Evaluating PM and Precursor Emission Inventories

1
Evaluating PM and Precursor Emission Inventories

• Background on PM and precursor emissions
• Sources of PM and precursor emissions
• Estimating man-made PM emissions
• The importance of emission inventory evaluation
• Emission inventory evaluation methods and tools

• Experience with the 1990 State Implementation
  Plan (SIP) base year (emissions) inventories
  brought to light deficiencies and inconsistencies
  in the inventory process now being used. In
  addition, the current leeway in selecting these
  processes has resulted in data sets of unknown
  quality and varying degrees of completeness.
• - Emission Inventory Improvement Program, 1997

2
What is Particulate Matter and How Does it Vary?
Husar, 1999

• What is Particulate Matter?
• How Does PM Vary?

3
What is Particulate Matter?

• What is particulate matter and where does it come
  from?
• The term Particulate Matter or aerosol, refers to
  liquid or solid particles suspended in the air.
  Depending on their origin and visual appearance,
  aerosols have acquired different names in the
  everyday language. Dust refers to solid airborne
  material, dispersed into aerosol from grainy
  powders such as soil. Combustion processes
  produce smoke particles, but the incombustible
  residues of coal are called flyash.
• Particles of size 2.5 um and smaller are referred
  to as PM2.5. PM2.5 is composed of a mixture of
  primary and secondary compounds. Primary
  particulate PM2.5 compounds consist of
  soil-related, inorganic, elemental, and organic
  carbonaceous particles emitted directly into the
  atmosphere by fossil fuel combustion, biomass
  combustion, and mechanical processes. Secondary
  particulate compounds are formed in the
  atmosphere via chemical reactions of particulate
  precursors (VOC, SO2, NOx, SOA, and NH3) emitted
  by many combustion and non-combustion sources.
  The principal types of secondary particles are
  ammonium sulfate and ammonium nitrate formed in
  the atmosphere from gaseous emissions of SO2 and
  NOx reacting with NH3.
• In the early days, air pollution had the
  appearance of both smoke and fog, so the term
  smog was created. In the open atmosphere, the
  visibility may often be reduced by regional haze,
  originating from various natural or anthropogenic
  sources. Neither water droplets of fog and
  clouds, snow, rain, sleet (hydrometors) nor dust
  particles larger than 100 um (blowing sand) are
  considered to be particulate matter.
• How does PM vary spatially, temporally, with
  particle size, and by chemical composition?
• As all pollutants, the ambient aerosol
  concentration patterns contain endless
  variability in space and time. However, unlike
  gaseous pollutants, particulate matter also
  depends on particle size, shape and chemical
  composition.
• The chemically rich aerosol mix arises from the
  multiplicity of PM sources, each having a unique
  chemical signature at the source. These chemical
  signatures are referred to as chemical speciation
  profiles. The primary aerosol chemical
  composition is further enriched by the addition
  of secondary species during atmospheric
  transport. The effective mixing in the lower
  atmosphere stirs these primary and secondary
  particles into an externally mixed batch with
  various degrees of homogeneity, depending on
  location and time. Lastly, repeated cloud
  scavenging and evaporation tends to mix the
  particles from different sources internally into
  particles with mixed composition.
• The result is a heterogeneous PM mixture that is
  probably unparalleled in the domain of
  atmospheric sciences. For instance it is common
  to find soot particles within sulfate droplets,
  or nitrate deposited on sea salt particles.

Husar, 1999
4
PM2.5 Background and Terminology (1 of 3)

• The term Particulate Matter (PM) or aerosol,
  refers to liquid or solid particles suspended in
  the air PM2.5 refers to particles of size 2.5 ?m
  and smaller.
• Dust refers to solid airborne material, dispersed
  into aerosol from grainy powders such as soil.
• Combustion processes produce smoke particles, but
  the incombustible residues of coal are called
  flyash.
• In the open atmosphere, the visibility may often
  be reduced by regional haze, originating from
  various natural or anthropogenic sources.

Husar, 1999
5
PM2.5 Background and Terminology (2 of 3)

• PM2.5 is composed of a mixture of primary and
  secondary compounds.
• Primary particulate PM2.5 consists of
  soil-related, inorganic, elemental, and organic
  carbonaceous particles emitted directly into the
  atmosphere by fossil fuel and biomass combustion,
  and mechanical processes.
• Secondary particulate compounds are formed in the
  atmosphere via chemical reactions of particulate
  precursors (VOC, SO2, NOx, SOA, and NH3) emitted
  by many combustion and non-combustion sources.

6
PM2.5 Background and Terminology (3 of 3)

• The principal types of secondary particles are
  ammonium sulfate and ammonium nitrate formed in
  the atmosphere from gaseous emissions of SO2 and
  NOx reacting with NH3.
• There is a direct relationship between the
  particle size and the atmospheric residence time
  of particles

7
Urban Sources of PM2.5 and Precursor Emissions

• Motor vehicle exhaust
• Paved road dust
• Fossil fuel fired boilers at electric utility
  plants
• Residential wood combustion
• Entrained geologic material

8
Characteristics of PM (1 of 2)

• Many sources of PM are seasonal (e.g.,
  residential wood combustion, dust sources, forest
  fires).
• Meteorology impacts the formation of secondary
  PM.
• The formation of secondary PM involving OH, O3,
  and H2O2, species which are normally present in
  the atmosphere but which are present in higher
  concentrations during periods of increased
  sunlight and temperature, peaks during the summer
  in most U.S. areas.
• The formation of secondary PM can be affected by
  the water content of the atmosphere higher water
  content facilitates the formation of certain
  secondary particles, consequently, PM also tends
  to be a problem in the winter.

EPA, 1996
9
Characteristics of PM (2 of 2)

• Primary and secondary PM2.5 have long lifetimes
  in the atmosphere (days to weeks) and travel long
  distances (hundreds to thousands of kilometers).
• PM2.5 particles tend to be uniformly distributed
  over urban areas, as a result, they are not
  easily traced back to their individual sources.
• Natural sources of background PM include wind
  blown dust from erosion and re-entrainment long
  range transport of dust from the Sahara desert
  sea salt particles formed from the oxidation of
  sulfur compounds emitted from oceans and
  wetlands the oxidation of NOx from forest fires
  and the oxidation of VOC compounds emitted by
  plants and trees.

EPA, 1996
10
PM and Precursor Emission Inventories

• Purpose of PM emission inventory development
• Used by regulatory community
• Air quality modeling support (model input)
• Exposure modeling support, health assessment
• Control cost analysis
• Regulatory control strategy development

The Clean Air Act requires state and local air
quality agencies to develop complete and accurate
inventories as an integral part of their air
quality management responsibilities. These air
emission inventories are used to evaluate air
quality, track emissions reduction levels, and
set policy on a national and regional scale... -
Emission Inventory Improvement Program, 1997
11
Emission Inventory Development (1 of 4)

• Estimating PM2.5 emissions is a complex process
  involving many data parameters.
• Uncertainties in emissions inventory estimates
  could range from about 10 for well defined
  sources (e.g., emissions from power plants) to an
  order of magnitude for widespread and sporadic
  sources (e.g., fugitive dust).
• General equation for estimating emissions gets
  complex when estimating PM
• E A x EF x (1-ER/100)
• Where
• E emissions EF emission factor
• A activity ER overall emission reduction
  efficiency ()

12
Emission Inventory Development (2 of 4)

• Spatial Allocation of Emissions Activities
• Emissions sources are spatially allocated to a
  region using the actual locations of the
  emissions sources, and/or using spatial surrogate
  data which are physical parameters that can be
  associated with emissions activities (e.g., acres
  of farmland might be the surrogate for emissions
  from farming operations)
• Temporal Allocation of Emissions Activities
• Emissions sources are temporally allocated by
  assigning a temporal profile, a distribution of
  emissions activity over a 24-hour period, to
  each source category.

13
Emission Inventory Development (3 of 4)

• Chemical Speciation of Emissions Sources
• In order to disaggregate PM emissions into
  individual chemical species, each emissions
  source category is assigned a speciation profile
  which provides a detailed chemical breakdown of
  the individual chemical species emitted from that
  source.
• Several sources of PM speciation data currently
  exist including
• EPAs SPECIATE (http//www.epa.gov/ttn/chief/soft
  ware.htmlspeciate)
• U.C. Davis (via California ARB)
  (http//www.arb.ca.gov/emisinv/speciate/pmucd.pdf)
  
• DRI (http//www.dri.edu)
• OMNI (via California ARB) http//www.arb.ca.gov/e
  misinv/speciate/pmomni.pdf

14
Emission Inventory Development (4 of 4)

• Challenges in developing PM2.5 emission
  inventories
• Limited PM2.5 activity and emission factor data
• Variable quality of existing PM2.5 data
• Estimation of secondary PM2.5

15
The Importance of Evaluating Emissions Estimates

• Reviewing and evaluating emissions inventories is
  critical because development of effective air
  pollution control strategies is predicated on the
  accuracy of the underlying emission inventory.
• Uncertainties associated with PM2.5 emissions
  estimates can range from 10 to an order of
  magnitude.
• Emissions data commonly contain errors in
  pollutant mass.
• Quality control and assurance of emissions
  estimates can significantly improve the quality
  of the data.

16
The Importance of Emission Inventory Evaluation

• Why bother evaluating emissions data?
• Emission inventory development is a complex
  process that involves estimating and compiling
  emissions activity data from hundreds of point,
  are, and mobile sources in a given region.
  Because of the complexities involved in
  developing emission inventories, and the
  implications of errors in the inventory on air
  quality model performance and control strategy
  assessment, it is important to evaluate the
  accuracy and representativeness of any inventory
  that is intended for use in air quality modeling.
  Furthermore, existing emission factor and
  activity data for sources of PM2.5 are limited
  and the quality of the data is questionable. An
  emission inventory evaluation should be performed
  before the data are used in photochemical
  modeling.
• What tools are available for assessing emissions
  data?
• There are several techniques used to evaluate
  emissions data including common sense review
  of the data, source-receptor methods such as
  CMB8, bottom-up evaluations that begin with
  emissions activity data and estimate the
  corresponding emissions, and top-down evaluations
  that compare emissions estimates to ambient air
  quality data. Each evaluation method exhibits
  strengths and limitations.
• Based on the results of the emissions evaluation,
  recommendations can be made on possible
  improvements to the emission inventory. Local
  agencies responsible for developing the inventory
  can then make revisions to the inventory data
  prior to air quality modeling.

17
Emission Inventory Evaluation Tools and Methods

• Use of mathematical techniques to evaluate
  emissions estimates
• Engineering judgement approach
• Bottom-up emissions evaluation
• Use of ambient air quality data to evaluate and
  reconcile emissions estimates
• Multivariate Techniques - Principal Component
  Analysis (PCA) receptor modeling, Chemical Mass
  Balance (CMB) receptor modeling
• Top-down emissions evaluation

18
Using Engineering Judgement to Evaluate Emissions
Estimates (1 of 2)

• Begin with knowledge of the region for which the
  emission inventory was developed (i.e., likely
  emissions sources, population, demographic
  characteristics).
• Review major sources of emissions and perform
  per-capita checks combined with conventional
  wisdom to evaluate emissions data.

• Provides a quick and inexpensive method to
  quality control emissions estimates
• Does not require extensive data
• Can quickly identify gross errors in emissions
  data

• Can identify errors in emissions data, gives no
  insight as to where errors emanate

19
Case Study Using Engineering Judgement to
Evaluate Emissions Estimates (2 of 2)

• Summer emission inventories often report
  significant emissions from seasonal sources such
  as residential fireplaces and wood stoves and
  snowmobiles.
• Residential fireplace and wood stove emissions in
  the summer?

Residential fireplaces and wood stoves are
large contributors to PM emissions in the
wintertime. Emissions contributions from this
source should be low in the summer months.
Taking the time to review seasonal emission
inventory data can catch errors like this.
20
Bottom-up Emission Inventory Evaluation (1 of 2)

• Method of assessing emissions data using census
  information and emissions activity data combined
  with emission factors to generate independent
  estimates to compare to existing data.
• This method is most useful when combined with the
  top-down evaluation when assessing large data
  sets. Top-down identifies problem categories,
  bottom-up used to investigate underlying
  information used to estimate categories.

• The emission estimates generated using this
  methodology can be very accurate if demographic
  and activity data are accurate.

• Extensive data requirements.
• Accuracy of emission factors
• Accuracy of activity data
• Time consuming.

21
Bottom-up Emission Inventory Evaluation (2 of 2)
Case Study Bottom-up Evaluation of Emissions
Activity Data (2 of 2)

• Mobile source emissions activity data for
  anonymous city
• Urban region with a total fleet of 366,699
  on-highway motor vehicles.
• Emissions data reported that 494 of these
  vehicles are heavy-duty diesel trucks (HDDTs).
• According to these figures, 0.1 of the vehicle
  fleet are HDDTs.

In other parts of the country with
similar characteristics, HDDTs make up
approximately 10 to 20 percent of highway
vehicles. HDDTs are significant contributors to
PM, consequently, errors in activity data can
lead to errors in emissions estimates.
Adapted from Haste et. al., 1998
22
Issues Associated With Emissions
EvaluationsUsing Ambient Data

• Ambient air quality data can be used to evaluate
  emissions estimates and source apportionment,
  however, the following issues should be
  considered
• Proper spatial and temporal matching of emissions
  estimates and ambient data.
• Ambient levels of background PM2.5.
• Meteorological effects on comparison.
• Comparisons only valid for primary PM2.5.
• Temporal resolution of ambient data (i.e.,
  24-hour average versus hourly ambient PM data)

23
Multivariate Analyses (1 of 2)

• Statistical procedures that can be used to infer
  mix of PM sources impacting a receptor location.
• Procedures including cluster, factor/principal
  component, regression, and other multivariate
  techniques available in statistical software
  packages.

• Simple statistical methods.
• Does not require speciation profile data.
• Ability to summarize multivariate data set using
  few components.
• Identifies unusual ambient samples.

• Does not apportion secondary aerosol.
• Analyst must infer how certain statistical
  species groupings relate to emissions sources
• Depends on correlation that can be driven by
  meteorology or co-location.

API, 1998
24
Multivariate Analysis Sample Output (2 of 2)

• Example analysis to be added

25
Overview of Receptor Models

• Receptor Models provide empirical relationships
  between ambient data at a receptor and PM
  emissions by source category. The fundamental
  principal of receptor modeling is that mass
  conversion is assumed and a mass balance analysis
  is used to identify and apportion sources of PM
  in the atmosphere. Receptor models are useful
  for resolving composition of ambient primary PM
  into components related to emissions sources.
• Three main types of receptor models
• Models that apportion primary PM using source
  information
• Models that apportion primary PM without using
  source information
• Models that apportion primary and secondary PM
• There are more than a dozen currently existing
  receptor models, however, EPAs OAQPS has only
  recognized CMB and PCA as part of their SIP
  development guidance documents.

API, 1998
26
The Chemical Mass Balance Receptor Model (1 of 4)

• The CMB uses the chemical and physical
  characteristics of gasses and particles measured
  at sources and receptors to both identify the
  presence of and to quantify source contributions
  to receptor concentrations.
• CMB calculations are based on fact that many
  chemical species in the atmosphere do not
  participate in rapid chemical reactions in the
  atmosphere, so they have the same chemical form
  as when they were emitted.
• These chemical species can be used in a
  three-step procedure to apportion the ambient
  pollutants to the sources from which they were
  emitted.

Watson et. al., 1998
27
The Chemical Mass Balance Receptor Model (2 of 4)

• Three step source apportionment using CMB
• Step 1 Measure chemical composition of
  emissions for the important sources of PM2.5
  (e.g., diesel engines, suspended road dust,
  coal-fired boilers). These chemical composition
  data are called source profiles. There are
  several existing libraries of source profile
  data.
• Step 2 Collect and analyze samples of chemical
  species in the ambient air.
• Step 3 Apply the CMB model to each ambient
  sample to determine the relative amounts of
  emissions from each type of source, which, when
  mixed together give the best agreement with the
  measured composition of the atmosphere.
• Each item of input data is accompanied by an
  estimate of its uncertainty. The CMB model
  combines these uncertainties to calculate the
  uncertainty in each output value that is
  attributable to the uncertainties in the input
  data.

Watson et. al., 1998
28
The Chemical Mass Balance Receptor Model (3 of 4)
Limitations

• User friendly model.
• CMB8 (version 8) operates in a Windows-based
  environment and accepts inputs and creates
  outputs in a wide variety of formats.
• Accepted by EPAs OAQPS for SIP development.
• Generates errors in source compositions accepted
  by OAQPS

• Requires source profile information.
• Results of CMB model are only as accurate as the
  speciation profile input data.
• Significant portion of PM2.5 mass is due to
  secondary compounds which are not apportioned by
  CMB.
• Can mis-specify emissions sources.
• Sensitive to collinearity.

29
Case Study Using CMB to Assess Emissions
Estimates and Source Apportionment (4 of 4)
Emission Inventory PM2.5 Source Apportionment
CMB PM2.5 Source Apportionment
Lurmann et. al., 1999 Watson et. al., 1998
30
Top-Down Emission Inventory Evaluation (1 of 6)

• Top-Down Emissions Evaluation method of
  comparing emissions estimates with ambient air
  quality data. Ambient/emission inventory
  comparisons are useful for examining the relative
  composition of emission inventories they are not
  useful for verifying absolute pollutant masses
  unless they are combined with bottom-up
  evaluations. The top-down method has
  demonstrated success at reconciling emissions
  estimates of VOC and NOx, however, using the
  top-down method for PM is currently being
  explored.
• Top-Down Approach for PM
• Compare morning (e.g., 700-900 am) ambient- and
  emissions-derived primary PM2.5 /NOx ratios.
• Early morning sampling periods are the most
  appropriate to use in these evaluations because
  emissions are generally high, mixing depths are
  low, winds are usually light, and photochemical
  reactions are minimized.

31
Top-Down Emission Inventory Evaluation (2 of 6)

• Ambient Data Requirements
• Select ambient monitoring sites dominated by
  fresh urban source emissions.
• Validate and process elemental PM data.
• Select early morning (e.g., 700-900 am) hourly
  data.
• Analyze meteorological data to determine the
  emission areas and elevated point sources that
  may influence ambient measurements.

• Emissions Data Requirements
• Evaluate emissions for same locations as ambient
  monitor.
• Process emissions data to get gridded, hourly
  PM2.5 data.
• Convert emissions data units to be compatible
  with ambient data units.
• Existing emissions processing software EPS 2.0,
  SMOKES, EMS-95

32
Top-Down Emission Inventory Evaluation (3 of 6)

• Analyses
• Compare ambient data with emission estimates from
  different wind quadrants surrounding the
  monitoring site.
• Compare ambient data with emission estimates with
  and without elevated point-source emissions. The
  inclusion of elevated point sources will depend
  on the meteorological conditions.
• Perform primary PM2.5/NOx ratios for day
  specific, weekday, and weekend data.
• Compare individual chemical species in the
  ambient air to chemical species in the emission
  inventory when speciation data is available.

33
Top-Down Emission Inventory Evaluation (4 of 6)
Wind Quadrant Definitions Used in the Top-down
Evaluation
34
Top-Down Emission Inventory Evaluation (5 of 6)

• Estimating primary PM2.5 in the ambient air using
  chemistry data
• Primary PM2.5 1.89Al 1.57Si 1.2K
  1.4Ca 1.43Fe EC 0.7(1.4OC)
• Where
• Al concentration of aluminum Si
  concentration of silica
• K concentration of potassium Ca
  concentration of calcium
• Fe concentration of iron EC elemental
  carbon
• OC organic carbon
• The multipliers in the equation account for the
  extra mass of oxygen in the crustal oxides and
  for the extra mass of hydrogen and oxygen with
  the organics.
• It is estimated that 70 to 90 percent of organic
  carbon is primary.

Kumar and Lurmann, 1996
35
Top-Down Emission Inventory Evaluation (6 of 6)

• Uncertainty Issues
• Proper temporal and spatial matching of emissions
  data and ambient data.
• Meteorological factors including temperature,
  wind speed, and inversion height.
• Level of ambient background PM and precursor
  concentrations due to transport and carry-over.
• Co-location of emissions sources.
• Uncertainties associated with primary versus
  secondary PM.
• Underlying assumption that emission inventory NOx
  estimates are reasonable.

36
Case Study Top-Down Emissions Evaluation (1 of 2)

• Analysis Objective
• Evaluate the consistency of gridded, hourly
  emission inventory with ambient PM and NOx data.
• Compare ambient data with emission estimates with
  and without elevated point-source emissions. The
  inclusion of elevated point sources will depend
  on the meteorological conditions.
• Perform primary PM10/NOx ratios for day specific,
  weekday, and weekend data.
• Identify areas of the emission inventory that
  appear to be inconsistent with the ambient air
  and make recommendations on possible improvements
  to the inventory.

Haste et. al., 1998
37
Case Study Top-Down Emissions Evaluation (2 of 2)
Top-Down comparison of ambient- and emissions
derived primary PM10/NOx in two anonymous cities.
Ambient Ratio
Emission Inventory Ratio
Comparison of the ambient- and emissions derived
PM10/NOx ratios in two anonymous cities are quite
different. It appears as though PM10 is
overestimated in the emission inventory by
approximately a factor of two. Recommendation
the PM10 portion of the inventory should be
investigated from the bottom-up.
Note that this example corresponds to PM10 a
similar comparison could be made for PM2.5
Haste et. al., 1998
38
Top-down Emission Inventory Evaluation
Limitations and Uncertainties

• Extensive data requirements.
• Uncertainties in the emission inventory
  carry-over to comparisons.
• Uncertainties in the ambient data measurements
  carry-over to comparisons.
• Comparison-related uncertainties
• include
• Proper temporal and spatial matching of
  emissions data and ambient data.
• Meteorological factors
• Level of ambient background PM concentrations
  and chemical reactions

• Provides a method to assess areas of an emission
  inventory that appear to be suspect improvements
  can be made prior to photochemical modeling.
• Can assess detailed chemical species composition
  between the inventory and ambient air if accurate
  PM species data is available.
• Can greatly improve emissions estimates.

39
Emission Inventory Data Sources

• Emissions data sources
• State and local air quality management agencies
• EPA National Emissions Trends Inventory
  http//www.epa.gov/ttn/chief/ei/
• Emission inventory improvement program guidance
  documents http//www.epa.gov/ttn/chief/eiip/techr
  ep.htm

40
Ambient Data Sources

• Ambient data sources
• AIRS Data via public web
  http//www.epa.gov/airsdata
• AIRS AQS via registered users
  register with EPA/NCC (703-487-4630)
• PM2.5 websites via public web

41
Meteorological Data Sources

• Meteorology data sources
• Meteorological parameters from NWS
  http//www.nws.noaa.gov
• Meteorological parameters from PAMS/AIRS AQS
  register with EPA/NCC (703-487-4630)
• Private meteorological agencies (e.g., forestry
  service, agricultural monitoring, industrial
  facilities)

42
Summary
43
References

• American Petroleum Institute (1998)/ Review of
  Air Quality Models for Particulate Matter,
  Technical Summary, Publication number 4669,
  March.
• Haste et. al., 1998 - personal communication
  Husar R. (1999) Draft PM2.5 topic summary
  available at http//capita.wustl.edu/PMFine/Workbo
  ok/PMTopics_PPT/PMDefinitions/sld001.htm
• Haste T.L., Chinkin L.R., Kumar N., Lurmann F.W.,
  and Hurwitt, S.B. (1998) Use of ambient data
  collected during IMS95 to evaluate a regional
  emission inventory for the San Joaquin Valley.
  Final report prepared for the San Joaquin
  Valleywide Air Pollution Study Agency and the
  California Air Resources Board, Sacramento, CA by
  Sonoma Technology, Inc., Petaluma, CA,
  STI-997211-1800-FR, July.
• Kumar N. and Lurmann F.W. (1996) Users guide to
  the speciated rollback model for particulate
  matter. Report prepared for San Joaquin
  Valleywide Air Pollution Study Agency,
  Sacramento, CA by Sonoma Technology, Inc., Santa
  Rosa, CA, STI-94250-1576-UG, September.
• Lurmann F.W., et. al., (1999) - personal
  communication
• Watson J.G., Fujita E.M., Chow J.C., Richards
  L.W., Neff W., and Dietrich D. (1998) Northern
  Front Range Air Quality Study. Final report
  prepared for Colorado State University,
  Cooperative Institute for Research in the
  Atmosphere, Fort Collins, CO by Desert Research
  Institute, Reno, NV.
• U.S. Environmental Protection Agency (1997)/
  Quality Assurance Committee Emission Inventory
  Improvement Program Introduction The Value of
  QA/QC, volume VI, chapter 1 January.
• U.S. Environmental Protection Agency (1996)/ Air
  Quality Criteria for Particulate Matter, chapter
  1, Executive Summary EPA 600/P-95/001aF, April.