1. Preparation of Fine Particulate Emissions Inventories
Chapter 2 - The National Emissions Inventory System (EIS) and Emission Inventory Tools
After this lesson, participants will be able to:
explain the National Emissions Inventory and the process by which it was developed.
describe current and future emissions inventory preparation tools.

2. What Will We Discuss in This Lesson
Overview of the NEI and Development Schedule
Key Source Categories for PM2.5
Opportunities for SLT Input
EI Development Tools
Process-based Emission Models

3. The National Emissions Inventory (NEI) and Emissions Inventory System (EIS)?
The NEI is
EPA’s national database of information on sources of air pollution and their emissions
used for analysis, modeling and development of priorities and other improvements
has been undergoing several changes that we will address to a limited extent as appropriate
The EIS is
EPA’s new system for collecting, storing, and accessing the NEI
currently available to states to input facility data
will be available to accept emissions data input very soon
Contact: or your EPA EI Coordinator
A quick note about the NEI and the EIS:
The NEI is the database of information on source contact information, location, throughputs and other activity data and emissions.
Its relied upon to support modeling, analyses related to AQ Standards-setting and to support EPA policy-setting.
Its housed in a new system, the EIS which is just now coming on line.

4. What’s Up With The NEI?
New – Air Emissions Reporting Requirements (AERR) 40 CFR Part 51, published December 17, See Rule at:
Changes in reporting schedules and data exchange processes
New EPA data system (some element/format changes)
Changes in technical requirements, terminologies and procedures – more discussion as this course unfolds
PM2.5,filterable AND condensible are BOTH required
For specific details/variations pertaining to your state’s data submittal procedures consult with EPA regional office
Lesson 5 contains more discussion of the AERR
Submission of data to the NEI is required by the new Air Emissions Reporting Rule, published in late 2008.
Applies to States and some tribes, if they are treated “As States” under Tribal Authority Rule and approved TIP
AERR harmonized NOx sip call and CERR requirements
Note changes to reporting schedules and data exchange processes.
Check with your EPA Regional Office for details.

5. NEI Development Schedule
2002 – Year of Emphasis (all categories)
2005 – Updates for Mobile and Fires
Note: Fires were updated each year ( )
2008 – Year of Emphasis (all categories)
Due May 31, 2010
2011 – Next “cycle” Year of Emphasis
Then Dec 31, 2012 for CY 2011

6. Information Included in the NEI
National tabulation of emissions of PM10, PM2.5, SO2, NOx, Pb, Ammonia, and VOC
Point sources by lat-long: 52,000 facilities, each containing multiple emission points
Over 4,500 types of processes represented
Area (non-point) & Mobile by County
400 categories of Highway & Non-Road Mobile
Over 300 categories of Area sources
Annual emissions, start/end dates, stack parameters, seasonal emissions
Also in the NEI
HAP emissions for over 6,000 types of processes (not required by AERR but likely to be - after rule making)
NEI information includes:
data on 52,000 point sources by latitude and/or longitude, with over 4,500 types of processes represented.
approximately 400 categories of highway and non-road mobile sources
approximately 300 categories of area sources in the NEI
annual emissions (area and mobile sources are allocated by county)
dates that sources started or stopped operations
stack parameters
HAP emissions for over 6,000 types of processes

7. Evolution of EPA’s National Emission Inventory
This graph represents a timeline showing the evolution of the National Emissions Inventory.
First NEI PM inventory National Acidic Precipitation Assessment Program, was preceded by NEDS.
Represented inventory for PM10 and was developed without input from the states (states have become involved in NEI development)
Early 1990s, minimal activity on developing PM inventories; National Particle Inventory was prepared in 1993.
In 1996 it was called the National Emissions Trends Inventory.
The NET was updated in 1999 and was renamed the NEI.
Integration of the National Air Toxics Assessment Inventory began with the 1999 NEI and was completed with the 2002 NEI.
Improving and updating the NEI is a continuing process. For example, the EIS is the new repository for the NEI
Notes: The NEI will soon be housed in EPA’s EIS.
NAPAP was preceded by NEDS in the early 1970’s.

8. Emissions Processor (eg SMOKE)
NEI Development ~ Cooperative
EPA
S/L/T
USERS
Preliminary
NEI for Base
Year 20XX
States &
Other
Stakeholders
Improved
Local
Activity
Levels &
Variables
Factor
and Model
Improvement s
Defaults for
Emissions-
Related
Databases
for Source
Levels
Existing
Point Source
Data
Growth
Factors for
Some
Categories
Emission
Factors and
Models
EPA Selects Data for “EPA Version” of Inventory
(Process Repeated Yearly for “A” Sources - Comprehensive every 3 Years)
NEI starts w/ prior NEI &
US EPA-prepared updates
Emissions Processor (eg SMOKE)
Inventory Used for Many
Purposes, e.g., Modeling
State / Local / Tribe Improve
and Supplement EPA Version
Speciation
factors
Multiple
Versions
Possible
The process of developing the NEI is the prior NEI and updates prepared by US EPA. At its heart are
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 some source categories.
This data is combined to form what is called the preliminary NEI, which is provided to state, local and tribal air agencies for refinement and improvement.
The preliminary NEI becomes the improved NEI by the use of improved EF’s and Emission Models and activity inputs
This is where State/local/tribal input is used by the EPA.
The S/L/T’s are involved in this process every year for type “A” sources, and a comprehensive EI every three years
Multiple versions are possible.
Emissions Processing is a step where the EI is prepared for use in models – more about that later.
Now, lets talk about how SLTs make the EI better.

9. Typical Source Categories of Primary PM2.5 Emissions
Fugitive Dust Sources (Crustal PM2.5)
Construction
Residential housing, commercial bldgs, roads
Mining and quarrying
Unpaved and paved roads (incl anti-skid materials)
Agricultural tilling
Beef cattle feedlots
Windblown dust
Note: These categories emit only “filterable” emissions
This list represents typical area source categories of fugitive dust sources of PM2.5 emissions.

10. Typical Source Categories of Primary PM2.5 Emissions (cont’d)
Open Burning Sources (Primary Carbonaceous PM2.5 )
Structure fires
Wild and prescribed fires
Agricultural field burning
Open burning
Residential municipal solid waste burning
Yard waste burning
Land clearing debris burning
Note: These categories emit both “filterable” and “condensable” emissions
This list contains typical area source categories of open burning sources of primary PM2.5 emissions.
Note, these categories emit both filterable and condensable PM2.5

11. Typical Source Categories of Primary PM2.5 Emissions (cont’d)
External/Internal Fuel Combustion (Carbonaceous PM Fine):
Residential wood combustion
Other residential fuel combustion
Industrial fuel combustion
Commercial/institutional fuel combustion
Note: These categories emit both “filterable” and “condensable” emissions
This list shows typical area source categories of primary PM emissions.
Note, these categories emit both filterable and condensable PM2.5

12. Typical Source Categories of Primary PM2.5 Emissions (cont’d)
Wildland and Agricultural Fires
Prior to 2002
Treated as area sources
Useless for real planning/impact analysis
Beginning in 2002, NEI treats fires as “Events” in time, space
Western states were leaders in 2002 NEI upgrade
EPA led development of new NEI “Events” category
More about this in Lesson 9
Wildfires:
2002 NEI began to include large fires as point sources for some areas of the United States.
Data on when they started and ended, and the location are essential to accurately model impact on air quality.
Smaller fires may continue to be treated as area sources and are allocated to county by using forested land area.
Emissions are assigned to months of the year using temporal allocation factors.
Treating fires as point sources is important. For example:
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 or when the fire occurred.
Consequently, it is impossible to determine if the particulate matter emissions in the Class 1 area are attributable to the fire.

13. Typical Source Categories of NH3 Emissions
Typical source categories of NH3 emissions include:
Animal husbandry
Agricultural fertilizer application
Agricultural fertilizer manufacturing
Wastewater treatment
Autos, Fires
This list presents typical area source categories of ammonia sources.

14. Specific Opportunities for Input from Federal / State / Local / Tribes
Industrial point sources
Fugitive dust
ONLY IF important part of strategy
(Check ambient measurements)
Residential Open Burning
Household waste, yard waste (volumes & burning practices)
Regulations & their effectiveness, local surveys of burn activities
Several specific PM2.5 combustion categories need better emissions models and emissions data.
Includes wildland burning of forests and rangeland and agricultural burning.
Data on location, dates, acreage burned and fuel loadings for the larger fires are needed.
Fugitive dust may or may not need scrutiny.
Another source category that may need better emissions data is residential open burning (household and yard waste).
Data on the volumes, burning practices, regulations and their effectiveness, and local surveys of burn activities are needed.   

15. Specific Opportunities for Input by Federal / State / Local / Tribes (cont'd)
Residential Wood Combustion
Fireplaces, Wood Stoves
local surveys of fuel burned, fireplace vs. wood stoves, wood stove types, local regulations
Data from local surveys of fuel burn are needed for residential wood combustion, fireplaces and wood stoves.
This includes data on whether the wood is being burned in a fireplace or a wood stove.
Saschamarkus 2009

16. Specific Opportunities for Input by Federal / State / Local / Tribes (cont’d)
Construction Debris & Logging Slash
Regulations & their effectiveness, local surveys of burn activities
Wildland and Agricultural Burning
Forests, rangeland & especially private & state / tribal burners
(acreages burned, fuel loadings for largest fires, dates, locations)
Data on the effectiveness of regulations for construction debris and logging slash are needed. This includes local surveys of burn activities.
For agricultural field burning, data on acreages, fuel loadings, and timing of the burn events are needed. We’ll talk about this more tomorrow

17. Questions
Kyle & Kelly Adams 2006

18. Overview: Inventory Preparation Tools
PM Activity Data
Emission Models
Process-based Emissions Models
Spatial Characterization & Location
Now we’ll talk about six tools available for EI Prep…
Emission Factors
PM Calculator & Activity Data
Emission Models
Process-based Emissions Models
Spatial Characterization & Location
Receptor Models

19. Inventory Preparation Tools
Activity Data
Area source activity data often available
But, in “obscure” local or internet sources- search engine time needed!
Trade Associations
Federal Agencies
County Gov’ts
More discussion of the NEI, EIS, CHIEF, AP-42, & Point Source EI Development Tools in Lesson 5
The PM Calculator, Version 2.0, is applicable to point sources only and requires the user to input uncontrolled emissions (either total filterable particulate or filterable PM10) for each source, the source category classification (SCC) and the control device, if any. The program will then calculate controlled emissions for filterable PM2.5 and filterable PM10 for each point source.
Note: The calculator is no longer supported by EPA, but it remains on the CHIEF website.
Activity data for nonpoint and mobile sources is often available on the web,
may require some intensive searching and followup telephone calls.
you might need to gather data locally, using local databases or surveys.

20. Inventory Preparation Tools (cont’d)
Emissions Models
TANKS
NONROAD
Ammonia (under dev by CMU & UC-Davis)
and
Fires (from Satellites & Recordkeeping)
Windblown Dust (under dev by ORD & WRAP)
Contact
Some of the inventory preparation tools are actually emissions models:
Process-based emission models.
TANKS—a model used for estimating storage tank emissions of VOCs.
NONROAD—a model used to estimate emissions from non-road vehicles.
BEIS—a model used to estimate biogenic emissions.
Ammonia – CMU Model available, Process-based under development at UC-Davis
Fires – SMARTFIRE
Windblown Dust – WRAP? And ORD
Potential for Fugitive Dust & Res wood

21. Inventory Preparation Tools (cont’d)
Process-based Emission Models
“emissions models on steroids”
Use realistic location & time relevant values (when possible) for e.g.,
wind, temperature
RH, vegetation types & moisture
soil type & moisture
Possible linkages:
MM5 &/or Weather Research and Forecast (WRF) model
GIS coverages
Can estimate emissions “on-the-fly” in Emissions Processor
Currently ~ BEIS3 only true process model
Process-based emission models models consider space- and time-sensitive emissions in an effort to reflect real world conditions such as:
wind,
temperature,
relative humidity,
vegetation type,
soil type, and moisture.
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.
Another model under development will estimate fire emissions by considering relative humidity, moisture, and wind speed.
These models would link to various databases such as the meteorological data processor (MM5). They would also link to GIS coverage of soil and vegetation types, and would contain emission algorithms that respond to these variables.
BEIS 3 model contains some aspects of process-based emission models, such as temperature. Consideration of temperature is critical for estimating biogenic emissions
Estimates “on-the-fly” in the Emissions Processor.

22. Inventory Preparation Tools (cont’d)
Process-based Emissions Models (cont’d)
Specific examples (some under development):
MOVES
Fires
Location-specific fuel characteristics
Research version estimating fuel moisture
Ammonia (under development)
temperature-driven release, moisture
There are a number of other variables that could be included in process-based emission models.
Some examples are listed here and include readily available data such as soil moisture, temperature and heating degree days
Real potential to improve the EI
Keep abreast of literature for release dates

23. Inventory Preparation Tools (cont’d)
Process-based Emissions Models (cont’d)
Wish List
Fugitive Dust (under development)
e.g., soil moisture, tilling schedule/type, nearby vegetation
Residential Wood Burning (why not?)
temperature or HDD
Evaporative Loss (why not?)
Temperature
There are a number of other variables that could be included in process-based emission models.
Some examples are listed here and include readily available data such as soil moisture, temperature and heating degree days
Real potential to improve the EI
Keep abreast of literature for release dates

24. Inventory Preparation Tools (cont'd)
Spatial Characterization & Locator Aids
Satellites
Google Earth (GE)
GIS
Example use of GE to support NAAQS
Confirmed locations of key Lead point sources
Many mis-located – ¼ mile to 7000 miles
Usefulness limited in large facilities
Confirmed monitor proximity to major highways
Other tools include special characterization and locator aids such as, Geographic Information Systems and Global Positioning Systems.
In Mexico and Canada, satellites are being used to locate fires.
Satellites are limited in their ability to locate certain fires, such as those below a certain size or masked by cloud cover.
Satellites have been used to characterize landscapes and land use, locate point sources, document windblown dust events and fire events and document areas surrounding AQ monitors.
GE useful to QA locations, count stacks, see population exposure, dispersion obstructions
GIS to locate facilities, stacks

25. Emissions Processing (EP)
Overview
EP’s prepare traditional EI (e.g., NEI) for grid models
Reformatted
Spatially
Temporally
Chemical species
SMOKE
Emissions processing (EP)
EP is needed to reformat and spatially and temporally and chemically resolve the EI for input to models.
After developing NEI emissions data, it is processed by an EP modeling system, such as the Sparse Matrix Operator Kernel Emissions (SMOKE) system or CONCEPT. CONCEPT is open source ~ potentially more transparent, but needs more documentation.
The EP applies factors to the emissions data when the inventory is to be used by air quality dispersion modelers.
Emissions modeling depends on speciation factors, temporization factors, and species allocation factors.

26. Emissions Processing (cont’d)
Input
“Reformatted” point source EI
Mobile Source EI
Annual, county-level area source EI
Annual point & event data (except for CEM data)
Outputs
Gridded, hourly emissions file
EC, OC, SO4, Nitrates, PM-Other/PM-Fine
Model-ready
“Factor-driven” manipulation of “input data”
Emissions processing (EP)
The data flow is from the NEI to the emissions processor and then into the air quality model.
It begins with the (usually) annual EI in a special format
Area source data is input to the emissions processor as an annual county level inventory.
Point source data is input as annual data, located by latitude and longitude.
CEM data feeds into the emission processor through a separate database.
The output is a gridded, hourly emissions file, speciated into elemental carbon, organics, primary sulfates, and primary nitrates, ready for modeling’
Outputs are obtained thru application of numerous “factors”
Factors are used to drive this allocation process.

27. Emissions Processing (cont'd)
“Factors” in the Emissions Processor
Spatial Allocation Factors
Temporal Allocation Factors
Speciation Profiles
Emissions processors contain default factors and profiles but these can be overridden with locally derived factors:
county to grid allocation factors
temporal allocation profiles
speciation profiles
The emissions processor transforms annual or county level inventory into a gridded, hourly emission file.
The data is speciated into EC, POA, Primary SO4, Primary Nitrate and Other (crustal materials/fugitive dust and unidentified species).
It is then ready to be used as input to a dispersion model.
Micky 2009

28. Emissions Processing (cont'd)
Spatial/Temporal Allocation Factors
County-to-Grid Spatial Allocation Factors
Temporal Allocation Profiles (hourly & seasonal)
Examples
Census tracts
Gridded land use
Vegetation “coverages”
Operating schedules
Ambient temperature
The emissions processor uses county to grid allocation factors to transform annual or county level inventory into a gridded emission file.
It also uses typical temporal emissions distributions to allocate the annual EI into daily and hourly emissions.
Discuss examples

29. Emissions Processing (cont'd)
Speciation Profiles (“factors”)
Estimate of the EC, OC, SO4, NO3, Crustal, and OC “Matter” portions of each PM2.5 source’s emissions
All PM2.5 sources assigned to 1 of ~ 70 “profiles”
Note: Speciated composition of PM2.5 is stored with modeling files, NOT in the NEI
Speciation profiles recently updated (Reff et al)
Speciate 4.2
Basis of Speciation profiles
Update by E.H. Pechan & Associates, Inc.
Emissions processor default to a set of ~ 70 speciation profiles to parse the PM-2.5 into its key components:
EC, Oc, SO4, no3, CRUSTAL and OC “Matter”
speciation profiles derived from SPECIATE 4.2

30. Emissions Processing (cont’d)
What is SPECIATE?
An EI Tool that derives OCM, EC and Crustal from PM2.5
Database of emissions profiles by species for each of ~ 70 source categories
Includes metals, ions, elements, organic and inorganic compounds, in consistent units
Three categories in SPECIATE:
Particulate matter (PM), VOC, Other gases
(e.g., Hg, NO/NO2/HONO, semi-volatile organics)
Housed in Microsoft Access database
Regular updates
Contact:
SPECIATE is a database of speciation factors by source category.
It derived an estimate of the OC, EC and Crustal components from
contains both separate test results and composites that may be used as the speciation profiles.
Each SCC is assigned or “mapped” to one of these composites.
These “mapped” profiles are then used by the EP to estimate the chemical composition of the emissions.
A similar process is done for VOC.
What does a speciation profile look like

31. Emissions Processing (cont’d)
Example PM-2.5 Speciation Profile ~ Rice Burning
Here is an example of a composite speciation profile for Rice & wheat straw burning
It was composited from 78 separate tests.
Both the 78 tests and the composite are found in SPECIATE
Note it has some metals and ions as well as the usual EC, OC, SO4 and NO3.
Why can a profile sum to less or more than 100%?
imprecision, failing to analyze all components

32. Summary The NEI NEI depends on local input
How it started, what is in it, how it is developed, tools
NEI depends on local input
Very important for some categories
“Local opportunities” to improve the EI
Process-based Emissions Models
More accurate, especially for daily EI
Emissions Processing
Making the EI “model ready”
Lets review what we’ve just discussed
Has its roots in the early NEDS and NAPAP EI’s
Includes 10’s of 1000’s of sources – criteria and HAP pollutants
Needs input from F/S/L/T
Fires and Fug dust – 2 special cases (fires now as Events, Dust near-source removal which we’ll discuss later)
Prep tools
Process-based models

33. Questions?
Richard0