ports in these studies
include: 1) the number of vessels in each category, 2) the vessel characterization,
including propulsion size (horsepower), capacity tonnage, and engine age, and 3) the
number of hours at each time-in-mode associated with cruising, reduced speed zone,
maneuvering, and hotelling.
Preparation of Fine Particulate Emissions Inventories
4-10
-------
These studies also contain data on the number of trips and the tons of cargo handled
by vessel type for the top 95 deep-sea ports and the top 60 Great Lake and inland
river ports. Based on the data calculated for a typical port, more detailed activity can
be estimated for these ports. These reports also describe how the typical port
inventories were developed and how they can be scaled that to a port activity in a
specific area.
4.3.4 Emission Factors
Horsepower-based emission factors are available for use with activity data on the
number and size of engines. There are also fuel-based emission factors available for
use with activity data on fuel consumption. EPA has been performing studies to
develop updated emission factors as part of their rulemaking activities, such as the
Category 3 engine final rulemaking that was published in 2003.
Table 4-6 presents EPA recommended PMi0 emission factors that EPA has developed
for specific categories of commercial marine engines, on a gram per kilowatt-hour
basis for Category 1 and Category 2 engines (i.e., small commercial marine vessel
engines).
Table 4-6. Small Commercial Marine Vessel PMi0 Emission Factors
Engine Category
PM10 [g/kW-hrl
Category 1: 37-75 kW
0.90
Category 1: 75-225 kW
0.40
Category 1: 225+ kW
0.30
Category 2: (5-30 1/cylinder)
0.32
EPA recommended PMio emission factors for the larger engines are listed in Table 4-
7. These factors are listed by the different modes of operation.
Table 4-7. Large Commercial Marine Vessel PMio Emission Factors
Mode: Engine
PM10 [g/kW-hrl
Cruise and Reduced Speed Zone: 2-stroke
1.73
Cruise and Reduced Speed Zone: 4-stroke
1.76
Maneuvering: 2-stroke
2.91
Maneuvering: 4-stroke
2.98
Hotelling: 2-stroke
0.32
Hotelling: 4-stroke
0.32
All Modes: Steam Generators
2.49
Emission factors in grams per gallon of fuel consumed are also available from
Procedures for Emissions Inventory Preparation, Volume IV: Mobile Sources, EP A-
450/4-8l-026d (Revised), U.S. EPA, OAQPS, July 1989. As with aircraft category,
PM2 5 emissions from commercial marine vessels are estimated to be 92% of the
PMio emissions.
Preparation of Fine Particulate Emissions Inventories
4-11
-------
4.4 LOCOMOTIVES
The SCCs representing the locomotive categories that are currently used in the NEI
are listed in Table 4-8. This includes larger Class I line haul locomotives that travel
through many states, as well as the smaller Class II and III line haul locomotives that
tend to operate in a smaller area. The NEI also has information on passenger Amtrak
trains, commuter trains, and switchyard operations.
Table 4-8. Locomotive SCC
SCC
Type
2285002006
Diesel Class I Line Haul
2285002007
Diesel Class II/III Line Haul
2285002008
Diesel Passenger (Amtrak)
2285002009
Diesel Commuter
2285002010
Diesel Switchyard Locomotives
4.4.1 NEI Method
The PM emission factors in that are used in the NEI for the line haul and yard
operations are listed in Table 4-9.
Table 4-9. NEI Locomotive PM Emission Factors
Type
PM10
PM2.5
Line-Haul
6.7 g/gallon
6.03 g/gallon
Yard
9.2 g/gallon
8.28 g/gallon
The activity data are based on a national estimate of the gallons of distillate fuel oil
consumed. This national fuel consumption is allocated among four of the five
categories of railroads to develop a national activity value for these four categories
(i.e., Class I, Class II/III, Passenger, and Commuter). Switchyard operation activity is
estimated by multiplying the national Class I fuel consumption by the estimated line-
haul percentage of the total fuel consumption. In other words, the fuel consumption
estimates for Class I line-haul locomotives are assumed to include switchyard fuel
consumption. This assumption is based on the fact that the larger line-haul railroads
are the ones that tend to operate in a switchyard.
The allocation of the activity data to the county level is based on a ratio of county to
national rail activity. This rail activity is measured as a product of density (gross tons
per mile) for each rail line and mileage for the associated rail line in the county.
Mileage for each rail line in the county is measured using a GIS database that is
available from the Bureau of Transportation Statistics.
Detailed documentation on the procedures used to develop criteria and HAP pollutant
locomotive emission estimates for the 1999 NEI can be found at the following web
address:
Preparation of Fine Particulate Emissions Inventories
4-12
-------
ftp://ftp.epa.gov/EmisInventory/finalnei99ver3/criteria/documentation/nonroad/99non
road_voli_oct2003 .pdf.
4.4.2 NEI Improvements
The first step in improving the NEI locomotive emission estimates is to examine the
NEI data for reasonableness. If the NEI data does not represent emissions in a
specific area, more representative fuel consumption at the local or state level should
be obtained. Also, because the NEI makes an assumption to estimate switchyard
emissions, an improvement could be made by obtaining information on the actual
switchyard activity in the study area.
Preparation of Fine Particulate Emissions Inventories
4-13
-------
Review Exercises
1. Which of the following types of equipment are denoted by a four-digit SCC instead of the
usual seven-digit designation?
a. Logging
b. Railroad Maintenance
c. Underground Mining
d. Recreational Equipment
2. Which of the following pollutants does not have an emission factor included in the
NONROAD model?
a. CO
b. C02
c. Ammonia
d. VOC
3. Which of the following is a measure of equipment activity that is used by the NONROAD
model to estimate exhaust emissions?
a. Load factor
b. Horsepower
c. Equipment population
d. All of the above
4. For gasoline and diesel engines, the NONROAD model assumes that PM2.5 is
PMio.
a. equal to
b. one half of
c. 0.08 times
d. 0.92 times
5. National aircraft emissions in the NEI are allocated to the county level based on .
a. population
b. airport LTO data
c. the number of airports
d. All of the above
6. The first step in estimating aircraft emissions for a specific airport in the NEI involves
a. estimating the time-in-mode for the aircraft fleet
b. defining the fleet make-up for the airport
c. determining the mixing height
d. determining the number of LTO at the airport
Preparation of Fine Particulate Emissions Inventories
4-14
-------
7. The NEI methodology for commercial marine vessels allocates emissions to a
single county.
a. port
b. underway
c. both port and underway
d. neither port nor underway
8. Which of the following is not an example of activity data that characterizes a commercial
marine vessel?
a. Propulsion size
b. Time-in-mode
c. Capacity tonnage
d. Engine age
9. The NEI allocates national fuel consumption to develop a national activity for all the
locomotive categories except .
a. Class I
b. Passenger
c. Commuter
d. Switchyard
10. Which of the following types of data is least likely to be obtained in a survey of a railroad
company, especially for smaller companies?
a. fuel consumption
b. total miles of track
c. gross tonnage
d. number of locomotives
Preparation of Fine Particulate Emissions Inventories 4-15
-------
Review Answers
1. b. Railroad Maintenance
2. c. Ammonia
3. d. All of the above
4. d. 0.92 times
5. b. airport LTO data
6. c. determining the mixing height
7. a. port
8. b. Time-in-mode
9. d. Switchyard
10. c. gross tonnage
Preparation of Fine Particulate Emissions Inventories
4-16
-------
Chapter 5: Point Source Inventory
Development
LESSON GOAL
Demonstrate, through successful completion of the chapter review exercises, a
general understanding of the issues associated with identifying point sources for
inclusion in an emissions inventory, including the form of the particulate matter and
the particle size. You should be able to describe the methods for estimating
emissions and be able to articulate the overlap issues associated with point and
nonpoint source emission inventories.
STUDENT OBJECTIVES
When you have mastered the material in this chapter, you should be able to:
1. Explain the difference between filterable and condensable particles.
2. Explain the difference between primary and secondary particulate matter.
3. Identify the form of particulate matter that States must report to EPA.
4. Describe the approach for calculating the particle size of particulate matter from
specific source categories.
5. Identify the available tools for estimating particulate matter emissions.
6. Identify and explain the overlap issues that exist between point and nonpoint source
emission inventories.
Preparation of Fine Particulate Emissions Inventories
5-1
-------
Chapter 5: Point Source Inventory
Development
5.1 IDENTIFYING POINT SOURCES
Point sources are stationary sources that are included in a point source inventory.
Total plant or facility emissions for a given pollutant is usually the criterion for
deciding if a specific source should be included in a point source inventory or an
nonpoint source inventory. These criteria are defined by either state, local or tribal
regulations or policy, or the reporting thresholds contained in the Consolidated
Emissions Reporting Rule (CERR).
5.1.1 Filterable versus Condensable PM
Filterable PM is particles that are directly emitted as a solid or liquid at stack or
release conditions and captured on the filter of a stack test train. Filterable PM may
be PMio or PM2.5 Condensable PM is material that is in the vapor phase at stack
conditions but condenses and/or reacts upon cooling and dilution in the ambient air to
form a solid or a liquid particulate immediately after discharge from the stack.
Condensable PM is almost always PM2.5 or less.
Combustion sources typically emit both filterable and condensable emissions.
Examples include boilers, furnaces and kilns, and both reciprocating internal
combustion engines and turbines. Fugitive dust sources emit filterable emissions
only. Examples of fugitive dust sources include storage piles and unpaved roads at
industrial sites.
5.1.2 Primary versus Secondary PM
Primary PM is the sum of the filterable and the condensable PM. All primary
particles are emitted directly from a stack. Secondary PM is particles that form
through chemical reactions in the ambient air after dilution and condensation has
occurred. Secondary PM is formed downwind of the source. Precursors of
secondary PM include SO2, NOx, ammonia and VOC. The secondary PM should not
be reported in the emission inventory, just the precursor emissions.
5.1.3 Ammonia Sources
Sources of ammonia emissions fall into three broad categories: industrial processes,
use of ammonia as a reagent in NOx control (e.g., selective catalytic reduction or
selective non-catalytic reduction), and refrigeration losses. Examples of industrial
Preparation of Fine Particulate Emissions Inventories
5-2
-------
processes that emit ammonia include combustion sources, ammonium nitrate and
phosphate production, petroleum refining, pulp and paper production, and beet sugar
production. These industrial processes represent the more significant contributors of
ammonia emissions from industrial processes as reported in the 2000 Toxics Release
Inventory (TRI).
5.1.4 Resources
Resources for identifying point sources of fine PM and ammonia include Volume II
of the Emissions Inventory Improvement Program (EIIP) guidance document for
point sources, AP-42 emission factors document, and existing inventories such as the
NEI and the TRI (for ammonia).
5.2 REPORTING PM
When States report their PM25 primary (PM2.5-PRI) emissions to the EPA, either
PM2.5 primary or the PM2 5 filterable and PM condensable components individually
can be reported. All PM condensable is assumed to be in the PM25 size. When
States report their PMi0 primary emissions to the EPA, either PMi0 primary or the
PMio filterable and PM condensable components individually can be reported.
Reporting should be done by using the NIF 3.0 PM pollutant code extensions that
identify the forms of the PM. This includes PRI for primary filterable, FIL for
filterable, and CON for condensable. The database management system will need to
be updated to record these pollutant code extensions.
The form of the PM should be verified to ensure that PM emissions that are recorded
as PMio or PM2 5 are correctly identified as filterable, condensable, or primary
emissions. This verification may require an examination of the emission factors on
which the emissions are based. If the emissions were reported by facilities, the
verification will require that States contact the facilities to ask them what emission
factors were used to calculate the emissions. Alternatively, if the emissions estimates
provided by the sources are based on stack test data, the States will need to ask them
what method was used to measure the emissions in order to determine the form of
PM.
Examining the test method used to collect the data can identify the form of the PM.
EPA's Reference Method 5 series is used to measure total PM filterable emissions.
Most of the AP-42 emission factors are based on Method 5 and, therefore, represent
PM-filterable. Method 17 is similar to the Method 5, however it is infrequently used.
Method 201/201A is designed for PMi0 filterable. In order to calculate or measure
the PMio filterable or the PM2 5 filterable, a particle size analysis of the total PM must
be conducted to develop the size fractions or cut points for PMi0 or PM2 5. This
information is used to develop particle size specific emission factors in AP-42.
However, most of the emission factors in AP-42 are for filterable emissions, although
there are some condensable emission factors for combustion sources. The filterable
Preparation of Fine Particulate Emissions Inventories
5-3
-------
and condensable emission factors need to be summed to obtain a PM primary
emission factor. There are some exemptions so it is important to always understand
the form of the PM that the emission factor represents.
For condensable, Preliminary Method 4 is being developed by the EPA to measure
both PM2.5 filterable and condensable. Method 202 is a method for condensable PM,
but it is not used frequently, mainly because regulations do not require sources to
measure condensable emissions.
5.3 PARTICLE SIZE
AP-42 provides particle size distribution data and particle size specific emission
factors. Some of the source categories (e.g., combustion) in AP-42 have particle size
specific emission factors for PM and, for those categories, that data should be used
first. Appendix B1 should be used for source categories that do not have particle size
specific emission factors. Appendix B1 contains particle size distribution data and
particle size emission factors for selected sources. It is based on documented
emissions data available for specific processes. In the event that Appendix B1 does
not have particle size data for the source category of interest, Appendix B2 should be
used. Appendix B2 contains generalized particle size distributions that are based on
data for similar processes. These distributions are approximations and should only
be used in the absence of source specific particle size distribution data.
Prior to consulting AP-42, any source specific data at the local or state level should
be examined. Any information reported by a source is going to be the best data. If
source provided information does not exist, the hierarchy of resources from AP-42
discussed above should be used. AP-42 chapters are not always clear on what source
test methods were used to develop the particle size data, so the background
information documents that were used to develop the chapters for AP-42 may need to
be consulted. AP-42 is available in EPA's Clearinghouse for Inventories and
Emission Factors (CHIEF) website at www.epa.gov/ttn/chief/.
5.4 EMISSION ESTIMATION TOOLS
5.4.1 Factor Information Retrieval System (FIRE)
The Factor Information Retrieval System (FIRE) is a compilation of emission factors
from AP-42 and other documents. It is an electronic database that is available on the
CHIEF web site. EPA is in the process of developing a more complete set of PM10
and PM2.5 filterable and PM condensable emission factors that will be incorporated
into FIRE.
Preparation of Fine Particulate Emissions Inventories
5-4
-------
5.4.2 The PM Calculator
The PM Calculator is a tool developed by EPA to calculate uncontrolled and
controlled filterable PM2.5 and PM10 emissions using AP-42 particle size data. For
example, it can be used to calculate the PM2.5 filterable emissions based on the PM10
filterable emissions contained in an inventory. It can also be used to calculate PM10
and PM2.5 from the total PM filterable emissions. The calculator only deals with the
filterable emissions (i.e., it does not address the condensable portion) and is for point
sources only. Although it contains over 2300 SCCs, it is limited in that it is based on
AP-42 particle size data that is not available for many sources. As a result, many
times it uses the generic particle size data that is contained in Appendix B2 of AP-42
or other sources. It is also available on the CHIEF web site.
5.5 POINT AND NONPOINT SOURCE OVERLAP
ISSUES
For categories included in the point and nonpoint source emission inventories, the
total point activity must be subtracted from the total state activity to obtain a total
nonpoint source activity. Using the fuel combustion category as an example, the
point source activity is the fuel throughput from the point source inventory. Total
activity is the statewide fuel throughput obtained from the state or local government
agency or from the state energy data reports published by the Interior Energy
Administration in the U.S. Department of Energy.
Ideally, the point source subtraction is based on activity data. For example, the point
source fuel throughput for a given year (e.g., 2002) is subtracted from the total
statewide fuel consumption for the same year. However, in a lot of cases, the activity
data for performing that calculation may not be available. In this case, an emissions
based calculation is acceptable. Under the emissions based approach, the total
source category activity and the point activity need to be on the same control level.
This control level should be an uncontrolled emissions basis because the total
statewide activity represents uncontrolled sources. In this case, it is important to
ensure that the point source emissions represent uncontrolled levels. It is also
important to check the uncontrolled emissions to make sure that they seem
reasonable.
5.5.1 Geographic Adjustment
The geographic level of the point source adjustment that is used to calculate the
nonpoint source activity is an issue when surrogate activity data (e.g., employment
and housing populations) is used to allocate total state activity to the county level.
For example, the EIIP method recommends using employment for specific SIC codes
to allocate total statewide natural gas combustion to the county level for fuel
combustion at industrial and commercial institutions. However, summing the point
source throughput for a given county and subtracting it from the total activity for the
Preparation of Fine Particulate Emissions Inventories
5-5
-------
county may produce negative results, indicating the point source consumption fuel
use is higher than that calculated for the nonpoint sources. This can be an artifact of
the allocation data that was used. The preferred approach is to sum up the point
source fuel throughput consumption on the state level and subtract it from the total
consumption for the state prior to doing the county-level allocation. It is also
preferable to obtain activity data such as employment data for the point sources
included in the inventory. In this way it is possible to make point source adjustments
to the surrogate allocations to account for the amount of employment that is
associated with the point sources.
5.5.2 QA/QC
It is also recommended that the county level nonpoint source estimates be reviewed
for reasonableness after the adjustment has been made. Adjustments should be based
on the experience of the agency personnel. For example, if the allocation method
places nonpoint source activity in a county for which it is known that there is no
activity, that county should be excluded from the allocation. Also, if all of a
county's activity is covered by the point source emission inventory, the nonpoint
source emissions should be zero.
5.5.3 CERR Reporting
If the point emission inventory includes sources with emissions below the CERR
point emission inventory reporting thresholds, the emissions for these small sources
can be included in the nonpoint source emissions. However, in this case it is
important to avoid double counting in the nonpoint source inventory. This can be
done by subtracting the total point source activity from the total state activity before
rolling up the small point source data to add to the inventory. In this way the
emissions for the small point sources in the area are not double counted.
Preparation of Fine Particulate Emissions Inventories
5-6
-------
5.6 ADDITIONAL MATERIALS
A suggested reading list for preparing point source inventories for fine PM is listed in
Table 5-1.
Table 5-1. Reading List
Stationary Source Control Techniques Document for Fine Particulate
Matter, EPA/OAQPS, Oct. 1998
(http://www.epa.gov/ttn/oarpg/tl/meta/m32050.html)
Emission Inventory Guidance for Implementation of Ozone and
Particulate Matter National Ambient Air Quality Standards (NAAQS)
AND Regional Haze Regulations, EPA, OAQPS
(http://www.epa.gov/ttn/chief/eidocs/publications.html)
Introduction to Stationary Point Source Emission Inventory
Development, EIIP Vol. 2, Chapter 1, May 2001
How to Incorporate Effects of Air Pollution Control Device Efficiencies
and Malfunctions into Emission Inventory Estimates, EIIP Vol. 2,
Chapter 12, July 2000
Preparation of Fine Particulate Emissions Inventories
5-7
-------
Review Exercises
1. particulate matter is particles that are directly emitted as a solid or liquid at stack
or release conditions and captured on the filter of a stack test train.
a. Condensable
b. Primary
c. Secondary
d. Filterable
2. Which of the following should not be reported in a particulate matter emission inventory?
a. Primary PM
b. Secondary PM
c. Secondary PM precursors
d. Filterable PM
3. What should be reported if it cannot be determined whether the emissions represent the
PM2.5 or PMio size fraction?
a. total primary PM
b. filterable total primary PM and condensable PM, individually.
c. both a and b
d. None of the above
4. Which EPA test method is designed for measuring PMio filterable?
a. Method 5
b. Method 201/201A
c. Method 17
d. Preliminary Method 4
5. In estimating particle size, which of the following sources of particle size data should be
consulted first?
a. AP-42
b. Appendix B-l
c. Appendix B-2
d. State and local data
6. Which of the following can be used to estimate PM2.5 emissions based on the PMio
emissions?
a. AP-42
b. FIRE
c. The PM Calculator
d. CHIEF
Preparation of Fine Particulate Emissions Inventories
5-8
-------
7. The best approach to calculating a total nonpoint source activity level is to base it on
a. activity data
b. emissions data
c. geographical location
d. population data
8. The geographical level of the point source adjustment is important when data
used to allocate total state activity to the county level.
a. employment
b. population
c. square mileage
d. All of the above
9. Adjustments made to the county level nonpoint source estimates should be based on
a. data obtained from the sources
b. the experience of the agency personnel
c. data obtained from surveys
d. All of the above
10. What is he appropriate action when the point source inventory includes sources with
emissions below the CERR reporting thresholds?
a. Nothing
b. Include those sources in the nonpoint source inventory
c. Include those sources in both the nonpoint source and pint source inventories
d. Create a special category for these sources in the point source inventory
Preparation of Fine Particulate Emissions Inventories
-------
Review Answers
1. d. Filterable
2. c. Secondary PM precursors.
3. c. both a and b.
4. b. Method 201/201A
5. d. State and local data
6. c. The PM Calculator
7. a. activity data
8. d. All of the above
9. b. the experience of the agency personnel
10. b. Include those sources in the nonpoint source inventory
Preparation of Fine Particulate Emissions Inventories
5-10
-------
Chapter 6: Nonpoint Sources
LESSON GOAL
Demonstrate, through successful completion of the chapter review exercises, a
general understanding of the approach for identifying nonpoint sources for inclusion
in an emissions inventory; the methodologies for estimating emissions from nonpoint
sources; and the reconciliation of fugitive emissions data with ambient data.
STUDENT OBJECTIVES
When you have mastered the material in this chapter, you should be able to:
1. Identify the different sources of data for identifying nonpoint sources for inclusion in
an emissions inventory.
2. Explain PM one pagers and identify the information that they contain.
3. Identify typical nonpoint source categories of PM emissions.
4. Describe the general methodology for estimating PM emissions from nonpoint
sources.
5. Explain the concepts of rule effectiveness and rule penetration.
6. Explain the mechanisms that lead to the disparity between fugitive emissions data
and ambient data.
7. Explain the issues with modeling fugitive dust with both Gaussian and grid models.
Preparation of Fine Particulate Emissions Inventories
6-1
-------
Chapter 6: Nonpoint Sources
6.1 OVERVIEW
A nonpoint source is any source that is a stationary source that is not included in the
point source inventory. It should be noted that for emission inventory development
purposes, EPA has traditionally used the term "area sources" to refer to stationary air
pollutant emission sources that are not inventoried at the facility-level. The
Consolidated Emissions Reporting Rule (CERR) specifies reporting thresholds for
point and area sources of criteria air pollutants, which vary depending on the pollutant
and the attainment status of the county in which the source is located (see
http://www.epa.gov/ttn/chief/cerr/index.html). The Clean Air Act (CAA) also
includes a specific definition of area sources of Hazardous Air Pollutants (HAPs) for
the purpose of identifying regulatory applicability. In particular, the CAA defines an
area HAP source as "any stationary source . . . that emits or has the potential to emit
considering controls, in the aggregate, less than 10 tons per year of any HAP or 25
tons per year of any combination of HAPs." Sources that emit HAPs above these
thresholds are categorized as "major sources." To reduce confusion between these
two sets of area source definitions, EPA has adopted the term "nonpoint" to refer to
all criteria air pollutant and HAP stationary emission sources that are not incorporated
into the point source component of the NEI.
Throughout this Chapter there are references to CHIEF, EIIP chapters, EIIP One-
pagers, and the PM2.5 Resource Center. Table 6-1 lists the web site address for each
of these references.
Table 6-1. Web Address for References Cited in this Chapter
Reference
Web Address
CHIEF
www.epa.gov/ttn/chief
EIIP Chapters
www.epa.gov/ttn/chief/eiip/techreport/volume03/index.html
EIIP One-pagers
www.epa.gov/ttn/chief/eiip/pm25inventory/areasource.html
PM2.5 Resource
Center
www.epa.gov/ttn/chief/eiip/pm25inventory/index.html
6.1.1 Identifying Nonpoint Sources
Volume III of the EIIP Area Source Guidance lists the PM fine categories for which
the EIIP guidance is available (see Table 6-2). AP-42 and existing emission
inventories also can help identify nonpoint source categories that are sources of fine
PM and ammonia emissions. Specific existing inventories include the National
Preparation of Fine Particulate Emissions Inventories
6-2
-------
Emissions Inventory, the Toxics Release Inventory, and any inventories developed
through the efforts of a regional planning organization or state and local agencies.
Table 6-2. Key Chapters of Volume III of the EIIP Area Source Guidance for
Sources of PM Emissions
Chapter
Topic
2
Residential Wood Combustion
16
Open Burning
18
Structure Fires
24
Conducting Surveys for Area
Source Inventories
The EIIP also has "area source category method abstracts" for charbroiling, vehicle
fires, residential and commercial/institutional coal combustion, fuel oil and kerosene
combustion, and natural gas and liquefied petroleum gas combustion.
The PM2.5 Resource Center, which is available on the CHIEF website contains "PM
one-pagers," which contain an overview of the NEI methods and summarize nonpoint
source NEI methods for specific categories of PMio, PM2.5, and ammonia. These
overviews provide the source category name and SCC, the pollutants of most
concern, current NEI method, and how state, locals, and tribal agencies can improve
on the NEI method, uncertainties and shortcomings. They also contain activity
variables used to calculate the emissions, current variables and assumptions used in
the methods, suggestions for improving the variables, and where to find additional
information and guidance for the categories. The open burning categories covered by
the one-pagers include residential yard waste for leaves, household waste, residential,
nonresidential, and road construction land clearing waste, structure fires, wildfires
and prescribed burning, and managed or slash burning. Fugitive dust categories
covered by the one-pagers include paved and unpaved roads, residential construction,
and mining and quarrying. One-pagers also exist for residential combustion (i.e.,
fireplaces, woodstoves, and other residential home heaters that burn natural gas or
fuel oil).
6.1.2 Typical Source Categories
Table 6-3 identifies typical area source categories grouped by fugitive dust sources of
filterable PM emissions, open burning nonpoint source categories of carbonaceous
fine PM, external and internal fuel combustion nonpoint sources of carbonaceous fine
PM, and ammonia nonpoint sources.
Preparation of Fine Particulate Emissions Inventories
6-3
-------
Table 6-3. Typical Nonpoint Source Categories
Source Category
Typical Source Categories
Fugitive Dust
Construction
Mining and Quarrying
Paved and Unpaved Roads
Agricultural Tilling
Beef Cattle Feed Lots
Open Burning
Open Burning (residential municipal solid waste, yard waste,
and land clearing debris)
Structure Fires
Prescribed Fires
Wildfires
Agricultural Field Burning
Fuel Combustion
Residential Wood Combustion
Other Residential Fuel Combustion
Industrial Fuel Combustion
Commercial/Institutional Fuel Combustion
Ammonia
Animal Husbandry
Agricultural Fertilizer Application
Agricultural Fertilizer Manufacturing
Waste Water Treatment
Figure 6-1 shows the relative contribution of the nonpoint particulate matter source
categories based on the 2001 National Emissions Inventory. Figure 6-2 shows similar
data for the ammonia source categories.
Preparation of Fine Particulate Emissions Inventories
6-4
-------
Figure 6-1. PM2.5 Emissions in the 2001 NEI
PM-2.5 Emissions in 2001 El
6% Fugitive Dust -
Construction & Misc
All Other (Total)
0.2% Commercial
Cooking
2% On-road Vehicles
4% Non-road Vehicles &
Engines
Residential Heating
2% Agricultural Burning
7% Other Burning
11% Ind. Processes
15% Fugitive Dust -
Agriculture
25% Fugitive Dust -
Roads
9% Fuel Combustion -
Utility
5% Fuel Combustion -
Industrial & Commercial 7% Forest Fires
Figure 6-2. NH3 National Emissions
1
Animal Husbandry
-
Fertilizer Application
Highway Vehicles
Industrial Processes
Waste Disposal
Oilier
0'
Zo
1 1 1 1 1 1 1 1
10% 20% 30% 40% 50% 60% 70% 80%
Preparation of Fine Particulate Emissions Inventories
6-5
-------
6.1.3 Estimating Emissions
Nonpoint source inventories are prepared and reported by the 10-digit SCC source
classification code. Also, actual emissions, not allowable or potential emissions are
reported for the NEI. EPA's master list of SCCs are available on the CHIEF website at
www.epa.gov/ttn/chief/codes/index.html#scc. This is a dynamic list that can be updated
(with EPA's approval) to add SCCs. For example, SCCs should be added if there are
several subcategories within a general nonpoint source category and a state or local
agency is estimating emissions at that level.
Emissions from nonpoint sources are calculated by multiplying the activity data with
the emission factor, control efficiency data, rule effectiveness, and rule penetration. It
should be noted that EPA guidance specifically excludes applying default RE/RP
assumption values for PM inventories. It is highly recommended that the EIIP
methods be followed since these were developed with state and local input and they
reflect the most current standardized procedures for preparing emission inventories.
The EIIP provides preferred and alternative methods for collecting activity data and
the use of emission factors, and contains suggested improvements on existing
inventory methods. Equation 6-1 is a summary of the emission estimation equation.
Equation 6-1. Nonpoint Source Emission Estimation Equation
CA = (EFa)*(Q)*[(1-(CE))(RP)(RE)]
where: Ca = Controlled nonpoint source emissions of pollutant A
EFa = Uncontrolled emission factor for pollutant A
Q = Category activity
CE = % Control efficiency/100
RP = % Rule penetration/100
RE = % Rule effectivenss/100
Activity data is obtained from various published sources of data or surveys.
However, the use of use national, regional and state level activity data requires
allocation to the counties using county-level surrogate indicator data. As a result, the
use of a survey is the preferred approach to obtain the local estimates of activity (i.e.,
a bottom-up approach, rather than a top-down approach).
Emission factors for PM and ammonia can be obtained from FIRE and AP-42.
Alternatively, the emission factor ratio or particle size multiplier approach can be
used. This involves calculating the PM2.5 emissions from the PMi0 emissions using
the ratio of PM2.5 to PMio emission factors in AP-42. However, the use of state,
local, and tribal emission factors are preferred over any other approach because they
are always specific.
Control efficiency is the percentage value representing the amount of a source
category's emissions that are controlled by a control device, process change,
Preparation of Fine Particulate Emissions Inventories
6-6
-------
reformulation, or a management practice. They typically are represented as the
weighted average control for a nonpoint source category.
Rule effectiveness (RE) is an adjustment to the control efficiency to account for
failures and uncertainties that affect the actual performance of the control method.
Rule penetration (RP) represents the percentage of the nonpoint source category that
is covered by the applicable regulation or is expected to be complying with the
regulation.
6.1.4 Spatial and Temporal Allocation
The available national, regional, or state-level activity data often require allocation to
counties or subcounties using surrogate indicators. As such, state, local, and tribal
agencies should review emission estimates developed in this manner for
representativeness. The available temporal profiles to estimate seasonal, monthly, or
daily emissions for specific categories may be limited so states are encouraged to
reflect local patterns of activity in their emission inventories. For example,
residential home heating emissions from fuel oil combustion can be allocated to the
county level by using the number of households in each county in the state.
6.1.5 El Development Approaches
The approaches that are available to state, local, and tribal agencies for developing an
emissions inventory include developing an emissions inventory following the EIIP
procedures; comparing the state, local, tribal activity data and assumptions to the NEI
defaults and replacing the defaults, as necessary; or using the NEI default estimates.
The triage approach to improving the emissions inventory involves considering the
importance of each NEI category and examining the potential impact on air quality,
considering emissions, receptor modeling, and other available information.
Improvements should be made to those categories that are determined to be important
using the suggestions and references provided in this training course. This includes
reviewing the available guidance and deciding what approaches are doable in the near
term and longer term.
6.2 RECONCILING FUGITIVE DUST EMISSIONS WITH
AMBIENT DATA
As discussed in Chapter 1, the main sources of crustal materials are unpaved roads,
agricultural tilling, construction, and wind-blown dust. There is a huge disparity
between the crustal data in an emissions inventory and the ambient air quality data.
The amount of crustal material on the ambient filters is much less than one would
Preparation of Fine Particulate Emissions Inventories
6-7
-------
expect given the large estimates of fugitive dust emissions in the NEI. This apparent
anomaly is explained by the fact that fugitive dust has a low transportable fraction.
The data presented in Figure 6-3 show that PM2.5 inventories in the States included in
the VISTAS area have fugitive dust in the 20-40% range. The rest of PM in the
inventory is from sources that are primarily carbonaceous. Comparing this data with
the data presented in Figure 6-4 shows that the ratio of crustal PM2.5 emissions to
total carbonaceous matter emissions does not match with the ratio of crustal to total
carbonaceous PM2.5 based on the ambient data.
Figure 6-3. Fugitive Dust Emissions in VISTAS States
Preparation of Fine Particulate Emissions Inventories
6-8
-------
6.2.1 Fugitive Dust Removal Processes
In the process of developing models the concept of a stilling zone underneath the canopy
of vegetation was recognized. Within the stilling zone (the bottom three-fourths of the
height of the vegetation) the air is very still and it lends itself to gravitational settling and
impaction and filtration by the vegetation.
In the western part of the country it is common to see wind breaks. These are basically a
row of trees or other tall vegetation designed to slowdown the wind speed on the leeward
side of the downwind side. The overall objective is to prevent the wind from catching the
soil and picking it up and eroding it. Another important feature of windbreaks is the
entrapment effect involving the transmittance of dust through a wind break. Research
shows that the dust that goes through a wind break is about the same as the optical
transmittance of light through a wind break and the remainder is trapped in the
windbreak.
6.2.2 Capture and Transport Fraction
Capture fraction is the portion of fugitive dust emissions that are removed by nearby
surface cover and transport fraction is the portion that is transported out of the source
area. The capture fraction plus the transport fraction together sum to the fugitive dust
emissions inventory.
Figure 6-5 shows a graph that plots a capture fraction value (from zero to 1) and the type
of vegetation qualitatively described as going from densely forested to barren. The test
data plotted on this graph suggest that there is a relationship between the amount of
vegetation and the capture fraction. This data suggest that tall leafy dense vegetation has
a high capture fraction and the short sparse scattered vegetation has a low capture
fraction. This conceptual model has yet to be integrated with air quality models, but it
does allow one to assign capture fractions to different types of vegetation as shown in
Table 6-4.
Preparation of Fine Particulate Emissions Inventories
6-9
-------
Figure 6-5. Capture Fraction Conceptual Model
Table 6-4. Capture Fraction Estimates
Surface Cover Type
CF (Estimated)
Smooth, Barren or Water
0.03-0.1
Agricultural
(N
O
1
o
Grasses
0.2-0.3
Scrub and Sparsely Wooded
0.3-0.5
Urban
0.6-0.7
Forested
0.9-1.0
By using land use databases that contain data on the fractional land use in six different
areas (barren and water, agriculture, grass, urban, scrub and sparse vegetation, and forest)
it is possible to do a computation of the capture fraction. As shown in Table 6-5, the
capture fraction for a given area is the summation of capture fraction by land use type
times the county fractional land use amount. The transport fraction is equal to one minus
the capture fraction. For example, the transport fraction from the source in Churchill
County, Nevada is much higher than the amount that gets away from the source in
Preparation of Fine Particulate Emissions Inventories
6-10
-------
Oglethorpe County, Georgia. The main difference is the amount of trees in those two
areas. In general, the transport fraction is fairly low in those areas of the country that are
very heavily forested, or in cities with a lot of buildings.
Table 6-5. Example Capture Fraction Calculations
Land Use
Type
Barrer
&
Water
Agriculture
Grass
Urban
Scrub
& Sparse
Vegetation
Forest
CF
TF
CF
.03
.15
.2
.6
.3
.95
Fractional
Land Use in
Churchill
Co., NV
.33
.03
.2.
0
.36
.05
0.23
0.77
Fractional
Land Use in
Oglethorpe
Co., GA
0
.1
.14
0
0
.76
0.76
0.24
6.2.3 Modeling Issues
There are modeling issues associated with using this approach to account for different
transport characteristics of dust in different parts of the country. Gaussian models
actually have removal mechanisms built in to them to accommodate capture fraction
through the use of empirical coefficients. Unfortunately, there is limited data and
guidance on how to apply these coefficients, so they are rarely used.
Grid models on the other hand are not equipped to handle particle transport. One issue
with grid models is that they tend to remix particles within the lowest layer during each
time step and this results in an underestimation of the removal by gravitational settling.
Within a time step of the model particles have had a chance to settle down, but not settle
out. In the next time step they are remixed into the whole lower mixing cell, so they may
never get out. Also, in the initial grid (i.e., grids no smaller than 4 km square) removal
processes, even gravitational settling, are ignored. This is a very significant omission
unless the grid is very small. However, modeling very small grids is not really practical.
6.2.4 Summary
Transport fractions should not be used to reduce the emission inventory nor with
Gaussian models. They can be used with grid models with the proper caveats. Because
there are other issues with the inventory, there will not be instantaneous agreement
between the fugitive dust emissions and the ambient data. For example, there are issues
with applying the unpaved road factors properly. The transport fraction concept is
evolving and over time grid model modifications could eliminate the need for this
approach.
Preparation of Fine Particulate Emissions Inventories
6-11
-------
Crustal material is a relatively small part of PM2 5 in the ambient air. Fugitive dust is
released near the ground and surface features often capture the dust near its source.
Finally, the capture/transport fraction concept provides a useful way to account for near
source removal when used with grid models. This area of research offers many
opportunities to improve model performance.
Preparation of Fine Particulate Emissions Inventories
6-12
-------
Review Exercises
1. Which of the following is a source for obtaining information for identifying nonpoint
sources for inclusion in an emissions inventory?
a. EIIP Area Source Guidance
b. AP-42
c. Toxics Release Inventory
d. All of the above
2. Which of the following is not found in the PM one-pagers for specific categories of nonpoint
sources?
a. An overview of the NEI methods
b. National emission estimates
c. Approaches for improving the NEI results
d. Activity variables
3. Typical fugitive dust categories of emissions include construction, mining, paved
and unpaved roads, agricultural tilling, and beef cattle feed lots.
a. ammonia
b. carbonaceous fine PM
c. filterable PM
d. All of the above
4. Which type of emissions are reported for PM in the NEI?
a. Actual
b. Allowable
c. Potential
d. All of the above
5. Which of the following data is used in estimating emissions from nonpoint sources?
a. control efficiency
b. rule effectiveness
c. rule penetration
d. All of the above
6. represents the percentage of the nonpoint source category that is covered by an
applicable regulation.
a. Rule effectiveness
b. Control efficiency
c. Rule penetration
d. Activity data
Preparation of Fine Particulate Emissions Inventories
6-13
-------
7. The area that comprises the bottom three-fourths of the height of vegetation underneath a
canopy of vegetation is called the zone.
a. dropout
b. inversion
c. laminar
d. stilling
8. The fraction is the portion of fugitive dust emissions that are removed by nearby
surface cover.
a. capture
b. transport
c. suspended
d. trapped
9. Transport fractions can be used with the proper caveats.
a. to reduce the emissions inventory
b. with Gaussian models
c. with grid models
d. All of the above
Preparation of Fine Particulate Emissions Inventories
6-14
-------
Review Answers
1. d. All of the above
2. b. National emission estimates
3. c. filterable PM
4. a. Actual
5. d. All of the above
6. c. Rule penetration
7. d. stilling
8. a. capture
9. c. with grid models
Preparation of Fine Particulate Emissions Inventories
6-15
-------
THIS PAGE INTENTIONALLY LEFT BLANK
Preparation of Fine Particulate Emissions Inventories 6-16
-------
Chapter 7: Fugitive Dust Area
Sources
LESSON GOAL
Demonstrate, through successful completion of the chapter review exercises, a
general understanding of the methods used in the NEI to estimate PM
emissions from agricultural tilling, paved and unpaved roads, and construction
activities.
STUDENT OBJECTIVES
When you have mastered the material in this chapter, you should be able to:
1. Explain how PM emissions are calculated for agricultural tilling operations.
2. Identify methods for improving the NEI emissions for agricultural tilling
operations.
3. Explain how PM emissions are calculated for paved and unpaved roads.
4. Identify methods for improving the NEI emissions for paved and unpaved
roads.
5. Explain how PM emissions are calculated for residential, commercial, and
road construction activities.
6. Identify methods for improving the NEI emissions for residential,
commercial, and road construction activities.
Preparation of Fine Particulate Emissions Inventories
7-1
-------
Chapter 7: Fugitive Dust Area
Sources
This Chapter addresses fugitive dust emissions from the following area
sources: agricultural tilling, paved roads, unpaved roads, and residential,
commercial, and road construction activities.
7.1 AGRICULTURAL TILLING
7.1.1 NEI Method
The SCC that is contained in the National Emissions Inventory for agricultural
tilling emissions is 2801000003. For this category the NEI contains estimates
of filterable PMi0 and PM2.5. There are no condensibles associated with this
category.
The activity data for the NEI was obtained from the Conservation Technology
Information Center (CTIC), which publishes a national crop residue
management survey every two years that contains county level activity data.
The NEI used the data from the 1998 survey. This database provides acres of
crops tilled in each county by crop type and by tilling method. The five tilling
methods included in the database include no till, mulch till, rich till, zero to
15% residue, and 15-30% residue.
The emission factor in the NEI is expressed as the mass of the total suspended
particulate per acre tilled. The emission factor is comprised of a constant of
4.8 pounds per acre pass of PM, the silt content of the surface soil, the number
of tillings per year, which is broken into conservation and conventional use,
and the particle size multiplier to calculate the PMi0 or the PM2.5 from the PM
emissions.
The silt content values that are used for various soil types in the NEI are listed
in Table 7-1. These soil types are assigned to counties by using the USDA
surface soil and county level maps to match the soil types to counties.
Preparation of Fine Particulate Emissions Inventories
7-2
-------
Table 7-1. NEI Silt Content Values
Soil Type
Silt Content (%)
Silt Loam
52
Sandy Loam
33
Sand
12
Loamy Sand
12
Clay
29
Clay Loam
29
Organic Material
10-82
Loam
40
Table 7-2 shows the number of tillings that are assumed by crop type for both
conservation and conventional use. The no till, mulch till, and ridge till
methods come from the county level inventory from the CTIC and are
grouped into the conservation use category. The acres reported for the zero to
15 percent residue and the 15 to 30 residue are grouped into the conventional
use category. As can be seen from the data in Table 7-2, the conventional use
category has more tilling passes per acre than the conservation use.
Table 7-2. Number of Tillings in NEI
Crop
Conservation Use
Conventional Use
Corn
2
6
Spring Wheat
1
4
Rice
5
5
Fall-Seeded Small Grain
3
5
Soybeans
1
6
Cotton
5
8
Sorghum
1
6
Forage
3
3
Permanent Pasture
1
1
Other Crops
3
3
Fallow
1
1
Preparation of Fine Particulate Emissions Inventories
7-3
-------
Equation 7-1 presents the equation that is used in the NEI for calculating total
PM emissions from agricultural tilling operations.
Equation 7-1. Agricultural Tilling Emission Estimation Equation
E = c*k*s06*p*a
where: E = PM emissions, lbs per year
c = constant 4.8 lbs/acre-pass
k = dimensionless particle size multiplier (PMi0 = 0.21; PM2.5 =
0.042)
s = silt content of surface soil (%), defined as the mass fraction of
particles smaller than 75 //m diameter found in soil to a
depth of 10 cm
p = number of passes or tillings in a year
a = acres of land tilled
This equation has been used to estimate PM emissions from agricultural
operations in the NEI prior to 1999. Since 1999 the number of acres tilled for
each of the five tillage types has been estimated based on a linear interpolation
of national level data available for 1998, 1999 and 2002. Using 1998 as the
basis, national growth factors were developed by tillage type for 1998, 1999
and 2002. These growth factors were applied to county level emissions for
1998 to estimate county level emissions for 1999 and 2002. Finally, the NEI
emission calculation assumed no controls.
7.1.2 Improving the NEI
One way to improve upon the NEI method is to use crop-specific acreage and
tilling practice data from the state or local agency or tribal authority. In
addition, if State or local emission factors exist, they should be used.
Another improvement is to perform a field study to determine the local silt
content percentage of the surface soil. The silt values that are used in the NEI
are based on limited data and represent averages for the entire country. Local
or state conditions may exist that warrant improving the NEI silt content
values. Finally, the development of crop calendars to determine the time and
frequency of the activities (e.g., land preparation, planting and tilling) will be
an improvement over the NEI data.
7.1.3 CARB Study
This discussion is based on the report "Computing Agricultural PMio Fugitive
Dust Emissions Using Process Specific Rates and GIS" by Patrick Gaffney
and Hong Yu and presented at the National Emissions Inventory Conference
in San Diego during April 2003 (download from the CHIEF web site).
Preparation of Fine Particulate Emissions Inventories
7-4
-------
The California Air Resources Board (CARB) prepared a statewide PMio
inventory for land preparation activities and harvest activities at the county
level. The goals were to obtain current crop-specific acreage data, develop
crop-specific temporal profiles or crop calendars, and to develop emission
factors for all crops.
In developing the inventory CARB obtained county level crop-specific
acreage data from the California Department of Food and Agriculture. This
department generates the crop data every year by county, and it includes over
200 crops and 30 million acres.
CARB also developed crop calendars for the 20 most important crop types
with importance based on the acreage and the potential emissions associated
with each crop type. The crop calendars were used to define the temporal
periods of farming operation activities for each of the crop types. Figure 7-1
is an example of a crop calendar for corn. These types of calendars are very
informative in terms of identifying when specific activities occur. As an
example, stubble disking for corn occurred in November and December with
one pass across the field. In contrast, the NEI assumes these emissions are
annual and does not apply any temporal adjustments.
Figure 7-1. Example Crop Calendar for Corn
Farming
Operations
Crop
Cycles
Per
Year
Passes
Per
Crop
Cycle
Fraction
of
Acreage
Per
Cycle
Passes During Month
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Land
Preparation
Stubble Disc
1
1
1.0
Finish Disc
1
1
1.0
List &
Fertilize
1
1
1.0
Mulch Beds
1
1
1.0
Planting
1
1
1.0
Cultivation
1
2
1.0
Harvesting
1
1
1.0
Prior to preparing the statewide PMio inventory for land preparation activities
and harvest activities CARB used the AP-42 tilling emission factor of 4.0 lbs
PMio/acre-pass for all land preparation activities. For harvesting, CARB only
estimated emissions for three crop types for which emission factors were
available. In order to improve over the past approach CARB conducted field
testing over a seven year period to develop emission factors for several
different types of activities that are crop specific and operation specific.
Preparation of Fine Particulate Emissions Inventories
7-5
-------
These new data allowed CARB to develop the crop calendar that was
discussed above.
Table 7-3 presents the land preparation emission factors that CARB developed
for five different types of activities. These emission factors were used as
surrogates for other land preparation activities such as wheat cutting where
specific factors were not available.
Table 7-3. Land Preparation Emission Factors
Land Preparation
(lbs PMio/acre-pass)
Root Cutting
0.3
Discing, Tilling, Chiseling
1.2
Ripping, Sub soiling
4.6
Land Planning & Floating
12.5
Weeding
0.8
Table 7-4 presents the harvest emission factors that CARB developed for
three types of crops. These factors were assigned to over 200 crop types and
adjusted using a division factor that was developed in consultation with the
agricultural industry within the state. For example, wheat harvesting was
assigned to another crop type, and then adjusted with a division factor. These
adjusted factors were considered to be the upper limit of the emission factors
for other crop types.
Table 7-4. Harvest Emission Factors
Harvest
(lbs PM10/acre-pass)
Cotton Harvest
3.4
Almond Harvest
40.8
Wheat Harvest
5
7.2 PAVED ROADS
7.2.1 NEI Method
The SCC that is contained in the National Emissions Inventory for paved road
emissions is 2294000000. For this category the NEI contains emission
estimates for PMio and PM2.5.
Preparation of Fine Particulate Emissions Inventories
7-6
-------
7.2.1.1 Activity Data
The activity data used for the NEI for paved roads is vehicle miles traveled
(VMT) on paved roads. Paved road VMT is estimated by subtracting the state
and road type-level unpaved road VMT from the total state road type-level
VMT. It is important to note that because the Federal Highway
Administration uses different methodologies to calculate unpaved road VMT
and total road VMT, there are a few instances (principally in western states)
where the unpaved road VMT is higher than the total VMT. In this case, the
unpaved VMT is simply reduced to equal the total VMT, and the paved roads
are assumed to be zero.
The NEI estimates monthly paved road VMT by applying temporal allocation
factors that were developed for the 1985 NAPAP study to the annual paved
road VMT estimate.
7.2.1.2 Emission Factors
The December 2003 version of the emission factor equation in AP-42 only
estimates PM emissions from resuspended road surface material. PM
emissions from vehicle exhaust, brake wear, and tire wear are estimated using
EPA's MOBILE6 model and are subtracted from the emission factor equation.
Equation 7-2 presents the formula for calculating the paved road emission
factor for all vehicle classes. It should be noted that the NEI used the pre-
December 2003 version of the emission factor equation for estimating paved
road emissions.
Equation 7-2. Paved Road Emission Factor Equation
PAVED = PSDPVD*(PVSILT/2)065*(WEIGHT/3)15 - C
where: PAVED = paved road dust emission factor for all vehicle classes
combined (grams per mile)
PSDPVD = base emission factor for particles of less than 10
microns in diameter (7.3 g/mi for PMi0)
PVSILT = road surface silt loading
WEIGHT = average weight of all vehicle types combined (tons)
C = emission factor for 1980's vehicle fleet exhaust, brake wear,
and tire wear
The road surface silt loading varies according to the 12 functional roadway
classifications that are contained in the NEI. For example, the silt loading for
county maintained class roads is one gram per square meter. However for
road types with an average daily traffic volume (ADTV) of less than 5,000
vehicles per day the silt loading is 0.2 grams per square meter. For road types
exceeding the 5,000 ADTV (i.e., freeways) the silt loading is 0.015 grams per
square meter. The national average vehicle weight is 6,360 pounds. Section
13.2.1 of AP-42 contains more information on determining appropriate silt
loading factors.
Preparation of Fine Particulate Emissions Inventories
7-7
-------
Since the amount of fugitive dust emissions is a function of the amount of
rain, the NEI makes an adjustment for precipitation. This is accomplished by
multiplying the emission factor by a rain correction factor that is calculated by
the formula in Equation 7-3. The precipitation data for the NEI was taken
from one meteorological station representative of an urban area for each state.
In this manner, the NEI developed emission factors on a monthly basis at the
state and the road type level for the average vehicle fleet.
Equation 7-3. Precipitation Adjustment Equation
Correction Factor =1- (p/4N)
where: p = the number of days during the averaging period with greater
than 0.01 inches of precipitation
n = the number of days within the averaging period (e.g., 365 for
annual)
7.2.1.3 Emission Calculations
Equation 7-4 shows the formula used in the NEI to calculate PMi0 emissions
from paved roads from resuspended road surface material. PM emissions
from vehicle exhaust, brake wear, and tire wear are estimated using EPA's
MOBILE6 model. PM2.5 are estimated by multiplying the PMio emissions by a
particle size multiplier of 0.25.
Equation 7-4. Paved Road Emission Calculation Equation
FM = VMT *FF
J-^1V-Ls,r,m v 1VJ-L s,r,m s,r,m
where: EM = PMio emissions (tons/month)
VMT = vehicle miles traveled (miles/month)
EF = emission factor (tons/mile)
S = State
R = road type class
M = month
Equation 7-5 shows the equation for allocating the monthly paved road
emissions at the state level to the county level.
Equation 7-5. County Level Allocation Equation
PVDEMISx.y = PVDEMISsty *VMTx.y/ VMTst.y
where: PVDEMISx,y = paved road PM emissions (tons) for county x and
road type y
PVDEMISst,y = paved road PM emissions (tons) for the entire
state and road type y
VMTx,y = total VMT (106 miles) in county x and road type y
VMTSt,y = total VMT (106 miles) in entire State for road type y
Preparation of Fine Particulate Emissions Inventories
7-8
-------
7.2.1.4 Controls
The NEI methodology assumes that controls are only in place for urban and
rural roads in serious PM non-attainment areas and for urban roads in
moderate PM non-attainment areas. A control efficiency of 79% is applied in
these areas. This value corresponds to vacuum sweeping on paved roads
twice per month. There is also an accounting of rule penetration that varies by
road type and the non-attainment area classification.
7.2.2 Improving the NEI
One method to improve the NEI is to obtain VMT data for both paved and
unpaved roads. This is preferable to the NEI approach of subtracting the
unpaved road VMT from the total VMT.
Also, local registration data may be available that represents the average
weight of the vehicles. This is preferable to the use of the NEI default value,
particularly since this variable is weighted most heavily.
Another option is to perform sampling to refine the value used for silt content.
However, this can be resource intensive and should only be used if enough
samples can be collected to give a good representation of the roads in the
inventory area.
7.3 UNPAVED ROADS
7.3.1 NEI Method
The SCC that is contained in the National Emissions Inventory for unpaved
road emissions is 2296000000. For this category the NEI contains emission
estimates for PMio and PM25. There is no condensable material so the PM
filterable (PM-FIL) is equivalent to PM primary (PM-PRI).
7.3.1.1 Activity Data
The activity data used by the NEI for unpaved roads is state level unpaved
road VMT data that is available from the Federal Highway Administration.
This data is allocated to counties by population. Due to the availability of
specific activity for the local classes this calculation is done differently for
urban and rural local functional classes (i.e., county maintained road types)
than it is for the state and federally maintained roads.
Equation 7-6 shows the equation for calculating the vehicle mile traveled by
road type.
Equation 7-6. Unpaved VMT Calculation Equation
Unpaved VMTRoadtype = MileageRoadtype *ADTV*DPY
Preparation of Fine Particulate Emissions Inventories
7-9
-------
where: Unpaved VMTRoadtyPe = road type specific unpaved VMT
(miles/year)
MileageRoadtype = total number of miles of unpaved roads by
functional class (miles)
ADTV = Average daily traffic volume (vehicle/day)
DPY = number of days per year
The non-local functional classes of roads tracked by the Federal Highway
Administration include rural minor collector, rural major collector, rural
minor arterial, rural other principal arterial, urban collector, urban minor
arterial, and urban other principal arterial. Because there are no estimates of
average daily traffic volume for the non-local roads, it is estimated from local
urban and rural VMT and mileage data for the local roads (see Equation 7-7).
Equation 7-7. ADTV Calculation Equation
ADTV = VMT/Mileage
where: ADTV = average daily traffic volume for State and federally
maintained roadways
VMT = urban/rural VMT on county-maintained roadways
(miles/year)
Mileage = urban/rural state-level roadway mileage of county-
maintained roadways (miles)
The total state unpaved VMT by road type is calculated by adding the non-
local functional class VMT to local functional class VMT. The total state
unpaved VMT is temporally allocated by month using NAPAP temporal
allocation factors.
7.3.1.2 Emission Factor
Similar to the AP-42 emission factor equation for paved roads, the unpaved
road emission factor equation only estimates PM emissions from resuspended
road surface material. PM emissions from vehicle exhaust, brake wear, and
tire wear are estimated separately, using EPA's MOBILE6, and are subtracted
out of the emission factor equation. It should be noted that the vehicle
exhaust, brake wear, and tire wear component is relatively much less for
unpaved roads than for paved roads.
Equation 7-8 shows the AP-42 empirical equation that is used to calculate the
unpaved road emission factor. It has some of the same variables as the paved
road equation, but they are weighted differently. For example, there is more
weight given to surface material silt content.
Equation 7-8. Unpaved Road Emission Factor Equation
EF = [k*(s/12)*(S/30)° 5]/[(M/0.5)°2] - C
where: EF = size specific emission factor (pounds per VMT)
Preparation of Fine Particulate Emissions Inventories
7-10
-------
k = empirical constant (1.8 lb/VMT for PM10-PRI, 0.27 for
PM2.5-PRI)
s = surface material silt content (%)
M= surface material moisture content (%)
S = mean vehicle speed (mph)
C = emission factor for 1980's vehicle fleet exhaust, brake wear,
and tire wear
Table 7-5 summarizes the NEI default emission factor input values and the
source of the values. The web address for the surface materials silt content
values links to a database for unpaved roads that provides all the supporting
documentation that was used, including a database of state level silt content.
It should be noted that the calculation of unpaved road emissions in the NEI
used the pre-December 2003 AP-42 emission factor equation. This equation
considers mean vehicle weight and, therefore, it is listed in Table 7-5. Also, it
should be noted that the precipitation data is obtained from one
meteorological station that is representative of rural areas since unpaved road
activity is expected to be occurring in rural areas.
Table 7-5. NEI Default Emission Factor Input Values
Input
Source of Values
Surface Material Silt
Content(s)
Average state-level sources available at
ftp://ftp.epa.gov/EmisInventory/
finalnei99ver2/criteria/documentation/
xtrasources/
Mean Vehicle Weight
(W)
National average value of 2.2 tons (based on
typical vehicle mix)
Surface Material
Moisture Content
(Mdiy)
1 percent
Number of days
exceeding 0.01 inches
of precipitation (p)
1. Precipitation data from one meteorological
station in state is used to represent all
rural areas of the state
2. Local climatological data available from
National Climactic Data Center at
http://www.ncdc.noaa.gov/oa/ncdc.html
7.3.2 Improving the NEI
Short of developing independent estimates, the NEI defaults should be
reviewed for representativeness. Also, local data should be used when
possible for the activity and emission factor. If resources are limited, the
Preparation of Fine Particulate Emissions Inventories
7-11
-------
focus should be on collecting data that represents local precipitation as well as
actual local VMT estimates.
7.4 CONSTRUCTION
7.4.1 Overview
The SCCs that are contained in the National Emissions Inventory for the
construction category are shown in Table 7-6. The NEI contains emission
estimates for PMio and PM2.5 and there are no condensibles, so PM-PRI is
equal to PM-FIL. The relative contribution of these three different types of
construction to the 1999 NEI is listed in the last column of Table 7-6.
Table 7-6. SCCs for Construction
Category
SCCs
% Contribution
Residential
2311010000
5
Commercial
2311020000
40
Road
2311030000
55
7.4.2 Residential Construction
7.4.2.1 NEI
The NEI uses the number of acres disturbed per year as the activity data for
residential construction. Since direct estimates of the number of acres
disturbed are generally not available, the value for this activity is estimated
through the use of housing start data that is available from the Bureau of the
Census. These data are available as regional monthly housing unit start
values. Data is also available at a national level for housing unit starts for the
various classifications of housing. These classifications include 1-unit houses,
2-unit houses, 3-4 unit houses, and 5+ unit housing. These housing
classifications are important because there are different numbers of acres
disturbed for each type of housing. The regional housing unit starts for each
of these categories is estimated using the fraction that is available at a national
level as shown in Equation 7-9.
Equation 7-9. Regional Housing Unit Starts Estimation Equation
Regional HS = Total Regional HS*(National HS by Category/Total National
HS)
where: HS = Housing Starts
Preparation of Fine Particulate Emissions Inventories
7-12
-------
Since these regional housing starts are on a monthly basis they are summed to
obtain an annual total. The next step is to allocate these regional housing
starts data to the county level. This is accomplished by using data on the
annual number of building permits in each county for each housing unit
classification. It should be noted that the building permit data should not be
used to estimate housing starts but only to allocate housing starts to the
county. This is because many times a building permit is issued but the
dwelling is never constructed. In short, the housing start data is a more
accurate estimate of what is really being constructed.
Also, the regional housing start data actually represents the number of units
that were started. However, the number of structures is a better activity
indicator of the number of acres that are disturbed. For example, the activity
data for an apartment building with multiple units should reflect the structure
as a whole (i.e., the number of acres disturbed in the building of the structure
and not for each unit). Table 7-7 shows the correlation between residential
structure starts and housing unit starts.
Table 7-7. Relationship Between Housing Units and Residential Housing
Structures
Housing Unit Starts
Residential Structure
Starts
1-unit
1 unit per structure
2- unit
2 units per structure
3-4 unit
3.5 units per structure
5+ unit
Region specific units per
structure as calculated
from building permits data
Equation 7-10 shows the equation for estimating the number of county
residential housing structure starts based on the regional number of structure
starts.
Equation 7-10. Residential of Structure Starts Estimation Equation
County SS = Regional SS*(County Bldg. Permits/Regional Bldg. Permits)
where: SS = Structure Starts
The number of acres disturbed and the duration of the construction activity
vary depending on the size and type of the structure. The assumed values for
both acres disturbed and duration are listed in Table 7-8. The basis behind
these assumptions can be found in Estimating Particulate Emissions from
Construction Operation, 1999.
Preparation of Fine Particulate Emissions Inventories
7-13
-------
Table 7-8. Assumed Values for Residential Construction
Type of Structure
Acres Disturbed
Duration of
Construction
1-unit
]A acre per building
6 months
2-unit
1/3 acre per building
6 months
Apartments
'/2 acre per building
1 year
The number of apartment structures is estimated by adding the number of 3-4
unit buildings and the number of 5+ unit buildings. Also, the number of 1-
unit houses needs to be estimated separately for houses with a basement and
those without a basement. This is because building a house with a basement
requires that additional dirt be moved and this must be accounted for in the
emission factor equation. The number of 1-unit houses without basements is
estimated by multiplying the regional number of 1-unit structures by the
regional percentage of one-family houses with basements and subtracting the
product from the total number of 1-unit houses.
The amount of dirt moved for 1-unit houses with basements is estimated by
multiplying the assumed average basement depth of 8 feet by the assumed
value of 2,000 square feet of dirt moved per structure. An additional 10
percent is added to this value to account for footings and other back-filled
areas adjacent to the basement.
Table 7-9 shows the emission factor data that the NEI uses to estimate the
emissions on an acre-per-month basis. Also, PM2.5 is assumed to be 20% of
PMio.
Table 7-9. NEI PMio Residential Construction Emission Factors
Housing Category
Emission Factor
(tons/acre/month)
1-unit housing with basement
0.011 (plus 0.059 tons/cubic
yard of on-site cut/fill)
1-unit housing without basement
0.032
2-unit housing
0.032
Apartments
0.11
Equation 7-11 shows the equation that NEI uses to estimate PMio emissions
from 1-unit residential structures with basements and Equation 7-12 shows the
equation used for one-unit structures without basements, as well as all two-
unit structures. The same equation is used for apartments with the exception
Preparation of Fine Particulate Emissions Inventories
7-14
-------
that the emission factor of 0.11 tons/acre/month is used instead of the 0.032
tons/acre/month value.
Equation 7-11. PMio Emission Estimation Equation for 1-unit
Residential Structures with Basements
Emissions = (EF*B*f*m) + 0.059 tons PMio/1000 cubic yards of cut/fill
where: EF = Emission factor (0.011 tons PMi0/acre/month)
B = number of housing starts with basements
f = buildings-to-acres conversion factor (1/4 acre per building)
m = duration of construction activity (months)
Equation 7-12. PMio Emission Estimation Equation for 1-Unit
Residential Structures without Basements and 2-Unit Residential
Structures
Emissions = (EF*B*f*m)
where: EF = Emission factor (0.032 tons PMio/acre/month)
B = number of housing starts with basements
f = buildings-to-acres conversion factor (1/4 acre per building)
m = duration of construction activity (months)
Controls in PMio non-attainment areas are accounted for by applying a control
efficiency of 50% for both PMio and PM2.5 emissions for all PMio
nonattainment areas. There is no adjustment made for attainment areas. The
50% value represents best available control methods on fugitive dust
construction activities in the nonattainment counties.
In addition to accounting for the control measures, other adjustments are
applied to the emission estimates for all three construction categories. These
adjustments are for soil moisture content and silt content. Emissions are
adjusted for soil moisture content by using average Precipitation Evaporation
(PE) values according to Thornthwaite's Precipitation Evaporation Index.
Equation 7-13 shows the formula for making this adjustment. This
adjustment accounts for precipitation and humidity in a certain area and, as
can be seen in the equation, the higher the PE the smaller the adjustment.
Equation 7-13. Soil Moisture Level Adjustment
Moisture Level Corrected Emissions = Base Emissions * (24/PE)
where: PE = Precipitation Evaporation value for county
Emissions are adjusted for the dry silt content in the soil of the area being
inventoried. Equation 7-14 shows the formula for making this adjustment.
Preparation of Fine Particulate Emissions Inventories
7-15
-------
Equation 7-14. Silt Content Adjustment
Silt Content Corrected Emissions = Base Emissions * (s/100)
where: s = % dry silt content in soil for area being inventoried
7.4.2.2 Improving the NEI
Obtaining local data for new housing starts, or permits for additions or
modifications to existing homes would be an improvement over the use of the
NEI defaults. Another improvement is to develop a buildings-to-acres
conversion factor for acres disturbed per construction unit as well as obtaining
data on the seasonality of residential construction practices. Finally, obtaining
local information on soil moisture content, silt content, and control
efficiencies would be an improvement over the NEI default values.
7.4.3 Commercial Construction
7.4.3.1 NEI
Similar to the residential construction category, the NEI uses the number of
acres disturbed each year as the activity representing fugitive dust emissions
from commercial construction. The NEI developed a top-down inventory by
using national level activity data on the dollar value of commercial
construction. These data were then allocated to the county level.
The allocation of the national level expenditure data was performed by using
two data sources: Annual Average Employment for SIC 154, Data Series
ES202, Bureaus of Labor Statistics, 1999 and Annual Average Employment
for SIC 154, Marketplace 3.0, Dunn & Bradstreet, 1999. Two data sources
were used because there were some data missing in the first database, and the
Dunn & Bradstreet database was used to fill in the gaps. Specifically, the
county proportion of the state total from the Dunn & Bradstreet database was
applied to the state total from the BLS data base to estimate employment for
counties where data were missing.
The dollar value activity data were converted to acres disturbed using a
conversion factor of 1.6 acres/106 dollars. This conversion factor was applied
to the estimated county-level construction valuation data.
The PM10-PRI emission factor for commercial construction is 0.19 tons per
acre month. The PM2.5 is assumed to be 20% of the PMi0.
Equation 7-15 shows the emission formula used in the NEI for calculating the
PM emissions from commercial construction.
Preparation of Fine Particulate Emissions Inventories
7-16
-------
Equation 7-15 Emission Estimation Equation for Commercial
Construction
Emissions = (EF*$*f*m)
where: EF = Emission factor (0.19 tons PMio/acre/month)
$ = dollars spent on nonresidential construction (millions)
f = dollars-to-acres conversion factor
m = duration of construction activity (assumed 11 months)
The emissions calculated from Equation 7-15are adjusted to reflect control
measures that are in place in PMio non-attainment areas. In addition to
accounting for the control measures, adjustments are applied for soil moisture
content and silt content using Equation 7-13 and Equation 7-14, respectively.
7.4.3.2 Improving the NEI
Improving the NEI results can be done by obtaining local information on
number of acres disturbed per construction event or per construction dollar
spent. Also information on location, average duration, and seasonality of
commercial construction practices would be an improvement over the NEI
default values. Finally, local information on soil moisture content, silt
content, and control efficiency would result in improved emission estimates.
7.4.4 Road Construction
7.4.4.1 NEI
The NEI uses the number of acres disturbed as the activity data indicator for
road construction. State level expenditure data for capital outlay for six road
construction classification are available. These classifications include:
Interstate, urban
Interstate, rural
Other arterial, urban
Other arterial, rural
Collectors, urban
Collectors, rural
Because some of the activities that are included in the total state level
expenditure data do not contribute to PM emissions, it was necessary to
remove the expenditures for these activities. These activities include minor
widening, resurfacing, bridge rehabilitation, safety, traffic operation and
control, and environmental enhancement and other.
To obtain the activity data in terms of acres disturbed it was necessary to first
convert the expenditure data to mileage and then to acreage. The NEI
estimated the miles of new road constructed by applying conversion factors of
$4 million dollars per mile of interstate, and $1.9 million dollars per mile for
other arterial and collector roads. These conversion factors were based on
Preparation of Fine Particulate Emissions Inventories
7-17
-------
information obtained from the North Carolina Department of Transportation.
The NEI then applied the conversion factors in Table 7-10 to convert to acres
disturbed per mile of road activity level.
Table 7-10 Road Construction Conversion Factors
Classification
Conversion Factor
(acres/mile)
Interstate, urban
15.2
Interstate, rural
15.2
Other arterial, urban
15.2
Other arterial, rural
12.7
Collectors, urban
9.8
Collectors, rural
7.9
The estimated acres disturbed are summed across all of the road types to
estimate the total acres disturbed. The NEI allocates these state-level
estimates of acres disturbed to the county-level by using housing start data.
This is the same data that was developed for the residential construction
category. This assumes that new road development is directly proportional to
new housing starts.
The PM10-PRI emission factor for road construction is 0.42 tons per acre
month. The PM2.5 is assumed to be 20% of the PMi0.
Equation 7-16shows the emission formula used in the NEI for calculating the
PM emissions from road construction.
Equation 7-16 Emission Estimation Equation for Road Construction
Emissions = (EF*$*fl*f2*d)
where: EF = Emission factor (0.42 tons PMio/acre/month)
$ = State expenditures for capital outlay on road construction
fl = dollars-to-miles conversion factor
f2 = miles-to-acres conversion factor
d = duration of roadway construction activity (assumed 12 months)
The emissions calculated from Equation 7-16are adjusted to reflect control
measures that are in place in PMi0 non-attainment areas. In addition to
accounting for the control measures, adjustments are applied for soil moisture
content and silt content using Equation 7-13 and Equation 7-14, respectively.
Preparation of Fine Particulate Emissions Inventories
7-18
-------
7.4.4.2 Improving the NEI
Obtaining information on location and timing of road construction practices in
the area is one way of improving on the NEI results. Also, obtaining local
data on the number of miles constructed and the number of acres disturbed per
project or per mile of road constructed is better than using the NEI default
values that are based on expenditure data. Also, local data on the duration of
the projects and information on private road construction activity (not
included in the NEI) would represent improvements. Finally, obtaining
information for making adjustments for soil moisture content, silt content, and
control efficiency would be an improvement over the NEI default values.
Preparation of Fine Particulate Emissions Inventories
7-19
-------
Review Exercises
1. Which of the following variables is not included in the NEI emissions methodology
for estimating emissions from agricultural tilling operations?
a. silt content of soil
b. acres of land tilled
c. control measures
d. number of passes
2. Which of the following would be an improvement over the NEI emissions methodology
for estimating emissions from agricultural tilling operations?
a. use of corn calendars
b. performing a field study to determine silt content
c. use of crop-specific acreage
d. All of the above
3. In the paved roads category, the NEI contains emission estimates for
a. PMio
b. PM25
c. Condensable PM
d. A and B
4. Which of the following is used as the activity data for paved roads in the NEI?
a. total miles of road
b. vehicle miles traveled
c. road type class
d. average vehicle weight
5. The assumed control measure for paved roads in the NEI is
a. wetting of the road
b. the use of dust suppression materials such as oil
c. vacuum sweeping
d. All of the above
6. Which of the following sources of emissions from unpaved roads are estimated by
EPA's MOBILE6.2 model?
a. vehicle exhaust
b. tire wear
c. brake wear
d. All of the above
Preparation of Fine Particulate Emissions Inventories 7-20
-------
7. In estimating the amount of dirt moved for 1-unit houses with basements, an
additional percent is added to the amount of dirt removed for the
basement to account for footings and other back-filled areas adjacent to the
basement.
a. 5
b. 10
c. 15
d. 20
8. A Precipitation Evaporation value represents high precipitation and
humidity and results in a adjustment to the base emissions estimate.
a. larger, larger
b. smaller, larger
c. larger, smaller
d. smaller, smaller
9. Which of the following activities need to be removed from State-level road
construction expenditures when developing an activity level for road construction
activities?
a. Resurfacing
b. Bridge rehabilitation
c. Minor road widening
d. All of the above
10. Which construction category requires a two-step conversion to obtain the activity
data of number of acres disturbed?
a. commercial
b. residential
c. road
d. All of the above
Preparation of Fine Particulate Emissions Inventories
7-21
-------
Review Answers
1. c. control measures
2. d. All of the above
3. d. AandB
4. b. vehicle miles traveled
5. c. vacuum sweeping
6. d. All of the above
7. b. 10
8. c. larger, smaller
9. d. All of the above
10. c. road
Preparation of Fine Particulate Emissions Inventories
-------
Chapter 8: Ammonia Emissions From
Animal Husbandry
LESSON GOAL
Demonstrate, through successful completion of the chapter review exercises, a
general understanding of the issues associated with estimating ammonia
emissions from animal husbandry operations and some of the efforts that are
being undertaken to address these issues.
STUDENT OBJECTIVES
When you have mastered the material in this chapter, you should be able to:
1. Explain the problems that have been identified with the ammonia emission
estimates in the NEI.
2. Explain the six-step process for improving the ammonia emissions estimates
from animal husbandry operations.
3. Explain the concept of manure management trains.
4. Identify the improvements that are being made to the NEI for the animal
husbandry category.
5. Explain the differences between the 1999 NEI and the 2002 NEI with respect
to ammonia emissions from animal husbandry operations.
Preparation of Fine Particulate Emissions Inventories
8-1
-------
Chapter 8: Ammonia Emissions From
Animal Husbandry
8.1 OVERVIEW
Almost 5 million tons a year of ammonia are emitted nationally and as shown
in Figure 8-1 animal husbandry is the largest contributor to ammonia
emissions nationally.
Figure 8-1. NH3 - Precursor to Ammonium Sulfate and Nitrate
Animal Husbandry
Fertilizer Application
Highway Vehicles
Industrial Processes
Waste Disposal
Other
n 1 1 1 1 1 r
0% 10% 20% 30% 40% 50% 60% 70% 80%
In addition to being the largest contributor to ammonia emissions nationally, it
is important to address ammonia emissions from animal husbandry because
inverse modeling suggests that ammonia emissions may be overestimated.
Inverse modeling involves doing a complete chemical transformation and
transport modeling of an area and accounting for all of the ammonia through
transformation and deposition processes. Comparing these results to the
ammonia that has been found in the ambient air indicates that ammonia may
be overestimated nationally. The interim improvements described in Section
8.2 below address many of these shortcomings.
Additionally, problems with the old NEI have been identified. For example,
there are probable errors in the emission factor selections, especially for beef.
Also, the NEI does not use information on variability of emissions due to
different manure handling practices within a given animal industry, nor does it
make total use of the National Agricultural Statistic Service (NASS) data on
different animal populations by weight. The NEI also does not take
Preparation of Fine Particulate Emissions Inventories
8-2
-------
temperature into account, which would greatly increase the temporal variation
in ammonia emissions.
Moreover, EPA's water emission effluent guidelines project has provided
some new information on animal production and waste handling practices.
Also, the National Academy of Sciences, at the behest of the agricultural
community, has reviewed EPA's inventory work, and recommended a long-
term data-gathering effort.
8.2 IMPROVING THE NEI
EPA has recently prepared a report that provides a basis for making interim
improvements to the NEI. It provides improved data on populations,
practices, and emissions. It is the beginning of a switch-over to a process-
based framework that is a consistent and transparent way of estimating
emissions that would allow for partial updating as better data becomes
available. This technique provides a lot of motivation and a structure for
making data-collection improvements. It also provides an opportunity to
educate users about the data limitations and the proper use of the data. The
goal is for the higher animal production states to begin to adopt and offer
improvements to the NEI using this new method.
Table 8-1 lists the six steps that comprise this new methodology for estimating
ammonia emissions from animal husbandry operations. Each of these six
steps is addressed in detail in the following paragraphs.
Table 8-1. Overview of New Estimation Methodology
Step 1
Estimate Animal Populations
Step 2
Identify Manure Management Trains
(MMT)
Step 3
Estimate Amount of Nitrogen
Excreted
Step 4
Identify Emission Factors
Step 5
Estimate Ammonia Emissions
Step 6
Estimate Future Ammonia Emissions
Step 1
The first step in this process is estimating average animal populations by
animal group, state, and county. This step uses the 2002 NASS data for state-
level populations, and the 1997 census of agricultural to apportion the state-
level NASS data to the county level. However, there are some privacy issues
with regard to animal populations. For example, a county with only one large
Preparation of Fine Particulate Emissions Inventories
8-3
-------
facility would create an industrial privacy issue since that facility will not
want their competition to know how many animals they are raising.
Step 2
The second step is using Manure Management Trains (MMT) for each animal
group to estimate the distribution of the animal population. Fifteen manure
management trains have been identified. Figure 8-2 shows an advanced
manure management train, one of several such trains for the dairy industry.
This manure management train begins with the amount of nitrogen excreted
by dairy cows. The train traces the manure through the different handling
options and shows how much is handled in different ways. The train also
shows the nitrogen and ammonia emissions at the various handling points.
For example, there is nitrogen loss in the flush barn and the lagoon, and
ammonia loss in the dry lot. There are other trains that provide similar
information for other farm industries. These trains characterize a type of
industry, and the general way that manure would be handled in a facility.
Preparation of Fine Particulate Emissions Inventories
8-4
-------
Preparation of Fine Particulate Emissions Inventories
8-5
-------
Some of the variables that affect the different trains include the way the
animals are housed, the waste storage methods, and the land application
methods that are used. For example, the non-feedlot outdoor confinement
(e.g., pasture) is one of the trains for swine, dairy, and beef. The MMTs
represent different pathways for the escape of ammonia into the air. In
applying the MMT approach to estimate the 2002 ammonia inventory, the mix
of MMTs is assumed to vary by state, but not within a state. Animal
population is allocated among the applicable trains. For example, in a given
state 20% of the hogs may be handled using manure management train 3,
another 60% may be using manure management train 7, and the rest of them
may be using manure management train 14. Finally, it should be noted that
the final stage on every train is land application.
Step 3
The third step is estimating the amount of nitrogen excreted from the animals
using each type of MMT. This step involves looking at typical animal
weights and data on the amount of nitrogen per thousand kilos of live weight.
The data on the nitrogen amounts can be obtained from NRCS Agricultural
Waste Management Field Handbook. Another useful source of information is
land grant university researchers and local agricultural extension agents. It is
important to include experts in the agricultural industry in the inventory
development efforts.
Step 4
Step four involves identifying or developing the emission factors for each
component of each manure management train. Some of these factors are in
pounds per animal, and some are percent air release of the input ammonia.
These factors are used to determine the amount of ammonia that goes to the
next stage of the manure train process. Under this approach, the air emissions
could never be higher than the original manure content. Also, using this
approach sets the stage for applying temporal profiles and process-related
variables such as moisture and rainfall.
Step 5
The next step involves applying this methodology to estimate annual ammonia
emissions from each animal group by MMT. This includes tracking the
ammonia release through each manure management train for each animal type
for each county and calculating ammonia releases to the air and transfers to
the next stage. This whole process assumes no air emission controls at this
time, but control assumptions could be added later. Emissions are summed
up to animal type and county, but the database is preserved with full detail for
transparency so that changes and improvements can be made.
Step 6
The last step involves estimating ammonia emissions for future years.
Preparation of Fine Particulate Emissions Inventories
8-6
-------
Other improvements that are being made to the NEI for animal husbandry
operations are to incorporate emission estimates for sheep, ducks, goats, and
horses. Additional data sources are being examined to provide recently
available data on manure production and excretion rates by animal type and
weight. Finally, EPA is examining ways to better address special, seasonal,
and regional differences in emissions.
Carnegie Mellon University has prepared a model for estimating ammonia
emissions from agricultural activities, humans, wastewater treatment,
wildfires, domestic and wild animals, transportation sources, industrial
activities, and soils. The Carnegie Mellon model includes an improved
methodology for fertilizer application when compared to the methodology
used in previous versions of the NEI. EPA is evaluating the methodologies
used for other source categories in the Carnegie Mellon model.
8.3 COMPARISON OF THE 1999 AND 2002
AMMONIA NEIs
A comparison of the 1999 NEI version 3 with the 2002 NEI version 1 shows
that there are some significant differences in the ammonia emissions (See
Table 8-2). As shown on this chart, about half of the emissions from all
animals come from calves and cattle. Also, total ammonia emissions from
animal husbandry operations decreased significantly from 3.4 million in 1999
to 2.3 million in 2002.
Table 8-2. NH3 - Comparison of '99 and '02 NEIs
Animal
Group
1999 NEI
2002 NEI
Population
Emission
Factor
Ib/head/yr
Emissions
Tons/year
Population
Emission
Factor
Ib/head/yr
Emissions
Tons/year
Cattle and
Calves
Composite
100,126,106
50.5
2,476,333
100,939,728
23.90
1,205,493
Hogs and
Pigs
Composite
63,095,955
20.3
640,100
59,987,850
14.32
429,468
Poultry
and
Chickens
Composite
1,754,482,225
0.394
345,325
2,201,945,253
0.60
664,238
TOTAL
1,917,704,286
N/A
3,461,758
2,362,863,831
N/A
2,299,199
Preparation of Fine Particulate Emissions Inventories
8-7
-------
Review Exercises
1. Which of the following statements about ammonia emissions from animal
husbandry operations is false?
a. Animal husbandry operations are the largest emitter of ammonia nationally.
b. Inverse modeling suggests that ammonia emissions may be underestimated.
c. There are probable errors in some of the ammonia emission factors in the NEI.
d. The NEI does not take temperature into account in estimating ammonia
emissions.
2. The characterize(s) the general way manure is handled by a specific facility.
a. NASS data
b. NEI
c. MMT
d. All of the above3. Improvements are being made to the NEI to account for
ammonia emissions from .
a. sheep
b. ducks
c. goats
d. All of the above
4. Which type of livestock emits the most ammonia on a per animal basis?
a. cattle
b. pigs
c. poultry
d. horses
5. Which type of livestock emits the most ammonia on a yearly basis?
a.
cattle
b.
pigs
c.
poultry
d.
horses
Preparation of Fine Particulate Emissions Inventories
8-8
-------
Review Answers
1. b. Inverse modeling suggests that ammonia emissions may be underestimated.
2. c. MMT
3 d. All of the above
4 a. cattle
5 a. cattle
Preparation of Fine Particulate Emissions Inventories
8-9
-------
THIS PAGE INTENTIONALLY LEFT BLANK
Preparation of Fine Particulate Emissions Inventories 8-10
-------
Chapter 9: Combustion Area
Sources
LESSON GOAL
Demonstrate, through successful completion of the chapter review exercises, a
general understanding of the methodologies for calculating emissions from
residential wood combustion, residential and land clearing debris burning,
agricultural field burning, and wildland fires.
STUDENT OBJECTIVES
When you have mastered the material in this chapter, you should be able to:
1. Explain the method used in the MANE-VU study to estimate emissions from
residential wood combustion sources.
2. Explain the method used in the NEI for calculating emissions from residential
wood combustion sources.
3. Identify the difference between the MANE-VU method and NEI method for
estimating emissions from residential wood combustion sources.
4. Explain the method used in the NEI for estimating emissions from residential open
burning.
5. Identify the different types of residential open burning.
6. Identify ways in which the NEI method for estimating emissions from residential
open burning can be improved.
7. Explain the method used in the NEI for estimating emissions from land clearing
debris burning.
8. Identify ways in which the NEI method for estimating emissions from land
clearing debris burning can be improved.
Preparation of Fine Particulate Emissions Inventories
9-1
-------
9. Explain the general method for estimating emissions from agricultural field
burning.
10. Explain the general method for estimating emissions from wildland fires.
11. Identity some of the efforts underway to improve the methods for estimating
emissions from wildland fires.
Preparation of Fine Particulate Emissions Inventories
9-2
-------
Chapter 9: Combustion Area Sources
This Chapter covers three types of combustion area sources: residential wood
combustion, residential/land clearing debris burning, agricultural field burning, and
wildland fires.
9.1 RESIDENTIAL WOOD COMBUSTION
9.1.1 MANE-VU Emissions Inventory
The MANE-VU View Regional Planning Organization conducted a residential
wood combustion survey to develop an emissions inventory for the year 2002. The
approach of using a survey is the EIIP preferred method for this category. The
objective of the MANE-VU project is to prepare a 2002 inventory based on a
survey of household equipment usage and wood consumption patterns. The survey
method is a stratified random sampling approach. The data collected for each
household consists of wood consumption at the equipment level for both real wood
and artificial logs, the type of real wood, and the temporal activity to calculate
monthly, weekly, and daily emissions.
9.1.1.1 Sampling Frame
The sampling was designed to address major sources of variability in wood
consumption activity. These sources of variability include the location and type of
housing, the heating demand expressed as heating degree days (HDD), and the
availability of wood.
Housing data from the 2000 census was used to stratify the sample by four
categories: urban, suburban, rural single family, and other homes. The other homes
category includes multi-family units such as apartments, condominiums, and mobile
homes. The rural single-family category was stratified into forested versus non-
forested areas using USGS-GIS data. Total annual heating degree days were used
to further stratify the sample into three zones: low, medium and high.
Table 9-1 is a sample frame shown in a grid. Within each cell the number 61 is the
minimum sample size that was determined based on calculations for the precision
desired from the survey. The numbers in parentheses represent the number of
surveys that were actually collected or completed. Surveys for which the
respondents did not categorize correctly were removed from the sample.
Preparation of Fine Particulate Emissions Inventories
9-3
-------
Table 9-1. Sample Frame
Geographic
Zone
Rural-Forested
Rural
Non-Forested
Suburban
Urban
Single-
Family
Other
Single-
Family
Other
Single-
Family
Other
Single-
Family
Other
High HDD
Cell 1
61
(173)
Cell 2
61
(64)
Cell 3
61
(87)
Cell 4
61
(66)
Cell 5
61
(61)
Cell 6
61
(72)
Cell 7
61
(69)
Cell 8
61
(69)
Low HDD
Cell 9
61
(150)
Cell 10
61
(62)
Cell 11
61
(118)
Cell 12
61
(69)
Cell 13
61
(76)
Cell 14
61
(67)
Cell 15
61
(75)
Cell 16
61
(62)
Med HDD
Cell 17
61
(87)
Cell 18
61
(60)
Cell 19
61
(91)
Cell 20
61
(64)
Cell 21
61
(71)
Cell 22
61
(60)
Cell 23
61
(63)
Cell 24
61
(68)
9.1.1.2 Survey Instrument
The survey instrument is a questionnaire developed to gather the activity data on
indoor equipment (fireplaces, woodstoves, pellet stoves, furnaces, and boilers), and
outdoor equipment (fire pits, barbeques, fireplaces, and chimineas). A pilot survey
was conducted to test the questionnaire. Based on the pilot survey, questions were
rephrased to clarify the questions in order to collect the information that was needed
to characterize the activity. The survey was conducted using computer-assisted
telephone interviewing with over 1,900 surveys being completed across all 24 cells.
9.1.1.3 Data Reduction
After completion, the surveys were quality assured to make sure that the data
collected made sense. Also, the user fraction (i.e., the fraction of the total
household population that burns wood in indoor and outdoor equipment), the annual
activity (i.e., cords of wood by equipment and wood types), and temporal data were
summarized for each cell. Finally, statistical analyses were conducted to identify
significant differences between cells for the user fraction and annual activity.
9.1.1.4 Results and Observations
Table 9-2 is the same as Table 9-1 with the exception that the grid cells have the
fraction of indoor wood burning equipment on a percentage basis. In some cases
the fractions add up to more than 100% because some houses were using more than
one piece of equipment. It should be noted that the rural forested areas within a
high heating demand zone has a higher diversity of equipment and more households
are using wood burning equipment than the urban areas.
Preparation of Fine Particulate Emissions Inventories
9-4
-------
Table 9-2. Sample Frame
Geographic
Zone
Rural-Forested
Rural
Non-Forested
Suburban
Urban
Single-
Family
Other
Single-
Family
Other
Single-
Family
Other
Single-
Family
Other
High HDD
Cell 1
Cell 2
Cell 3
Cell 4
Cell 5
Cell 6
Cell 7
Cell 8
FP=34
FP=75
FP=43
FP=33
FP=36
FP=0
FP=80
FP=100
WS=67
WS=75
WS=76
WS=67
WS=64
ws=o
WS=30
ws=o
F/B=21
F/B=0
F/B=7
F/B=0
F/B=18
F/B=0
F/B=0
F/B=50
PS=4
PS=0
PS=0
PS=0
PS=0
PS=0
PS=0
PS=0
Low HDD
Cell 9
Cell 10
Cell 11
Cell 12
Cell 13
Cell 14
Cell 15
Cell 16
FP=60
FP=100
FP=61
FP=50
FP=70
FP=67
FP=90
FP=100
WS=65
ws=o
WS=54
WS=50
WS=35
ws=o
WS=10
ws=o
F/B=5
F/B=0
F/B=4
F/B=0
F/B=0
F/B=0
F/B=0
F/B=0
PS=2
PS=0
PS=4
PS=0
PS=5
PS=33
PS=0
PS=20
Med HDD
Cell 17
Cell 18
Cell 19
Cell 20
Cell 21
Cell 22
Cell 23
Cell 24
FP=55
FP=60
FP=59
FP=100
FP=81
FP=50
FP=100
FP=0
WS=66
WS=60
WS=45
ws=o
WS=27
WS=50
ws=o
ws=o
F/B=7
F/B=0
F/B=0
F/B=0
F/B=8
F/B=0
F/B=0
F/B=0
PS=7
PS=0
PS=9
PS=25
PS=4
PS=0
PS=0
PS=0
FP = Fireplace; WS = Woodstove; F/B = Furnace/Boiler; PS = Pellet Stove
Totals do not always add to 100 since some respondents use more than one type of equipment. Values in bold are derived from
responses that were identified as wood consumption outliers (equipment could be mis-categorized by respondent).
Another observation is that rural areas have a higher percentage of stoves and
furnaces and boilers than urban areas. Urban and suburban areas have a lower
diversity of equipment types and a higher percentage of fireplaces than rural areas.
With respect to heating demand, rural areas have a higher percentage of stoves and
furnaces in the higher HDD zone, and rural areas have a higher percentage of
fireplaces in the lower HDD zone.
For indoor equipment, because of the sample size of the survey, it was hard to find
households that burned wood in urban areas. However, the urban sample size was
not increased (due to budget constraints and priorities) to obtain a representative
sample for three instead of two HDD zones. As a result, emissions were not
calculated for each piece of indoor equipment in urban areas. Rather, in order to
maintain precision, the equipment and fuel-based survey results were used to
estimate average emissions (pound of PM2.5 per household per year) and a
household-based statistical model was used to estimate emissions for each cell for
indoor equipment.
Because there was enough data collected to maintain the sample frame precision,
emissions were estimated for outdoor equipment using the survey results. The
emissions are the product of the fraction of outdoor equipment users per cell, the
annual activity, and the emission factor. This is the first attempt to estimate
Preparation of Fine Particulate Emissions Inventories
9-5
-------
emissions from outdoor wood burning equipment at the household level and is a
tremendous improvement over the NEI, which only includes indoor equipment.
9.1.1.5 Emission Inventory Development
Emissions were estimated for all criteria pollutants and precursors, and several
dozen toxic air pollutants. They were estimated at the census track level and
summed to the county, state and region. Emissions were temporally allocated to
support modeling using profiles that were developed from the survey.
9.1.1.6 Lessons Learned
The survey instrument for regional surveys should be tailored to suit the usage
patterns on rural and suburban and urban areas. It is difficult to find wood burners
in the urban areas, and the sample size may need to be increased to locate these
sources. For indoor equipment, to keep resources manageable, the use of
statistically derived emissions based model (household level) instead of an
equipment specific method should be considered. The concern with this MANE-
VU approach, however, is that it aggregates emissions for different types of wood
burning equipment, which should be disaggregated in order to conduct a control
strategy analysis.
9.1.1.7 Documentation
Documentation for the MANE-VU project can be obtained at
www.manevu.org/pubs/index.asp. This contains the work plan, including the
equations for calculating the sampling precision.
9.1.2 NEI
The NEI categorizes fireplaces into four SCCs and woodstoves into three SCCs as
shown in Table 9-3. A description of the equipment associated with each SCC is
also included in Table 9-3.
Table 9-3. NEI SCCs for Residential Wood Combustion
SCC
Combustion Source
FIREPLACES
2105008001
Without Inserts
2104008002
With Inserts; Non-EPA Certified
2104008003
With Inserts; Non-Catalytic, EPA Certified
2104008004
With Inserts; Catalytic, EPA Certified
WOODSTOVES
2104008010
Non-EPA Certified
Preparation of Fine Particulate Emissions Inventories
9-6
-------
2104008030
Catalytic, EPA Certified
2104008050
Non-Catalytic, EPA Certified
The pollutants included in the NEI for residential wood combustion include PMio
primary, PM2.5 primary, NOx, CO, SOx, and HAPs. The emission factors that are
used for residential wood combustion represent primary emissions. There is no
breakout of the filterable and condensable portions of the emission factor for this
category.
9.1.2.1 Emission Factors
The emission factors used in the NEI for fireplaces without inserts (pounds
pollutant per ton of dry wood) are obtained from AP-42 except for PM and CO
which are obtained from Houck, J.E. et al, Review of Wood Heater and Fireplace
Emission Factors. The PM2.5 emission factor is assumed to be the same as the
PM10 primary emission factor. The emission factors for all pollutants from
woodstoves and fireplaces with inserts are obtained from AP-42.
9.1.2.2 Emission Estimation Methodology
The NEI developed separate national wood consumption estimates and, therefore
emission estimates, for fireplaces with inserts, fireplaces without inserts, and
woodstoves to account for the different emission factors and different usage
patterns. The methodology is different for fireplaces without inserts than it is for
fireplaces with inserts and woodstoves. As such, these are discussed separately in
the following sections.
9.1.2.2.1 Fireplaces without Inserts
The first step in estimating emissions from fireplaces without inserts is to determine
the number of homes with fireplaces in the United States. These data can be
obtained from the US Department of Census (DOC). These data need to be
adjusted to account for the fact that some homes have more than one fireplace
(multiply by 1.17) and for the fact that not every home burns wood (74% burn
wood, 26% burn gas).
After making the adjustments to account for multiple fireplaces and those that burn
wood, the number of fireplaces not being used (42% not used) and the number of
fireplaces with inserts are subtracted. Fireplaces with inserts are treated in the same
manner as woodstoves and are discussed in section 9.1.2.2.2.
Based on DOC data the NEI separated fireplaces without inserts into 2 categories;
those used for heating and those used for aesthetics. The amount of wood burned in
each device is determined by assuming wood consumption rates of 0.656 cords
burned /unit/year for fireplaces used for heating and 0.069 cords/unit/year for
fireplaces used for aesthetics. In 1997, EPA estimated that 2.94 million cords of
Preparation of Fine Particulate Emissions Inventories
9-7
-------
wood were burned in the former and 0.483 million cords of wood were burned in
the latter.
Once the national wood consumption for fireplaces without inserts is calculated it is
necessary to allocate it to 1 of 5 climate zones based on temperature, and
demographics/population (i.e., the number of single-family home). Within each
climate zone, wood consumption is allocated to individual counties.
Table 9-4 shows the climate zones defined by the ranges of heating degree day and
cooling degree day values as well as the amount of national consumption that is
allocated to each zone.
Table 9-4. Climate Zones
Climate Zone
Percent of Wood Consumed
1 (>7000 HDD)
36
2 (5500-7000 HDD)
19
3 (4000-5499 HDD)
21
4 (<4000 HDD and <2000 CDD)
15
5 (<4000 HDD and >2000 CDD)
9
The census data classifies counties as either urban or rural. A county is classified as
urban if 50 percent of the county's population is located in cities and towns and it is
classified as rural if less than 50 percent of the population is located in cities and
towns. The total wood consumption for all the urban counties are summed for each
climate zone, and the same is done for the rural counties. The data is adjusted if the
percentage proportion between urban and rural areas does not match the percentage
in the number of units that are reported in the 2001 census. This data is shown in
Table 9-5. For example, if the total wood consumption for woodstoves in climate
zone 1 is 60 percent for rural and 40 percent for urban, then each urban and rural
county within zone 1 receives a percent increase or decrease in cordwood
consumption to obtain the correct percent split to reach the 65 percent rural and 35
percent urban split for zone 1.
Table 9-5. Urban/Rural Apportionment Data
Type
Rural
Urban
Woodstoves
65%
35%
Fireplaces with Inserts
43%
57%
Fireplaces without Inserts
27%
73%
Preparation of Fine Particulate Emissions Inventories
9-8
-------
Finally, AP-42 factors are used to determine county emissions from fireplaces
without inserts.
9.1.2.2.2 Fireplaces with Inserts and Woodstoves
The first step in estimating emissions from fireplaces with inserts and woodstoves is
to determine the number of woodstoves and inserts in the United States. These data
are obtained from the DOC and are adjusted for the fact that some homes have more
than one stove. Also, units used for main heating purposes are considered different
from units that are used for other heating purposes (e.g., aesthetic).
The total cords of wood consumed by the residential section for 1997 are obtained
from the Energy Information Administration (EIA). Since this value does not
include consumption for aesthetic purposes, it is necessary to subtract the cords of
wood used in fireplaces for aesthetic purposes.
Using the same approach that was used for fireplaces without inserts, the national
wood consumption for fireplaces with inserts and for woodstoves is allocated to 1
of 5 climate zones (see Table 9-4). Within each climate zone, the wood
consumption is allocated to the individual counties using the relative percent of
detached single family homes in the county to the total number of detached single
family homes in the entire climate zone.
After allocating to the climate zones, the wood consumption in each zone is
summed and compared the urban and rural split. The total is adjusted until the
desired split is achieved. For woodstoves, the split is 69 percent rural and 31
percent urban. For inserts, the split is 50/50. For example, if the total wood
consumption for woodstoves in climate zone 1 was 60 percent for rural 40 percent
fir urban, then each urban and rural county with that zone would receive a percent
increase or decrease in cordwood consumption to obtain the correct percent split to
reach the 69 percent rural and 31 percent urban split.
Wood consumption for woodstoves and fireplaces with inserts are allocated to one
of the three SCCs as shown in Table 9-6. Fireplaces without inserts are recorded on
one SCC, so there is no need to allocate to SCCs.
Table 9-6. Apportionment for Woodstoves and Fireplaces with Inserts
Type of Device
Percent of Total Wood
Consumption
Non-Certified
92
Certified Non-Catalytic
5.7
Certified Catalytic
2.3
Preparation of Fine Particulate Emissions Inventories
9-9
-------
Once the amount of wood consumed per residential wood combustion type is
obtained, AP-42 emission factors are used to calculate emission estimates.
9.1.2.3 Seasonal Adjustment
When the NEI method was developed the seasonal activity was allocated by climate
zone. The seasonal throughput percentages assigned to each climate zone are listed
in Table 9-7. Zone five is the warmest zone, so all the activity was placed into the
winter category. Summer has no activity with the NEI default method, and the
activity is distributed across the seasons for zones two, three and four.
Table 9-7. Apportionment for Woodstoves and Fireplaces with Inserts
Climate
Zone
Winter
Spring
Summer
Fall
5
100
0
0
0
4
70
15
0
15
3
50
25
0
25
2
40
30
0
30
1
33.33
33.33
0
33.33
9.1.2.4 Improving the NEI
One approach to improving on the NEI method is to conduct a local survey, or
allocating emissions within the seasons. It is preferable to use local data and the
preferred collection method is to do a local or statewide survey. The EIIP provides
an alternative method that uses the bureau census data and the EIA data method.
Any assumptions other than 100% for rule effectiveness and rule penetration should
be incorporated into the emissions estimation methodology since the NEI method
does not account for the effect of state and local rules. Finally, the residential wood
combustion section of the EIIP series (Chapter 2 of Volume III) contains
information on conducting a survey.
9.1.3 Comparison of the MANE-VU and NEI
The MANE-VU inventory is a bottom-up methodology and the NEI is top down.
MANE-VU provides better estimates by geographic area and census. It also
accounts for differences in housing type (single versus multi-family homes).
MANE-VU provides better estimates of usage patterns based on heating demand,
and it includes outdoor equipment not included in the NEI estimates. It also
provides some temporal data that can be used to allocate emissions.
Preparation of Fine Particulate Emissions Inventories
9-10
-------
9.2 Residential/Land Clearing Debris Burning
9.2.1 Residential Open Burning
Residential open burning includes household waste burning and yard waste
burning, which includes brush waste and leaf waste.
9.2.1.1 NEI
Table 9-8 lists the SCCs and the pollutants for residential open burning that are
included in the NEI.
Table 9-8. Residential Opening Buring SCCs and Pollutants
Category
SCCs
Pollutants
Residential Municipal
Solid Waste Burning
2610030000
PM10, PM2 5, CO, NOx,
VOC, S02, 32 HAPs
Residential Leaf Burning
2610000100
PMio, PM2 5, CO, VOC, 6
HAPs
Residential Brush
Burning
2610004000
PMio, PM2 5, CO, VOC, 6
HAPs
The first step in developing activity data for residential municipal solid waste is to
estimate the rural population by county by applying percentages of rural and urban
population from the census data. The second step is to multiply the rural
population by a per capita household waste factor of 3.37 pounds per person per
day. Once the total waste generated is estimated, the amount of waste burned is
estimated by assuming that 28% of the household waste generated is burned. The
final step is to account for burning bans. Ideally this is done by knowing exactly
which areas have instituted a burning ban and the time period over which the ban
applies. However, the NEI assumes that if a county has an urban population that
exceeds 80% of the total population the amount of waste burned is zero.
The activity data for yard waste is estimated in a similar manner to household waste
using a per capita waste factor for yard waste generation of 0.54 pounds per person
per day. However, since different types of yard waste materials have different
emission factors it is necessary to estimate the percentage of total yard waste that
corresponds to leaf, brush, and grass waste. The NEI assumed that 25% was leaf
waste, 25% was brush waste, and 50% was grass waste. The amount of waste
burned is estimated by assuming that 28% of the total leaf and brush waste is
burned and that 0% of the grass waste is burned. One additional adjustment is
made to the amount of yard waste burned to try to account for the variation in
vegetation among the counties. This is done by using an estimate of the percent of
Preparation of Fine Particulate Emissions Inventories
9-11
-------
the forested acres per county that was obtained from the biogenic emissions land
cover database from the Biogenic Emission Inventory System (BEIS) as shown in
Table 9-9. For example, if the BEIS data indicates that a county has less than 10%
forested acres, the NEI assumes that there is no yard waste generated.
Table 9-9. Vegetation Adjustment Values
Percent Forested Acres per County
Adjustment for Yard Waste Generated
< 10
Zero out
>=10 and <50
Multiply by 50%
>=50
Assume 100%
The final step is to account for burning bans in the same manner that was used for
household waste. Once the activity data is estimated, emissions are calculated by
the use of Equation 9-1. A 100% CE is assumed for counties that have an urban
population greater than 80% of the total population. The NEI also assumes that RE
and RP are 100% for these areas. The NEI assumes that all other counties are
uncontrolled.
Equation 9-1. Emission Estimation Formula for Household and Yard Waste
Burning
E = A*EF*(1-CE*RP*RE)
where: E = Controlled emissions (lbs pollutant/year)
A = Activity (tons of waste burned/year)
EF = Emission factor (lbs/ton waste burned)
CE = % Control efficiency/100
RP = % Rule penetration/100
RE = % Rule effectiveness/100
There is an EIIP document for open burning and it contains an alternative approach
for estimating emissions for yard waste. This approach involves obtaining records
of burning permits or violations and data (or assumptions) on typical volumes and
material composition.
9.2.1.2 Improving the NEI
The open burning EIIP (Volume III, Chapter 16) contains alternative methods for
estimating activity data for this category. Another approach is to use the NEI
methodology coupled with state or local estimates of the per capita waste
generation and the amount or percentage of waste burned. Also, state/local data on
the months when yard waste is burned would be an improvement since the NEI
does not make any temporal adjustment for yard waste burning. Some of the
sources for this type of information include the Solid Waste agency, the Air
Preparation of Fine Particulate Emissions Inventories
9-12
-------
Agency, the Health Department, the Solid Waste Management agency, and through
the use of local surveys.
The NEI can also be improved by obtaining better estimates of control measures
that are applied to open burning. This involves identifying the rules that limit or
prohibit open burning and the organization that enforces those rules (e.g., fire
marshal, health department). For areas that have burning prohibitions, a rule
effectiveness survey can be performed to estimate the compliance rate with the rule.
This is critical in rural areas where there are few complaints about open burning.
Also, rule penetration is critical since many open burning rules have exemptions
that are listed (e.g., firefighting training activities, recreational campfires). Rule
penetration is also important for seasonal bans.
9.2.1.3 MANE-VU Example
This example examines the development of a 2002 residential open burning
inventory for the MANE-VU states. This was developed by a multi-state Regional
Planning Organization and followed the procedures in the EIIP document (i.e.,
conducting a survey) to obtain activity data.
A survey instrument was developed to collect data on the number of households
that burn waste, the burn frequency, the amount burned, and the seasonal nature of
the burning. Three separate surveys were performed for residential municipal solid
waste, brush waste and leaf waste. The data collected from these surveys were used
to estimate emissions for each survey area and to estimate default activity data for
those areas not included in the surveyed areas.
Equation 9-2 shows the equation that was used to estimate the amount of waste
burned based on the data collected from the surveys.
Equation 9-2. Equation for Estimating Mass of Waste Burned
Wt = HH * Bt * M
where: Wt = Mass of waste burned per time period
HH = Number of households that burn
Bt = Number of burns per time period
M = Mass of waste burned
In addition to collecting data to estimate activity data, a control database was
developed that established area-specific control efficiency, rule effectiveness, and
rule penetration. Because rule effectiveness and rule penetration can vary
significantly depending on enforcement and the rule applicability, a rule
effectiveness survey was conducted to determine the level of compliance with the
state or local open burning prohibitions. This data was also used to estimate default
RE values for use in the non-surveyed areas.
Using the activity data and the control information, emissions were estimated for all
criteria pollutants and precursors as well as for several HAPs. The emissions were
estimated at the census track level and then summed to the county, state, and
Preparation of Fine Particulate Emissions Inventories
9-13
-------
regional level. Finally, the data on the occurrence of the burning activities were
used to temporally allocate the emissions to support modeling using profiles that
were developed from the survey.
A number of lessons were learned from conducting the survey including that
separate surveys should be performed in targeted areas where leaf burning is
significant. In addition, household waste and yard waste surveys should be
performed separately simply to reduce the length of the survey. Another lesson
learned is that a larger sample may have allowed for greater geographic distinction.
In addition, a regional survey provides greater consistency that allows for easier
comparison of emission estimates from different areas. Finally, better accounting
of controls results in a decrease of the NEI emissions.
9.2.2 Land Clearing Debris Burning
Land clearing debris burning is covered under SCC 2610000500. The NEI contains
emission estimates for PMio, PM2.5, CO, VOC, and 6 HAPs from this category.
9.2.2.1 NEI
The activity data for this category is the same as that used for the construction
category (i.e., the number of acres disturbed for the different types of construction
categories). A loading factor is applied to the number of acres disturbed to produce
an estimate of the amount of material that is being burned. Weighted county-
specific loading factors were developed based on the acres of hardwood, softwoods,
and grasses. The average loading factors (Table 9-10) are multiplied by the percent
contribution of each type of vegetation class to the total county land area. The
average loading factors for hardwood and softwoods were adjusted by an additional
1.5 to account for the mass of tree below the surface. It should be noted that the
emission factors presented in Table 9-10 reflect this adjustment.
Table 9-10 Fuel Loading Factors
Fuel Type
Fuel Loading
(tons/acre)
Hardwood
99
Softwood
57
Grass
4.5
9-3 shows the formula for developing the loading factors.
Equation 9-3. Equation for Estimating Fuel Loading Factor
Lw = Fh*Lh + FS*LS + Fg*Lg
Lw = County-specific weighted loading factor
Fh = Fraction of county acres classified as hardwoods
Equation
where:
Preparation of Fine Particulate Emissions Inventories
9-14
-------
Lh = Average loading factor for hardwoods
Fs = Fraction of county acres classified as softwoods
Ls = Average loading factor for softwoods
Fg = Fraction of county acres classified as grasses
Lg = Average loading factor for grasses
Emissions are estimated from the activity data as shown by Equation 9-4. This
formula multiplies the activity data, the number of acres of land, and the county-
specific loading factor. Since the loading factor does not vary by the types of
construction, the number of acres cleared for all three types of activities (residential,
commercial, and road construction) are summed before the loading factor is
applied. The NEI assumes that all the fuel loading on the land cleared is burned and
that no controls or bans are in place.
Equation 9-4. Equation for Estimating Emissions from Land Clearing Debris
Burning
E = A * LF * EF
where: E = Emissions (lbs pollutant/year)
A = Number of acres cleared per county
LF = County-specific loading factor (tons/acre)
EF = Emission factor (lbs pollutant/ton)
9.2.2.2 Improving the NEI
The NEI does not take into account data on burning practices or controls, so a good
place to begin to improve the NEI is to review the EIIP section on open burning.
The EIIP methods rely on a direct measure of mass of waste or debris burned,
which may be obtainable from local officials that track this activity for permitting
purposes. Also, obtaining a better estimate of the acres cleared for the fugitive dust
construction category would improve the inventory for the land clearing debris
burning category. Other approaches for improving the NEI include:
Developing an improved loading factor.
Identifying specific counties with burning bans.
Specifying counties where wastes are burned.
Obtaining state or local estimates of the percentage or amount of waste
burned per construction event (the NEI assumes that the fuel loading
associated with the land that is cleared is being burned).
9.2.2.3 Northern Virginia Example
This Northern Virginia area study involved a RE survey to determine the level of
compliance with rules for land clearing debris burning and residential waste
burning. The objective of the study was to develop a defensible RE value for use in
the State Implementation Plan.
The study reviewed the existing conditions of the open burning rules to determine
the time period of the ban and the exemptions that apply. A survey of local open
Preparation of Fine Particulate Emissions Inventories
9-15
-------
burning officials responsible for tracking and enforcing open burning rules was
conducted. The survey form was derived from an EPA questionnaire that is
available from the rule effectiveness guidance. Responses to the questions on the
survey were assigned a specific point value that adds up to a maximum of 100
points. This point value is considered equivalent to the RE percentage value. If all
the questions were answered with the highest rating, an RE value of 100% was
assigned. The RE values were analyzed by county as well as for the five-county
region and a regional RE value of 93% was estimated. Although not done in this
case study, separate RE values could be developed for urban and rural area in cases
where there are significantly different population densities.
Some of the lessons learned from this study are that the local officials tend to defer
to the county or state level officials for enforcing the open burning rules. Also, in
developing an annual emissions inventory, it is important to note that RE may be
high only for the time period that the ban is in effect. In this case, the duration of
the ban (RP) needs to be taken into account if it is less than annual or seasonal.
Also it is important to account for when the ban is taking place and if it overlaps
with when the activity occurs. For example, a ban in place for the summer months
for brush waste burning will have minimal impact if the majority of the brush
burning occurs in the fall.
9.3 Agricultural Field Burning
9.3.1 Introduction
Agricultural burns create particulate matter pollution and their inventory is
important to the overall particulate matter air quality analysis. The SCC for
agricultural burning is 2801500000 and EPA encourages States to inventory both
PM10 and PM2.5-PRI. Since agricultural burning is a combustion process, both
condensibles and filterables are included in the PM-PRI estimate.
EPA develops emission estimates for most source categories in the NEI and then
the States submit any improved information that they have for those particular
categories. However, for agricultural burning EPA does not at this time prepare an
estimate of emissions from agricultural burning. In this case EPA encourages each
State to develop their own inventories and submit them. In 1999 ten States
(Alabama, California, Delaware, Georgia, Idaho, Kansas, Maine, Oregon, Texas,
and Utah) developed their own agricultural burning inventory. In general, these
States developed the inventories by characterizing the activity or acres of the crop
burned, the loading factor, the ton of biomass of vegetation per acre burned, and the
emission factor in terms of pounds of PM2.5 per ton.
9.3.2 Wheat Stubble Burning Example
This study involves wheat stubble burning and uses county-specific data. The
activity data that was obtained are the acres of wheat burned by month. This was
obtained from burn permits that are usually issued by the county fire department.
Preparation of Fine Particulate Emissions Inventories
9-16
-------
Also, the fuel loading for wheat stubble was obtained from the county agricultural
extension office. The emission factors are from a study done by CARB (Jenkins,
B.M. et al., Atmospheric Pollutant Emission Factors from Open Burning of
Agricultural and Forest Biomass by Wind Tunnel Simulations, Volume 2, Results,
Cereal Crop Residues, California Air Resources Board Project Number A932-126).
The emission factors are 8.82 pounds per tons of wheat stubble burned for PMio and
8.34 for PM2.5. The spatial resolution for this inventory is the county and the
temporal resolution is monthly. Equation 9-5 shows the formula for calculating
PM2.5-PRI emissions. This calculation would be repeated for each month during
the burning season and summed to give an annual emissions estimate. It should be
noted that if the number of acres burned per fire is larger than 100 acres; the
specific latitude and longitude of the fire should also be obtained.
Equation 9-5. Equation for Estimating Emissions from Agricultural Burning
E = A * LF * EF
where: E = Emissions (lbs pollutant/month)
A = Number of acres burned per month
LF = Loading factor (tons/acre)
EF = Emission factor (lbs pollutant/ton)
9.3.3 Improvements
EPA encourages all states to develop their own agricultural burning inventory. For
fires larger than 100 acres EPA suggests that they be located at a specific latitude
and longitude and the stop and start date and time of the fire be recorded. Smaller
fires should be lumped into overall monthly acreage like in the previous case study
example. Obtaining information on agricultural burning requires coordination with
the burners and the permitting authorities. In order to develop an agricultural
burning inventory, states needs to build a system and a relationship with the burners
and permitting authorities. Chances are pretty good that the first time a State tries
to obtain this information they will find that records are not kept or are not kept in a
way that can easily be understood.
The local acres of crops burned are obtained from burn permits or from a survey of
county agricultural extension offices or perhaps a combination of both. It is
important that States verify that the burns actually occurred. Often a burner will get
a permit to burn a lot more acreage than they actually are able to burn in a particular
day. In many cases a burner is limited by the weather or other factors that keep
them from burning the acreage that they are permitted to burn. Finally, States need
to obtain local fuel loading data. This is preferably obtained from the local county
agricultural extension office or the local Natural Resources Conservation Service
Center. This is highly preferable to using the national defaults that are available in
Chapter 2.5 of AP-42.
Preparation of Fine Particulate Emissions Inventories
9-17
-------
9.4 Wildland Fires
Fires have become a major issue in both visibility impairment and in creating high
concentrations of PM2 5 that could result in health problems. The problems have
been mainly in the West, but also wildfires from the Southeast, the Central States,
Canada, and Mexico have become a concern. EPA's wildland burning inventory
includes both wild and managed burns. The typical agencies that burn are the
National Park Service, the United States Forest Service, the Bureau of Land
Management, the United States Fish and Wildlife Service, State & Tribal Forests,
and private burners. Wildland fires are categorized into two types: wildfires and
managed or prescribed burns. Prescribed burns are those burns that are ignited
intentionally for habitat improvement of the wildlife; for managing the overall
under growth and understating of the forest; and to reduce the risk of wildfires later
on by removing the fuels from the forested area.
9.4.1 NEI
It should be noted that this discussion focuses on the technique for estimating
emissions from wildfires; however, emissions from prescribed or managed fires are
estimated in a similar fashion. The pollutants that are included in the NEI inventory
for wildland fire emissions are PMi0, PM2.5, NOx, CO, VOC, S02, and about 30
HAPS. The emissions factors for estimating fire emissions, and the state-specific
fuel consumed per acre burned are found in the NEI documentation. The technique
is to merge the factor and fuel consumption information with annual activity data
obtained at either the state or regional level from the main burning agencies. Most
of the federal burners keep fairly good records of the burns that they conduct mostly
because these fires end up being watched and/or fought by personnel. Some states
also provide burn data as do some private burners.
The data obtained from the burners is at the state level or regional level and it is
allocated to the state or county level using the amount of forested area in a state. In
other words, since the NEI does not have data on the exact location of the fires, the
amount of acreage that was burned during a year in a particular state is allocated
across the state to the forested lands.
The NEI allocates the emissions annually and the emissions processor allocates the
emissions diurnally and monthly. This allocation is important because certain areas
of the country have different fire seasons and fire seasons are different for
prescribed burns and managed burns.
9.4.2 Improving the NEI
In order to improve wild land fire emissions, national and regional databases and
models must be improved. Fires need to be treated as events (i.e., specify the area
burned, when it was burned, and where it occurred). In addition, large fires need to
be entered into the databases as point sources with a particular location (lat/long)
Preparation of Fine Particulate Emissions Inventories
9-18
-------
and a start date, end date, and the time of day. National regional models and
databases need to be developed and refined to improve the pre-burn fuel loading
information. The information in AP-42 is very general, very dated, and averaged
over large regions of the country. Finally, the use of fuel consumption models
needs to be to refined and expanded and guidance on estimating the impact of
mitigation measures on emissions needs to be provided.
There is a Memorandum of Understanding (MOU) in effect between the EPA,
Department of Interior, and the United States Department of Agriculture to develop
a fire events database. It is a broad scope MOU that covers fire management
activities including ways to improve the national databases. There is a similar
effort (NEISGEI) being conducted at Washington University in St. Louis. There
currently exists a database for recording fire events in the Pacific NW called the B-
RAINS system. Although these types of projects are moving toward real time data
collection, quality assurance and data sharing, there is much more work needed in
these areas.
EPA is also investigating the potential use of satellites to improve wildland fire
inventories. EPA has funded a report entitled Overview of Using Satellites in AQ
Management. There is also collaboration going on with NASA to take advantage of
their skills in aerial surveillance with satellites. There are several interagency
groups working on the use of satellites including the National Interagency Fire
Center (a jointly funded effort of all the Federal burners) in Boise, Idaho, the
Missoula Fire Research Center, and Salt Lake City. Another project includes
CAMFER, which is a project underway at University of California Berkeley.
9.4.3 Emission Estimation Tools and Inventories
EPA recently published a report entitled Fire Emission Estimation Methods
(available on the CHIEF web site) that contains a lot of good background
information on wildland fire emission estimation. In addition, there is a lot of
ongoing work to improve emission estimation tools for wildland fires. The US
Forest Service has ongoing work on the development of fuel consumption and fire
behavior models at the Fire Sciences Lab in Missoula and also at the Pacific NW
Research Station in Corvallis. Also, there is also a lot of emission factor testing
occurring in the Fire Sciences Lab in Missoula.
There is also collaboration going on between all the different burn agencies, EPA,
and the Regional Planning Organizations (RPO). The Western Regional Air
Partnership (WRAP) conducts a fire emissions joint forum and EPA and the burn
agencies participate in that forum. There is a RPO project to refine the 2002
wildland fire emissions inventory. There was a national fire emissions workshop
held in May of 2004 that focused on the latest ideas and methodologies for
estimating fire emissions. Also, the US Forest Service with assistance and funding
from EPA is developing a geographic coverage of the fuel types and fuel conditions
for burning at a 1km resolution. A map of the country that will be useful in GIS
Preparation of Fine Particulate Emissions Inventories
9-19
-------
systems will be developed out of this project. Finally there will be further work on
developing an emissions model that will estimate fire emissions in real time using
real time meteorological data. Output from this model will be fed directly into the
grid models for estimating ambient air concentrations associated with fire
emissions.
The emissions model that is under development is the Wildlands Fire Emissions
Model. It will interface with SMOKE and OpEMs (the emissions model that is
under development by the RPO), and the CMAQ modeling system. The user will
need to input fire locations, durations, and size of the fire (i.e., the blackened area of
the fire). The model components, which will be drawn from the Blue Sky system
being developed in the Pacific NW, are:
1. A fuel loading default that will use either the national fire danger rating
system or, as it becomes available, the FCC map.
2. Fuel moisture will be calculated using actual metrological data for the
period during, and immediately before the fire. This is a significant
improvement over the past and an important improvement since fuel
moisture is critical in determining the amount of fuel that will burn and the
emissions from that fuel.
3. Fuel consumption models are being built into the model. Both the
CONSUME / FOFEM are such models that have recently been improved
significantly. The CONSUME model is developed in the Corvallis lab and
the FOFEM has been developed by the Missoula Fire Lab. These models
compliment each other and have strengths and weakness that, when used
together properly, give a pretty good handle on fuel consumption.
4. The emission heat release and plume rise is being handled through the EPM
model and the modified Briggs plume rise equation. There is an
improvement to the EPM model called FAR, which is about to be released
in beta test form.
The output of the model will be a gridded hourly emission estimate and plume
characteristics. The output will be able to be interfaced with grid models to provide
a regional scale estimate of the effects of fires. For instance, this new wildland fire
model will be able to estimate the NOx plume from a wildland fire and the effects
of that increased NOx on ozone formation. The integration, testing, and release of
the model are anticipated for late 2004.
Preparation of Fine Particulate Emissions Inventories
9-20
-------
Review Exercises
1. The NEI methodology for residential wood combustion made adjustments to the national
number of fireplaces to account for .
a. fireplaces that burn gas
b. fireplaces without inserts
c. fireplaces used for aesthetic purposes
d. All of the above
2. Which type of residential wood combustion is not allocated to different SCCs?
a. woodstoves
b. fireplaces with inserts
c. fireplaces without inserts
d. All of the above
3. The NEI methodology for residential municipal solid waste burning assumes that if a
county has an urban population that exceeds percent of the total population, the
amount of waste burned is zero.
a. 50
b. 75
c. 80
d. 90
4. The NEI methodology for residential municipal solid waste burning assumes that
percent of household waste generated is burned.
a. 18
b. 28
c. 38
d. 48
5. The land clearing debris burning load factors for are adjusted by an additional
1.5 to account for the mass below the surface.
a. hardwoods and grasses
b. softwoods and grasses
c. hardwoods and softwoods
d. All of the above
6. The activity data for land clearing debris burning is the same that is used for the
category.
a. agricultural burning
b. unpaved roads
c. agricultural tilling
d. construction
Preparation of Fine Particulate Emissions Inventories
9-21
-------
7. Which of the following variables are not used in the NEI to estimate emissions from land
clearing debris burning?
a. number of acres cleared
b. county-specific loading factor
c. emission factor
d. rule effectiveness
8. For which of the following categories did the NEI not develop a methodology?
a. agricultural field burning
b. agricultural tilling
c. wood stoves
d. land clearing debris burning
9. Which of the following is not a source of variability in wood consumption activity in the
MANE-VU study?
a. type of housing
b. heating degree days
c. moisture content of wood
d. availability of wood
10. In estimating emissions from wildland fires, the data obtained from the burners is
allocated to the state or county level using .
a. the number of burn permits issued
b. the number of acres burned
c. the amount of forested land in a state
d. All of the above
Preparation of Fine Particulate Emissions Inventories 9-22
-------
Review Answers
1. d. All of the above
2. c. fireplaces without inserts
3. c. 80
4. b. 28
5. c. hardwoods and softwoods
6. d. construction
7. d. rule effectiveness
8. a. agricultural field burning
9. c. moisture content of wood
10. c. the amount of forested land in a state
Preparation of Fine Particulate Emissions Inventories
-------
THIS PAGE INTENTIONALLY LEFT BLANK
Preparation of Fine Particulate Emissions Inventories 9-24
------- |