Science and Technical Support Work Group
Conceptual Model for Ozone,
Participate Matter and Regional
Science
confirm
(Assumptions
predict -> for
modify <
confirm^
Engineering
Policy
Federal Advisory Committee Act for Ozone, Participate Matter and
Regional Haze
-------
450R97004
Science and Technical Support Work Group
Conceptual Model for Ozone, Particulate Matter
and Regional Haze
February 19,1997
Submitted to:
Ozone Policy and Strategies Group
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Prepared by:
Science Applications International Corporation
Integrated Environmental Services Division
615 Oberlin Road, Suite 300
Raleigh, North Carolina 27605
EPA Contract No. 68-D3-0030
SAIC Work Assignment No. HI-107
OP LDBRARY
-------
About the Cover
The cover, designed by Dr. Harvey Jeffries (UNC-Chapel Hill) STSWG Co-Chair, is a
schematic representation of the many interactions and feedback mechanisms required
between science, engineering and regulatory policy which are considered in the design of
the Conceptual Model for the. integrated implementation of a revised ambient air quality
standard for ozone and paniculate matter and a regional haze/visibility rule.
Disclaimer
This report has not been reviewed and approved for distribution by the. U.S.
Environmental Protection Agency's Office of Air Quality Planning and Standards. Thus,
the assumptions, findings, conclusions, judgements and views presented herein are those
of the individual authors, and should not be interpreted as necessarily representing official
U.S. Environmental Protection Agency policies. Mention of trade names or commercial
products are not intended to constitute endorsement or recommendation for use.
-------
Acknowledgments
The Science and Technical Support Work Group (STSWG) is appointed in support of the Federal Advisory
Committee Act's review of the Ozone and Particulate Matter National Ambient Air Quality Standards
(NAAQS) and a review of the Regional Haze Rule, hi addition to the Work Group members listed below,
numerous U.S. EPA staff have contributed to the compilation of this document. The members of this Work
Group are:
Tom Helms (Co-Chair)
U.S. EPA/OAQPS
Harvey Jeffries (Co-Chair)
UNC-Chapel Hill
David Sanders (Work Group Lead)
U.S. EPA/OAQPS
Diana Andrews
KYDept. of Environmental Protection
John Cabaniss
Assoc. of International Automobile Manufacturers
Glenn Cass
California Institute of Technology
Kirit Chaudhari
VA Dept. of Environmental Quality
David Chock
Ford Research Laboratory
John Core
WESTAR
Ellis Cowling
North Carolina State University
Bruce Hill
Appalachian Mountain Club
Jay Hudson
Santee Cooper
David Kelly
Navajo EPA
Dave McNeill
UTDept. of Environmental Quality
Tom Moore
AZDept. of Environmental Quality
Rich Poirot
VTDept. of Environmental Conservation
S.T. Rao
NY Dept. Of Environmental Conservation
Jay Rosenthal
U.S. Navy
Mark Scruggs
National Park Service
Jay Turner
Washington University
Manop Vanichchagorn
LA Dept. Of Environmental Quality
Dan Weiss
Cinergy Corporation
Jeffrey West
GPU Generation Company
Steve Ziman
Chevron Research & Technology Company
ill
-------
IV
-------
Table of Contents
Section Page
ES Executive Summary ES-1
1.0 Introduction 1-1
2.0 Current Environmental State 2-1
2.A Ozone Air Quality Characterization 2-1
2.B Particulate Matter Air Quality Characterization 2-13
2.C Visibility/Regional Haze Air Quality Characterization 2-38
3.0 Processes: How the State is Created, Sustained and Maintained 3-1
3.A Atmospheric Chemistry 3-1
3.B Primary and Secondary Emission Processes : 3-10
3.C Meteorology 3-13
3.D Deposition/Removal 3-18
3.E Linearity/Nonlinearity 3-18
3.F Commonalities, Disconnects and Integrable Elements 3-19
4.0 Current Tools to Address and Implement Current State of
Knowledge 4-1
4. A Monitoring Technologies, Meteorology, and Network Design 4-1
4.B Emission Estimates and Inventories 4-19
4.C Air Quality Models 4-24
5.0 Time-Distance Considerations Relevant to "Transport" and Regions of
Influence 5-1
6.0 Current Needs Based on Relevant Issues and Identified Information
Gaps 6-1
7.0 Integration of Numerical Models and Ambient Monitoring
Data for Effective Air Quality Management 7-1
8.0 Developing a Working and Responsive Science-Policy Continuum 8-1
9.0 References 9-1
-------
List of Figures
igure
Figure 1
Figure 2.1
Figure 2.2
Figure 2.3
Figure 2.4
Figure 2.5
Figure 2.6
Figure 2.7
Figure 2.8
Figure 2.9
Figure 2.10
Figure 2.11
Figure 2.12
Figure 2.13
Page
Schematic of the process-response relationships discussed
within the Conceptual Model 1-4
Hourly frequency distributions for the maximum 3-month
period in 1993 2-3
Hourly ozone data with an 8-hour average superimposed for
a hypothetical urban area for three days in July 2-4
Metropolitan area ozone trends adjusted for meteorological
variability for the period 1984-1993 2-6
Spatial distribution of counties with a 1 -hour daily maximum ozone
concentration, 1 expected exceedance design values greater than
0.12 ppm based on 1991-1993 air quality data 2-7
Spatial distribution of counties with average annual fifth highest
8-hour daily maximum design values greater than 0.08 ppm based
on 1991-1993 air quality data 2-8
Spatial distribution of counties with highest 3-month SUM06
exposure index values greater than 25 ppm-hours in 1990
based only on daylight hours , 8:00 am - 8:00 pm LSI 2-9
Idealized bi-modaj particle distribution diagram 2-15
Major constituents of particles measured at sites in the eastern (left)
and western (right) United States 2-17
Patterns of zinc, arsenic, sulfur (sulfate), and selenium in the
United States 2-18
Areas designated nonattainment for the current PM-10 NAAQS 2-24
Average PM-10 mass concentration (in ug/m3) for each site in the
IMPROVE network 2-26
Trend in annual mean PM-10 concentrations and the PM-10 annual
mean concentration trends by location, for the period 1988 - 1995 2-27
Quarterly averaged fine particle estimates from the combination
of IMPROVE, NESCAUM and FAA data 2-28
VI
-------
List of Figures (continued)
Figure Page
Figure 2.14 Average fine mass (PM-2.5) concentration (in ug/m3) for each site in the
IMPROVE network 2-29
Figure 2.15 Average coarse particle mass concentration (in |ig/m3) for each site in the
IMPROVE network 2-30
Figure 2.16 Average fine sulfate aerosol concentrations (in ug/m3) for each site in the
IMPROVE network 2-31
Figure 2.17 Average fine nitrate aerosol concentrations (in ug/m3) for each site in the
IMPROVE network 2-32
Figure 2.18 Average fine organic aerosol concentration (in ug/m3) for each site in the
IMPROVE network .- 2-33
Figure 2.19 Average fine elemental carbon concentrations (in ug/m3) for each site in the
IMPROVE network 2-34
Figure 2.20 Average fine soil aerosol concentrations (in ug/m3) for each site in the
IMPROVE network 2-35
Figure 2.21 Average annual visibility impairment in deciviews calculated from total
(Rayleigh included) reconstructed light extinction for the three-year
period, March 1992 through February 1995, for IMPROVE sites 2-39
Figure 2.22 Average winter visibility impairment in deciviews calculated from total
(Rayleigh included) reconstructed light extinction for the three-year
period, March 1992 through February 1995, for IMPROVE sites 2-40
Figure 2.23 Average spring visibility impairment in deciviews calculated from total
(Rayleigh included) reconstructed light extinction for the three-year
period, March 1992 through February 1995, for IMPROVE sites 2-41
Figure 2.22 Average summer visibility impairment in deciviews calculated from total
(Rayleigh included) reconstructed light extinction for the three-year
period, March 1992 through February 1995, for IMPROVE sites 2-42
Figure 2.23 Average autumn visibility impairment in deciviews calculated from total
(Rayleigh included) reconstructed light extinction for the three-year
period, March 1992 through February 1995, for IMPROVE sites 2-43
Figure 2.26 Quarterly averaged extinction coefficients based on the combination
of IMPROVE, NESCAUM and FAA data 2-44
vii
-------
List of Figures (continued)
ure Page
Figure 2.27 Spatial variation in the (a) average relative humidity and the (b)
sulfate relative humidity correction factor 2-48
Figure 3.1 Details of the OH-chain cycle 3-2
Figure 3.2 Schematic of the OH-chain VOC and NOX oxidation cycle 3-4
Figure 3.3 Linkages between oxidant chemistry and fine particulate (FP) formation 3-21
Figure 4.1 Schematic of an Air Quality Simulation Model (AQSM) 4-26
Vlll
-------
List of Tables
Table
Table 2.1
Table 2.2
Table 2.3
Table 2.4
Table 2.5
Table 2.6
Table 2.7
Table 2.8
Table 3.1
Table 3.2
Table 4.1
Table 4.2
Table 4.3
Table 4.4
Table 4.5
Table 4.6
Number of original nonattainment areas not meeting selected
standards based on 1991 -1993 air quality monitoring design values
Statistics of violations of a given ozone air quality standard for counties
whose long-term means of corresponding design values are less than
the level of the standard
2-10
2-12
Particle size fraction terminology 2-16
Annual average concentrations and chemical composition from the
IMPROVE monitoring sites 2-19
Annual summer and winter concentrations from the IMPROVE monitoring
sites 2-20
PM-10 and PM-2.5 regional background levels 2-37
Average natural background levels of aerosols and light extinction 2-46
Dry particle light extinction efficiency values used in 1996 analysis of
IMPROVE data 2-47
Constituents of atmospheric fine particles (^ 2.5um) and their major
sources 3-5
Constituents of atmospheric coarse particles (> 2.5um) and their major
sources
3-6
WMO observation levels for lower tropospheric soundings for operational
and research purposes 4-10
WMO observation accuracies for lower tropospheric soundings for
operational and research purposes 4-11
Meteorological variables that can be measured with various upper air
monitoring systems 4-12
Typical vertical ranges and resolutions for upper air monitoring systems 4-14
FIRE emission factors 4-23
Summary of eulerian Air Quality Simulation Models 4-32
IX
-------
-------
CHAPTER 1
Introduction
Emerging air quality management control policies for ozone (03), paniculate matter (PM) and
regional haze (RH) rely on technical information and scientific knowledge gleaned from many diverse
sources. An increasing amount of interaction among a diverse community of air quality professionals (i.e.
scientists, engineers and regulators) accompanies any effort toward integrating programs across pollutants
and over wide geographic regions. Compounding the differences among programs and specialties is die
overriding motivation to understand similarities and overlaps to optimize technical resources and identify
windows of opportunity for successful integrated air quality management. The objective of the Conceptual
Model is to establish a common frame of reference for the technical information and methodologies that
underpin implementation programs designed to improve subsequent dialogues addressing program
integration and its accompanying science-policy interface issues.
The Conceptual Model, was developed by the Federal Advisory Committee Act's (FACA) Science
and Technology Support Work Group (STSWG) to address the state-of-the-science pertaining to integrated
pollutant implementation for ozone, paniculate matter and regional haze. The following chapters of this
document are organized by descriptions of the following topics:
Existing environmental state (tlirough summaries of measured ambient air quality data);
Physical/chemical processes which characterize air quality; and,
Scope of monitoring and modeling analysis and emission inventory programs to characterize and
predict air quality phenomena.
The information presented in the following chapters is reference oriented, as it covers a broad scope of topics,
but additionally provides direction and insight to the major national/regional programs and special field
studies/programs which collectively form the technical foundation for ozone, paniculate matter and regional
haze programs.
This initial version of the Conceptual Model, which currently contains four chapters, will not
adequately address some of the technical issues of concern because of inherent uncertainties in the current
state-of-science. To this end, much of the focus of the FACA STSWG will be directed toward a full
exploration of what is known (and what is not known) over the coming months with an objective assessment
of the uncertainties attendant in the data and tools used for air quality analyses. Accordingly, suggestions
regarding additional information sources, topics, and issues requiring attention and technical corrections are
strongly encouraged and will be considered in future revisions of this and other documents submitted to the
docket by the FACA STSWG. Feedback in these areas is critical for providing a meaningful and useful
reference document and crediting those individuals and groups responsible for the technical and scientific
work that supports both national and regional air quality programs.
CHAPTER 1: Introduction 1-1
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
The Need for a Conceptual Model
In order to develop an adequate understanding of the formation of ozone, particulate matter and
regional haze in different areas of the country, it would be useful to develop conceptual models of the relevant
processes that lead to the formation of each. Further, it would be useful to develop a conceptual model of
how monitoring networks, data analyses, emissions inventories, modeling analyses, and environmental
assessment studies can be best utilized in the subsequent implementation programs. In developing the initial
model, we will encounter many more questions than the available scientific and technical information to
answer them. Thus, the conceptual model should be viewed as a "work in progress", which will be constantly
evolving as new and better scientific information becomes available, is validated and then implemented in
air quality management programs across the United States.
Consequently, these conceptual models will provide the necessary feedback mechanisms to identify
additional needed information and to verify the usefulness of existing information. While these feedback
mechanisms will be useful for establishing research agendas, they will also identify the "act as if we know
how it works" assumptions that will become necessary to formulate and apply any practical computational
models within a time frame shorter than that needed to advance fundamental scientific knowledge. The entire
process of creating conceptual models is iterative in nature; but, provides an overall framework for increasing
the scientific and technical information needed to provide a basis for sound air quality planning and
management. In addition, conceptual models can codify the processes independent of any air quality model,
and thus provide an independent mechanism to evaluate the performance of such computational air quality
tool(s).
The Formulation of a Conceptual Model
In formulating the conceptual model, the STSWG accepts the following foundational principles: (1)
in the environment there are pollutants which are harmful to human health and welfare; (2) that there is a
need to reduce the concentration of these pollutants to acceptable levels as codified in the ambient air quality
standards, (e.g. NAAQS and the national Class I visibility goal (NVG)); 3) compliance with these NAAQS
and other air quality goals is testable using ambient measurement data; and, 4) that human activities do
influence or even create the harmful levels of these pollutants. The social goal desired is to modify human
activities so as to reach safe levels of these pollutants in the ambient atmosphere. We accept that some form
of regulation will be necessary to achieve this modification of human activities. In the specific cases of
ozone, PM-fine (i.e. particulate matter ^ 2.5 ^m in diameter), and regional haze, we accept that these are
pollutants that have affects on human health and overall environmental welfare. We also accept that ambient
levels of these pollutants need to be maintained at or below specific air quality standards (i.e. NAAQS), but
that the levels and forms of these standards are outside the scope of this Work Group's activities.
While there are specific emission sources, transport phenomena, and transformation processes
unique to ozone, particulate matter, and regional haze, we accept that there are also common processes that
link the three. That is, ozone, PM-fine, and regional haze have a sufficiently common origin such that it
makes sense to define atmospheric and emissions data which suggest an integrated set of modified human
1-2 CHAPTER 1: Introduction
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
activities, if any, that will achieve the necessary reductions and maintenance of ambient concentrations of
these pollutants at or below the current NAAQS..
In the following chapters of this document, detailed information is contained on the following:
Chapter 2: Current Environmental State
Chapter 3: Processes: How the State is Created, Sustained and Maintained
Current Tools to Address and Implement the Current State of Knowledge
Chapter 4:
Chapter 5:
Chapter 6:
Chapter 7:
Chapter 8:
Time-Distance Considerations Relevant to Transport and Regions of Influence (To be
completed at a later date)
Current Needs Based on Relevant Issues and Identified Information Gaps (To be completed
at a later date)
Integration of Numerical Models and Ambient Monitoring Data for Effective Air Quality
Management (To be completed at a later date)
Developing a Working and Responsive Science-Policy Continuum (To be completed at a
later date)
Chapter 9: References (To be updated with each revision of the document)
It is through the compilation of these chapters that the STSWG is striving to fulfill its charge from the FACA
Subcommittee by attempting to form a conceptual framework for the understanding of the pertinent and
relevant scientific issues surrounding the implementation of a revised ozone and paniculate matter NAAQS
and a regional haze/visibility rule. As stated earlier, this document represents a "work in progress", which
will be continually updated in future months as the STSWG uses rapidly evolving science to address the
issues related to ozone, particulate matter and regional haze. Figure 1 is a schematic which represents the
suite of process-response relationships which are to be considered within the Conceptual Model.
CHAPTER 1: Introduction
1-3
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
rmMbrtf
location
T
^ geogenic A
^ emissions j *
^ biogenic ^\ >
^ emissions j
T
ecosystem
t t
Air Qualit
H Nino effects
General circulation
Seasons .s^
Weather systems^
meteorology ^
Ail' mass
Stability
Solar insolation
Surface condition
1 1
^ mixing ^
^ volume J
\
pnmary
concentrations
1
/anthropogenic j
^ emissions J
\
4, , i
y Conceptual Model
^Transport (advection by wind)
Vei-tical shear «w "tw
Low-level jets ***< **»«
Orographic effets
Lake/sea breeze
Physical Transformation
Phase change (gas-particle partition^
Breakup /agglomeration
Chemical Transformation
Photolysis
Oxidaaon(OH.O3.NO3)
Neutralization
Chemical Production
Odd oxygen (Ox)
Deposition/Rem ova 1
Radical termination
Surface reaction/absorption
Rajnout/Washout/Qoud venting
Global 2000-10,000 m w».
Regional 1000-1500 m VCTUi<*
Meso 800-1000 m OTtr'Ultu"
Local 200 - 800 m""*
Surface 10 -200m
primary & («*», **C*K
> secondary ££^°
3mDl6n( trasfoTnatiim*
concentrations
/
tions ^policies ^^attainmenn exposure
\tes!L~^ |
f 1 . NAAOS. effects
1
Goals
Figure 1: Schematic diagram of the process-response relationships discussed within the Conceptual
Model.
1-4
CHAPTER 1: Introduction
-------
CHAPTER 2
Current Environmental State
2.A Ozone Air Quality Characterization
This section provides an overview of ozone air quality in the continental United States, covering to
the extent practicable, diurnal and seasonal time frames, rural and urban environments and recent trends.
Materials were extracted and condensed from EPA's Criteria Document for Ozone (EPA, 1996a), the EPA
Air Quality Trends Report (EPA, 1996b), and the EPA Staff Paper on ozone (EPA, 1996c).
Surface Ozone Concentrations
Ozone is a naturally occurring, trace constituent of the atmosphere. Within the scientific community
there is controversy regarding how much of the ambient ozone measured at ground-level is natural and how
much is photochemically produced from anthropogenic and biogenic precursor species. Estimates of the
natural component of O3 vary widely in the scientific literature, and there is no standardized terminology
regarding the concept of O3 background concentrations. A recent survey of scientific literature revealed a
plethora of terminology used to characterize background ozone: baseline ozone, clean air background, global
background, urban background, North American background and regional surface background.
Based on this review of currently available literature, it is obvious that "natural" 03 background is
a multidimensional and complex concept. Background O3 concentrations vary by geographic location,
altitude and season. For the purposes of this document, background ozone is defined as the ozone
concentrations that would be observed in the U.S. in the absence of anthropogenic precursor emissions of ' '
volatile organic compounds (VOCs) and oxides of nitrogen (NOJ in North America. During the summertime
ozone season, in the U.S., daily 1-hour maximum background ozone concentrations are typically between ,
/0.03 to 0.05 parts per million (ppm). Part of this background is due to natural sources and part of it is due ' _a/f* "V-
to long-range transport of anthropogenic precursor emissions. ^^ / f P
The natural component of the background ozone concentration originates from three sources:
stratospheric ozone, which is transported down to the troposphere through tropopause folding along the polar
jet stream; ozone formed from photochemically-initiated oxidation of biogenic and geogenic methane and
carbon monoxide; and, photochemically-initiated oxidation of biogenic VOC.. The magnitude of this natural
part cannot be precisely determined for two reasons. First, that part resulting from long-range transport of
anthropogenic precursor emissions is not known. Second, because NOX plays an important role in the
oxidation of methane, carbon monoxide, and biogenic VOC., it is not possible to determine the exact amount
of ozone that would have been formed due solely to natural NOX emissions. However, some rough estimates
have been made.
. On the basis of O3 data from isolated monitoring sites in the U.S. (EPA, 1996a), a reasonable
/ estimate of the annual average Q, background concentration near sea level is 0.020jo_(X035 ppm. This
V estimate includes a 0.005 to 0.015 ppm contribution (averaged over time) from stratospheric intrusions into r ~>*.
the troposphere and a 0.01 ppb contribution from photochemically-initiated oxidation of methane and carbon
CHAPTER 2: Current Environmental State 2-1
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
monoxide. The remainder is due to the photochemically-initiated oxidation of biogenic VOC. and long-range
transport.
Similarly, an estimate of the background 03 concentration at orhsar coastal locations in the U.S. for
a 1-hour daily maximum during the summer is usually in the range of 0.03 tx> 0.05 ppm. At clean sites in the Q -,
Western U.S., the maximum annual hourly values are in the range oKTOl)Olo07075 ppm(Ip?A, 1996a).
IM***^
Hourly Ozone Air Quality Distributions
Figure 2.1 presents histograms of the hourly O3 concentrations for the 3-month peak summer period
at sites in varying geographic locations and source-receptor environments: an urban site in Chicago; a site
downwind of Chicago; a site downwind of Atlanta; and, a site at higher elevation in Albuquerque. All hourly
concentrations are displayed. These charts demonstrate that there are distinct differences among varying
sites, with the downwind sites exhibiting a greater frequency of higher concentrations^ ^
, Figure 2.2 shows the typical relationship between hourly ozone concentrations and the corresjgonding
"T\" running 8-hour averages. The individual bars correspond to hourly values and the line shows the
corresponding 8-hour averages. Each 8-hour average is associated with the start hour of the 8-hour period.
For example, the 8-hour average from 4 p.m. to midnight is plotted at 4 p.m. It should be noted that the next
8-hour period, from 5 p.m. to 1 a.m., is plotted at 5 p.m. but it contains hourly values from two different
days. If the 8-hour daily maximum is selected from all twenty-four 8-hour averages starting within the day,
it is possible to have daily maxima from adjacent days that have some hourly values in common. This is
unlikely for the typical urban diurnal pattern shown in Figure 2.2; but, it may occur for sites with less
pronounced diurnal patterns, such as rural, higher elevation sites. An alternative approach which has received
some attention is to select the midpoint of the 8-hour interval as the time of observation. This approach has
the following advantages: (1) the 8-hour time series will not appear displaced when compared with the 1-hour
time series; and, (2) if the 1-hour time series is truly unimodal per day, then the chances are that the 8-hour
time series will also be unimodal per day. However, at this time, the current proposed NAAQS is based on
assigning the averaged value to the first hour of the averaging period.
V""
\ Another point to note with 8-hour averages is that the 8-hour daily maximum for a particular day
can actually be higher than the 1-hour maximum. Again, this would be uncommon but can happen because
the daily maximum 8-hour average could contain up to seven hourly values from an adjacent day.
Such elevated 03 levels may be occurring at higher altitudes due to the intrusion of stratospheric ozone.
However, not all of the clean sites are located at high altitudes. An estimate for background 8-hourjiaily lj/i** "
maximum O3 concentrations during the summer (based on diurnal profiles) is also in the range of 0.03 to "
0.05 ppm (Kelly et al., 1982,1984). * 3^ $
fPfi
In recent work completed byAltshuUer and LeFohn (1996), maximum hourly ozone concentrations
were found to range from/0^050 to O..098_pj)m at inland clean sites and range from 0.044 to 0.080 ppm at
coastal clean-srtesTTAdditionally, Altshuller and LeFohn (1996) have shown that based on data collected at
eir clean sites, the annual maximum 8-hour concentrations range from 0.035 to 0.079 ppm.
2-2 CHAPTER!: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Chicago -Cook Co., IL
Max 3-Month Period
Number of hours
500
400
300
200
100
o
I"
' |-j
'
p
m
U U lT~lr*iA D n n n o n
0 .02 .04 .06 .08 .10 .12 .14 .16
.01 .03 .05 .07 .09 .11 .13 .15 .17
Ozone concentration, ppm
Albuquerque, NM
Max 3-Month Period
Number of hours
500
400
300
200
100
o
H
1 n
Innnnnnnnnn
0 .02 .04 .06 .08 .10 .12 .14 .16
.01 .03 .05 .07 .09 .11 .13 .15 .17
Ozone concentration, ppm
Chicago - Downwind Kenosha Site
Max 3-Month Period
Number of hours
200 -
100 r
oDL
1
n_
JUDna^.JLJLJL 0
0 .02 .04 .06 .08 .10 .12 .14 .16
.01 .03 .05 .07 .09 .11 .13 .15 .17
Ozone concentration, ppm
Atlanta, GA
Max 3-Month Period
Number of hours
" n
II ILJLJLJunnx n o o o
0 .02 .04 .06 .08 .10 .12 .14 .16
.01 .03 .05 .07 .09 .11 .13 .15 .17
Ozone concentration, ppm
Figure 2.1: Hourly frequency distributions for the maximum 3-month period in 1993 (from EPA,
1996c).
CHAPTER 2: Current Environmental State
2-3
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Concentration, ug/m3
0.2
0.15
0.1
0.05
July6
July?
JulyS
1-HOUR
DATA
8-HOUR
AVERAGE
OVERLAPPING
PERIODS
MDN 6AM NOON 6PM MDN 6AM NOON 6PM MDN 6AM NOON 6PM MDN
Figure 2.2: Hourly ozone data with an 8-hour average superimposed for a hypothetical urban area for
three days in July (from EPA, 1996c).
2-4 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
/
Ozone Air Quality Trends
The interpretation of recent 03 trends is difficult to ascertain due to the large temporal variation that
results from confounding factors of meteorology and emissions changes. Peak O3 concentrations typically
occur during hot, dry, stagnant summertime conditions. T/nus, Summer 1988, as the third hottest summer I " °/:
on record since 1931, was highly conducive to 03 formation with peak Q levels cpjnparaWe-toJhDse
recorded in the earlier peak year of 1983 (EPA, 1996c). /Meteorological conditions/iffl991 and 1993 wer
also highly conducive to 03 formation, especially in the/eastern half of the country, aftheugh the magnitude Q °|
and frequencyofexceedances were significantly less than those recorded JrfT98 8 TERA, 1996c). In contrast,
the years^r9^9and 1992j saw meteorological conditions that were generally~net-asj£onducive to 03 formation.
These cRahges in meteorological conditions have led to large year-to-year differences in peak 03
concentrations. In response to the National Academy of Sciences recommendations (NAS, 1991), EPA has
developed a statistical model that adjusts for meteorological variability to detect the underlying 03 trend (Cox
an^ ^^ 1993). Fig016 2-3 presents the meteorologically adjusted, and unadjusted, ten year O3 concentration
trends in 43 metropolitan areas. Thg 99th percentile daily maximum 1-hour concentration^declined^Lpercenr^^
ier_vear or 12 percent since 1984. The national trend in the composite mean of the annual second highest
laily maximum 1-hour concentration at 509 sites is shown for comparison, which also declined by 12 percent
between 1984 and 1993. ^_
-
Spatial Distributions of Air Quality Concentrations
This section provides a brief overview of how O3 concentrations vary across the country with respect
to alternative standards. Figure 2.4 displays a map of those counties with 1 -hour daily maximum, I expected
exceedance 03 design values greater than 0.12 ppm based on 1991 -1993 air quality monitoring data. Figure
2.5 shows the spatial distribution of counties with average annual fifth highest 8-hour daily maximum O3
design values greater than 0.08 ppm, based on 1991-1993 data. Figure 2.6 depicts those counties with a 3-
month SUM06 03 exposure index value greater than 25 ppm-hours in 1990. These SUM06 values are based
only on the daylight hours, 8:00 am - 8:00 pm Local Standard Time (LST). However, it should be noted that
because a number of counties in the U.S. have no ozone monitoring data, their ozone air quality is unknown,
and is not plotted in these figures. In each of these maps, the county air quality status was determined by the
peak design value site in each county. For the current standard, the design value is simply the fourth highest
daily maximum 1-hour concentration measured during 1991-1993, since if the fourth highest value is reduced
to the level of the standard, there will be only three days above the level of the standard, or 1 exceedance per
year. Because both the proposed primary and secondary standards are concentration based standards, the
design value is identical to the standard test statistic.
Ozone Air Quality Status with Respect to Standards
Ambient air quality data from 1991-93 have been used to estimate the air quality status with respect
to the current and the proposed primary and secondary standards. Table 2.1 presents the air quality status
of the original 98 areas designated nonattainment for 03 under the Clean Air Act Amendments of 1990 with
respect to these alternative standards. The time and space distributions of the proposed standards could be
significantly different than those of the existing O3 standard. Thus, a comparison of the number of
CHAPTER 2: Current Environmental State 2-5
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Concentration, pom
M« Adjutlad Tfand-43 MSA's
(09DI paicantiu daily max l.hr cone )
National Composite M*an Ozoni Trind
{Annual 2nd Dally Max 1-hr)
Actual (43 MSA'S)
(99th Parcantila)
Mat Adjustad (it MSA's)
(99th P.re.nlil.)
National (509 srtas)
(2nd Daily Max 1-hr)
1994 1985 1996 198? 1988 1989 1990 1991
1992 1993
Figure 2.3: Metropolitan area ozone trends adjusted for meteorological variability for the period 1984-
1993 (from EPA, 1996c).
2-6 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Figure 2.4: Spatial distribution of counties with a 1-hour daily maximum ozone concentration, 1
expected exceedance design values greater than 0.12 ppm based on 1991 -1993 air quality
data (from EPA, 1996c).
CHAPTER 2: Current Environmental State 2-7
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Figure 2.5: Spatial distribution of counties with average annual fifth highest 8-hour daily maximum
design values greater than 0.08 ppm based on 1991-1993 air quality data (from EPA,
1996c).
2-8 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Figure 2.6: Spatial distribution of counties with highest 3-month SUM06 exposure index values greater
than 25 ppm-hours in 1990 based only on daylight hours, 8:00 am - 8:00 pm LST (from
EPA, 1996c).
CHAPTER 2: Current Environmental State 2-9
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Table 2.1: Number of original nonattainment areas not meeting selected standards based on 1991-
1993 air quality monitoring design values.
OZONE AIR QUALITY STANDARD
Primary standard
1-hour, 1 exceedance, 0.12 ppm
8-hour, avg annual 2nd max, 0.08 ppm
avg annual 5th max, 0.08 ppm
Secondary standard
3-month SUM06 > 25 ppm-hours
(8:OOam-8:OOpmLST)
Number of Areas with Monitors
Number of
Non-
Compliant
Areas
39
85
64
70
97
2-10 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
nonattainment areas between existing and proposed standards is by definition incomplete until monitoring
networks have been adjusted, and the ambient data analyzed.
Ozone Air Quality Relationship Among Standards
Although previous sections have shown that there are significant differences in air quality
distributions among monitoring sites with differing environments, several patterns emerge from examining
typical relationships. For example, the average ratio of 8-hour to 1-hour design values for a 1 exceedance
per year standard is 0.86 based on 1991-1993 data. This ratio has been fairly stable over time increasing
slightly from 0.8 1 in 1980 to 0.86 in 1993. This increase in the ratio is expected as the upper range of the
concentrations decreases. The ratio is also fairly consistent across EPA Regions, with the median ratio
ranging from 0.80 to 0.88.
In terms of exceedances (i.e., the number of days with daily maximum 8-hour average concentrations
greater than 0.08 ppm), sites meeting an average annual 2nd highest daily maximum 0.08 ppm standard
average 1.2 exceedances per year and 2.3 exceedances in the worst year of three. Sites meeting an average
annual 5th highest daily maximum 0.08 ppm standard average 3.0 exceedances per year and 5.4 exceedances
in the worst year of three. In the worst year of three, 95% of sites meeting an average annual 2nd highest
daily maximum 0.08 ppm standard have seven or fewer exceedances, while 95 % of sites meeting an average
annual 5th highest daily maximum 0.08 ppm standard have 12 or fewer exceedances. These results are based
on data collected during the period 1991-1993.
Ozone Variability and Its Implications with Respect to the Standard
Meteorology fluctuates from year to year as does the ozone concentration being influenced by
meteorology. The attainment test for the current ozone air quality standard stipulates that an area is violating
the standard if any monitor in the area registers four or more exceedances of the level of the standard in three
years. For many areas, because of year-to-year fluctuations of the ozone concentration, which is not
accounted for by the test, flip-flops in the attainment status will occur. Unless a statistical test of attainment
'or violation is used to avoid flip-flops and secure a long-term attainment status, an area must effectively. _-<
s~ToTtfeF4he-desrgffvaIue well below the level of the standard to make room for the impact of meteorological I
Table 2.2 shows the statistics of violation of a given ozone air quality standard for counties whose
long-term means (for three-year periods ending in 1985-1995) of the corresponding design values are already
less than the level of the standard. The forms and levels of the standard considered are shown in the Table
2.2. Forty to 60 percent of these counties actually violate the standard at least once over the eleven three-year
periods. The long-term mean below which no violation of a given standard is observed is also presented in
the table. This value is the "effective" target of the design value to assure compliance of the standard for the
eleven consecutive three-year periods based on the present attainment test. As the length of time increases
and more data become available, this value will likely decrease. The range of the observed clean-site or
background ozone concentrations for the corresponding form of the standard are shown in the last column
CHAPTER 2: Current Environmental State 2-11
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Table 2.2: Statistics of violations of a given ozone air quality standard for counties whose long-term
means of the corresponding design values are less than the level of the standard (from
Chock, etal., 1997).
Design Value
(Form of the std)
4th in 3 yrs**
(Ih, lex,>125ppb)
8h, ami. 5th, >85 ppb
8h, ann. 2nd, >85 ppb
8h, ann. 3rd, >85 ppb
8h, ann. 3rd, >80.5 ppb
Total # of
counties w/
mean <
std
164
116
57
76
39
#of
counties w/
1 violation
of the std
14
12
7
4
3
# of counties
w/2
violations of
the std
8
14
4
9
4
#of
counties w/
3 violations
of the std
25
12
4
11
6
#of
counties
w/>3
violations
of the std
21
24
22
17
6
Total # of
counties w/
> 1
violation of
the std
68
62
37
41
19
%of
counties
w/> 1
violation
of the std
41%
53%
65%
54%
49%
Long-term
mean (ppb)
below which
no violation
was observed
107.7 '
75.7
76.8
78.1
74.3
Natural or
remote-area
background O,
concentration
range (ppb, 3-
yr avg) for the
corresponding
design value
53-92
38-65
44-68
41-66
41-66
269 counties satisfy the requirement of having 11 consecutive 3-year periods of data. The long-term mean is the mean of the design
values for the 11 consecutive, overlapping 3-year periods. The present attainment test is used.
The long-term mean is the average of 10 overlapping 3-year periods in 268 counties. The design values for the period 1992-1994 were
not available.
2-12 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
of the table. If the "effective" target of the design value is close to the background range, there is concern
that the standard may not be achievable (Chock, et al., 1997).
2.B Particulate Matter Air Quality Characterization
This section defines the various subclasses of particulate matter (PM) and then briefly discusses
some of the available information on recent PM concentrations and trends. Materials for this section are
extracted from the EPA PM Criteria Document (EPA, 1996d), EPA PM staff paper (EPA, 1996e) and the
EPA Trends Report on Air Quality (EPA, 1996b). The most comprehensive discussion on PM air quality
characterizations is available in Chapter 6 of the PM Criteria Document.
Overview of Particle Composition and Size Distributions
Particulate matter consists of discrete particles in the condensed (liquid or solid) phase spanning
several orders of magnitude in size, ranging from clusters of 0.005 um in diameter to coarse particles on the
order of 100 um. Particle size diameter is the aerodynamic diameter, defined as the diameter of a spherical
particle with equal settling velocity but a material density of 1 g/cm3, effectively normalizing particles of
different shapes and densities. In addition to characterizations by size, particles can be described by their
formation mechanism or origin, chemical composition, physical properties, and by sampling/measurement
technique. This latter point on monitoring is critically important, given: (1) the complexities of measuring
monitoring and comprehensive (space, time and chemical composition) air quality characterizations, and (2)
difficulty in identifying causative components associated with adverse health.
PM Ambient Air Quality Data Collection
As defined by the existing Federal Reference/Equivalent Methods (FR/EM) for PM-10 and the
proposed FR/EMs for PM-2.5, ambient aerosol in solid form below these cut points will continue to not be
both captured and retained for gravimetric and/or subsequent chemical analyses. Thus, these current
sampling method(s) not only limit the definition of ambient PM aerosol, but prevent the fullest analysis of
health/welfare cause and effect as well as crippling the development of controls appropriate to the sources
by constraining the choices. No perfect method for capturing and determining all the chemical characteristics
of ambient PM aerosol is currently in use, and that method may not presently exist. The tools for attempting
the technical integration of PM-fine with 03 do not appear to be evolving, and the EPA program support for
correcting this deficiency is unclear. The exact air quality control and health/welfare protection impacts are
not clear at this time, but a policy to encourage development of new, improved ambient PM measurement
techniques needs to be considered in light of the needs of state and local air quality planning requirements.
Sources of Primary PM Emissions & Secondary PM Precursors
A variety of anthropogenic, biogenic and geogenic sources contribute to PM concentrations,
including fuel combustion (from vehicles, power generation, and industrial facilities), residential fireplaces,
agricultural/forest burning, construction and demolition activities, wind blown dust, road dust, fugitive
CHAPTER 2: Current Environmental State 2-13
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
emissions, and biogenic precursors. All of these source generation activities are coupled with
"natural/pristine" background levels, atmospheric transformations, deposition/removal and meteorological
Q K/\ ^ ^processes to produce a given "state". The major chemical constituents of PM are sulfates, nitrates,
\ ^JJ^carbonaceous compounds (both elemental and organic carbon compounds), acids, ammonium ions, metals,
water, and crustal materials. The relative fractions of these components vary from place to place and over
time.
Multi-modal Size Distributions
Particles typically exhibit a molti-modal distribution^hich varies with location and time. Based on
particle size and formation mechanlsrnTparticles can be dassified into two fundamental modes (fine and
coarse). An idealized mass distribution of the fine and coarse modes (Figure 2.7) can be subdivided further
into nuclei (<. 0.1 um) and accumulation (0.3 to 0.7 um) modes. Nuclei or ultra fine particles tend to exist
as disaggregated particles for very short periods of time (minutes) and rapidly^coagulate into accumulation
mode particles.
A site-specific valley in the modal distribution, distinguishing fine and coarse modes, lies between
1.0 um to 3.0 um where minimum mass occurs between the two modes. A clear choice of cut points does not
exist, given the overlap that occurs between the modes. Most ambient measurements of fine particle mass
in the U.S. have used instruments with cut points of 2.5 or 2.1 um. Table 2.3 provides size-cutoff related
terminology used to describe particulate matter.
Composition of Fine and Coarse Fraction Particles
Fine and coarse mode particles generally have distinct chemical composition, solubility, and acidity.
Fine mode PM is composed of varying proportions of sulfates, acids, nitrates, elemental carbon, organic
carbon compounds, trace metals, crustal elements and water. Coarse fraction constituents are primarily
composed of crustal elements: silicon (Si), aluminum (Al), iron (Fe), and potassium (K). Biological materials
(bacteria, pollen, and spores) also appear in the coarse mode. The chemical composition of PM-10, the sum
of both modes, is more heterogeneous than either mode alone.
Figure 2.8 combines the available published data on the chemical composition of PM-2.5 and coarse
fraction particles in U.S. cities by region. In addition to the larger relative shares of crustal materials in the
West, total concentrations of coarse fraction particles are generally higher in the arid areas of the Western
and Southwestern U.S. The fraction of PM-2.5 due to sulfate is greater in the East, and the nitrate fraction
is much larger in the West. Isopleths drawn from the IMPROVE data for selected metals (see Figure 2.9)
illustrate the elevated sulfate levels in the East. Tables 2.4 and 2.5 summarize composition data from the
IMPROVE database by region and season, respectively.
A material balance on the chemical composition of atmospheric fine particle samples typically shows
that the fine particle mass consists of a combination of organic compounds, elemental (black) carbon, sulfate
ion, nitrate ion, ammonium ion, sea salt, soil dust and water. These chemically distinct particle constituents
2-14 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Coarse-Mode Particles
0.2 0.5 1.0 2 5 10 20
Aerodynamic Particle Diameter (Da), urn
Total Suspended Particles (TSP) *
PM10
50 100
PM
Figure 2.7: Idealized bimodal particle distribution diagram (from EPA, 1996e).
CHAPTER 2: Current Environmental State 2-15
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Table 2.3: Particle Size Fraction Terminology
Fraction
Description
Fine particles
PM-2.5
Coarse particles
Coarse fraction of PM-10
PM-10
Total Suspended
Particles (TSP)
Fine mode particles which are generally formed through chemical reaction,
nucleation, condensation of gases, and coagulation of smaller particles;
contains most numerous particles and represents most surface
Particles with an upper 50% cut point of aerodynamic diametersh less than
2.5 um, a measurable approximation for fine particles2
Coarse mode particles which are mostly mechanically generated through
crushing or grinding
Particles with an upper 50% cut point of aerodynamic diameters between
2.5 \im and 10 um
Particles with an upper 50% cut point of aerodynamic diameters less than
10 urn, including fine fractions and part of the general coarse mode.
(NOTE: The characterization of the mass concentration for this cut point
is highly dependent upon the sampler model design and the individual
unit's operation characteristics.)
Particles with an upper 50% cut point of aerodynamic diameters
less than approximately 50 urn; highly wind speed dependent
'Aerodynamic diameter is defined as the diameter of a spherical particle with equal settling velocity but
a material density of 1 g/cm3. This normalizes particles of different shapes and densities.
2PM-X indicates an upper cut point with a 50% cut point of X /^m diameter. Because samplers have a
collection efficiency that varies around the 50% cut point, not all particles less than X ^m diameter will
be collected and some particles greater than X /im diameter will be collected.
2-16 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
PM2.5 Mass Apportionment
- Minerals 4.3%
Unknown 22.8%
EC 3.9%
PM2.5 Mass Apportionment
EC 90% > ' Minerals 7.6% -
34.1%
OCxl.4 44.6%
OCX 1.420.9%-
NO; 1.1% -^ v (Nh£ ). 10.7%
Nitrate based on 2 studies
EC 29.6%
00X1.450%
NOj 23.7%
Nitrate based on 2 studies; OC and EC based on 4 studies
Reconstructed sum = 103.9%
Figure 2.8: Major constituents of particles measured at sites in the eastern (left) and western (right)
United States (from EPA, 1996d).
CHAPTER 2: Current Environmental State 2-17
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Figure 2.9: Patterns of zinc, arsenic, sulfur (sulfate), and selenium in the United States (from EPA,
1996d).
2-18 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Table 2.4: Annual average concentrations and chemical composition from the IMPROVE monitoring
sites (from Malm, et al., 1994).
Region
Northwest*
Southwest"
California Coastal Mountains'
Transitional Region4
Appalachian Mountains'
Annual Average Concentrations, jig/m9 and Compost
Total
15
5
3
5
2
PM-2.5
3.55
3.91
4.99
5.15
10.91
(NHJjSO, /
% PM-2.5
0.88 / 25
1.28/33
1.41/28
1.97/38
6.33 / 58
Organics/
% PM-2.5
1.63/46
1.38/35
1.95/39
2.01 / 39
2.97/27
PM Coarse
4.46
5.62
8.85
6.54
6.24
PM-10
8.0
9.5
13.8
1.1.7
17.2
a Cascades (1), central Rocky Mountain (5), Great Basin (1), Northern Rocky Mountains (1), Sierra Nevada (1), Sierra Humboldt
(2), and Colorado Plateau (4)
b Colorado Plateau (3), Sonora Desert (2)
c Same as subregion
d Western Texas (2), Northern Great Plains (1), Boundary Waters (2)
CHAPTER 2: Current Environmental State 2-19
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Table 2.5: Annual summer and winter concentrations from the IMPROVE monitoring sites (from
Malm, etal., 1994).
Subregion
Central Rockies
Colorado Plateau
Coastal Mountains
Sonora Desert
West Texas
Northern Great
Plains
Boundary Waters
Appalachian
Mountains
No of Seasons of
Region of U.S. Sites the Year
NW 5 annual
summer
winter
NW-SW 7 annual
summer
winter
NW 3 annual
summer
winter
SW 2 annual
summer
winter
Transitional to 2 annual
east summer
winter
Transitional to 1 annual
east summer
winter
Transitional to 2 annual
east summer
winter
Eastern U.S. 2 annual
summer
winter
PMr,
3.3
4.8
2.0
3.4
4.1
2.9
5.0
4.5
5.6
4.4
5.6
3.2
5.4
6.6
3.6
4.5
5.6
3.4
5.3
6.2
5.2
10.9
16.6
6.5
(NH4),S04
0.8
1.0
0.5
1.1
1.3
0.9
1.4
1.9
0.9
1.5
2.1
1.2
2.1
2.5
1.5
1.5
1.8
1.2
2.0
2.2
2.0
6.3
10.5
3.0
Organics
1.5
2.4
0.9
1.2
1.6
1.1
1.9
1.4
2.3
1.5
1.8
1.1
1.5
1.7
1.1
1.5
2.2
1.1
2.1
3.1
1.4
3.0
4.4
2.0
PM
Coarse
4.8
7.5
3.0
4.7
6.4
3.2
8.9
10.7
7.7
6.0
7.6
3.3
7.5
7.4
5.1
6.3
9.7
3.9
5.7
8.2
3.2
6.2
11.2
3.1
2-20 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
provide a very strong indication of the sources from which the fine particle mass originated. In addition,
small quantities of trace metals and hydrogen ion may be present in fine particles at concentrations too low
to account for most of the mass but at concentratioas high enough to produce concern over the potential
for undesirable effects (e.g. strong acid aerosols; heavy metals toxicity).
Direct emissions of particle-phase material from sources are referred to as primary emissions in
order to distinguish them from the secondary fine particles that are produced from gas-phase precursors
by atmospheric chemical reactions. Primary fine particles typically arise from combustion, from
specialized industrial processes, and from sources of geological dust. Black elemental carbon particles
are entirely due to primary emissions from combustion sources. In many cases, elemental carbon particle
concentrations are dominated by diesel soot and wood smoke. Elemental carbon particles often act as a
good indicator of atmospheric dilution and transport because they are not produced by atmospheric
chemical reactions and are generated at particles sizes (typically about 0.2 jim particle diameter) that do
not deposit out of the atmosphere easily but instead disperse in much the same way as carbon monoxide
(CO). Primary organic carbon-containing particles are emitted directly from more than 70 different types
of mobile, stationary and fugitive source types. Prominent sources of fine organic aerosols include motor
vehicle exhaust, food cooking (e.g. meat charbroliers), wood combustion, the organic carbon content of
paved road dust, plus smaller contributions from airborne plant fragments, stationary oil and gas
combustion, aircraft engines, fine particle tire dust, brake dust and cigarette smoke, to name but a few
of the many possible sources. Fine particle organic compounds plus elemental carbon typically contribute
25-40% of the fine particle mass concentrations measured in cities.
A further contribution to primary fine particle concentrations consists of the fine particle fraction
of the geological dust emissions to the atmosphere. While most of the mass of soil and road dust emissions
to the atmosphere is found in coarse particles having aerodynamic diameter greater than 2.5 ^m, a few
percent of the crustal dust emissions occur in the fine particle size range. It is not uncommon to find that
perhaps 10% of an airborne fine particle sample is composed of geological material, with fugitive paved
road dust being the most important source of this type in cities.
Sulfate-containing particles can be emitted directly from combustion sources; typically abouv3%
of the sulfur content of fuels is emitted directly as primary particulate sulfate. However, the majorityo
the sulfates in the atmosphere arise, from gas-to-particle conversion processes in which S02 gas produced
-iBy combustion ot sulfur bearing fuelsjs oxidkedJn.the-atmosphere to form sulfuric-acid. .ammonium
bisulfate or ammonium sulfate particles. SO2 oxidation can occur by both homogeneous gas phase
chemical reactions and heterogeneousprocesses within cloud or fog droplets. The gas phase homogeneous
oxidation of SO2 occurs primarily due to reaction between S02 and the hydroxyl radical, with additional
contributions from ozone-olefin reactions products. The heterogeneous conversion of SO2 to form sulfates
with liquid water droplets can occur due to reaction of dissolved S02 with dissolved hydrogen peroxide,
or ozone. In the presence of catalytic metals (Fe, Mn, and others), dissolved SO2 can be oxidized at
appreciable rates by dissolved oxygen. Which of the various SO2 oxidation pathways dominates depends
on the details of the local accumulation of co-pollutants such as ozone, hydrogen peroxide, hydrocarbons,
ammonia, and catalytic metals. The important point to observe is that, given so many competitive
pathways, SO2 oxidation usually is ongoing due to at least one or more of the available routes, generally
CHAPTER!: Current Environmental State 2-21
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
producing overall conversion rates of a fraction of a percent per hour to 10% per hour or more. Over
a period of several days, much of the S02 in the atmosphere will be converted to sulfate if it is not
removed first by wet or dry deposition at the earth's surface.
The reaction products of SO2 in the atmosphere are non-volatile and will not return to the gas
phase once formed. The speciation of the paniculate sulfates as sulfuric acid, ammonium bisulfate,
ammonium sulfate, or some combination of the above, is largely determined by the amount of ambient
ammonia available to neutralize the acid sulfate species. In the Eastern United Stated, unneutralized acid
sulfate particles are frequently reported, while in California and in other western locations the sulfate
aerosols are commonly found to be fully neutralized, with excess ammonia remaining in the gas phase.
Aerosol nitrates are produced by reactions involving atmospheric nitric acid vapor. Nitric acid
is formed in the atmosphere as a further reaction product of N02. During the daytime, NO2 oxidation
occurs by reaction with the hydroxyl radical. At night, NO2 can react to form nitric acid by pathways
based on formation of the nitrate radical, N03, followed by production of N2O5, which then can hydrolyze
to produce nitric acid. The nitrate radical can also react with certain organic species to produce nitric
acid. As nitric acid concentrations build-up in the atmosphere in the presence of free ammonia gas, a
critical point is passed at which the product of the nitric acid vapor concentration times the ammonia
concentration in the gas phase exceeds the value of the equilibrium dissociation constant for ammonium
nitrate aerosol. Since ammonium nitrate is a solid or liquid phase species, gas-to-particle conversion
occurs at this point. The value of the equilibrium dissociation constant for ammonium nitrate is extremely
seasitive to temperature and humidity, with more ammonium nitrate formation from the same quantity of
ammonia and nitric acid vapor expected at low temperature and at high relative humidity. While heavy
photochemical smog episodes greatly favor rapid nitric acid vapor production, such photochemical smog
episodes often occur on days with very high temperatures that do not necessarily favor ammonium nitrate
production from nitric acid vapor. Instead at high temperatures the nitric acid vapor may well stay in the
gas phase where it is available for relatively rapid deletion by dry deposition at the ground. Fine particle
nitrate concentrations in the vicinity of 100 ^m/m3 over 24-hour averaging times have been observed in
the eastern end of the South Coast Air Basin that surrounds Los Angeles during late October, a time of
year with moderate photochemical smog but generally lower values of the equilibrium dissociation constant
for ammonium nitrate. Just as ammonium nitrate formation can be encouraged by high nitric acid levels,
high ammonia levels, low temperatures, and high relative humidity, ammonium nitrate aerosols once
formed can be destroyed (returned to the vapor phase) if one or more of the key factors driving
ammonium nitrate formation is removed from an air parcel. Thus ammonium nitrate contained in air
parcels transported into the desert downwind of Los Angeles could be expected to evaporate if the gas
phase nitric acid or ammonia that surrounds the particles is removed by dry deposition, or if temperatures
increase or relative humidities decline, all of which are likely to occur in a desert environment. If the
aerosol studied is a mixed salt composed of various proportions of sulfate, nitrate, ammonium, chloride,
etc., all found in the same particles, then the above discussion of ammonium nitrate equilibrium with nitric
acid and ammonia only needs to be replaced by a more complex description based on multi-component
aerosol thermodynamics.
2-22 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
A second common route for aerosol nitrate formation involves nitric acid (and possibly
attach on the NaCl in sea salt particles to form NaN03 aerosol and to release gaseous chloride species to
the atmosphere. While most sea salt particles are in the coarse particle mode and are larger than 2.5 /*m
in diameter, one should still look for the formation of some fine particle NaNO3. Further, the
corresponding release of gaseous chloride species from the aerosol is another transformation route that
can affect fine particle concentration and composition.
The final major route for formation of fine particles by atmospheric chemical reactions involves
secondary organic aerosol production. Secondary organic aerosol production is the subject of much
ongoing research. Two routes for secondary organics formation seem possible. The traditional view has
been that secondary organic aerosol formation will occur when atmospheric chemical reactions produce
organic compounds such as aliphatic dicarboxylic acids that have extremely low vapor pressures. Once
these low vapor pressure products saturate the gas phase at the level of a few parts per thousand (ppt), any
further chemical production will result in aerosol production. More recently, it has been realized that
secondary organic aerosol production also can occur if organic gases dissolve into an existing organic
liquid phase in the atmospheric particle complex. This organic liquid phase could be present due to
primary organic aerosol emissions or due to previous secondary organic aerosol production. Because of
the difficulties of distinguishing primary organic aerosols from secondary organic aerosols in an
atmospheric sample, and due to the general lack of much organic aerosol data in the Eastern U.S., much
less is known about the relative importance of secondary organic aerosol formation than is known about
primary particle emissions and sulfate plus nitrate formation. In the Los Angeles area where data are
available that can be used to draw some preliminary conclusions, it appears that secondary organic aerosol
accounts for 15-30% of the total organic aerosol (hence 5-10% of the entire fine particle mass
concentration) on an annual average basis, but can account for the majority of the organic aerosol present
over short periods during severe photochemical smog episodes.
Paniculate Matter Concentrations
Correlations Between PM2,5 and Coarse Fraction Mass
Ambient daily fine and coarse fraction mass concentrations generally are not well correlated. This
is plausible considering the differences between the two modes with respect to sources and
chemical/physical attributes. Current analyses based on various data sets show R2 values between daily
PM-2.5 and PM-10-2.5 mass to be 0.13 for non-rural sites and 0.21 when rural sites included. Fine and
coarse fractions should correlate in certain instances where the fine mode fraction is the major fraction
ofPM-10(EPA, 1996d).
PM-10 Concentrations and Trends
State and local air pollution control agencies have been collecting and reporting PM-10 mass data
to EPA's AIRS database since mid-1987. A geographic distribution (Figure 2.10) of the 83 PM-10
nonattainment areas (as of September, 1994) shows an overwhelming number of western nonattainment
CHAPTER 2: Current Environmental State 2-23
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
MNCATU RELATIVE tOZ
CT AFFECTED PtVULATtGN
Figure 2.10: Areas designated nonattainment for the current PM-10 NAAQS.
2-24 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
areas, relative to more populated eastern areas. Many of the highest values occur in both small
communities and the large urban areas of the western states with fugitive dust sources, and in mountain
valleys where the topography and seasonal meteorological episodes tend to confine and concentrate wood
smoke and other emission sources. Figure 2.11 shows isopleths of the three-year average PM^IO aerosol
mass concentration measured during the period March 1992 through February 1995 for the IMPROVE
network. The highest concentrations occur in the eastern United States where most regions exhibit
concentrations in excess of 10 /ig/m3 (Sisler, et al., 1996).
National PM-10 trends from 1988 to 1995 can be readily examined in Figure 2.12. The figure
shows the site-to-site variability and the trend in the ninetieth percentile 24-hour PM-10 concentrations.
The national average of annual mean PM-10 concentrations decreased 22% over the eight year period.
Annual average PM-10 concentrations ranged from 20 to 30 fig/m3 for most U.S. regions by 1995.
Fine Particle Concentrations and Trends
PM-2.5 concentration data are more limited than for PM-10. Additionally, it should be noted that
no PM-2.5 data, which have been characterized using the proposed fme-PM FR/EM, have been collected
to this date. From 1983 to 1993, less than 50 sites reported data to AIRS in any given year. Quarterly
U.S. maps (Figure 2.13) of PM-fine based on aggregations IMPROVE, Northeast States Coordinated Air
Use Management (NESCAUM) and airport visibility data show the regional character of elevated fine
particle levels in the Eastern U.S. and California as well as a strong seasonality. Higher levels prevail
in the East during summer; high wintertime levels exist in southern California. These data generally do
not include urban concentrations and only reflect regional non-urban concentrations. Hence much urban
area spatial detail is not available and care should be exerted when interpreting such diagrams. Figure
2.14 shows isopleths of the three-year average fine aerosol concentrations (i.e. PM-2.5) from the
IMPROVE network measured between March 1992 and February 1995. There is a strong gradient in
PM-fme concentrations from southern California (maximum of 9 /*g/m3) to southern Oregon, Nevada,
Utah, western Colorado, and Wyoming (minima 2.7 -3.1 /ig/m3). As was observed for the distribution
of PM-10, fine aerosol concentrations also increase towards the eastern U.S. Figures 2.15 - 2.20 show
isopleths of coarse mass, fine sulfate, fine nitrate, organic, elemental carbon, and soil aerosol
concentrations for the same three-year period from the IMPROVE network. These figures show the
general tendency for concentrations of fine particles (i.e. those most important to determining visibility)
to be highest in the eastern U.S. and in southern California, and lowest in the relatively unpopulated
regions of the western United States. The largest single component of the fine aerosol in the east is
sulfate, while in the Pacific Northwest it is organics, and in southern California it is nitrate. In general,
the largest mass fractions of the fine aerosol are sulfates and organics (Sisler, et al., 1996).
National PM-2.5 trends are not available because of the limited number of sites measuring PM-2.5
and the sampling period at most sites is restricted to a few years. The development of national trends is
further hindered because a variety of sampling frequencies and techniques have been used for PM-2.5.
Visibility data have been used as a surrogate to estimate fine particle trends because the extinction
coefficient (B^,) is directly related to fine particle mass (EPA, 1996d). Sufficient visibility data are
available to produce national trends from 137 U.S. sites (principally airports) since 1948. In the
CHAPTER 2: Current Environmental State 2-25
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
10
PM10
ANNUAL
8.5
4.2 Denoli N.P.
Figure 2.11: Average PM-10 mass concentration (in /xg/m3) for each site in the IMPROVE network
(Sisler, etal., 1996).
2-26 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Concert rst ion, ugAn3
35|
30
25
20
15
10
5
Rural(1Z7stK) Suburban {378 stes) Uten f«8gtasj
jTBOhftrffintile
f-Msai
^-lOthftrcentile
955Sitffi
88 89 90 91 92939495
Concentration, ug/m3
70
60
50
40
30
20
10
0
8889909192939495
Figure 2.12: Trend in the annual mean PM-10 concentrations and the PM-10 annual mean
concentration trends by location, for the period 1988-1995 (EPA, 1996b).
CHAPTER 2: Current Environmental State 2-27
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Figure 2.26: Quarterly averaged extinction coefficients based on the combination of IMPROVE,
NESCAUM and FAA data (from Husar, 1996).
2-44 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
PM2.5
ANNUAL
1.8 Denoli N.F.
Figure 2.14: Average fine mass (PM-2.5) concentration (in
network (Sisler, etal., 1996).
for each site in the IMPROVE
CHAPTER 2: Current Environmental State 2-29
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
3.5
COARSE MASS
ANNUAL
2.3 Denali N.P.
Figure 2.15: Average coarse particle mass concentration (in
network (Sisler, etal., 1996).
for each site in the IMPROVE
2-30 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
AMMONIUM SULFATE
ANNUAL
0.5 Denoli N.P
Figure 2.16: Average fine sulfate aerosol concentrations (in /ig/m3) for each site in the IMPROVE
network (Sisler, etal., 1996).
CHAPTER 2: Current Environmental State 2-31
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
AMMONIUM NITRATE
ANNUAL
0.1 Denali N.P.
Figure 2.17: Average fine nitrate aerosol concentrations (in /ig/m3) for each site in the IMPROVE
network (Sisler, et al., 1996).
2-32 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
ORGANICS
ANNUAL
3.0
2.5
2.5
1.0 Denali N.P.
Figure 2.18: Average fine organic aerosol concentrations (in /ig/m3) for each site in the IMPROVE
network (Sisler, et al., 1996).
CHAPTER 2: Current Environmental State 2-33
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
ELEMENTAL CARBON
ANNUAL
0.4
0.3
0.1 Denoli N.P,
Figure 2.19: Average fine elemental carbon aerosol concentrations (in jig/m3) for each site in the
IMPROVE network (Sisler, et al.f 1996).
2-34 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
FINE SOIL
ANNUAL
0.2 Denoli NP.
Figure 2.20: Average fine soil aerosol concentrations (in
network (Sisler, etal., 1996).
for each site in the IMPROVE
CHAPTER 2: Current Environmental State 2-35
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
northeastern states, winter haze shows a 25 percent decrease while in the southeastern states, there is a
40 percent increase in winter haze (NAPAP, 1991). For the National Acid Precipitation Assessment
Program (NAPAP) analyses, the Northeast was defined as Indiana, Ohio, Pennsylvania, New York,
Kentucky, West Virginia and the New England States, and the Southeast was defined as states south of
the Ohio River and east of the Mississippi River. The summer haziness in the Northeast shows an increase
up to the mid-1970s followed by a decline. In the Southeast, there was an 80 percent increase in summer
haziness, mainly occurring in the 1950s and 1960s (NAPAP, 1991).
Visibility has been monitored with more precision by the IMPROVE network from 1987 to
present. In eastern remote locations, air quality data from 1982 to 1992 showed roughly a 3 percent
increase in sulfate mass concentration during the summer and a smaller negative (although not statistically
significant) trend in the winter (Eldred and Cahill, 1994). Western visibility monitoring through the
IMPROVE network has not shown any trends for the period.
Background Levels
Natural sources contribute to both fine and coarse particles in the atmosphere. Natural (i.e.
background) PM concentration data are defined as the distribution of PM concentrations that would be
observed in the U.S. in the absence of anthropogenic emissions of PM and precursor emissions of volatile
organic compounds (VOC.), NOX, and SOX in North America. When scientists are asked to separate
complicated, overlapping natural and anthropogenic emission sources when conducting source
apportionment of ambient PM concentration data for the purposes of developing controls, the regulatory
result is often driven by policy decisions like the May 1996 EPA Natural Events Policy.
Background levels of PM vary by geographic location and season. The natural component of the
background arises from physical processes of the atmosphere that entrain fine particles of crustal material
(i.e., soil) as well as emissions of organic particles resulting from natural combustion sources such as
wildfire. In addition, certain vegetation can emit fine organic aerosols as well as their precursors. The
exact magnitude of the natural portion of PM for a given geographic location can not be precisely
determined because they are difficult to separate from the long-range transport of anthropogenic particles
or precursors. Only broad estimates for longer averaging times can be developed at this time, as shown
in Table 2.6. Both the magnitude and composition of these regional average mass loadings are quite
variable for the temporal and spatial scales represented by the monitoring sites used to construct these
averages. Great care should be taken in assuming the accuracy and/or precision of these data for use in
site-specific, shorter-timeframe regulatory conclusions which might be made when interpreting the table
below.
As noted in the estimates, there is a definite geographic trend to these levels with the lower value
applicable to the Western U.S. and the somewhat higher value applicable to the Eastern U.S. The Eastern
U.S. is estimated to have more natural organic fine particles and more water associated with hygroscopic
fine particles than the West.
2-36 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
TABLE 2.6: PM-10 and PM-2.5 regional background levels (from EPA, 1996e).
PM-10 , annual average
PM-2.5 , annual average
Western U.S. (/ig/m3)
4-8
1-4
Eastern U.S. G*g/m3)
5-11
2-5
Source: EPA, 1996d, page 6-44. The lower bounds of the above ranges are based on
compilations of natural versus human-made emission levels, ambient measurements in remote
areas, and regression studies using human-made and/or natural tracers (NAPAP, 1991; Trijonis,
1982). The upper bounds are derived from the multi-year annual averages of the clean remote
monitoring sites in the IMPROVE network (Malm et al., 1994). It is important to note, however,
that IMPROVE data used here reflect the effects of background and anthropogenic emissions from
within North America and therefore provide conservative estimates of the upper bounds.
CHAPTER 2: Current Environmental State 2-37
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
The range of background concentrations can be higher on an episodic basis. Specific natural
events such as wildfires, volcanic eruptions, and dust storms can lead to very high levels of PM.
Disregarding such large and unique events, an estimate of range of typical background on a daily basis
can be obtained from reviewing various multi-year data as well as special field studies. On very clean
days, IMPROVE daily measurements are less than 1 /ig/m3 of PM-2.5. On some days atmospheric
conditions are more conducive to accumulation and formation of PM from both natural and anthropogenic
emissions sources. Upper bound estimates of daily background under these conditions can be made using
field data, such as short-term special studies. One study measured daily concentrations of up to
approximately 12/Ag/m3 PM-10 in remote clean areas of the Eastern U.S. (Wolff etal., 1983).
2.C Visibility/Regional Haze Air Quality Characterization
Overview of Current Visibility Conditions
Annual average visibility conditions vary regionally across the U.S. The rural East generally has
higher levels of impairment than remote sites in the West, with the exception of the San Gorgonio
Wilderness, Point Reyes National Seashore, and Mount Rainier, which have annual average levels
comparable to rural sites in the Northeast. Higher averages in the East are due to generally higher
concentrations of anthropogenic fine particles and precursors, higher background levels of fine particles,
and higher average relative humidity levels. Visibility conditions also vary significantly by season of the
year. With the exception of remote sites in the northwestern U.S., visibility is typically worse in the
summer months. This is particularly true in the Appalachian region, where average extinction in the
summer exceeds the annual average by 40%. Figures 2.21 - 2.25 summarize the annual and seasonal
average visibility impairment for the three-year period March 1992 through February 1995 from the
IMPROVE network data (Sisler, et al., 1996).
Figure 2.26 presents 1990 quarterly averages (75th percentile) of monitored visibility levels based
on a fusion of IMPROVE, NESCAUM and Federal Aviation Authority (FAA) data nationally. Clear
regional and seasonal patterns are evident. The eastern U.S. is strongly affected by summertime visibility
degradation. Western U.S. (i.e. southern California) visibility degradation is apparent in summer and
winter seasons (Husar, 1996).
Estimated Background Levels of Fine Particles and Associated Light Extinction
Total light extinction is determined by the combined effects of fine particles from background
(i.e., geogenic and biogenic) sources, Rayleigh scattering (i.e., the degree of light extinction that would
be found even in a particle-free atmosphere), and fine particles from anthropogenic sources. Chapter 6
of the Criteria Document identifies several alternative definitions of "background" levels of fine particles.
For the purposes of this document, background PM is defined as the distribution of PM concentrations that
would be observed in the U.S. in the absence of anthropogenic emissions of PM and precursor emissions
of VOCS, NOX, and SOX in North America. Previously, Table 2.6 showed the range for annual average
regional background PM-fine mass to be 2-5 ^g/m3 and 1-4 jig/m3 in the east and west, respectively. The
2-38 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
VISIBILITY IMPAIRMENT l!\
ANNUAL
17
DECIVIEWS (DV)
15
8 Denali N.P.
Figure 2.21: Average annual visibility impairment in deciviews calculated from total (Rayleigh
included) reconstructed light extinction for the three-year period, March 1992 through
February 1995, for IMPROVE sites (Sisler, et al., 1996).
CHAPTER 2: Current Environmental State 2-39
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
17
15
7 Denali N.P.
Figure 2.22: Average winter visibility impairment in deciviews calculated from total (Rayleigh
included) reconstructed light extinction for the three-year period, March 1992 through
February 1995, for IMPROVE sites (Sisler, et al., 1996).
2-40 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
20
17
20
9 Denali N.P.
Figure 2.23: Average spring visibility impairment in deciviews calculated from total (Rayleigh
included) reconstructed light extinction for the three-year period, March 1992 through
February 1995, for IMPROVE sites (Sisler, et al., 1996).
CHAPTER!: Current Environmental State 2-41
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
20
20
20
10 Denoli N.P.
Figure 2.24: Average summer visibility impairment in deciviews calculated from total (Rayleigh
included) reconstructed light extinction for the three-year period, March 1992 through
February 1995, for IMPROVE sites (Sisler, et al., 1996).
2-42 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
17
15
13
7 Denali N.P.
Figure 2.25: Average autumn visibility impairment in deciviews calculated from total (Rayleigh
included) reconstructed light extinction for the three-year period, March 1992 through
February 1995, for IMPROVE sites (Sisler, et al., 1996).
CHAPTER 2: Current Environmental State 2-43
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Figure 2.26: Quarterly averaged extinction coefficients based on the combination of IMPROVE,
NESCAUM and FAA data (from Husar, 1996).
2-44 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
lower bounds of these ranges, taken from estimates in the 1990 report of the National Acid Precipitation
Assessment Program, are based on compilations of natural versus human-made emission levels, ambient
measurements in remote areas, and regression studies using human-made and/or natural tracers (NAPAP,
1991; Trijonis, 1982). The upper bounds are derived from the multi-year annual averages of remote
monitoring sites in the IMPROVE network (Malm et al., 1994). It is important to note, however, that
IMPROVE data used here reflect the effects of background and anthropogenic emissions from within
North America and therefore provide conservative estimates of the upper bounds. Table 2.7 from the
NAPAP report includes estimates of annual average background levels of fine particles by aerosol
constituent, as well as their related contributions to light extinction, expressed in inverse megameters
(Mm"1) (NAPAP, 1991). On an hourly or daily basis, however, background concentrations will vary
considerably depending on seasonal, meteorological, and geographic factors.
Role of Humidity in Light Extinction
As mentioned previously, humidity plays a significant role in the impairment of visibility by fine
particles, particularly in the East, where annual average relative humidity levels are 70-80% as compared
to 50-60% in the West (Sisler et al., 1993). Humidity is especially important (depending on air trajectory)
along the California coastal regions and offshore waters. Table 2.7 accounts for relative humidity effects
by assigning an extinction efficiency for water associated with aerosols, while extinction efficiencies found
in Table 2.8 are modified by a relative humidity adjustment factor in calculating total extinction. The
adjustment factor represents (1) the hygroscopic nature of the aerosol constituent, and (2) the average
annual humidity for the relevant location (Sisler et al., 1993).
Because annual average relative humidity is higher in the East, the same ambient concentration
of sulfate, for example, will on average lead to greater light extinction in an eastern location rather than
a western one. The top map in Figure 2.27 illustrates the regional variability of annual mean relative
humidity nationwide. The bottom map depicts the variability of the relative humidity correction factor
used for sulfates in the IMPROVE analysis (Sisler et al., 1993). For example, when corrected for
humidity, the overall extinction efficiency for sulfates in the East may exceed 11-12 m2/g, whereas the
extinction efficiency for sulfate in the West may be one-third to one-half of that.
Rayleigh Scattering
Rayleigh scattering represents the degree of light extinction found in a particle-free atmosphere
caused by the gas molecules that make up "blue sky" (e.g., N2, 02, CO2 )(U.S. EPA, 1979). The concept
of Rayleigh scattering can be used to establish a theoretical maximum horizontal visual range in the earth's
atmosphere, absent visibility-impairing particles. At sea level, this maximum visual range would be
approximately 330 kilometers (equivalent to light extinction of 10-12 Mm"1). Table 2.7 shows the
contribution to total light extinction from "Rayleigh scattering." While Rayleigh scattering can be used
in calculating this maximum visual range, it is analagous to a baseline or boundary condition and should
not be considered to contribute to visibility impairment. Rather, only fine and coarse particles from
natural and anthropogenic sources should be treated as factors in determining visibility impairment.
CHAPTER 2: Current Environmental State 2-45
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Table 2.7: Average natural background levels of aerosols and light extinction (from EPA", 1996e).
Average
Concentration
East
Average
Concentration
West
Error
Factor
Extinction
Efficiencies"
(m'/g)
Extinction
Contributions
-East
(Mm'1)
Extinction
Contributions
-West
(Mm'1)
Fine Particles (s 2.5 jim)
Sulfates
Organics
Elemental
Carbon
Ammonium
Nitrate
Soil Dust
Water
0.2
1.5
0.02
0.1
0.5
1.0
0.1
0.5
0.02
0.1
0.5
0.25
2
2
2-3
2
1.5-2
2
2.5
3.75
10.5
2.5
1.25
5
0.5
5.6
0.2
0.2
0.6
5.0
0.2
1.9
0.2
0.2
0.6
1.2
Coarse
Particles (2.5 -
10 /mi)
3.0
3.0
1.5-2
0.6
Rayleigh Scatter
Total
1.8
12
26±7
1.8
11
17±2.5
The extinction efficiencies are based on the literature review by Trijonis et al: (1986 and 1988).
All the extinction efficiencies represent particle scattering, except for elemental carbon where the
10.5 m2/g value is assumed to consist of 9 nvVg absorption and 1.5 m2/g scattering. Note that the
0.6 nrVg value for coarse particles is a pseudo-coarse scattering efficiency representing the total
scattering by all ambient coarse particles (2.5 /*m) divided by the coarse particle mass between
2.5 and 10
2-46 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Table 2.8: Dry particle light extinction efficiency values used in 1996 analysis of IMPROVE data
(from EPA, 1996e).
Aerosol Constituent
Sulfates
Nitrates
Organics
Soil Dust
Coarse Particles
Extinction
Efficiency
(in m'/g)
S.O^RH)
3.0/(RH)
4.0
1.0
0.6
/(RH) is the relative humidity correction factor. It
is the ratio of wet scattering divided by dry
scattering.
CHAPTER 2: Current Environmental State 2-47
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
(a) Annual mean relative humidity.
(b) Sulfate relative humidity correction factor FT.
Figure 2.27: Spatial variation in the (a) average relative humidity and the (b) sulfate relative humidity
correction factor (from EPA, 1996e).
2-48 CHAPTER 2: Current Environmental State
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Table 2.7 illustrates that estimated extinction contributions from Rayleigh scattering plus
background levels of fine and coarse particles, in the absence of anthropogenic emissions of visibility-
impairing particles, are 26 plus or minus 7 Mm"1 in the East, and 17 plus or minus 2.5 Mm'1 in the West.
These equate to a visual range in the East of 150 plus or minus 45 kilometers and 230 plus or minus 40
kilometers in the West. Excluding light extinction due to Rayleigh scatter, annual average background
levels of fine and coarse particles are estimated to account for 14 Mm'1 in the East and about 6 Mm"1 in
the West. Major contributors that reduce visibility from the Rayleigh maximum (approximately 330 km
visual range) to the ranges noted above are naturally-occurring organics, water, and suspended dust
(including coarse particles) in the East, and suspended dust, organics, and water in the West. In these
ranges of fine particle concentrations, small changes have a large effect on total extinction. Thus, one can
see from Table 2.7 that higher levels of background fine particles and associated humidity in the East
result in a fairly significant difference between estimates of naturally-occurring visual range in the East
and West.
CHAPTER 2: Current Environmental State 2-49
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
2-50 CHAPTER 2: Current Environmental State
-------
CHAPTER 3
Processes: How is the State
Created, Sustained and
Maintained?
This section discusses pollutant emissions, atmospheric conditions, and chemical reactions that
contribute to the formation and persistence of ozone, fine particulate matter, and regional haze in the
atmosphere. These processes can be very complex and may lead to controversy over the most effective
strategies to control each type of pollutant.
3.A Atmospheric Chemistry
Complex chemical reactions dominate the formation of ozone, fine particulate matter, and regional
haze. The following section discusses the formation of ozone near the earth's surface, primary particles as
well as secondary fine particulate matter and other visibility impairing compounds. In addition, a brief
discussion of the fundamentals of visibility is presented.
Ozone
Ozone is both a natural component of the earth's atmosphere and an air pollutant. Ozone is formed
in the stratosphere from molecular oxygen in air and intense ultraviolet solar radiation. Its formation
consumes almost all radiation below 300 nm and thus prevents harmful ultraviolet radiation from reaching
the earth's surface. Ozone is slowly transported from the stratosphere to the earth's surface, primarily
through tropopause folding. Thus, a natural background of 03 exists near the earth's surface that has been
estimated to be near 20 ppb due solely to stratospheric sources. Current background is about 40 ppb and
the increase has been attributed to global pollution including fossil fuel use and agriculturally related biomass
burning. Additionally, current research indicates that it may be slowly increasing. Continental background
O3 (i.e., Eastern US and Western Europe) may be much higher due to region-wide, slow chemical production
processes (similar to those in urban areas and dependent upon widely dispersed NOJ as well as due to direct
transport of urban produced O3.
The general explanation for the formation of ozone in or near urban areas is that combustion sources
that use air as an oxidizer will produce oxides of nitrogen (NO, N02) when temperatures are above 2500°K.
In addition, incomplete combustion results in the emission of raw components and oxygenated organic
compounds from the fuels. In sunlight these are sources of free radicals (e.g., *OH, HO2% RO*, R02«) that
oxidize "emitted" volatile organic compounds to carbonyls, CO, and CO2, while simultaneously oxidizing
NO to NO2 and recreating the free radical (Figure 3.1). Each free radical is cycled up to five times. High
background levels of CO also can assist in creating and sustaining high O3 episodes. The N02 reacts with
CHAPTER 3: Processes: How the State is Created, Sustained and Maintained 3-1
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
total
fe rr rrntrrl mill h> fOHl -4 i
t
II II
[NO2] [CO] [HCHO][RR'CO][CH4] [
I 1 1 |
T mmmmm] -1'O.IJ
Old
t 11
[CO] or [CO2] V T
[H02] ^-
[OH]
re-created [OH] from propagation
iew[OH]*« photo|ysisof°3
i i i
HC^tHC^.-tHCn [RJ
1
^0][RRCO] "
1
|+o2
[R02] -^ new [RO2] -^ photoly;
| of RR'C
- >-NO^
-"^ )
y r.^.5
+NO2
r
oNOj||v
[RR'CO] u .
r, ,« i ^ photolysis
- new [HO2] -« ot HCHO & RRCO
W WfRO2
new [OH] from RRCO photolysis
is
Figure 3.1: Details of the OH-chain cycle. The gray boxes represent the production of termination
products (Jeffries, 1997).
3-2 CHAPTER 3: Processes: How the State is Created, Sustained and Maintained
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
sunlight to recreate NO and to produce 03. After the first oxidation of NO to NO2 , every subsequent
operation of the cycle produces an 03 molecule with an efficiency of greater than 90%. hi current chemical
reaction mechanisms, a typical nitrogen is cycled 3 to 5 times (Figure 3.2). Some of the 03 produced reacts
with organics and with sunlight to produce more free radicals to maintain the cyclic oxidation process. This
represents a powerful positive feedback process on the formation of more 03, given available NOX.
The carbonyls produced in the organic oxidation also react with sunlight to produce more free
radicals. As the cycle operates, N02 is converted into inorganic and organic nitrates, a form of nitrogen that
cannot cycle. In addition, this process also removes free radicals. A system that converts all nitrogen oxides
to such products cannot create any more O3. Since NO2 reacts rapidly with free radicals, in situations that
have a limited supply of radicals, NO2 can compete with the VOCs for the limited free radicals, resulting in
virtually no production of 03. Thus, different mixtures of VOCs and NOX can result in different ozone levels
such that the total system becomes nonlinear. That is, large amounts of VOC and small amounts of NOX
make O3 rapidly, but are quickly limited by removal of the NO^ VOC reductions, under these circumstances,
show little effect on 03. Large amounts of NO and small amounts of VOC (which usually implies smaller
radical source strengths) result in the formation of inorganic nitrates, but little 03. In these cases, reduction
of NOX results in an increase in 03. Some combination of VOC and NOX is optimum at producing 03.
The preceding is a static description. In reality, physical processes compete with chemical processes
and change the outcomes in complex ways. The existence of feed back and non-linearity in die
transformation system confounds the description. Competing processes determine the ambient concentration
and there are an infinite set of process magnitudes that can give rise to the same ambient concentrations and
changes in concentrations. Lack of any direct measurement of process magnitudes results in the need to use
inferential methods to confirm any one explanation of a particular O3 concentration.
Primary Particles
Unlike the other criteria pollutants (SOj, NOj, CO, 03, Pb), which are well defined chemical entities,
ambient PM-fine is comprised of a complex mixture of chemical constituents. Because of this fact, sources
of each constituent of the atmospheric aerosol must be considered in turn. Since participate matter is
composed of both primary and secondary constituents, emissions of both the primary components and the
gaseous precursors must be considered.
Tables 3.1 and 3.2 summarize the discussion presented earlier in Chapter 2 on anthropogenic and
natural sources for the major primary and secondary aerosol constituents of fine and coarse particles.
Anthropogenic sources can be further divided into stationary, area, fugitive, and mobile sources. Stationary
sources include fuel combustion for electrical utilities and industrial processes; construction and demolition;
metals, minerals, petrochemicals and wood products processing; mills and elevators used in agriculture;
erosion from tilled lands; waste disposal and recycling; and, fugitive dust from paved and unpaved roads.
Mobile (e.g. transportation-related) sources include direct emissions of primary PM and secondary PM
precursors from highway and off-highway vehicles and non-road sources. Also shown are sources for
precursor gases whose oxidation forms secondary particulate matter. The latter topic will be discussed in
greater detail in the following section. In general, the nature of sources of particulate matter shown in Table
CHAPTER 3: Processes: How the State is Created, Sustained and Maintained 3-3
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
^ 1 / + H20
jrO-"
^~\? F
r^
JnftiatloB
\x, , f, s ii/
JflSial Species. V
\^«-battos -
q
MO . i
Emission ^
1
Nr = number of OH cycles = Q/q = 1 / (l-Pr)
PrQ
r^ ^^S*^"Ai'
muinilinriiMinii ''""'' '"..".... .%- ! s'^
' ; ^ Propagation : , '- \ -
^^b i vfS'j in J j*j-i1 "*
*qf j
NOtoN02 oA"
_, ^ oxidation Or
& Reaction i
^^ Nitrogen Products ^
Loss of NO and NO2
PnE
prc
>J
1
03
photolysis
* Total O3
Production
^ Af
03 -
Reaction
Loss
J
i
y
NetO3
Production
Nn = number of cycles = E/e = 1 / (l-Pn)
H = organic processes Q = inorganic processes
Figure 3.2: Schematic of the OH-chain VOC and NOX oxidation cycle. The gray boxes include mostly
organic processes and the white boxes include mostly inorganic processes (Jeffries, 1997).
3-4 CHAPTER 3: Processes: How the State is Created, Sustained and Maintained
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Table 3.1:
Constituents of atmospheric fine particles (< 2.5 um) and their major sources.
Sources
Primary
Aerosol
Species
so<-
NO,
Minerals
NH«+
Organic
Carbon (OC)
Elemental
Carbon
Metals
Bioaerosols
Natural
Sea Spray
N/A
Erosion,
Re-
entrainment
N/A
Wild Fires
Wild Fires
Volcanic
Activity
Viruses,
Bacteria
Anthropogenic
Fossil Fuel
Combustion
Motor Vehicle
Exhaust
Fugitive Dust, Paved,
Unpaved Roads; Fuel
Combustion; Agriculture
and Forestry
Motor Vehicle
Exhaust; Fertilizer
Application
Open Burning;
Prescribed Fire;
Wood Burning;
Cooking; Motor
Vehicle Exhaust; and,
Tire Wear
Motor Vehicle
Exhaust; Wood Burning;
Prescribed Fire; and,
Cooking
Fossil Fuel Combustion;
Smelting; and Brake Wear
N/A
Secondary
Natural
Oxidation of Reduced Sulfur
Oases Emitted by the Oceans
and Wetlands; and, SO2 and
H2S Emitted by Volcanism
and Forest Fires
Oxidation of NO, Produced by
Soils, Forest Fires, and
Lightning
N/A
Emissions of NH3 from Wild
Animals and Undisturbed Soils
Oxidation of
Hydrocarbons Emitted
by Vegetation (e.g.
terpenes, waxes); and,
Wild Fires
N/A
N/A
N/A
Anthropogenic
Oxidation of SO2 Emitted
from Fossil Fuel
Combustion
Oxidation of NOX Emitted
from Fossil fuel
Combustion; and Motor
Vehicle Exhaust
N/A
Emissions of NH3 from
Animal Husbandry,
Sewage; and, Fertilized
Land, Point Sources
Oxidation of
Hydrocarbons
Emitted by Motor
Vehicles; Open
Burning; Prescribed
Fire; and. Wood
Burning
N/A
N/A
N/A
CHAPTER 3: Processes: How the State is Created, Sustained and Maintained 3-5
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Table 3.2:
Constituents of atmospheric coarse particles (> 2.5 jim) and their major sources.
Sources
Primary
. Aerosol
Species
Minerals
Miscellaneous
Ions
Organic
Carbon (OC)
Metals
Organic
Debris
Bioaerosols
Natural
Erosion,
Re-entrainment
Sea Spray
N/A
Erosion; Re-
entrainment;
and, Organic
Debris
Plant and Insect
Fragments
Pollen; Fungal
Spores;
Bacterial
Agglomerates
Anthropogenic
Fugitive Dust; Paved,
Unpaved Roads; Agriculture
and Forestry
Road Salting
Tire Wear and Asphalt Wear
N/A
N/A
N/A
Secondary
Natural
N/A
N/A
N/A
N/A
N/A
N/A
Anthropogenic
N/A
N/A
N/A
N/A
N/A
N/A
3-6 CHAPTER 3: Processes: How the State is Created, Sustained and Maintained
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
3.1 is very different from that for paniculate matter shown in Table 3.2. A large fraction of the mass in the
fine size fraction is derived from material that has been volatilized in combustion chambers and then
recondensed to form primary fine PM, or has been formed in the atmosphere from precursor gases as
secondary PM. Since precursor gases and fine particulate matter are capable of traveling great distances, it
is difficult to identify individual sources of the constituents shown in Table 3.1. The PM constituents shown
in Table 3.2 have shorter residence times in the atmosphere, so their impacts tend to be more localized. Only
major sources for each constituent are listed in these tables.
Natural sources of primary PM include aeolian processes across undisturbed land, sea spray and
plant debris, wildfires, earthquakes, volcanism and insect debris. The oxidation of a fraction of terpenes
emitted by vegetation and reduced sulfur species from anaerobic environments leads to secondary PM
formation. Ammonium (NH/) ions which are crucial for regulating the pH of particles are derived from
emissions of ammonia (NH3) gas. Source categories for NH3 have been divided into emissions from
undisturbed soils (natural) and emissions which are related to human activities (e.g. fertilized lands, domestic
and farm animal waste). Clearly, fuel wood burning is an anthropogenic source of PM, whereas wildfires
would be a natural source. Forest fires have been included as a natural source because of the lack of
information on the amount of prescribed burning or accidental fires caused by humans. Future policy by the
EPA will likely provide guidance on burning, which may alter and/or better define and segregate the
inventory of forest fire emissions into anthropogenic and natural categories to assist in regulating ambient
air quality.
Secondary Fine PM
The generation, transformation, transport and removal of particles involves a complex interaction
of numerous mechanical, biological, physical and chemical processes. Although large particles (diameters
> 2.5u) share some attributes with fine particles, there are several fundamental differences between fine and
large particles that allow for separate treatments. Perhaps the most important difference between fine and
large particles is that the major fraction of large particle mass is generated from primary emissions of
materials, whereas the dominant component of fine particle-mass is formed via secondary reactions and
condensation processes. While this simple picture appears to reflect Eastern U.S. sulfate laden aerosols, both
the complexity of aerosols and the scarcity of particle mass speciated data are causes for concern regarding
our ability to formulate basic process understandings of aerosols.
Organic Aerosol Formation
The understanding of secondary-formed organic aerosols is less clear relative to inorganic aerosols.
The principal available secondary organic aerosol model (Pandis et al., 1992) assumes that organic
compounds with relatively low vapor pressures (semi-volatile organic compounds, SVOC) condense or
nucleate when their concentrations exceed saturation concentrations. This gas-particle partitioning is
typically described by equilibrium relationships of a solute between aqueous and vapor phases (e.g., Henry's
law). Secondary organic aerosols are derived mostly from aromatics (toluene, xylene, benzene compounds)
and various C7 and greater olefins, especially the biogenic terpenes (e.g., Smoky mountain haze). These
volatile organics initially react with hydroxyl radicals, and subsequently form lower vapor pressure
CHAPTER 3: Processes: How the State is Created, Sustained and Maintained 3-7
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
condensable products. Estimates of aerosol yields of individual VOC species (emissions) have been
developed (Grosjean and Seinfeld, 1989) and are used to parameterize organic aerosol formation in models.
Primary organic aerosols generally dominate the total organic aerosols fraction, although summertime
secondary organic formation could be significant Measurement analyses typically provide estimates of total
organics and, therefore, do not distinguish primary and secondary organic aerosol components.
Sulfate Formation and Sulfate-Nitrate-Ammonia System
Sulfur dioxide, emitted from many of the same combustion related sources releasing NOx, is
oxidized by »OH to form sulfuric acid (H2S04), assuming an ammonia free environment. Intermediate steps
within this oxidation generate hydroperoxy radicals, thus continuing the cycling of radicals (i.e., this is not
a termination step). In addition to the gas phase oxidation of sulfur dioxide, significant oxidation of SOX
occurs in clouds, fogs and available water droplets. Sulfur dioxide is soluble in water and exists in
equilibrium with various reduced-form sulfur species, including bisulfite (HS03+) which is oxidized to sulfate
in the presence of dissolved hydrogen peroxide (H202) or ozone (in typical acid environments, oxidation
through H2O2 is favored) or oxygen, catalyzed by dissolved iron and manganese. On an annualbasis, gas
and aqueous oxidation pathways produce comparable sulfate yields; gas phase oxidation is enhanced during
summer months.
An equilibrium system exists among several gas-phase and liquid-phase nitrogen and sulfur species
where ammonia plays a potentially important role in formation of ammonium nitrate particles. Gaseous
ammonia is highly soluble and can act as a neutralizing agent creating a spectrum of sulfate based
ions/compounds ranging in acidity from sulfuric acid to neutralized ammonium nitrate. Ammonia
preferentially reacts with sulfates and stoichiometrically requires 2 equivalents, versus 1 for nitric acid, for
neutralization. Excess ammonia would be available to establish equilibria among ammonia and nitric acid
gases and ammonium nitrate (solid/liquid phases). Ammonium nitrate is relatively unstable, and reforms
nitric acid and ammonia gases at elevated temperatures (e.g. > 10°C). Because of the abundance of sulfate
and temperature effects, the Eastern U.S. is believed to be relatively nitrate free in the summer. Ammonium
nitrate typically is a winter time issue; it can be the dominant winter time aerosol component in southern
California. Unfortunately, this instability creates substantial measurement problems. For reasons of
consistency, filter samples often are equilibrated at room temperatures and humidity before weighing,
resulting in loss of excess water (a dominant component) and semi-volatile compounds (e.g., organic
aerosols, ammonium nitrate). Additional losses occur during sample transport and storage.
The dominance of sulfate aerosols in the Eastern U.S. suggests a limiting effect on ammonium nitrate
formation due to ammonia, although nitrate hot spots probably exist in areas not monitored with local high
ammonia and NOX emissions. NOX emissions could lead to increasing nitrate aerosol levels if greater relative
reductions of SOX emissions (to NOJ occur in future years. Southern California aerosols are dominated by
nitrates and organics (relative to sulfates), making this consistent with larger NO,/SOX inventories and
available ammonia sources.
Fine particle growth often is affected by available water. Sulfate and nitrate-based particles are
strongly hygroscopic and much of the particle bound mass, especially at high (e.g., > 95%) relative
3-8 CHAPTER 3: Processes: How the State is Created, Sustained and Maintained
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
humidities, consists of water. The ammonium-nitrate-sulfate system is deliquescent. Deliquescent particles
undergo a rapid growth phase after transitioning from solid to liquid phases at high relative humidities. The
combination of high sulfate levels and relative humidity collectively act to impair visibility more in the
Eastern than the Western U.S. (i.e. water accretion shifts particle size distributions toward more effective
light scattering regimes). Fine particles often are highly water soluble, providing an eventual pathway for
removal through precipitation. In addition, many atmospheric gases (e.g., nitric acid and hydrogen peroxide)
which have important roles in particle formation are extremely soluble and play significant roles in aqueous
phase oxidation reactions that generate particles.
Regional Haze/Visibility
Light scattering and absorption by particles and gases contribute to visibility extinction. Particles
are the major cause of visual impairment in polluted atmospheres, with minor contributions due to light
absorption by N02. Particle scattering dominates (compared to absorption), especially in rural
environments. However, light absorption by soot, usually associated with diesel fuel combustion, is
extremely efficient and a significant component of the light extinction budget in urban environments.
Absorption and scattering are functions of particle composition (refractive index), size distribution and
concentration. Light extinction is strongly dependent on particle size distribution with peak efficiencies,
depending on species, in the .1 to lu range (similar to the visible light spectrum, .4 - .7u ). Thus,
characterization of particle size distributions by component (i.e., sulfate, nitrate, organic carbon, elemental
carbon) provides the basis for developing light extinction budgets. Modeled visibility calculations typically
are developed through post-processing of simulated particle fields. Sulfate, nitrate and organic carbon
aerosols effectively scatter light and have similar scattering (mass basis) efficiencies, although organic
aerosols appear to be somewhat more efficient on a mass unit basis. Water droplets are very effective light
scatterers but, with the exception of fogs, most "pure" water is in the vapor phase. Hygroscopic aerosols
(sulfates and ammonium nitrate), which accrete water, change both the aerosol size distribution and refractive
index which, in turn affects scattering, particularly at relative humidities greater than 70%. As particles
accrete water, their mean diameters shift upward, often into very efficient light scattering ranges. Thus, it
is the impact of water on the total aerosol that creates "haze" conditions rather than homogeneous water
droplets acting alone. With typically stronger early morning (high relative humidity) impacts, relative
humidity exhibits large temporal variations . However, secondary formation tends to peak in the afternoon,
during a diurnal low in the relative humidity. Therefore, it is more reasonable to view humidity effects on
an average basis (i.e., high humidity days impart greater visual impairment).
Fundamental understanding of visibility extinction is based on Rayleigh and Mie scattering theories,
both of which involve relationships between particle (and gas) dimensions and incident wavelengths.
Rayleigh scattering occurs in all directions (forward and backward) and is limited to atmospheric gases
(mostly Nj) and very small particles, with diameters considerably less than the spectral range of visible light
(~ .4 - .7 u). Clean, background conditions are referred to as "Rayleigh" atmospheres with visibility up to
330 km (i.e., some natural extinction exists). Blue skies result from dominant Rayleigh scattering of
wavelengths in the blue region (~ .5 u) by atmospheric N2, which comprises nearly 80% of all atmospheric
gases. Very small particles (<. 1 u) preferentially scatter blue light, leaving larger wavelength red hues
CHAPTER 3: Processes: How the State is Created, Sustained and Maintained 3-9
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
relatively unaffected from a line-of-sight perspective during clean condition (very small particles)
sunset/sunrise periods.
Mie scattering (forward and sidewards) by particles with diameters similar to the visible light
spectrum is the dominant visibility impairment mechanism. Thus, most efficient scattering is associated with
particles on the order of. 1 \i. Mie theory provides a complete solution for the interaction between a polarized
plane wave and a dielectric sphere with a complex refractive index. Thus, for a real refractive index, only
scattering can occur; however, for a complex index, both scattering and absorption are possible, and the
coefficients depend upon both the real and imaginary components. Low- and medium-level clouds consist
of water droplets, which allow for the direct application of Mie'e scattering theory (Goody, 1995).
3.B Primary and Secondary Emissions Processes
The formation of ozone, fine paniculate matter, and regional haze are dependent on both direct
emissions and secondary interactions. Ozone is produced through a series of complicated interactions
between HC and NOX in the presence of sunlight. The dominant component of the fine particle mass is
formed from secondary reactions and condensation processes. Secondary pollutants, such as sulfate aerosols
and nitrate aerosols are also important in the formation of regional haze. The following section discusses
the sources of these emissions and their contribution to the formation of ozone, fine particulate matter, and
regional haze.
Development of Ozone and Precursor Emission Inventories for 1990 CAA Requirements
Ozone and precursor emission inventory methodology can be summarized by listing its four major
emissive components: point, area, mobile and biogenic sources. Point sources are large, stationary
individually inventoried emission points generally using two methods to identify and estimate emissions.
One method estimated emissions from fossil fuel steam electric utility units by applying an emission factor
(i.e., pounds of pollutant emitted per ton of fuel consumed) to the fuel consumed by each utility unit in the
U.S. based on Department of Energy utility fuel consumption data. Another method estimates all remaining
point sources from the 1985 National Acid Precipitation Assessment Program (NAPAP) emission inventory,
which uses bottom up projection methodologies to estimate emissions. These point source emissions are
projected from 1985 to 1990 using the Bureau of Economic Analysis (BEA) growth factors for each
industrial segment.
Area sources are emission sources that are too small and/or too numerous to be individually
inventoried. As with point sources, some of these emissions are projected from the 1985 NAPAP inventory
using BEA growth factors for each industrial segment. Two segments of area sources estimate emissions
using different methods. Solvent VOC emissions are estimated from the total quantity of solvents consumed
in the U.S. by assuming that all solvents are eventually emitted to the atmosphere. Then these emissions are
apportioned to each county based on county employment figures for specific industry groups using County
Business Pattern statistics. On the other hand, area source emissions from residential wood combustion are
estimated using EPA's wood burning model. This model estimates wood consumption for all counties in the
United States based on survey data from the northeast and several western states. However, other states
3-10 CHAPTER 3: Processes: How the State is Created, Sustained and Maintained
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
(most notably California) which are developing detailed inventories for modeling exercises have developed
more robust estimation methods.
Emissions from on-road vehicles are estimated by applying emissions factors to vehicle activity
levels. The activity level is the total mileage driven or vehicle miles traveled (VMT) (one VMT = one mile
driven by one vehicle). The annual VMT data for each State are obtained from the Federal Highway
Administration's Highway Performance Monitoring System data base. State VMT totals are apportioned
to individual counties based on population. County CO, VOC and NOX mobile source emission factors are
obtained from EPA's MOBILESa model from the Office of Mobile Sources (OMS). The MOBILESa model
includes both evaporative and exhaust emissions. Non-road mobile sources include commercial marine
vessels, lawn and garden equipment, and construction equipment. Emissions from these non-road mobile
sources are estimated from a detailed 1990 emission inventory prepared for 27 nonattainment areas by the
OMS. These inventories were combined, classified by equipment and engine type, and distributed to the
appropriate non-road source categories to create national county-level emissions. Aircraft and railroad non-
road mobile source emissions are projected from the 1985 NAPAP inventory.
Natural sources (e.g. biogenic and geogenic) also contribute significant amounts of VOC and some
NOX to the emission inventory. Various plant species and the soil produce these emissions, which are
estimated using EPA's Biogenic Emissions Inventory System - Version 2 (BEIS2). County land use and
meteorological data were input to this model, producing estimated biogenic emissions by county. Currently,
this inventory is being used by the EPA in national computer modeling simulations to analyze various options
for a revised ozone NAAQS. Individual states prepared bottom up inventories, required by the Clean Air
Act (CAA), for all ozone nonattainment areas for the year 1990. These State Implementation Plan (SIP)
inventories have two applications. First, because the CAA requires specific VOC emission reductions, the
SIP inventories and follow-up periodic inventories are used to determine if the required VOC reductions were
achieved. Second, the States also use the SIP inventories for some nonattainment areas as input to
photochemical grid models to predict future-year ground-level ozone concentrations and to evaluate various
control strategies needed to attain the ozone NAAQS.
Emission estimation methodologies vary among the 33 states and the District of Columbia. The
most significant differences occur in point source emission calculations. EPA guidance requires that all VOC
sources emitting 10 tons/year and NOX and CO sources emitting 100 tons/year to be inventoried indivfdually
as point sources. These point sources are located either in the nonattainment area or in a surrounding 25 mile
wide buffer zone. Emission estimation methods range from Continuous Emissions Monitoring System
(CEMS) emissions data to the application of emission factors and activity levels determined from survey
responses. Although EPA requires states to use the EPA MOBILES a model to estimate on-road mobile
source emissions, application of this model varies significantly, because portions of the MOBILESa model
permit the user to either accept built-in default values or to provide location-specific values. Some states
conduct extensive studies to develop specific inputs while others rely primarily on the default values. Non-
road mobile and biogenic emissions are primarily estimated using a top down methodology. Area sources
are estimated by area source category for nonattainment areas. Estimation methodologies vary, but are often
based on a per capita or per employee emission factor. The activity level ( e.g., population or number of
employees) can be determined locally or obtained from national census data. SIP inventories are submitted
CHAPTER 3: Processes: How the State is Created, Sustained and Maintained 3-11
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
to the EPA Regional Offices (RO), which uses a detailed checklist to review each inventory based on a set
of minimum requirements. The States submitted 97 ozone nonattainment inventories, including 357 of the
total 3141 United States counties.
In the eastern United States, transport from one nonattainment area to other areas has been
recognized as a significant problem. The strategy of treating each nonattainment area independently,
therefore, has been unsuccessful. The Ozone Transport Assessment Group (OTAG), consisting of 38 eastern
states, was formed to address this problem. OTAG's approach is to construct a regional emission inventory,
use this inventory as input to a photochemical grid model, and develop a region-wide attainment strategy.
The emission inventory, based on 1990 emissions, will be a blend of EPA's national top down inventory and
SIP inventories for VOC, NOX and CO. Because SIP data are presumed to be more accurate than EPA's top
down results, the SIP data will be used where available. For those portions of the 38 states not included in
the SIP inventories, EPA's top down inventory will be used.
Particulate Matter and Regional Haze Inventories
Emission inventory requirements are the similar for both particulate matter and regional haze. The
atmospheric constituents that are important to the formation of both PM and RH are: VOC, NOH S02, NH3,
PM-10 and PM-2.5. Much of the top down inventory methodology for VOC and NOX applies to S02, PM-10
and PM-2.5. The main differences are for on-road mobile sources, fugitive dust sources and ammonia
sources. Emissions of PM-10, PM-2.5 and S02 from on-road mobile sources are estimated using EPA/OMS1
particulate model (PART5). PART5 estimates emission factors from vehicle exhaust, brake wear, tire wear
and re-entrained road dust from paved and unpaved roads. These emission factors are then used with VMT
to estimate on-road mobile source emissions. Unique PM source categories in this inventory include wind
erosion, agricultural tilling, and construction activities. Wind erosion emissions from agricultural land are
estimated using Department of Agriculture crop land data, meteorological data, and a wind erosion model.
Agricultural tillage emissions are estimated based on the total area planted, crop type, soil parameters, and
the number of tillings per year. Construction emissions are estimated based on an emission factor, the area
of land under construction (determined from the cost of construction), and the average duration of
construction activity. Ammonia emissions are thought to be dominated by animal husbandry activities and
emission factors are applied to animal populations from the Department of Agriculture's'agriculture census.
Bottom up SIP inventories, required by the CAA, are also prepared for PM-10 nonattainment areas.
One difference between PM-10 and ozone SIP inventories is the relative size of the nonattainment areas.
PM-10 nonattainment areas are generally much smaller than the ozone areas. Also, some PM-10 areas are
dominated by a single source category (e.g., residential wood combustion or fugitive dust). Therefore, the
extent and scope of these inventories are more limited than the ozone inventories. Because PM-10 inventories
are seldom input to computer modeling simulations, no additional quality assurance work associated with
modeling occurred. Therefore, EPA regards current PM-10 SIP inventories to be less rigorous than ozone
SIP inventories. The States have submitted a total of 70 PM-10 nonattainment inventories.
The Grand Canyon Visibility Transport Commission (GCVTC) developed an emission inventory
for an 11 western state region to evaluate regional haze. This emission inventory combines both top down
3-12 CHAPTER 3: Processes: How the State is Created, Sustained and Maintained
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
and bottom up methodologies and is based on EPA's top down ozone inventory. PM and visibility pollutants
have been added using a top down methodology. Certain source categories (e.g., prescribed burning) received
special emphasis in the preparation of the inventory. Each of the 11 states were participants in the inventory
and had to agree with the inventory values for their state. California and Oregon supplied complete data sets
so that none of the original top down values were used. Most of the other states supplied primarily point
source data. While this inventory is a major step forward, many limitations of trying to assemble a
comprehensive inventory for an aggregate indicator like PM-fme were noted when modeling of these data
were attempted. Most notable were problems with applying a coarse (50 km by 50 km) grid cell size to the
data. For example, spatial accuracy of emissions data and precision of fugitive (area source) emissions data
often made the modeling results at the receptor insensitive to large modeled changes in emissions patterns.
Thus, future PM-fine inventories to predict site-specific compliance with the PM NAAQS and/or the NVG
will require numerous improvements in both detail and quality.
3.C Meteorology
Meteorological processes are among the strongest forcing functions affecting air quality at all
atmospheric temporal and spatial scales, including urban and regional ozone, fine particles, and visibility
impairment. The strength of the various meteorological processes can change significantly from day to day,
imposing a confounding signal on the effects of emissions changes on air quality. It is important to
understand these processes as they strongly affect the chemical reactions and rates that can lead to a
deterioration in air quality. They are also primarily responsible for the transport and dilution of directly
emitted pollutants. Ultimately, to ascribe improvements in air quality to emissions reductions over the long
term, the meteorological signal must be removed from the air quality trend. This requires an understanding
of how the meteorological processes affect that signal.
Historically, only local scales of meteorological processes were considered to be important, as air
quality deterioration was thought to be a local problem, hi fact this philosophy governed the incentive to
build taller stacks and chimneys on industrial and power plants in the 1960's and 1970's, in order to emit air
pollutants at higher levels in faster moving air streams and remove the material more quickly from the local
area. Later studies showed that pollutant lifetimes were sufficiently long such that long-range transport could
cany pollution hundreds of miles from the source area and still have an impact on ground level air quality.
The "tall stack movement" was seen to have improved local air quality but at the expense of regional air
quality, which was brought on by pollutant transport. Today we understand better that meteorological
processes on the larger scale help organize the smaller scale processes and that all scales are important to
overall air quality issues.
The largest scale of atmospheric motion is the global scale. Exchange between the northern and
southern hemispheres is on seasonal or longer time scales. The transport of material around the northern
hemisphere was first studied quantitatively using a tracer of opportunity, fallout from nuclear tests. This time
scale ranges from weeks to months depending on the latitude and season. The most important scale for
regional transport, say over North America, is the scale of the typical high and low pressure system, which
is seen on daily weather maps. The spatial extent of these systems is on the order of 1000 km or larger. This
is known by meteorologists as the synoptic scale. Smaller scale weather systems are classified under the
CHAPTER 3: Processes: How the State is Created, Sustained and Maintained 3-13
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
general term, mesoscale, which can cover systems from a few kilometers in horizontal extent to a few
hundred kilometers. The most dramatic examples of mesoscale systems range from the size of an individual
tornado to the much larger hurricane. Finally, the smallest scales of relevance here are the mesoscale and
microscale associated with local features such as small obstacles (hills, buildings, etc.), which cause
perturbations in the local flow regime. All of these systems are involved in the transport of pollutants.
Physical features associated with these systems can often influence the chemical transformations of species
in the atmosphere.
Boundary Layer Processes
The lowest 1-2 km of the troposphere is known as the planetary boundary layer (PBL), where much
of the air pollutant burden resides. This layer is also called the mixing layer or mixed layer. In this layer both
horizontal and vertical transport of pollutants occurs. This height represents the vertical extent to which
pollutants are actively mixed and has a strong diurnal pattern superimposed upon the changes "due to the
passages of synoptic scale or mesoscale weather systems. Under cloud-free conditions, the ground is heated
by sunlight during the day causing bubbles of warm air (thermals) to rise. The result of the rising buoyant
thermals is a turbulent flow which mixes pollutants upward from surface-based sources or downward from
elevated sources. Another form of turbulent mixing occurs because of wind shear, the rate at which wind
speed and direction change with height. This change of wind with height is caused by frictional drag of the
earth's surface on the wind as well as differential advection. Under conditions of weak solar heating and
strong winds this latter form of turbulence dominates. With weak winds and strong heating the former form
of turbulence dominates. On any given day, a mixture of the two forms of turbulence exists. At night, again
under clear skies, the ground cools and a temperature inversion (temperature increases with height) develops.
Shear driven turbulence is then opposed by negatively buoyant air resulting in a shallower mixing layer with
much weaker mixing. This type of layer is referred to as a surface-based stable layer, hi analyzing these
stable layers, the height of the mixed layer is often less important than the time required for emissions from
a near-ground-level source to mix out of the stable layer.
Most sources of pollutants emit within the mixed layer. Their ground level concentrations are
inversely proportional to this height. For secondary species such as ozone and fine particles, their precursor
chemicals are directly affected by the mixing height. Diurnal variations in thermal heating and turbulence
lead to diurnal variations in the mixing height. It can vary considerably over space and time. "Typical"
values of the mixing height range from 50-100 m at night to 1200-2000 m in the afternoon. This diurnal
cycling effects a higher concentration of air pollutants emitted near the ground at night and a dilution during
the day. However, the rising mixing height also entrains air pollutants from aloft that may have been emitted
there by taller stacks or carried there by long range transport. As photochemical reactions occur, the
concentrations of secondary pollutant species typically increase during the day in the deeper mixed layer.
hi areas where the mixing height does not grow very deep during the day, such as when upper level inversions
restrict the growth, the potential for much deteriorated air quality is high (e.g. this is the classic Los Angeles
smog setup). As the sun sets and the thermal energy source from the ground is cut off, convective motions
decrease and a more stable temperature structure develops near the ground. The air near the ground becomes
"decoupled" from the air several hundred meters above the ground after sunset. Often a jet of higher speed
winds forms above the surface-based stable layer. This jet can carry air pollutants long distances downwind
3-14 CHAPTER 3: Processes: How the State is Created, Sustained and Maintained
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
during the overnight period, allowing entrainment at a different location the next day. In addition, mesoscale
eddies and larger scale offshore flows also can transport pollutants long distances above the shallow
nocturnal surface layer (Stull, 1988 and Garratt, 1994).
Synoptic Scale Processes
Smaller scale atmospheric processes act within the context of the larger, or synoptic scale (1000 km
or greater). When the atmosphere is dynamically active on the synoptic scale, with the frequent passage of
frontal zones and air mass exchanges, dilution is strong and air quality is usually good. When air masses
slow down, or stagnate, as sometimes happens during the warm season, synoptic scale subsidence (sinking
of the air) can occur, increasing the potential for air pollution episodes by decreasing the mixing height. The
"Bermuda high" situation in eastern North America refers to the semi-permanent ridge of high pressure just
east of the North American continent in the warm season. When this ridge grows stronger, as it occasionally
does, it moves onto eastern North America, bringing warmer temperatures, sinking air motions, and
southwesterly winds, all inducements for greater air pollution potential in the region. The southwesterly
winds also align the major urban source areas of the East Coast increasing the transport aspect of air
pollution. On the west coast, when the Pacific subtropical high prevails over the southwestern U.S. (creating
increased low-level subsidence) strong inversions and smog build-up generally occur.
Other synoptic scale processes are also relevant. The barrier between the troposphere and the
stratosphere is known as the tropopause. The strong temperature inversion here usually prevents much air
exchange across the tropopause, except along the path of the jet stream where tropopause folding generally
occurs. In this region, intrusions of ozone-rich stratospheric air into the troposphere take place. Also,
seasonal and latitudinal variations in the stratospheric ozone column can affect the amount of ultraviolet
radiation received near the earth's surface. The less ozone in the column the more radiation available to drive
photochemical reactions. This effect, however, is marginal compared to the impact of clouds on UV
radiation.
Complex Regimes
There are a number of meteorological regimes where the interaction of boundary layer and larger
scale processes are quite complex because of the local environment. One of these regimes is the land/water
interface near coastal regions or the shores of very large lakes. In these areas a mesoscale (100-500 km)
circulation can develop, driven by the relatively large air/water temperature contrasts. In the warm season
the daytime air temperature over the land is greater than the water, allowing for water-to-land winds and
rising air over the land. This circulation may enhance or retard the synoptic scale flow features. Occasionally
it will locally overcome the synoptic influences. The internal mesoscale boundary layer may have restricted
mixing, vertical wind shears, and temperature, wind, and moisture discontinuities with the synoptic scale
regime above it. At night, the effect lessens and the land temperature becomes cooler than the water,
allowing for weak land-to-water transport, and causing local recirculation to occur. Some of the major urban
areas of the United States with air pollution problems are located in areas affected by these type circulations,
including Los Angeles, Houston-Galveston, Chicago, and New York. These mesoscale processes must be
CHAPTER 3: Processes: How the State is Created, Sustained and Maintained 3-15
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
diagnosed carefully when sources areas are embedded in land-water circulations, since the effects on transport
and dilution can be quite complex.
Another complex regime exists in areas of complex, varying terrain: mountains and valleys. Here,
the local thermal and mechanical forcing of winds can change the flow patterns from the overall synoptic
scale. Mountain/valley circulations can preferentially channel air along certain trajectories and through
certain passes. Diurnal variations can evolve from upslope flows during the daytime and downslope at night,
forced from thermal heating and temperature contrasts. The scale of these motions can be quite small or
rather large, depending upon the extent of the terrain variations. Again, sources of pollution caught up in
these circulations can be greatly affected. Urban areas such as Los Angeles, Fresno, Bakersfield, and Denver
or rural areas where power plants or industrial facilities are located can set up situations for this
meteorological and emissions interaction. In some instances, blocking effects can cause eddy formation
during nighttime hours, which strengthens the inversion. Examples of this phenomena are the Bakersfield
and Fresno eddies, and the eddies formed just offshore of Santa Barbara and Santa Monica. Also, downwind
from the Appalachian mountains in the northeastern U.S., mechanically-driven perturbations often develop
within a given synoptic regime, critically influencing the low-level wind fields in determining the magnitude
and location of maximum ozone concentrations.
Clouds
Clouds encompass several meteorological processes that are relevant to air quality. First, clouds are
accompanied by rising air motions that contribute to their existence. These vertical motions add to the
dilution of the Planetary Boundary Layer (PEL), and indeed, when clouds are convectively active and grow
in vertical structure themselves they are effective at "venting" the boundary layer, and can introduce polluted
air to levels above the PEL where they may remain for extended time periods and can be transported by upper
level winds. Clouds affect the solar radiation available to drive meteorological, biological, and chemical
processes relevant to local and regional air quality. Direct impacts of reduced solar radiation are lower
photolysis rates to drive photochemical reactions (UV-spectrum) and lower emissions of isoprene, a
naturally-occurring reactive organic compound and ozone precursor (visible-spectrum). Reduced solar
radiation impinging on the ground will also affect the energy balance, resulting in changes in temperature,
moisture, and winds. In some cases of spotty summertime cumulus clouds, reflections between clouds may
lead to increased UV radiation available for photochemical reactions. The other major effect from clouds
is the moisture source for in-cloud aqueous chemical reactions, such as the transformation of sulfur dioxide
gas to sulfate aerosol particles. Pollutant removal occurs by washout of material within the cloud, and rainout
of material below the cloud during precipitation events. Rain and its associated downward vertical air
motions can be a very effective cleansing phenomenon for the PEL. Unfortunately, this cleansing of the air
also deposits the pollutants to the ground as in acid deposition.
Interactions of Meteorology and Other Processes
Temperature directly affects some source emissions processes, including evaporative organic
emissions from automobiles and other mobile sources, solvent evaporation, and increased natural emissions
of isoprene and monoterpenes, which contribute to the ozone and fine particle problems. Temperature can
3-16 CHAPTER 3: Processes: How the State is Created, Sustained and Maintained
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
indirectly affect emissions also by leading to increased use of automobiles, and residential and commercial
air conditioning on hot days. Temperature and moisture are key factors in many chemical rate constants that
determine the speed of photochemical reactions. Solar radiation, as discussed above, strongly influences
emissions and chemical processes.
Meteorological processes also act as the context for all other atmospheric processes that may lead
to air quality deterioration. That is to say, it is the forces of motion and energy in the atmosphere that cause
molecules to come together to participate in atmospheric chemistry or cause sufficient dilution to minimize
the impact on air quality. Meteorology also determines the course and lifetimes of pollutants emitted from
many major sources into the atmosphere. Understanding the fate of air pollution is not possible without a
thorough understanding of meteorological processes.
Data and Analyses
The U.S. National Weather Service operates a monitoring network of surface and upper air
meteorological observations. Hourly surface measurements are obtained at several hundred sites, and twice
daily upper air "soundings" are taken at a subset of these sites. An increasing array of radar wind profilers
are supplementing the balloon-borne sounding information. Data from weather radars and satellites provide
additional information and aid in interpreting the surface and upper air data. This data network is necessary
for characterizing the synoptic scale meteorological state of the atmosphere at any given time. There are still
gaps in the network (typical distances between upper air sites may be 400-600 km) such that accurate
determination of hourly resolved upper-air flow fields and the spatial variability of the mixed layer is not
completely possible. Special field studies that support increased meteorological measurements aloft help
alleviate these data gaps. Another problem area relevant to air quality is the stagnation case, where wind
speeds become light and variable through a great depth of the troposphere. In this case, the synoptic
observing network may not indicate enough of a coherent flow to make trajectory calculations meaningful.
Here, a stochastic approach to meteorological analysis may be used.
For the complex meteorological regimes discussed above there are rarely sufficient data to fully
characterize the atmospheric state. In these cases meteorologists rely more heavily on complex numerical
models that solve the basic equations of motion (e.g. hydrostatic and non-hydrostatic) to obtain a reasonable
atmospheric state. Models however tend to accumulate small errors that become significant at time "scales
beyond 48-hours into a model simulation. The use of data assimilation (e.g. Four Dimensional Data
Assimilation, FDDA), in which observations are dynamically ingested by the model during simulation, helps
correct these errors.
Concluding Remarks
The main thrust of this discussion was to point out the importance of meteorology in determining
air quality. Meteorology determines how primary precursor emissions are mixed and transported as well as
sets the conditions for chemical reactions to occur. The rate of conversion of precursor species into ozone
and fine particles and the further distribution of these two pollutants is primarily set by meteorological
conditions such as cloud cover and wind flow.
CHAPTER 3: Processes: How the State is Created, Sustained and Maintained 3-17
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
3.D Deposition/Removal
Deposition processes, while reducing the concentration of airborne pollutants, cause the introduction
of these materials into the terrestrial and aquatic ecosystems. Deposition of pollutants (both particles and
gases) during transport can occur through dry or wet processes. Dry processes include retention of airborne
pollutants by the Earth's surface or uptake by vegetation, both of which are enhanced by gravitational settling
and impaction due to turbulent mixing. Wet processes include washout of pollutants due to all forms of
precipitation (e.g., rain, fog, and cloud impaction). All forms of deposition will deplete airborne
concentrations of pollutants and increase the exposure of human and natural resources and materials to these
pollutants.
Models must include a theoretical treatment of these deposition processes if they are to be well-based
assessment tools. Most models approximate deposition by making use of a derived quantity, the deposition
velocity, or, in the case of washout by rain, the washout coefficient. The deposition velocity is defined as the
ratio of the deposition rate to the airborne concentration, and the washout coefficient is a function of rainfall
rate and the terminal velocity of the airborne pollutant. Therefore these deposition parameters must be
independently determined.
Correctly approximating the rate of deposition is a particularly difficult challenge for model
developers. Limited studies are available from which one can parameterize the deposition process. Terrain
effects and site-specific canopy uptake data (dependent on ambient conditions) are frequently not available.
For wet deposition, knowledge of the spatial and temporal occurrence of precipitation events and clouds is
required.
3.E Linearity/Nonlinearity
The relationship between changes in sulfur dioxide emissions and the resulting change in ambient
sulfate concentration and deposition is not necessarily one-to-one. NAPAP sponsored research indicates that
the formation and deposition of secondarily formed sulfate particulates from primary sulfur dioxide
emissions is governed by both linear and nonlinear processes. It appears that the clear air (gas phase)
conversion of sulfur dioxide to sulfate is probably linear. However, aqueous phase conversion associated
with clouds and precipitation scavenging can be nonlinear. The nonlinear cloud-based conversion Is
particularly important in the winter when concentrations of hydrogen peroxide (necessary for the formation
of sulfate) are low, thereby inhibiting the sulfate formation. Similarly, sulfur dioxide saturation of cloud
droplets in other seasons can diminish wet deposition of sulfur. NAPAP was less certain about the possible
reduction of dry deposition if the surface had become acidic due to previous deposition.
The above processes affect the concentration of secondarily formed sulfate particles. There are also
effects, related to the relative humidity and the composition (e.g., ammonium sulfate or sulfuric acid) and
characteristics (externally or internally mixed) of the chemical species that make up the sulfate particles.
These have significant implications for the interpretation and predictions of concomitant effects on visibility.
Hygroscopic particles (e.g., sulfates) will grow in the presence of water vapor to particle sizes that cause
greater extinction (visibility impairment) than their "dry" counterparts. The relationship between relative
3-18 CHAPTER 3: Processes: How the State is Created, Sustained and Maintained
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
humidity and extinction is highly nonlinear at relative humidities greater than about 70%. Internally and
externally mixed particles have different extinction efficiencies and will react differently to the removal of
one of its constituents.
The net result of these processes and characteristics is that one can not assume that a given reduction
in pollutant emissions will necessarily result in a like change (linearly related) in concentration, deposition,
or visibility impairment. More important, in the near term, emission reductions and/or ambient air quality
improvement relationships are poorly understood for aggregate indicators such as integrated ozone, PM-fine
and regional haze.
3.F Commonalities, Disconnects and Integrable Elements
EPA is currently pursuing a joint implementation strategy for a new ozone and fine particulate matter
NAAQS and a regional haze rule. While there are connections among source categories and precursor
emissions there are also significant differences in the processes responsible for the formation of fine
particulates and ozone and subsequent differences in ozone and fine particulate matter episodes. This section
discusses the commonalities, differences, and the elements needed for a successful integrated implementation
program.
Commonalities Among Ozone, Fine Particles, and Haze
Common "Direct"Precursors
NOX and certain VOC species can be oxidized and lead to the formation of nitric acid (a precursor
for particulate nitrate) and organic aerosols. However, the effect of VOC emissions on resultant ambient PM
aerosol measurement is highly dependent upon the sampling device, and is in general not well understood.
Interdependence on Radical Balances
The formation of ozone and secondarily formed particles is preceded by oxidation of their respective
precursors. The oxidation is carried out through hydroxyl and peroxy radicals. Thus any radical initiation,
propagation or termination process potentially affects formation of ozone and fine particles.
Common "Indirect" Precursors
Because of the radical interdependence, potentially all precursors are common precursors. In other
words, SO2 could be an ozone precursor; and the majority of VOC species, whose daughter products do not
eventually form organic aerosols, could be fine particle precursors. This "universal" pool of precursors does
not infer "common" response behavior. A reduction of a precursor that reduces a component of fine
particulates, potentially can result in increased ozone or an increase in another fine particle component.
Multiple nonlinearities and positive and negative feedbacks exist (analogous to NOx-ozone-VOC control).
Thus, integrated implementation is far from a straightforward exercise, and the tools available for such
assessment are yet to be evaluated.
CHAPTER 3: Processes: How the State is Created, Sustained and Maintained 3-19
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Common Source Categories
Based on a commonality of precursors, a particular source emitting one precursor can affect ozone
and fine particles, and a single source emitting multiple precursors can affect ozone and fine particles. The
earlier discussion of meteorological transport and deposition issues opens the possibility that" the same
precursor from the same source could be ambient air quality neutral, positive and/or negative in effect for
ozone, PM-fine and/or regional haze. Figure 3.3 shows the complex linkages between oxidant chemistry
and fine particle formation.
Co-episodicity
Assuming the above, co-episodes of high ozone and high fine particles can occur. Co-episodic
events are a subset shared processes. That is, some of the processes which explain ozone or fine particle
formation always operate, whereas concurrent "episodes" may occur rarely, sometimes or often. Most likely,
far more single pollutant episodes exist than co-episodic (ozone and secondary PM) events. Examples of
single-pollutant episodes include all of the non-summertime secondary PM events, often dominated by
ammonium nitrate formation, which is favored by cool temperatures (< 60 °F) over gaseous ammonia and
nitric acid. Of course, the "primary" particle episodes rarely (if ever) coincide with an ozone episode.
Research exploring the frequency and characterization of co-episodic and uni-episodic events, as defined by
ambient air quality measurements, would yield further insight into the underlying causes of events and
provide direction for integrated implementation opportunities.
Differences Among Ozone, Fine Particles, and Haze
Process Level Differences
Those processes generating the primary component of fine particles can be thought of as being
emissions based, and often are separated from the secondary chemistry responsible for ozone formation.
Emission based processes include a host of mechanically driven agricultural dust, fugitive and roadway dust
events, as well the generation of carbon soot from wood and fuel combustion, which form primary
condensable organic aerosols.
Episodic Differences
Most of the emissions-based process differences lend themselves almost directly to episodic
differences between ozone and particles. For example, the summertime wind-generated dust events in the
southwest do not correlate well with high ozone events. This lack of coincidence is expected given that the
high-wind meteorological conditions required to generate significant dust loadings are not favorable for
developing high ozone.
Many secondary-driven high particle events do not coincide with ozone. In many episodes where
ammonium nitrate is a significant particle component, the meteorology will not favor ozone formation since
ammonium nitrate is favored at relatively low temperatures and volatilizes to nitric acid and ammonia gases
3-20 CHAPTER 3: Processes: How the State is Created, Sustained and Maintained
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Nighttime N2OS
ihemJ5tiy_
Radical Pool
HO2-; RO2-
Clouds/Aqueous
Sink
Figure 3.3: Linkages between oxidant chemistry and fine particulate (FP) formation (Scheffe, 1997).
CHAPTER 3: Processes: How the State is Created, Sustained and Maintained 3-21
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
at higher temperatures. Thus, wintertime nitrate events do not coincide with ozone. Despite these episodic
differences, one could argue that nitrate and ozone share common components - the basic radical chemistry
cycles and NOX precursors (e.g., Figures 3.1-3.3). On balance, the frequency of co-episodic events is
probably far less than single event episodes.
Integration
Programmatic
What does integration mean from an implementation perspective? The connections among source
categories and precursor emissions (the independent variables in a control program) can be illustrated using
an air quality modeling approach. The emission bases underlying most current ozone modeling efforts
include nearly all of the source categories for aerosol formation, with certain notable exceptions in
agricultural and natural emissions areas. The most typical example is utility boilers which emit both SOX and
NOX. From a precursor perspective, NOX and certain VOC categories are obvious potential precursors for
ozone and aerosols. But any oxidant precursor (including all VOC and CO) needs to be considered as
potentially affecting aerosols, given the interplay between pollutants due to free radical chemistry cycles.
In formulating control strategies, one might start with the ozone management strategies developed as part
of ongoing regional analyses (e.g. OTAG), and attempt to quantify the future impact on secondary aerosols
due to combined CAA provisions for acid precipitation (Title 4) and ozone (Titles 1 and 2). The result of
this exercise would produce the residual aerosol (and haze) related air quality benefits from an ozone
precursor control oriented approach. This current exercise is planned with the RADM/RPM modeling suite
of air quality models. Although this example does not represent "full" integration given the unidirectional
information flow (ozone to particles), it does acknowledge similarities among programs, and avoids mistakes
and inefficiencies incurred from independent analysis. Ideally, fully integrated modeling systems (e.g.
MODELS-3) with sophisticated feedback treatment would be used to "integrate" multi-pollutant control
design efforts. In the meantime, an understanding of pollutant linkages should reinforce the development of
data systems (emissions, meteorological and air quality databases) and model interfaces with similar
attributes to both improve and economize the analysis efforts.
3-22 CHAPTER 3: Processes: How the State is Created, Sustained and Maintained
-------
CHAPTER 4
Current Tools to Address and
Implement Current State of
Knowledge
Both regulatory air quality management and research analysis activities are based on measured data
and analysis tools that are used to assess both the current state of air quality and the likely future response
to proposed or promulgated air pollution control rules. In general, the databases involve measurements of
ambient air quality, meteorological parameters, and to a lesser extent source emission rates. The tools used
to evaluate progress toward air quality improvement or future predicted conditions include techniques for
estimating emissions from sources that are not measured directly, and air quality models that seek to predict
future changes in air quality as a result of proposed changes in emissions characteristics due to control
requirements and regulations.
The standard methods used to develop the primary databases and the tools used to assess future
conditions are discussed in this chapter. The first section describes monitoring techniques for the primary
pollutants and precursors of concern to ozone, PM and haze programs. The discussion on monitoring
techniques addresses issues related to measurement of primary and surrogate species that represent a broad
class of related air pollutant species. The second section discusses current procedures for estimating both
baseline and future year emissions conditions. The features of air quality models are addressed in the third
section. Finally, some important issues that affect all air quality planning activities are discussed to lay a
foundation for a policy that seeks to address air quality management in an integrated approach that
recognizes the interrelation among ozone, PM and haze control programs.
4.A Monitoring Technologies, Meteorology, and Network Design
Monitoring Techniques
The discussion below describes methods and techniques currently in use for the measurement of
ambient concentrations of important air quality parameters and meteorological data to support analyses of
the formation and transport of those pollutants. This section includes discussions of reference methods for
the criteria pollutants, and accepted methods for other pollutant species of interest. There is also a brief
discussion of techniques used to monitor routine meteorological parameters.
Ozone
Routine Techniques Several techniques have been used to measure ozone and are described in
great detail elsewhere (NAS, 1991). Chemiluminescence is the EPA Reference Method and has been used
CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge 4-1
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
extensively for routine surface level ozone regulatory measurements. Chemiluminescence is produced from
ozone reacting with ethylene gas. In the presence of excess ethylene, the intensity of light produced is
proportional to the concentration of ozone. More recently, the Chemiluminescence technique has been
replaced by an equivalent method based on ultraviolet (UV) photometry, the principle of which is based upon
the absorption of UV light by ozone molecules. Both methods produce accurate measurements of ambient
ozone levels.
Nitrogen Oxides/Nitric Oxide/Nitrogen Dioxide (NOJNO/NOJ
Routine Techniques. Chemiluminescence is used for most NOX and NO measurements and is the
EPA Reference Method for measuring N02. The ambient monitors most commonly employed are configured
to provide all three measurements, hi the NO mode, the air sample reacts directly with ozone to produce a
Chemiluminescence intensity directly proportional to the concentration of NO. In the NOX mode, the air
sample passes through a converter which reduces N02 to NO. The total NOX (converted N02 plus NO in the
original sample) then reacts with ozone to produce a proportional Chemiluminescence signal. The two signals
(NO and NO2 plus NO) are compared electronically and the resultant signal is equivalent to the concentration
of N02 in the ambient sample. Chemiluminescent NO measurements are believed to be very reliable.
However, data for NOX and N02 may have a higher variability due to the need to rely on converter efficiency
and the stability of the monitor electronics to accurately evaluate the NOX and NO 2 concentrations.
Additionally, it has been shown that N0y species (e.g. peroxyacetylnitrate (PAN) and nitric acid) are also
converted, thus creating a positive bias in the reported NO2 measurements.
Non-routine techniques. Differential Optical Absorption Spectrometry (DOAS) is the only
equivalent certified technique for N02. DOAS produces high quality N02 (and several other atmospheric
gases) measurements, but at considerable initial capital costs. DOAS measures the absorption at a
characteristic wavelength (for each gas of interest) compared to absorption at less absorbing nearby
wavelengths, hi addition, optical techniques like DOAS sample over specified path lengths and, depending
on the instrument alignment, allows for varying degrees of greater volumetric and vertical profile
representation. Other techniques include luminol Chemiluminescence, photolytic (N02 - NO) conversion,
laser induced fluorescence (LIF), and tunable-diode laser absorption spectrometry (TOLAS). Photolytic
conversion has the potential for providing a low-cost, routine instrument. The air stream is irradiated with
a light source (~ 4000 nanometers (nm)) photolyzing N02 into NO and an oxygen atom; the NO is then
measured through Chemiluminescence to determine NO2 concentration. Difficulties associated with
inefficient absorption efficiencies and production of a low-cost, intense light source accompany diis
technology.
Total NOy
Total N0y is not required in routine measurement programs, although N0y essentially has become
a routine measurement in nearly all modern field programs like the Southern Oxidant Study (SOS) and the
North American Research Strategy for Tropospheric Ozone in the Northeast (NARSTO-NE). The basis for
total N0y is the chemiluminescence/molybdenuni (or gold converter) method modified to reduce the loss of
nitric acid, PAN and organic-nitrates by placing the converter as close as possible to the intake. The
4-2 CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
technology for total N0y measurements is available and relatively inexpensive and is believed to be reliable.
Remaining concerns for routine total N0y operation are calibration/QA techniques, available expertise and
associated operator difficulties.
Speciated NOy (nitric acid and PAN)
In addition to NO and N02, the remaining major N0y components include nitric acid (HN03), and
PAN. Other N0y components include various organic nitrates and aerosol nitrate. Nitric acid is measured
in the Clean Air Status and Trends Network (CASTNET) national monitoring network at roughly 35 sites,
and in intensive field programs like SOS, South Coast Air Quality Study (SCAQS), and San Joaquin Valley
Air Quality Study (SJVAQS)/Atmospheric Utility Signatures, Predictions and Experiments (AUSPEX)
Regional Modeling Adaptation Project (SARMAP). Techniques include filter pack, annular denuder,
denuder difference, TOLAS and several others. HN03 is highly soluble, reactive with various surfaces and
subject to temperature sensitive equilibrium with ammonium nitrate. These are all factors which result in
difficult and less than reliable HNO3 measurements. PAN can be measured reliably with electron-capture
gas chromatography (ECGC). However, PAN is far from a routine measurement requiring substantial
attention to calibration procedures and instrument operation.
Carbon Monoxide (CO)
Routine techniques. The reference method for the measurement of CO in ambient air is based on the
principle of absorption of infrared (IR) radiation by a photometer designed to respond only to those
wavelengths at which CO strongly absorbs. IR energy is passed through a cell containing the gas sample to
be analyzed and the quantitative absorption of energy by CO on the sample cell is measured by a detector
whose output signal is directly proportional to the concentration of CO in the ambient sample. This
technique has been demonstrated to produce highly accurate analytical results.
Nonmethane Hydrocarbons (NMHC)
Until recently, speciated NMHC were not included in routine air quality measurement programs and
sampling and analysis for speciated hydrocarbons was limited to various special field studies. The
Photochemical Assessment Monitoring Site (PAMS) program (which is discussed in the following section
on monitoring networks) requires near-continuous speciated NMHC measurements of roughly 56 target
species. Gas chromatography with flame ionization detector (GC-FID) is the standard analysis system in
use. Samples are collected either with canisters or through continuous GC operation. The more prevalent
continuous GCs are subject to water management/sample preparation difficulties which can result in losses
of the more polar-behaving compounds, including biogenic alkenes and alcohols (technically not a NMHC).
Operation of continuous GCs is not trivial, requiring substantial operator attention, calibration and QA
procedures.
CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge 4-3
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Carbonyls
Carbonyls are more difficult to measure than NMHC. In general, GC-FID is not very sensitive to
carbonyls and other methods must be applied. Common methods use cartridges filled with various solid
sorbents coated with 2,4-dinitro phenylhydrazine (DNPH) to collect a sample, and analyses are performed
with high pressure liquid chromatography (HPLC). The most common sorbents are silica gel (e.g. the PAMS
standard) and a modified silica gel, CIS, which is used more frequently in research studies. Both sorbents
are subject to various interferences, although very recent studies suggest that CIS may be a more reliable
sorbent. Carbonyl measurements remain difficult, are subject of substantial debate in the research and
monitoring communities, and pose significant challenges to routine networks like PAMS. The PAMS
program requires reporting of three carbonyls: acetone, acetaldehyde and formaldehyde. Several additional
higher molecular weight carbonyls are also captured with cartridge techniques and their cumulative mass,
as a fraction of total VOC, can be significant.
Peroxides and Radicals
Measurements of intermediate, fast-reacting radicals (OH, HO2, R02) and more stable peroxides
currently are restricted to research grade studies like SOS. However, these measurements provide valuable
information as indicators of NOX or VOC-limiting environments, and support 'in-situ' evaluation of chemical
mechanism behavior.
Particles
Total Suspend Particles (TSP)
Total Suspended Particles (TSP) are measured using a high-volume sampler as described in 40 CFR
Part 50, Appendix B. This sampler has a cut point of aerodynamic diameters that varies between 25 and 40
urn depending on the wind speed and direction. The method provides a measurement of the mass
concentration of total suspended paniculate matter in the ambient air sample. The measurement process is
non-destructive, and the size of the sample collected is usually adequate for subsequent chemical analyses.
Particles <. 10 fim (PM-W)
PM-10 consists of particles measured by a sampler that contains a size fractionator (classifier)
designed to have an effective cut point of 10 um aerodynamic diameter. This measurement includes the fine
mode and part of the general coarse mode and is an indicator for thoracic particles (i.e. particles that penetrate
to the trachea-bronchial and the gas-exchange regions of the lung).
Particles
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
efficiency curve and the size of the coarse mode particles present in the ambient sample. The Interagency
Monitoring of Protected Visual Environments (IMPROVE) sampling system consists of: (1) a size selective
inlet; (2) a cyclone to provide a particle size cutoff based on the flow rate; (3) collection substrates; (4) a
critical orifice that provides the proper flow rate for the desired particle size cutoff; and, (5) a vacuum pump
that produces the flow. The sampling system consists of four independent sampling modules, three modules
provide for the collection of particles less than 2.5 ^un in diameter, while the fourth system is a PM-10
sampler with a wind insensitive size selective inlet that collects particles less than 10 um. Filter media used
in the systems consist of teflon, nylon (denuded), and quartz, hi the IMPROVE analyses, coarse mass is
estimated gravimetrically by subtracting the PM-2.5 mass concentration from the total aerosol mass (PM-10)
concentration.
Visibility
Monitoring of protected visibility areas is conducted in the IMPROVE network on two
complementary fronts: (1) optical monitoring of visibility; and, (2) monitoring the concentration and
composition of aerosols in these areas. For the optical monitoring, two measurements are possible, extinction
(b^,) measured by transmissometers and scattering (bKat) measured by nephelometers. Transmissometers
are instruments calibrated to measure the atmospheric irradiance, at a wavelength of 550 nm, of a light source
after the light has traveled over a finite atmospheric path length. Integrating nephelometers measure the
scattering of light over a defined band of visible wavelengths from an enclosed volume of air. For the
IMPROVE network, the particulate monitors provide measurements of PM-10 and PM-2.5 mass. Chemical
and elemental analysis of the PM-2.5 fraction is used to identify the fine aerosol species.
Meteorology
The following is a brief discussion of the types of surface and upper air meteorological data currently
collected across the United States and the current methods used to collect this information (Crescenti, 1996).
A substantial network of surface meteorological instruments provides hourly wind velocity, temperature,
and humidity measurements. Most of these surface meteorological sites are associated with major airports.
However, some of these observations are made coincident with air quality monitoring stations, with the
PAMS locations being a typical example. This section on instrumentation is followed by more detailed
assessment of how these data are used in air quality management/control studies.
Surface Observations
Guidance for surface meteorological measurements is provided in several documents, which include
the On-Site Meteorological Instrumentation Requirements to Characterize Diffusion from Point Sources
(U. S. EPA, 1981); Guide to Meteorological Instruments and Methods of Observation (WMO, 1983);
Instructor's Handbook on Meteorological Instrumentation (NCAR, 1985); Ambient Monitoring
Guidelines for Prevention of Significant Deterioration (U. S. EPA, 1987a); On-Site Meteorological
Program Guidance for Regulatory Modeling Applications (U. S. EPA, 1987b); and Quality Assurance
Handbookfor Air Pollution Measurement Systems, Volume IV: Meteorological Measurements (U. S. EPA,
CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge 4-5
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
1989). Thus, for a detailed description of the various meteorological monitoring techniques, the reader is
referred to the above documents.
Surface-based meteorological instrumentation should not be mounted on or near solid structures such
as buildings, stacks, water storage tanks, grain elevators, and cooling towers since they may create significant
wind flow distortions. Instead, these instruments should be mounted on an open lattice 10-m tower, since
this structure creates the least amount of wind flow distortion. There are several types of open lattice towers
currently employed for meteorological observations: (1) fixed; (2) tilt-over; and, (3) telescopic. Regardless
of which type of tower is used, the structure should be sufficiently rigid and properly guyed to ensure that
the instruments maintain a fixed orientation at all times.
The objective of instrument siting (horizontal and vertical probe placement) and exposure (spacing
from obstructions) is to place the sensor in a location where it can make measurements that are representative
of the general state of the atmosphere in the region of interest. The selection of a site for a meteorological
tower should be made with an understanding of the regional geography. Ideally, a meteorological tower
should be located in an open, level area away from the influence of obstructions such as buildings or trees.
The area surrounding the site should have uniform surface characteristics and the specific site characteristics
should be well documented. This is especially important where complex terrain (i.e. terrain with significant
topographic features) may simultaneously introduce different meteorological regimes. Secondary
considerations such as accessibility and security must also be taken into account, but should not be allowed
to compromise data quality.
Although it may be desirable to collocate the surface meteorological measurements with the ambient
air quality measurements (i.e. PAMS data), this may not be possible at all monitoring sites without violating
one or more of the above criteria. Surface meteorological measurements in urban areas, where compliance
with the above guidance may be precluded by the close proximity of buildings and other structures, present
special difficulties, hi such cases, site selection requires an assessment of the likelihood that the data
collected at a given location will be valid for the intended application/analyses, hi all cases, specific site
characteristics should be well documented. This is especially important in areas where surface characteristics
and/or terrain are not uniform and whenever standard exposure and siting criteria can not be met.
Wind Speed and Wind Direction ~~
Horizontal wind speed (m/s) and wind direction (degrees clockwise from geographical north) are
essential to the evaluation of atmospheric transport and dispersion processes. Measurements of wind speed
and direction are also important in assessing atmospheric stability and turbulence. Wind speed is typically
measured with a cup or propeller anemometer; wind direction is measured with a vane. The standard height
for surface layer wind measurements is 10 m above ground level (WMO, 1983). It is important that the
tower be located in an area of level and open terrain. The wind sensor should be sited such that the horizontal
distance to an obstruction is at least ten times the height of the obstruction. An obstruction may be man-
made (e.g., building) or natural (e.g., trees).
4-6 CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Air Temperature
There are several types of temperature sensors currently available: (1) wire bobbins; (2)
thermocouples; and, (3) thermistors. Platinum resistance temperature detectors (RTD) provide accurate
measurements with a stable calibration over a wide temperature range and are among the more popular
sensors used in ambient monitoring. The temperature sensor should be mounted on the tower 2 m above the
ground and away from the tower a distance of at least one tower width from the closest point on the tower.
This height is consistent with World Meteorological Organization (WMO, 1983) and EPA standard
monitoring procedures. The measurement should be made over a plot of open, level ground at least 9 m in
diameter. The ground surface should be covered with non-irrigated short grass or, in areas which lack a
vegetation cover, natural earth. Concrete, asphalt, and oil-soaked surfaces should be avoided. As such, the
sensor should be at least 30 m away from any paved area. Other areas to avoid include large industrial heat
sources, roof tops, steep slopes, hollows, high vegetation, swamps, snow drifts, standing water, and air
exhausts (e.g., tunnels and subway entrances). The sensor should be located a distance from any obstruction
of at least four times the obstruction height.
Relative Humidity
Measurements of atmospheric humidity are essential to understanding chemical reactions in the
atmosphere. The relative humidity is defined (List, 1951) as the ratio of the ambient mixing ratio (w) to the
saturation mixing ratio (w^) at a given air temperature and barometric pressure, i.e.,
RH = 100
Other measures of atmospheric humidity include vapor pressure (hPa), dew point temperature (°C), specific
humidity (g/kg), and absolute humidity (g/m3). All variables except for the relative humidity provide a
complete specification of the amount of water vapor in the atmosphere. However, any of these variables can
easily be derived from the relative humidity given the ambient air temperature and barometric pressure.
There are various techniques for measuring atmospheric humidity. However, the Emergence of capacitive
thin-film technology is now producing sensors which are reasonably accurate, reliable, compact, and
inexpensive. Crescenti and Payne (1991) compared thin-film relative humidity sensors from two different
manufacturers and found that they performed quite well. These sensors are becoming more common as they
are easy to install and operate.
Barometric Pressure
Barometric pressure (hPa) is useful for examining trends in the weather on the order of several days
or more. It is also essential for the calculation of thermodynamic quantities such as air density, absolute
humidity, and potential temperature. There are numerous commercially available pressure transducers which
range widely both in price and performance. Most of these sensors are capable of measuring barometric
pressure with an overall accuracy of ±1.0 hPa over a range of 800 to 1100 hPa, a resolution of 0.1 hPa.
CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge 4-7
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Precipitation
The total amount of precipitation which reaches the ground is expressed as the depth to which it
would cover a plane horizontal to the earth's surface in a given period of time. There are several rain gauge
variations: (1) tipping-bucket; (2) weighing-bucket; (3) capacitive-siphon; and, (4) optical. The most
common are the tipping and weighing-bucket which are cylindrical in shape with a 20 cm diameter collection
orifice. Measurement accuracy for all types of rain gauges is influenced more by exposure than by variations
in sensor design. High winds generally cause an underestimation of precipitation. Therefore, efforts should
be taken to minimize the wind speed at the orifice, especially in open areas.
Solar Radiation
Solar (sometimes called shortwave) radiation is a measure of the electromagnetic radiation of the
sun and is represented as an energy flux (W/m2). Solar radiation measurements are used in heat flux
calculations, for estimating atmospheric stability, and in modeling photochemical reactions. The solar
spectrum is comprised of ultraviolet radiation (0.10 to 0.40 ^im), visible light (0.40 to 0.73 um), and near-
infrared (0.73 to 4.0 um) radiation. About 97% of the solar radiation incident at the top of the earth's
atmosphere lies between 0.29 and 3.0 um (WMO, 1983). A portion of this energy penetrates through the
atmosphere and is received at the earth's surface. The rest is scattered and/or absorbed by gas molecules,
aerosols, various particulates, cloud droplets, and ice crystals.
A pyranometer is an instrument used for measuring energy fluxes in the solar spectrum. The sensor
measures global solar (direct and diffuse) radiation when installed facing upwards in a horizontal plane
tangent to the earth's surface. The sensing element of the pyranometer is usually a thermocouple which is
protected by a clear glass dome to prevent entry of wavelengths outside the solar spectrum (i.e., longwave
radiation).
Longwave Radiation
Longwave radiation is a measure of terrestrial and atmospheric radiation and is also represented as
an energy flux (W/m2). Longwave radiation measurements are used in the calculation of the heat budget.
The longwave spectrum falls in the range of 3.0 to 100 um. A pyrgeometer is an instrument used for
measuring longwave energy fluxes. The sensor measures atmospheric radiation when installed facing
upwards in a horizontal plane tangent to the earth's surface. Similarly, the sensor measures terrestrial
radiation when facing downwards. Pyrgeometers should be sited with the same criteria used for solar
radiation measurements.
Ultraviolet Radiation (UV)
Ultraviolet radiation may be divided into three sub-ranges: UV-A (0.315 to 0.400 um), UV-B (0.280
to 0.315 um), and UV-C (0.100 to 0.280 um). Due to stratospheric absorption by ozone, UV radiation that
reaches the surface is usually limited to wavelengths longer than 0.28 \im (UV-A and UV-B ranges). The
most important photochemically active chemical species at these wavelengths are ozone, nitrogen dioxide,
4-8 CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
and formaldehyde. All three of these chemical species are important in the chemistry of ozone formation.
Ultraviolet pyranometers which have a spectral response spanning both the UV-A and UV-B (0.280 to 0.400
um) ranges are recommended for most applications. The same siting criteria used for solar and longwave
radiation measurements apply.
Upper Air Observations
Historically, there has been a shortage of vertical wind, temperature, and humidity profile data, which
are critical inputs for air quality models. Until the last 3 or 4 years, the twice daily National Weather Service
(NWS)/Federal Aviation Administration (FAA) rawindsondes, limited to major airport locations, and military
rawindsondes were the primary sources of such data. However, recent advances in remote sensing technology
combined with several field study efforts and the PAMS data have resulted in near routine operation of sodar
and radar/RASS systems to profile winds and temperatures on a near continuous basis. Additionally, the
National Oceanic and Atmospheric Administration (NOAA)/NWS Next Generation Radar (NEXRAD)
system is operational, providing an enormous national network of vertical profile data. Overall, the
implementation of upper meteorological monitoring devices over the last 2 or 3 years is probably the most
substantial relative improvement across all disciplines supporting air quality management.
Vertical profiles of wind speed and wind direction, temperature and humidity are needed for use in
three-dimensional atmospheric transport and dispersion modeling. Profiles of air temperature are highly
desired since this is a principle indicator of atmospheric stability. Other variables which can be measured
include vertical wind speed, relative humidity, and barometric pressure. EPA currently does not have any
specific guidance on measurement levels and accuracies for any upper air data. However, Tables 4.1 and 4.2
are World Meteorological Organization (WMO) guidelines which can be used as a starting point.
Profiles in the first several hundred meters of the atmosphere are very important. It is highly desired
to obtain profiles of at least 1000 m or to the top of the convective mixed layer (which can easily exceed
2000 m on summer afternoons). However, not all measurement systems are capable of an extended height
range. Wind, temperature, and humidity profile data may be obtained by nearby National Weather Service
(NWS) and military radiosonde sites.
Determination of the depth of the atmospheric boundary layer or mixed layer (i.e., mixing height)
is also an important upper air measurement. Reliable estimates of the mixing height are essential to
dispersion modeling because this is the depth through which vertical mixing of pollutants normally occurs.
The degree of dispersion within the mixed layer is primarily a function of atmospheric turbulence (i.e., wind
flow, surface heating). The mixing height can be determined based on air temperature, turbulence, and/or
aerosol concentration data.
There are a variety of platforms for measuring upper air meteorological data. These include aircraft,
tall towers, balloon systems, and ground-based remote sensors. As with any measurement system, each has
its advantages and disadvantages. The variables that can be measured with each upper-air system are
summarized in Table 4.3. Note that with the exception of aircraft and tower, no one upper air measurement
CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge 4-9
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Table 4.1: WMO observation levels for lower tropospheric soundings for operational and research
purposes (WMO, 1983).
Variable
Wind Speed and
Wind Direction
Air Temperature and
Relative Humidity
Interval (m)
50
100
200
300
20
50
100
Range (m)
0 to 300
400 to 600
800 to 1200
1500 to 3000
0 to 300
350 to 1000
1100 to 3000
4-10 CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Table 4.2: WMO observation accuracies for lower tropospheric soundings for operational and research
purposes (WMO, 1983).
Variable
Wind Speed
Wind Direction
Accuracy
±0.5 m/s
±10%
±10°
±5°
WS <; 5 m/s
WS > 5 m/s
WS s 5 m/s
WS > 5 m/s
Air Temperature
Relative Humidity
±0.2 °C
±5%
±1%
RH * 95%
RH>95%
CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge 4-11
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Table 4.3: Meteorological variables that can be measured with various upper air monitoring systems.
Variables include horizontal wind speed and direction (WS/WD), vertical wind speed (W),
air temperature (T), relative humidity (HUM), and barometric pressure (BP).
System WS/WD W T HUM BP
Aircraft / / / / /
Tower / / / / /
Radiosonde / / / /
Tethersonde / / / /
Radar / /
Sodar / /
RASS /
4-12 CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
system is capable of acquiring all of the variables listed in the table. Typical vertical ranges and resolutions
for these systems are presented in Table 4.4.
Aircraft
Aircraft (both airplanes and helicopters) are the ultimate mobile observation station. They are
capable of traversing large horizontal and vertical distances in a relatively short period of time. This platform
can be equipped with meteorological instrumentation and an assortment of chemical sensors. Traditionally,
aircraft are used for episodic field studies which often require extensive data sets for model evaluation.
Lenschow (1986) provides an excellent overview of aircraft measurements in boundary layer applications.
While an aircraft can provide detailed atmospheric observations over large areas, the total sampling time per
flight (typically 6 to 8 hours) is relatively short because of fuel considerations. Aircraft may also be subject
to Federal Aviation Administration (FAA) restrictions on flight paths over urban areas. In addition, the
operating cost for this type of platform is extremely expensive.
Tall Towers
In some instances it may be possible to use existing towers which may be located in monitoring areas
to acquire vertical profiles of atmospheric boundary layer data. Radio and television transmission towers,
which may be as tall as 600 m, can be equipped with 'in situ' meteorological sensors at many levels. An
advantage to using a tower is the ability to run an unattended data acquisition system. Also, data can be
collected under all weather conditions. However, the main disadvantage of using a tower is .the inability to
determine the mixed layer height during most of the day. When moderate to strong convective conditions
exist, the mixed layer height easily exceeds that of the tallest towers. Another disadvantage is the potentially
high cost of maintenance, especially during instances when the instrumentation needs to be accessed for
adjustments or repairs.
Balloon Systems
Balloon-based systems offer a relatively inexpensive means for upper-air meteorology
measurements. There are two types of balloon systems: (1) radiosonde (sometimes called rawinsonde); and,
(2) tethersonde. The radiosonde is reliable, robust, light weight, and relatively small. The radiosonde is
expendable, and can be mass produced at low cost. The radiosonde is comprised of sensors, a tracking
device, and a radio transmitter. This sensor package is suspended from a hydrogen or helium filled balloon
and is released at the surface. Air temperature is measured with a bimetallic strip, ceramic semi-conductor,
or a wire resistor. The relative humidity is measured with a carbon hygristor or a thin-film capacitive chip.
The barometric pressure is obtained with the an aneroid capsule. Ground-based radar is used to determine
horizontal wind speed and direction. The radiosonde is capable of easily traversing the depth of the
troposphere and reaching well into the stratosphere. A tethersonde system is comprised of a tethered balloon
with several sonde packages attached to the line. Variables measured include horizontal wind speed and
direction, air temperature, relative humidity, and barometric pressure. These data are telemetered to the
ground by radio or by conductors incorporated within the tethering cable. The tethersonde is capable of
reaching altitudes up to 1000 m. However, this system can only operate in light to moderate wind conditions
CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge 4-13
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Table 4.4: Typical vertical ranges and resolutions for upper air monitoring systems.
System Range (m) Resolution (m)
Aircraft 100 to 10,000 1
Tower 10 to 600 1
Radiosonde 10 to 10,000 5
Tethersonde 10 to 1,000 5
Radar 100 to 3,000 60 to 100
Sodar 50 to 1,000 25 to 50
RASS 100 to 1,500 60 to 100
4-14 CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
(5 m/s at the surface, 15 m/s aloft). A tethered balloon may also pose as an aviation hazard and is subject
to FAA regulations. A permit must be obtained for permission to operate such a system. Low cost is the
main advantage for these systems, as well as ease of transport and relatively low maintenance. The main
disadvantage for balloon systems is that they can be very labor intensive, especially if data are needed on an
frequent basis. In addition, vertical wind speed can not measured by either balloon system.
Ground-Based Remote Sensors
Ground-based remote sensors have become effective tools for acquiring upper-air information and
have played an increasingly important role in atmospheric boundary layer studies. However, there is a
distinct void in available guidance needed to help potential users in the regulatory community. Because of
their unique nature and constant evolution, EPA guidance for remote sensors is more generic than that which
already exists for many of the well established 'in situ' meteorological sensors. Efforts are underway to
provide more clearly defined guidance and standard operating procedures which will appear in the next
edition of the Quality Assurance Handbook for Air Pollution Measurement Systems, Volume IV:
Meteorological Measurements (EPA, 1989). Unlike 'in situ' sensors which measure by direct contact,
remote sensors do not disturb the atmosphere. Another fundamental difference is that remote sensors
measure a volume of air rather than a fixed point in space. The thickness of the volume is a function of the
pulse length and frequency used. The width of the volume is a function of beam spread and altitude. Siting
of these profilers is sometimes a difficult task. Artificial and natural objects located near the sensors can
potentially interfere with the transmission and return signals, thereby contaminating the wind velocity data.
There are two basic types of remote sensing systems used to acquire three-component wind velocity
profiles: (1) radar (radio detection and ranging); and, (2) sodar (sound detection and ranging). Radars (also
called wind profilers) transmit an electromagnetic signal (-915 MHz) into the atmosphere in a predetermined
beam width which is controlled by the configuration of the transmitting antenna. Sodars (also called acoustic
sounders) transmit an acoustic signal (~ 2 to 5 KHz) into the atmosphere in a predetermined beam width
which is also controlled by the transmitting antenna. The radar has a range of approximately 100 to 3000
m with a resolution of 60 to 100 m The sodar has a range of about 50 to 1000m with a resolution of about
25 to 50 m.
A radio acoustic sounding system (RASS) utilizes a combination of electromagnetic and acoustic
pulses to derive a virtual air temperature profile. A RASS usually consists of several acoustic antennas
placed around a radar system. The antennas transmit a sweep of acoustic frequencies vertically into the
atmosphere. As the sound pulses rise, the speed of the acoustic wave varies according to the virtual air
temperature. Concurrently, a radar beam is emitted vertically into the atmosphere. The radar beam will most
strongly reflect off the sound wave fronts created by the acoustic pulses. The virtual air temperature is
computed from the speed of sound which is measured by the reflected radar energy. The typical range of a
RASS is approximately 100 to 1500 m with a resolution of 60 to 100 m.
CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge 4-15
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Meteorological Data Requirements
While great strides have been made with profilers, in conjunction with radiosonde measurements to
sample the vertical profiles of meteorological variables needed in air pollution assessment and forecasting,
there are still frequently large gaps in data which complicate the process of determining when and where
transport of air pollutants occur. More data is needed to define the spatial continuity of air pollution masses,
their chemical changes with time, and particularly in Southern California, the spatial (both horizontal and
vertical) and temporal variations in wind direction and speed. Each Basin and valley has its own mesoscale
wind and thermal influence on local conditions which are strongly tied to the diurnal heating cycle; and these
variations are superimposed on the larger scale synoptic flow pattern. Where mountainous terrain is
prevalent, these slopes should be used to sample wind, temperature, humidity, ozone, PM, haze and
precursors to fill some of the voids in ambient data. Along coastlines, the data voids are particularly severe
because very little data is available over the ocean, yet the water has a profound impact on wind circulations
in all adjacent areas. In addition to the use of automatic weather and pollution monitoring stations over
complex terrain, ozonesondes are a potentially useful technique for determining the vertical profile of ozone.
Launched on a balloon (with conventional radiosondes in tandem), these soundings can be used to detect
rivers of pollution which remain trapped aloft within inversion layers, and flow to neighboring areas by winds
which can be dramatically different than the flow at the surface. To help fill in some of the voids at sea,
aircraft measurements and island stations are needed to help define the seaward transport of air pollutants
and their subsequent transport to other regions. Because of the large amount of horizontal and vertical
variability, the largest possible spatial domain must be used in modeling studies in order to get a clear picture
of transport.
Diurnal Changes
In regions characterized by land/sea breeze regimes, there is sometimes the tendency to simplify the
meteorology by assuming the wind is offshore at night and onshore in day, and to use wind statistics at two
times a day to represent these changes. However, to understand transport, it is essential to know how (and
through what directions) the wind changes in going from land breeze to sea breeze. For instance, in Southern
California, there is very frequently a period of southeast winds between the nighttime northerly land breeze
and the afternoon westerly sea breeze. The nighttime land breeze advects varying amounts of polluted air out
to sea, and then the southeast winds are responsible for transporting some (or in some cases, much) of the
polluted air mass along the coast to adjacent Basins. At certain times of day and year, these intervening winds
may actually be the predominant wind. These climatological features need to be documented (i.e. monitored),
and incorporated into modeling studies. Simplification of realistic wind conditions can lead to grave
misconceptions about where polluted air is coming from (e.g., confusion between 'local episodes' and
'transport days').
Vertical Shear
In some regions, surface winds may be completely unrepresentative of flow patterns that exist even
500 or 1000 feet above the surface. For example, in the southeasterly flow discussed previously, ozone
and/or precursors may be transported tens of miles or more to a neighboring area. However, a shallow
4-16 CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
westerly sea breeze may cut in underneath at the downwind site, thus making it appear that observed
pollutants were coming from a different direction, including the ocean. Some initial studies are underway in
Southern California to determine in what regions and how much of the time, the inversion layer may be
embedded within an air stream that is markedly different than the surface flow. Since it has been found that
inversions can trap ozone and transport it to downwind areas where it may later be fumigated back down to
the surface, detecting and understanding the local variability of the wind in the vertical is essential for
understanding and predicting transport.
Ozone versus Precursors and?articulates
In detecting and predicting when transport takes place, it is also critical that enough variables to
chemically describe the polluted mixture be simultaneously tracked. Visibility studies have been conducted
in the past which document the diurnal and large scale transport of particulates associated with "smoggy"
air masses responsible for high ozone levels inland. However, in areas with proximity to the ocean, ozone
levels within the polluted air may not be significant due to various influences such as reduced sun light, ozone
destruction and other processes. If only ozone is measured or recorded, erroneous conclusions can be
obtained regarding the presence of air pollution, but even more importantly, about the source of ozone in
downwind areas where die same polluted air mass may move back into heated, inland valleys to again
produce exceedances.Some of the confusion also leads to erroneous attributing of the offshore polluted air
transported to far away coastal locations as consisting of natural 'marine haze'.
Network Design
Considerations of Network Design for Exposure Assessment
Ambient air quality monitoring networks are designed primarily to collect a sufficient base line of
information to identify population areas that are exposed to concentrations of pollutants that are in excess
of the NAAQS and to measure progress toward reducing those population exposures. EPA requirements and
guidance for criteria pollutant monitoring, are codified in 40 CFR 58. These requirements have focused
predominantly on the single objective of determining compliance with the NAAQS. While this is an
important and desirable objective, other important information on the environmental state and ambient
pollution exposures, sources, formation and transport processes, and effects is neglected and/or precluded
by the "traditional" approach.
Measurement of an exceedance of any NAAQS, under this traditional approach, results in locally
unpleasant consequences for sources and regulatory agencies in the proximity of measured exceedances.
Conversely, there are no consequences for economically competitive sources which may contribute
substantially to exceedances at many sites, but are located in upwind, "attainment" political jurisdictions.
This situation produces strong disincentives to expand measurement networks and to locate additional
monitoring sites to characterize maximum human exposures, or otherwise provide useful information on
pollution sources, processes or effects.
CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge 4-17
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
NAAQS are typically established for the most easily quantified individual chemicals (e.g. ozone)
or broad categories of complex pollutant mixtures (e.g. particulate matter), which are taken as surrogates for
larger groups of related chemicals (e.g. photochemical oxidants), or for more specific chemical components
of a complex pollutant mixture (e.g. the sulfate, nitrate, acidic, organic, or trace metal components of
particulate matter). This approach precludes the collection of information that could more accurately define
actual pollutant exposures to injurious chemicals, and/or could support future development of revised
standards to achieve more efficient progress toward national health, welfare and environmental goals.
Ambient concentrations and effects of ozone, fine particles and regional haze result from emissions
from many different source types over broad regions, as modified by complex intervening atmospheric
chemical transformation and meteorological transport processes. Measurement of the resultant end products
of these complex emissions and processes provides limited information to support development of effective
emission reduction strategies to attain standards or to advance the scientific basis for development of such
strategies. Notable exceptions to this approach include the systematic measurements of some oxidant
precursor concentrations in the PAMS program, and comprehensive measurements of atmospheric optics,
scenes, and detailed aerosol composition in the IMPROVE program
Discussion of Existing Networks
National networks: SLAMS/NAMS/PAMS
State and Local Air Monitoring Stations (SLAMS) networks are the primary source for ozone
ambient air quality data used by the U.S. EPA and State and local air quality management agencies. Of the
981 total ozone monitoring sites reported to the U.S. EPA's AIRS database, 780 of these, or approximately
80%, are SLAMS ozone sites. Most of the SLAMS are located near or within urban areas. However,
roughly 10% are located in more rural settings, particularly those with transport assessment or
background concentration determination objectives.
National Air Monitoring Stations (NAMS) are subsets of the SLAMS. The NAMS were established
to provide consistent air quality data that could be used to assess national ozone trends and to make national
ozone policy decisions. Of the total number of SLAMS ozone monitoring sites, there are approximately 229
ozone NAMS sites in operation at this time. The 40 CFR 58 regulation requires that two NAMS sites be
established within any urban area with a population of 200,000 or more. One of these sites is located to
assess population exposure and the second is located to determine the maximum concentrations for that urban
area, hi addition, all PAMS monitoring sites (see below) require ozone measurements, and the CASTNET
sites include rural ozone measurements.
Photochemical Assessment Monitoring Stations (PAMS) networks, the most recent addition to the
ozone monitoring requirements in 40 CFR 58, are a subsets of SLAMS and include routine measurements
of ozone precursors as required by Section 182 (c)(l) of the 1990 Clean Air Act Amendments. PAMS
networks are currently required in 22 ozone nonattainment areas designated as serious, severe, and extreme
with respect to the current 1-hour ozone NAAQS.
4-18 CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
These monitoring stations collect ambient air measurements for a target list of approximately 60
volatile organic compounds (VOC) including several carbonyls, as well as oxides of nitrogen, ozone, and
both surface (10-m) and upper air meteorological measurements. PAMS is the first national program to
require near-continuous (hourly, or 3-hour averages) speciated VOC and NOX measurements.
PAMS requirements are designed to provide as much information as practicable on the roles of
ozone precursors, pollutant transport, and local meteorology in the photochemical process as well as to
establish a feedback loop as a "reality check" on proposed ozone control (State Implementation Plan or SIP)
strategies. Eventually, PAMS will provide a data base for evaluating the success of the control programs and
developing mid-course strategy corrections. Specific provisions of the Rule require the establishment and
operation of up to five PAMS in each affected area, depending on the population. The stations are designed
to sample representative air parcels upwind, within the central business district, and downwind of the urban
core of nonattainment areas.
4.B Emission Estimates and Inventories
The development of an emissions inventory is a complex process that often involves a variety of
specific procedures to prepare a complete comprehensive data base for specific applications. This section
outlines the fundamentals of the emissions inventory development process and is not meant to describe all
aspects of emissions inventories. Emissions inventory data are prepared to support regulatory efforts and
to serve as input for air quality simulation models. This section also addresses the types of emissions
inventories currently in use for these regulatory support and research activities. The role of emissions
inventory data for modeling exercises is increasing in importance and emissions inventories are directly
related to many of the issues being discussed in the overall implementation development process to support
the new NAAQS and Regional Haze programs.
Considerations for Use of Emissions Inventories in Modeling Analyses
A regional air quality modeling effort begins with an emissions inventory. An existing regulatory
emissions inventory is often the starting point for development of the modeling emissions inventory, which
must satisfy the specific needs of the selected air quality model. Therefore, modeling emissions inventory
tools are a superset of regulatory emissions inventory development tools.
The purpose of a regulatory emissions inventory is to estimate and report emissions of specified
primary pollutants by county. The regulatory emissions inventory tools include emissions factor estimation
software such as MOBILE for on-road mobile sources, Federal Aviation Administration Engine Emissions
Database (FAAEED) for aircraft, TANKS for storage tanks, Biogenics Emission Inventory System (BEIS)
for biogenic sources, and other specialized models. Emissions factors for many sources are simply looked
up in AP-42 or EPA's FIRE database. These emissions factors are multiplied by estimated activity levels
to produce estimated emissions. The emissions may be reported for an average day, an average weekday of
a specific season, or a worst-case scenario. The report includes emissions by source category for each
reported county.
CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge 4-19
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
The purpose of a modeling emissions inventory is to provide the emissions data at the level of detail
required by a selected air quality model. In terms of the top down and bottom up inventories (which are later
in this document) a modeling inventory inherently has additional uncertainties relative to any common
regulatory inventory. Some of the causes and consequences of these uncertainties are summarized below.
The purposes of the additional tools used in modeling inventory development are:
convert primary pollutant emissions into emissions of the pollutant species used by the air quality
model's chemical mechanism;
estimate emission factors for physical conditions of the modeling scenario (e.g., use actual
temperatures instead of average temperatures for evaporative emissions);
forecast activity levels from the year of the base emission inventory to the year of the modeling
scenario;
estimate emission factors for the year of the modeling scenario (e.g., run MOBILE for fleet and RVP
characteristics of the modeling scenario);
estimate activity levels for the time of the modeling scenario (e.g., weekday of season in modeling
scenario instead of average);
allocate daily emissions to each hour of the day; and,
spatially allocate county emissions to the three-dimensional grid cell structure in the air quality
model.
Development of a modeling emissions inventory adds to model generalization by using one
speciation or temporal profile to represent multiple emissions source categories, and by assuming all
emissions are constant within each hour. The development of a modeling emissions inventory also adds to
model distortion by speciating emissions to fit the chemical mechanism in the model, by using spatial
surrogates to locate emission sources, and by spatially smearing emissions into a large grid cell volume under
an assumption of complete and instantaneous mixing. Preparation of modeling emissions inventories adds
to model deletion by ignoring those sources that are not represented in the inventory (i.e., each ignored source
may be small, but the cumulative effect of all ignored sources is unknown), and often by ignoring other
chemical species such as methane.
Uncertainties in the modeling emissions inventory include all uncertainties in the base regulatory
emissions inventory (i.e., methods, measurements, and missing sources), as well as all uncertainties
associated with converting the base to the modeling emissions inventory. These conversion uncertainties
include:
measurements of speciation;
development of speciation profiles;
representativeness of speciated samples to unsampled sources in the same source category, in a
different source category, in a different geographic region, or in a different year;
growth factor estimates;
spatial surrogate errors (i.e., original data errors, calculation errors);
4-20 CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
applicability of spatial surrogates (i.e., validity of population surrogates for both 1990 and 2010,
and the adequacy of a surrogate for a specific source category); and,
applicability of temporal profiles (i.e., validity of EPA default temporal profiles for this modeling
domain, representativeness of one profile for all sources of one source category throughout all urban
and rural areas in the modeling domain).
Existing Emissions Inventories
Emissions inventories are used in several ways as tools in air quality management. The inventory
on its own provides useful information on the magnitude and distribution of emissions across all
anthropogenic and natural sources of emissions and the inventory can be tracked over time to present useful
information on the growth or reduction of emissions from the various anthropogenic and natural sources.
It is also a major input to atmospheric computer models that are used to predict ground-level pollutant
concentrations and to analyze various air pollution control strategy options.
Two basic types of emission inventories (e.g. top down and bottom up) are used in air pollution
management studies. The top down inventory is the simpler, and the less expensive of the two. When
combined to form national or regional inventories, inventory method consistency is a concern especially when
comparing emissions from one inventory area to another. Top down inventories are prepared from existing
data sources and use estimation techniques that are comprehensive but less specific than the bottom up
method. The geographical extent is usually large, often being national in scope. Top down inventories have
the advantages of being consistent in methodology, can cover very large geographical areas, are relatively
inexpensive and can be produced and modified quickly. The EPA uses top down inventories for regional
scale modeling and the tracking of emission trends over time.
Bottom up inventories are prepared using specific locally derived data and may cover an area as
small as a single point source or as large as a state. Bottom up inventories are characterized by developing
specific emissions data from as many specific sources as possible. This entails direct contact with emission
sources, particularly large point sources. Bottom up inventories are usually produced by State and local
environmental agencies, and are regarded as having greater accuracy than top down inventories. These
inventories are costly and time consuming to produce.
Ozone is a pollutant that is formed in the atmosphere from precursors that are emitted from various
sources. The precursors that are important in ozone formation are VOCs, NOX, and CO. These precursors
have been inventoried for a number of years and inventoried intensively since 1990. A national top down
inventory for these precursors has been developed by the EPA and is publicly available via anonymous FTP
at earthl.epa.gov under the directory pub/gopher/EmisInventory. The file names are: US90ARV2,
US90MBV2 and US90PTV2. In addition to the ozone precursors, this inventory also contains sulfur dioxide
(SOj), ammonia (NH3), PM-10,PM-2.5 and secondary organic aerosols (SOA). This top down inventory is
comprehensive, with emission estimates presented for all source categories for each county in the United
States.
CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge 4-21
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Emission Factors
As stated above, emission factors are a primary tool in the development of emission inventories.
Table 4.5 is a presentation, by pollutant, of the emission factors currently contained in the EPA's Factor
Information Retrieval Data Base (FIRE).
In addition to the number of emission factors for each pollutant, the table contains information on
the percentage of these factors that have the higher A or B data quality rating. This table also reveals that
there is good coverage for the ozone pollutants although the quality of VOC is low relative to NOX and CO.
The particulate matter and haze pollutants are less well covered. Good coverage exists for NO*, VOC and
S02. The coverage for PM-10 would also appear to be good. However, the great majority of the PM-10
emission factors were derived from the PM emission factors shortly after the EPA promulgated the current
PM-10 NAAQS. This effort involved applying existing particle size distribution data to the PM emission
factors to derive the new PM-10 factors. The data quality ratings for the new PM-10 emission factors were
not lowered as they should have been. Therefore, the coverage for PM-10 is still good but the confidence
in the factors is lower.
No emission factors for PM-2.5 are in this data base. The factors that were used in the EPA top
down inventory were derived from the PM emission factors using a calculator program that included particle
size distribution data for the various source categories. Confidence in these size distribution data is low,
especially for the PM-2.5 size interval. Subsequently, EPA is currently collecting field test data for PM-10
and PM-2.5 emissions from paved and unpaved roads, since both of these fugitive particulate source
categories are major emission sources. When these data are available, confidence in the fine particle emission
factors will improve. Emission factors for NH3 are not in the EPA FIRE data base. The NH3 factors that
were used in the EPA top down inventory are a compilation of NH3 factors based on recent studies primarily
from Europe. Subsequently, confidence in these factors is also low.
Data Sources and Tools
A list of the data sources and tools used for emissions inventory development that have been
discussed in this Section and in Chapter 3 is provided below:
Department of Energy utility fuel data;
1985 NAPAP emission inventory;
Bureau of Economic Analysis (BEA) growth factors;
Quantity of solvents consumed;
County Business Patterns; County employment for specific industry groups;
EPA wood burning model;
Federal Highway Administration's Highway Performance Monitoring System (HPMS) database;
County population;
EPA MOBILE mobile sources emission factor model;
EPA/OMS detailed 1990 nonroad mobile inventory for 27 nonattainment areas;
Biogenic Emissions Inventory System - Version 2 (BEIS2);
4-22 CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Table 4.5:
FIRE emission factors.
FIRE EMISSION FACTORS
POLLUTANT
CO
NOX
voc
S02
PM
PM-10
PM-2.5
NH,
NUMBER OF EMISSION
FACTORS
912
1232
2487
1358
2008
1212
0
0
PERCENT OF EMISSION
FACTORS RATED A OR B
24
17
10
27
27
23
-
-
CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge 4-23
-------
Science and Technical Support Work Group (STSWG) Conceptual Mode], February 1997
Land use data;
Meteorology data;
Point source surveys;
AP-42; ' -
EPA FIRE database;
EPA/OMS PARTS mobile sources PM emission factor model;
Department of Agriculture crop land area, area planted, crop type, soil parameters, times tilled per
year;
Department of Agriculture census of agriculture;
Wind erosion model;
Cost of construction;
Average duration of construction activity; and,
Recent European NH3 studies.
4.C Air Quality Models
This section addresses some of the issues related to the application of modeling analyses to
regulatory decision making. The discussion focusses on two principal modeling approaches that are referred
to here as air quality simulation models and receptor models. Although air quality simulation models are
used for the analysis of particulate matter issues, they are currently most commonly applied to studies of
urban and regional ozone episodes. Receptor models are used primarily to assess urban scale or smaller scale
PM issues. The section begins with a discussion of the basic issues related to model formulation, inputs and
performance. That is followed by a brief discussion of some of the problems encountered in modeling
analyses and the section ends with a summary of some of the features of existing air quality models.
Model Formulation and Applications
Basics
This section addresses urban, regional and super-regional scale modeling systems, most of which
use gridded Eulerian frameworks. Gridded models utilize a fixed-frame reference with respect to moving air
masses and can accommodate a physically realistic treatment of atmospheric mixing processes. In contrast,
trajectory or Lagrangian systems utilize a moving reference frame perspective tracking the movement (and
retaining the identity) of individual plumes. Trajectory models estimate particular plume concentration
distributions along a path; whereas Eulerian systems provide simultaneous calculations at every grid cell,
essentially capturing the fully coupled calculations of all interacting plumes (actually, air volumes). The
formulation of Lagrangian systems restricts the realistic treatment of atmospheric mixing, since plumes are
subject to degradation through wind shear and other physical phenomena and thus lose individual identity.
The Lagrangian treatment of individual plumes requires substantially fewer computational resources,
whereas the solution of fully coupled simultaneous systems (Eulerian) is far more demanding. The chances
for plume degradation are enhanced when trajectory models are applied on the regional scale, that is of
4-24 CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
interest in so many air quality issues and, therefore, Eulerian systems have become the most common
modeling approach for urban and regional studies.
Although many current efforts are dedicated to large scale reactive modeling systems, receptor and
non-reactive dispersion modeling tools perform important roles. For example, the demand for'observational
based analyses has and will expand the use of receptor modeling techniques for interpreting ambient data and
evaluating emissions. The strong probability of a fine particle standard requires attention to regional scale
reactive modeling systems. This is because the fraction of mass due to secondary formation increases with
decreasing size cutoffs, and the atmospheric residence times for small particles approach those of gases.
Nevertheless, primary particles are important components of particle mass regardless of the size cutoff.
Furthermore, the current PM-10 standards are not disappearing. Consequently, the application of receptor
and nonreactive, local-scale modeling techniques remains an important component of a comprehensive air
quality management effort.
Eulerian Models
All modeling systems invoke many approximations both in the description of physical and chemical
processes as well as in the solution of the system of mathematical equations used to represent the physics
and chemistry. The spatial extent, or "domains", of model applications often are characterized as being
of urban (100-500 km), regional (500-2000 km) or super-regional (> 2000 km) scales. Gridding refers to
the horizontal resolution used to delineate simulated air quality concentrations and provide detail on the
emissions distribution and meteorological variables (e.g., winds and temperatures). Typical grid resolution
ranges from 2-5 km and 20- 80 km for urban and regional applications, respectively. Most models produce
hourly outputs, which can be aggregated for other averaging times of interest (e.g., 8-hr, 24-hr, seasonal,
annual). Models are resolved vertically (typically 5-15 levels or more) to account for varying meteorology
and emissions and to approximate vertical mixing phenomena. Some modem systems accommodate nested
or variable scale gridding schemes which allow for detailed spatial treatment in urban areas (2- 8 km) and
less dense resolution in peripheral/rural portions (20-80 km) of the domain in order to optimize
computational resources and apparent precision. Over the past decade there has been a trend toward
increasing regional-scale (or mixed regional/urban) Eulerian modeling in recognition of the interaction
between regional/rural and urban areas and associated "transport" issues.
Models should be viewed as a "system" (see Figure 4.1), including the meteorological and emission
preprocessing models and the Air Quality Simulation Model (AQSM). Preprocessors assemble raw data
(i.e., emissions inventories and meteorological measurements) into the spatial and temporal fields required
by the AQSM. The AQSM calculates concentration fields of air quality species (i.e., compound, element,
free radical, or surrogate group) which are determined by the combined interactive effects of source
emissions, mixing processes (advection and dispersion), deposition and chemical transformations.
So-called "transport" (e.g. via advection and dispersion) into a modeling domain is quantified
through boundary conditions which are user-specified. The AQSM includes the chemical mechanism which
performs the chemical transformation calculations through highly condensed approximations. Typically,
chemical mechanisms will include 20-50 species and the order of 100 reactions to represent the thousands
CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge 4-25
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Air Quality Modeling System
Ambient Concentrations^
.Deposition levels
Figure 4.1: Schematic of an Air Quality Simulation Model (AQSM).
4-26 CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
of actual chemical reactions occurring in the atmosphere. Chemical mechanisms have been developed and
tested through combinations of fundamental reaction chemistry and smog chambers.
All gridded modeling systems share the general attributes discussed above. Differentiation among
modeling systems is dictated by the required species and averaging times of interest and the desired balance
between physical reality and computational resource demands. For example, the Urban Airshed Model
(UAM) and the Regional Oxidant Model (ROM) ozone models are generally applied for episodic ozone
events occurring over periods of 2 to 14 days where processes/events such as precipitation, cloud interactions
and particle formation may not be critically important to determining peak ozone concentrations. Hence,
unlike the Regional Acid Deposition Model (RADM), the UAM and ROM do not include cloud and
precipitation processes nor a characterization of sulfur and nitrogen chemistry (gas and aqueous phases)
which are required for precipitation and particle formation processes. Conversely, the Regulatory Modeling
System for Aerosols and Deposition (REMSAD) does not include the detailed oxidant chemistry of the UAM
or ROM, which would exact enormous computational burdens on performing "screening" types of analyses
over annual time scales.
Modern model applications utilizing gridded systems are extremely resource and data intensive
exercises. A series of sequential steps applicable to regulatory model applications include:
Establishing model domain, and characterization/selection of modeling episodes;
Raw data gathering and processing of model inputs;
Model testing, including component testing of emissions and meteorological preprocessors;
Development of emission control strategies and model application and interpretation of results; and,
Corroboration and evaluation of strategy results.
While guidance documents exist for ozone modeling (EPA, 1991), the application process is a mix
of art, science and engineering, and the relative attention to any particular step varies strongly among
applications. The potential quality of a modeling application often is constrained by available data. The
use of national emissions, meteorological and air quality data bases as sole data sources for modeling is a
much debated topic, and is often viewed as more critical than the internal differences among the AQSMs.
Consequently, most historical modeling applications have been associated with special intensive field
programs designed to augment national data bases. Equally perplexing are issues related to "appropriate"
use of models. For example, the current CAA and EPA guidance require modeled ozone attainment
demonstrations; whereas, the scientific community has questioned how a deterministic tool can be applied
to a probabilistic standard under highly stochastic meteorological and emissions phenomena. All of these
problems are compounded with secondary particulate modeling, given the need to characterize additional
physical/chemical processes (aqueous and gaseous phases, organic aerosols, size distributions and growth;
the shortage of ambient data for model evaluation, especially time-varying speciated data), and related
incommensurabilities between models and measurements.
CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge 4-27
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Particulate Air Quality Modeling
Currently, there are no routine methods available that can be used to assess the formation of
secondary PM-10 or PM-2.5 for either the 24-hour or annual standards. Conceptually, the most attractive
approach for the 24-hour standard is to build upon the framework of the Eulerian models now used primarily
for ozone assessment. This can be done by having the model produce hourly concentrations for ozone, other
oxidants, the nitrogen oxide species, gaseous organic species, and sulfur oxide species. These chemical
species can then be converted into the corresponding secondary aerosols by appropriate additions to the
chemical mechanisms, coupled with a physical module to describe the conversion and growth of gaseous
species to aerosols. At the present time, there is insufficient data to even evaluate these models to determine
if they are capable of reproducing the aerosol concentrations, much less determining the accuracy and
precision of the models. Furthermore, the data needs for such a model have not been described specifically,
therefore it is not clear what data would be required to use this type of model in a regulatory effort. Until this
type of model, or some other approaches are developed, it will be difficult to determine source receptor
relationships for pollutants such as ammonium nitrate, secondary organic aerosols, or ammonium sulfate.
Similarly, there is no recommended regulatory approach for modeling secondary aerosols relative
to the annual standard. Any of the conceptual models to perform this kind of assessment would necessarily
require a significant number of simplifying assumptions about the chemistry of secondary aerosol formation
and the role of meteorology. There have been suggestions that a suite of models could be used to provide
information that could be used to identify those emission sources which need to be controlled. For example,
one can use a speciated linear rollback approach which assumes that changes in paniculate (including
secondary paniculate) will be proportional to the relative contribution of the different sources of that
paniculate or its precursor. Another member of the model suite could be a trajectory model or a particle in
cell model. These models need both gaseous and paniculate emissions inventories, but they can use
simplifying assumptions of rates of formation of secondary sulfate and nitrate aerosols in place of a
sophisticated chemical mechanism. They can also use simple treatment of meteorology for advection and
dispersion in place of sophisticated meteorological models. This hybrid approach could be applied for either
the 24-hour or annual standards. However, such a hybrid approach contains numerous assumptions and
approximations, and needs to be tested against any and all available data to corroborate the findings.
Receptor Models
This section will describe issues surrounding the availability of air quality models for assessment
of both 24-hour and annual fine primary and secondary particle assessment. It will also address the need for
and availability of emissions, air quality and meteorological data for evaluation of model performance, tests
of sensitivity of emissions reductions relative to changes in ambient fine particle levels, and use for
assessment of alternative control. None of the described modeling approaches can be used to demonstrate
attainment.
Present methodology for assessment of PM-10 strategies are primarily based on the use of statistical
models which attempt to allocate the concentration of ambient monitored species to sources which emit some
combination of these species. This approach, referred to as receptor modeling, is used for directly emitted
4-28 CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
compounds that, in the aggregate, make up primary PM-10 and PM-2.5. The model assumes that no
chemical transformation occurs once the particles are emitted. EPA has endorsed the use of Chemical Mass
Balance (CMB), and related models to develop PM strategies. CMB is viable for sources such as road dust,
construction, wood stoves and primary mobile to name a few. CMB and related receptor models are not
viable for secondary particle components-those particles which arise due to chemical transformation in the
atmosphere, such as ammonium nitrate, ammonium sulfate and secondary organic aerosols.
Neither meteorology nor emissions are required for receptor models; consequently, CMB can not
give any indication about the spatial distribution of the potential culpable sources. CMB can be used to
assess strategies for both the 24 hour and annual standards, given the above restrictions, hi terms of being
able to carry out an assessment using receptor modeling techniques, it is critical that there be source profiles
available for each primary source type that contributes to the primary ambient PM. These profiles must be
speciated to allow separation of similar types of sources. There needs to be a unique combination of speciated
components which allow the source type to be well characterized. For example, nickel and vanadium are
unique tracers for crude oil. Heptanes (steroidal type compounds) may prove to be unique tracers for light
duty mobile sources now that lead is no longer in fuels. In addition, the profile should have been developed
through a robust set of measurements. At the present time, many profiles are not statistically robust, and
do not provide a sufficient characterization of the source. This is especially true for the family of mobile
sources and different sources of fugitive dust.
Summary of Uncertainties Affecting Modeling Analyses
The performance characteristics of any model generally depend on three factors: (1) model
assumptions; (2) input data; and, (3) numerical solutions. This discussion also addresses a fourth factor that
is often overlooked, the evaluation and purpose of the model. Many of the issues related to input data,
primarily emissions estimates, have been discussed in previous sections of this document, and will not be
repeated here. These issues will be discussed in more detail below.
Model Assumptions
Chemistry
Use of the Carbon Bond-IV (CBM-FV) mechanism, which is most commonly applied in urban
modeling studies, makes it difficult to predict the concentrations of speciated VOC except for the explicit
species like isoprene. Formaldehyde concentrations are frequently over predicted because the mechanism
treats some reactive olefins as formaldehyde. Aromatics chemistry is still not well known and as a result
CBM-IV tends to be less reactive than Statewide Air Pollution Research Center (SAPRC-90). SAPRC-90,
however, has problems defining the rate constants of the different lumped species, because these rate
constants depend on the constituents represented by the lumped species. Currently, this problem is treated
on an ad hoc basis.
Existing mechanisms have been evaluated against chamber results where initial NOX concentrations
are typically several times higher than ambient. These mechanisms must still be checked against experiments
CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge 4-29
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
for conditions of low NOX concentrations and low VOC/NOX ratios. These experiments are very difficult to
conduct and obtain reasonable results.
Meteorology
Currently, there is a large uncertainty in predictions of mixing height based on limited measurement
data. Furthermore, our knowledge of the mixing coefficients is still based on the homogeneous surface layer
assumptions which have never been shown to be valid in an urban environment. The nocturnal jet is ignored
by most regional models, and wind field predictions in the lower troposphere over complex terrain are not
good.
Scale Problems
Nonlinearity in chemistry substantially affects model results over the range of temporal and spatial
scales involved in the chemical and physical processes. Most modeling applications use rather large grid
sizes and time steps. As a result, the model's ability to predict pollutant concentrations deteriorates. For
example, the dissimilar distributions frequently observed in VOC and NOX emissions and concentrations,
can affect the size of the predicted N0x-control region depending on the grid size of the model used. Choice
of the vertical layer depths also impact the chemistry and the model predictions. For the lowest layer in the
models, the predicted concentrations of species emitted from surface sources are not reliable unless the layer
thickness is less than the mixing scales of the species. However, currently there is very little information
available on the specifics of this mixing scale.
Numerical solutions
This issue is often taken for granted, even in cases when serious problems can result, especially in
the solution of the advection equations. The impact on the model predictions can be significant (10 -20%
can be expected for the predicted peak concentrations.) All present air quality models have such difficulties,
however, there are efforts underway to resolve these problems.
Evaluation and purpose of the model
Even if the model were perfect, with no serious problems with the above three factors, it may still
perform poorly for a given application, and the results may be evaluated incorrectly. First, it is always more
difficult to predict the upper extreme values than the means. This is true for the daily max as well as for the
annual extremes of the daily max. Extreme values result from combinations of unusual events due to
fluctuations. The time scale of these fluctuations leading to the daily max may cause the effects to be
partially averaged out in the model. For annual extremes, however, these effects are not averaged out.
Therefore these fluctuations must be modeled faithfully, unless the model is based on a longer time scale
(e.g., daily). Such a model, however, may not be too useful for control planning.
The model may be tuned to predict a specific episode well, but the conditions leading to that episode
might not be repeated again. In that case, if the necessary controls predicted for that episode are not
4-30 CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
substantially valid for other episodes the modeling analyses may not provide accurate control strategies.
Similarly, predictions for the average conditions are not always valid for the extreme conditions. The
regulatory requirement of predicted extreme values has added a tremendous and almost unreasonable demand
on air quality models, given the uncertainties inherent in modeling analyses. Comparison of model outputs
with observations can also be problematic, because measurements are often representative of "only a limited
spatial scale relative to the model grid size. Longer averaging times of multiple hours could limit spatial
averaging problems except in areas near major NOX sources or major ozone sinks where is ozone is not
uniform locally. The scale of nonuniformity can be dramatically worsened by the wind shifts and
fluctuations.
Emitted species concentrations are not always uniform in the early morning because of poor mixing.
If the dominant mixing scale and the model grid resolution do not match, comparison with observations could
be rather meaningless. Even if they match, comparison may still be meaningless if the position of
measurements does not correspond to the relevant spatial scale. Third, it is often quite difficult to compare
the predicted peak concentrations with observations because the latter are restricted by the availability of the
monitors which may not be located near the predicted peak concentrations.
Despite numerous issues related to models and the model application process, the modeling system
remains an extremely powerful tool for:
1. Expanding/complementing the time and space constraints of ambient monitoring which often is
limited to a small number of surface observations;
2. Further expanding time by predicting future conditions; the very basis of control strategy analysis;
and,
3. Integrating emissions, meteorological and air quality data for explaining physical and chemical
phenomena that can not be described by data alone (or a single data category).
Description of Common Eulerian AQSMs
The following summaries of Eulerian AQSMs provides basic information regarding pertinent
references, applicability (pollutants and spatial scales), and chemical mechanisms. Comments regarding any
unintended mischaracterizations or omissions are greatly appreciated. One of the most comprehensive
modeling reviews was written by Seigneur and Saxena (1990) for the Electric Power and Research Institute
(EPRI), and along with assistance from numerous model developers provides the basis for the modeling
summary. The American Petroleum Institute (API) recently issued a Request for Proposals (RFP) for a
review of particle models. PM and ozone modeling are treated together, reflecting the commonality in
process treatment among the available Eulerian models. Visibility is a direct function of particle levels,
particle size distributions and particle composition; all properties that particle models strive to simulate.
Consequently, visibility is (can be) treated as a postprocessing step in any particle model under discussion.
A summary of Eulerian AQSMs is presented in Table 4.6.
CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge 4-31
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Table 4.6: Summary of eulerian Air Quality Simulation Models (AQSM).
Summary of Eulerian Air Quality Simulation Models
Model
UAM-IV
UAM-V
ROM
RADM-II
RPM
SAQM
DAQM
CIT
Pollutants/
scales/comments
03; urban; EPA
guideline model;
public domain
0,; variable scale;
PIG; urban and
rural; proprietary
03; regional;
public domain
Acid dep; O3;
regional;
public/research
secondary PM;
RADM physical
attributes;
research (not
evaluated)
o,
secondary aerosol;
PM,03
o,
Chemical Mechanism
CB-IV(full gas-phase
oxidant chemistry)
CB-IV(full gas-phase
oxidant chemistry)
CB-IV(full gas-phase
oxidant chemistry)
Lumped molecular (full gas-
phase oxidant chemistry
with S); aqueous chemistry;
aerosols (sulfate only)
reads full RADM fields;
adds nitrate-sulfate-
ammonium; highly
parameterized org-aerosol
treatment; bi-modal PM
CB-IV; SAPRC-90
Major Applications
State SIPs; 5-City;
SCAQS;NEU.S.
Corridor
LMOS; GMAQS;
OTAG
ROMNET;
MATRIX study;
State UAM SIPs
NAPAP; acid dep.
stnd. feasibility;
Ches. Bay N-dep;
EMEFS
base- 1990 and 20 10
CAA strategies
Aug. 3-6, 1990
Episode, S. Calif.
Rocky Mtn Front
Range; Colorado
Los Angeles
Principal references
Reynolds et al.
(1979); Tesshe and
McNally(1990)
Morris, etal. (1993)
Lamb (1983);
Schere and Wayland
(1989);Chuetal.
(1992)
Chang etal., (1987)
Binkowski and
Shankar(1995);
Venkatram(1990)
Middleton(1993)
McRae et al.
(1982);Russelletal
(1 985); Harley etal.
(1993)
4-32 CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Summary of Eulerian Air Quality Simulation Models
Model
URM
STEM-II
CALGRID
ADOM
REMSAD
MODELS-3
Pollutants/
scales/comments
03; aerosols;
multi-scales
O3; acid dep;
regional
O,; regional and
urban scale
03; acid dep; PM;
regional scale
PM; deposition;
national, regional
and urban scales
multi pollutant;
under
development
Chemical Mechanism
SAPRC-90
lumped molecular for VOC;
sulfur-nitrogen-
ammonia/ammonium
chemistry for acid precip
CB-IV; SAPRC-90
oxidant chemistry with wet
deposition
empirical reduced form CB-
IV oxidant chemistry - 03 is
required input
RADM-IITRPM
Major Applications
NEU.S.;
Mexico/U.S. border
Philadelphia; central
Japan; Kentucky,
NWU.S.
LMOS, MOCA
Eastern North
America; Northern
Europe
Contiguous U.S.
SESAUM/SOS
prototype
Principal references
Kumar and Russell
(1995)
Carmichael et al.
(1991)
Yamartino et al.
(1992)
Venkatram et al.
(1988)
SAI(1996)
Dennis, etal. ( 1995)
CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge 4-33
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
4-34 CHAPTER 4: Current Tools to Address and Implement Current State of Knowledge
-------
CHAPTER 5
Time Distance
Considerations Relevant to
Transport and Regions of
Influence
Chapter 5 of the Conceptual Model is currently under development and review by the STSWG.
This chapter will be provided to FACA Subcommittee members at a later date.
CHAPTER 5: Time Distance Considerations 5-1
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
5-2 CHAPTER 5: Time Distance Considerations
-------
CHAPTER 6
Current Needs Based on
Relevant Issues and
Identified Information Gaps
Chapter 6 of the Conceptual Model is currently under development and review by the STSWG.
This chapter will be provided to FACA Subcommittee members at a later date.
CHAPTER 6: Current Needs 6-1
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
6-2 CHAPTER 6: Current Needs
-------
CHAPTER 7
Integration of Numerical
Models and Ambient
Monitoring Data for
Effective Air Quality
Management
Chapter 7 of the Conceptual Model is currently under development and review by the STSWG.
This chapter will be provided to FACA Subcommittee members at a later date.
CHAPTER 7: Integration of Numerical Models and Ambient Data 7-1
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
7-2 CHAPTER 7: Integration of Numerical Models and Ambient Data
-------
CHAPTER 8
Developing a Working and
Responsive Science-Policy
Continuum
Chapter 8 of the Conceptual Model is currently under development and review by the STSWG.
This chapter will be provided to FACA Subcommittee members at a later date.
CHAPTER 8: Developing a Working and Responsive Science-Policy Continuum 8-1
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
8-2 CHAPTER 8: Developing a Working and Responsive Science-Policy Continuum
-------
CHAPTER 9
References
Altshuller, A.P. and A.S. LeFohn. 1996. Background ozone in the planetary boundary layer over the United
States. J. Air Waste Manage. Assoc., 46, p. 134-141.
Binkowski, F.S. and U. Shankar. 1995. The regional particulate matter model, 1: Model description and
preliminary results. J. Geophys. Res., 100, D12. p. 26191-26209.
Carmichael, G.R., L.K. Peters and R.D. Saylor. 1991. The STEM-II regional scale acid deposition and
oxidant model -1. An overview of the model development and applications. Atmos. Environ., 25A,
p. 2077-2090.
Chang, J.S., R.A. Brost, IS. Isaksen, S. Madronich, P. Middleton, W.R. Stockwell and C.J. Walcek. 1987.
A three-dimensional Eulerian acid deposition model: physical concepts and formulation. J. Geophys.
Res., 92, p. 14681-14700.
Chock, D.P., S.L. Winkler and S. Pezda. 1997. Attainment flip-flops and the achievability of the ozone air
quality standard. J. Air Waste Manage. Assoc., in press.
Chu, S.H., W.M. Cox, E.L. Meyer, N.C. Possiel, R.D. Scheffe, S.J. Roselle and K.L. Schere. 1992. A matrix
study of VOC and NOX reductions and ozone responses for the eastern U. S. Report to the Assistant
Administrator, Office of Air and Radiation, U.S. EPA.
Cox, W.C. and S.H. Chu. 1993. Meteorologically adjusted ozone trends in urban areas: a probabilistic
approach. Atmos. Environ., 27b, 4, p. 425-434.
Crescenti,G.H. 1996 Personal Communication to Richard Scheffe, OAQPS/EMAD/AQMG.
Crescenti, G.H. and R.E. Payne. 1996. Evaluation of two types of thin film capacitive relative humidity
sensors for use on buoys and ships. Seventh Symposium on Meteorological Observations and
Instrumentation, Amer. Meteor. Soc., New Orleans, Louisiana, p. 125-128.
Dennis, R.L., D.W. Byun, J.H. Novak, K.J. Galluppi, C.J. Coats and M.A. Vouk. 1995. The next generation
of integrated air quality modeling: EPA's MODELS-3. Atmos. Environ., Accepted for publication.
Eldred, R.A. and T.A. Cahill. 1994. Trends in elemental concentrations of fine particles at remote sites in
the United States of America. Atmos. Environ., 28, p. 1009-1019.
Garratt, J.R 1994. The Atmospheric Boundary Layer. Cambridge University Press, New York, New York,
316p.
CHAPTER 9: References 9-1
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Goody, R. 1995. Principles of Atmospheric Physics and Chemistry. Oxford University Press, New York,
New York, 324p.
Grosjean, D. and J.H. Seinfeld. 1989. Parameterization of the formation potential of secondary organic
aerosols. Atmos. Environ., 23, p. 1733-1747.
Harley, R.A., A.G. Russell, G.J. McRae, G.R. Cass, and J.H. Seinfeld. 1993. Photochemical modeling of
the Southern California Air Quality Study. Env. Sci. Technoi, 27, p. 378-388.
Husar, R.1996. Personal communication to Richard Scheffe, U.S. EPA/OAQPS/AQMG.
Jeffries, H. 1997. Personal communication to FACA Science and Technical Support Work Group (STSWG).
Kelly, N.A., G.T. Wolff, and M.A. Ferman. 1984. Sources and sinks of ozone in rural areas. Atmos.
Environ., 18,1251-1266.
Kelly, N.A., G.T.Wolff, and M.A. Ferman. 1982. Background pollutant measurements in air masses
affecting the eastern half of the United States -1. Air masses arriving from the northwest. Atmos.
Environ., 16,1077-1088.
Kumar, N. and A.G. Russell. 1995. Development of a computationally efficient reactive sub-grid scale
plume model and the impact in the northeastern United States using increasing levels of chemical
detail. J. Geophys. Res., Submitted for publication.
Lamb, RJ. 1983. A regional-scale (1000 km) model of photochemical air pollution - 1. Theoretical
formulation. EPA Report No. EPA-600/3-83-005.
Lenschow, D.H. 1986. Aircraft measurements in the boundary layer, in: Probing the Atmospheric Boundary
Layer. American Meteorological Society, Boston, Massachusetts.
List, RJ. 1951. Smithsonian Meteorological Tables. Smithsonian Institution, Washington, DC, 527p
Malm, W.C., J.F. Sisler, D. Huffinan, R. Eldred and T.A. Cahill. 1994. Spatial and seasonal trends in
particle concentration and optical extinction in the United States. J. Geophys. Res., 29, p. 1347-
1370.
Middleton, P.B. 1993. Denver Air Quality Modeling Study (DAQMS), paper presented at the Twelfth
Annual Meeting, Amer. Assoc. For Aerosol Res., Oak Brook, Illinois, October 11-15,1993.
McRae, G.J., W.R. Goodin and J.H. Seinfeld. 1982. Development of a second generation mathematical
model for urban air pollution. I. Model formulation. Atmos. Environ., 16, p. 679-696
9-2 CHAPTER 9: References
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Morris, R.E., M.A. Yocke and T.C. Myers. 1993. Application of the nested-grid Urban Airshed Model to
the Lake Michigan region. AWMA International Conference and Course, Tropospheric Ozone:
Nonattainment and Design Value Issues, October 27-30,1993, Boston, Massachusetts.
National Academy of Sciences. 1991. Rethinking the Ozone Problem in Urban and Regional Air Pollution.
National Research Council, National Academy of Sciences, National Academy Press, Washington,
DC.
National Acid Precipitation Assessment Program (NAPAP). 1991. Acid Deposition: State of Science and
Technology - Report 24, Visibility: Existing and Historical Conditions - Causes and Effects. Office
of the Director, Washington, DC.
National Center for Atmospheric Research. 1985. Instructor's Handbook on Meteorological Instrumentation.
NCAR Report No. NCAR/TN-237+IA, Boulder, Colorado.
Pandis, S.N., R.A. Harley, G.R. Cass and J.H. Seinfeld. 1992. Secondary organic aerosol formation and
transport. Atmos. Environ., 26A, p. 2269-2282.
Reynolds, S.D., T.W. Tesche and L.E. Reid. 1979. An Introduction to the SAI Airshed Model and its
Usages. System Applications, Incorporated, San Rafael, California, Report No. SAI-EF79-31.
Russell, A.G., G.J. McRae and G.R. Cass. 1985. The dynamics of nitric acid production and the fate of
nitrogen oxides. Atmos. Environ., 19, p. 893-905.
Scheffe, R. 1997. Personal communication to Science and Technical Support Work Group (STSWG).
Schere, K.L. and R.A. Wayland. 1989. EPA Regional Oxidant Model (ROM2.0): Evaluation on 1980
NEROS Data Bases. EPA Report No. EPA-600-3-89-057.
Seignour, C. and P. Saxena. 1990. Status of subregional and mesoscale models, .Volume 1: Air Quality
Models. Electric Power Research Institute Report No. EPRI EN-6649.
Sisler, J., W. Malm, J. Molenar, K. Gebhardt. 1996. Spatial and seasonal patterns and long-term variability
of the chemical composition of the haze in the U.S.: An analysis of data from the IMPROVE
Network. July, 1996.
Sisler, J., D. Huffman, and D. Lattimer. 1993. Spatial and temporal patterns and the chemical composition
of the haze of the United States: An analysis of data from the IMPROVE Network, 1988-1991. Fort
Collins, CO.
Stull, R.B. 1988. An Introduction to Boundary Layer Meteorology. Kluwer Academic Publishers,
Dordrecht, The Netherlands, 666p.
CHAPTER 9: References 9-3
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
Systems Applications International (SAI). 1996. Statistical support for the participate matter NAAQS.
Systems Applications International, San Rafael, California.
Tesche, T.W. and D.E. McNally. 1990. Photochemical modeling of two 1984 SCCCAMP ozone episodes.
J. Appl. Meteor., 30, p. 745-763.
Trijonis, J. 1982. Existing and natural background levels of visibility and fine particles in the rural East.
Atmos. Environ., 16, p. 2431-2445.
United States Environmental Protection Agency. 1996a. Air Quality Criteria for Ozone and Related
Photochemical Oxidants. Office of Health and Environmental Assessment, Research Triangle Park,
North Carolina, EPA Report Nos. EPA-600/P-93-004aF-cF.
United States Environmental Protection Agency. 1996b. National Air Quality and Emission Trends Report,
1995. Office of Air Quality Planning and Standards, Research Triangle Park, North Carolina, EPA
Report No. EPA-454/R-96-005, 168 p.
United States Environmental Protection Agency. 1996c. Review of the National Ambient Air Quality
Standards for Ozone - Assessment of Scientific and Technical Information: Staff Paper. Office of
Air Quality Planning and Standards, Research Triangle Park, North Carolina, EPA Report No. EPA-
452/R-96-007,480 p.
United States Environmental Protection Agency. 1996d. Air Quality Criteria for Particulate Matter.
National Center for Environmental Assessment, Office of Research and Development, Research
Triangle Park, North Carolina.
United States Environmental Protection Agency. 1996e. Review of the National Ambient Air Quality
Standards for Particulate Matter - Policy Assessment of Scientific and Technical Information: Staff
Paper. Office of Air Quality Planning and Standards, Research Triangle Park, North Carolina, EPA
Report No. EPA-452/R-96-013.
United States Environmental Protection Agency. 1991. Guideline for Regulatory Application of the Urban
Airshed Model. Office of Air Quality Planning and Standards, Research Triangle Park, North
Carolina, EPA Report No. EPA-450/4-91-013.
United States Environmental Protection Agency. 1989. Quality Assurance Handbook for Air Pollution
Measurement Systems. Volume IV: Meteorological Measurements. Research Triangle Park, North
Carolina, EPA Report No. EPA-600/4-90-003.
United States Environmental Protection Agency. 1987a. Ambient Monitoring Guidelines for Prevention of
Significant Deterioration (PSD). Research Triangle Park, North Carolina, EPA Report No. EPA-
450/4-87-007.
9-4 CHAPTER 9: References
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
United States Environmental Protection Agency. 1987a. On-site Program Guidance for Regulatory Modeling
Applications. Research Triangle Park, North Carolina, EPA Report No. EPA-450/4-87-013.
United States Environmental Protection Agency. 1981. On-Site Meteorological Instrumentation
Requirements to Characterize Diffusion from Point Sources. Research Triangle Park, North
Carolina, EPA Report No. EPA-600/9-81-020.
Venkatram, A. 1990. Development and application of the SARMAP Air Quality and Meteorological
Modeling System. California Air Resources Board, Sacramento, California.
Venkatram, A., P.K. Karamachandani, and P.K. Misra. 1988. Testing a comprehensive acid deposition
model. Atmos. Environ., 22, p. 737-747.
Wolff, G.T., N.A. Kelly, M.A. Ferman, M.L. Morrisey. 1983. Rural measurements of the chemical
composition of airborne particles in the eastern United States. J. Geophys. Res., C-Ocean Atmos.,
88, p. 10769-10775.
World Meteorological Organization. 1983. Guide to Meteorological Instruments and Methods of
Observation (Fifth Edition). WMO No. 8, Geneva, Switzerland.
Yamartino, R.J., J.S. Scire, G.R. Carmichael, and Y.S. Chang. 1992. The CALGRID mesoscale
photochemical grid model -1. Model formulation. Atmos. Environ., 26A, p. 1493-1512.
CHAPTER 9: References 9-5
-------
Science and Technical Support Work Group (STSWG) Conceptual Model, February 1997
9-6 CHAPTER 9: References
------- |