United States Atmospheric Sciences
Environmental Protection Research Laboratory
Agency Research Triangle Park NC 27711
Research and Development August 1987
&EPA PROJECT REPORT
A SENSITIVITY ANALYSIS AND PRELIMINARY
EVALUATION OF RELPAP IIWLVING
FIf-E AND COARSE PARTICULAR MATTER
-------
A SENSITIVITY ANALYSIS AND PRELIMINARY EVALUATION OF RELMAP
INVOLVING FINE AND COARSE PARTICULATE MATTER
by
Brian K. Eder
Meteorology and Assessment Division
Atmospheric Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
ATMOSPHERIC SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
The information contained within this document has been
funded by the United States Environmental Protection Agency.
It has been subjected to the Agency's peer and administrative
review and it has been approved for publication as an EPA
document. Mention of trade names or commercial products does not
constitute an endorsement or recommendation for use.
AFFILIATION
Mr. Eder is currently on assignment from the" National
Oceanic and Atmospheric Administration, United States Department
of Commerce.
-------
ABSTRACT
In response to the new, size discriminate federal standards
for Inhalable Particule Matter, the REgional Lagrangian Model of
Air Pollution (RELMAP) has been modified to include simple,
linear parameterizations which simulate the chemical and physical
processes of fine and coarse particulate matter.
Because these new, simplified parameters are only accurate
to a limited degree, they may be upgraded or replaced in the
future with more sophisticated parameters as further research is
conducted. As an initial step in this possible refinement,
RELMAP has been subjected to a sensitivity analysis in which the
effect of inducing a +/- 50% change in the three major
parameterizations (transformation rate and wet and dry deposition
rates) involving the simulation of fine and coarse particulate
matter has been examined. Simulated concentrations of fine and
coarse particulate matter proved to be most sensitive to the wet
deposition of fine and coarse particulate matter, respectively;
fine concentrations were somewhat sensitive to the transformation
rate of sulfur dioxide (S02) into sulfate (S04~), and less
sensitive to the wet deposition of SO^, and the dry deposition of
fine particulate matter and S02; and finally coarse
particulate concentrations were somewhat sensitive to the dry
deposition of coarse particulate matter.
In order to assess the model's abilities, and to determine
just how accurately these new parameters simulate the actual
physical and chemical processes of the atmosphere, RELMAP was
evaluated for the summer of 1980, using emissions data from the
NAPAP Version 5.0 emissions inventory, monitoring data from the
Inhalable Particulate Network and meteorological data from the
National Climatic Data Center. Unfortunately, several obstacles
limited the scope of this evaluation; the two most important
being the omission of open source emissions from the NAPAP
inventory, and the spatial and temporal incompatibility of the
IPN data. Given the nature of these deficiencies, it is not
surprising that RELMAP significantly underpredicted the
concentrations of fine and coarse particulate matter. The model
did, however, exhibit some skill in its simulation of the
concentrations, producing correlation coefficients of 0.53 and
0.33 for fine and coarse particulate matter, respectively.
-------
CONTENTS
Abstract iii
Figures vi
Tables ix
Acronyms and Symbols x
Acknowledgments xi
1. Introduction 1
2. Model Background 5
Transport and Diffusion 8
Transformation , 10
Dry Deposition 14
Wet Deposition 18
3. Sensitivity Analysis 22
Coarse Particulate Matter 26
Fine Particulate Matter 30
4. Preliminary Model Performance Evaluation 39
Deficiencies in the NAPAP Emissions Inventory....40
Deficiencies in the IPN Data Set 46
Model Evaluation for the Summer of 1980 54
5. Conclusions and Recommendations 60
References 67
Appendices
A. Fine Particulate Matter Concentrations 70
B. Coarse Particulate Matter Concentrations 74
-------
FIGURES
Figure Caption Pace
2.1 RELMAP's 45° x 30° longitude-latitude domain 6
2 . 2 Depiction of RELMAP parameterizations 7
2.3 Three layer vertical profile with nighttime
allocation of emissions 9
2.4 Bimodal probability distribution of particle size 11
2.5 Latitudinal variation in the composite transformation
rate of S02 to S04~ 13
2.6 Diurnal variation in the composite transformation
rate of S02 to S04~ 13
2.7 Land use categories used for dry deposition and
corresponding surface roughness lengths (z0) 15
3.1 RELMAP domain with sensitivity analysis transect 24
3.2.a Absolute sensitivity of coarse particulate matter
concentration to changes in the wet deposition
rate of coarse particulate matter 27
3.2.b Relative sensitivity of coarse particulate matter
concentration to changes in the wet deposition
rate of coarse particulate matter 27
3.3.a Absolute sensitivity of coarse particulate matter
concentration to changes in the dry deposition
rate of coarse particulate matter 28
3.3.b Relative sensitivity of coarse particulate matter
concentration to changes in the dry deposition
rate of coarse particulate matter 28
3.4.a Absolute sensitivity of fine particulate matter
concentration to changes in the wet deposition
rate of S02 31
3.4.b Relative sensitivity of fine particulate matter
concentration to changes in the wet deposition
rate of S02 31
vi
-------
3.5.a Absolute sensitivity of fine particulate matter
concentration to changes in the wet deposition
rate of fine particulate matter 32
3.5.b Relative sensitivity of fine particulate matter
concentration to changes in the wet deposition
rate of fine particulate matter 32
3.6.a Absolute sensitivity of fine particulate matter
concentration to changes in the dry deposition
rate of S02 33
3.6.b Relative sensitivity of fine particulate matter
concentration to changes in the dry deposition
rate of S02 33
3.7.a Absolute sensitivity of fine particulate matter
concentration to changes in the dry deposition
rate of fine particulate matter 34
3.7.b Relative sensitivity of fine particulate matter
concentration to changes in the dry deposition
rate of fine particulate 34
3.8.a Absolute sensitivity of fine particulate matter
concentration to changes in the transformation
rate of S02 to S04 = 35
3.8.b Relative sensitivity of fine particulate matter
concentration to changes in the transformation
rate of S02 to
4.1 Area and point source emissions of TSP emitted
within the RELMAP domain 41
4.2.a Point source fractionalization of TSP emitted
within the RELMAP domain 43
4.2.b Area source fractionalization of TSP emitted
within the RELMAP domain 43
4.3 Inhalable particulate network sites used in the
preliminary model evaluation 43
4.4 Temporal depiction of the observed and simulated
fine particulate matter concentrations for
Hartford, Conn., for the summer of 1980 52
4.5 Temporal depiction of the observed and simulated
coarse particulate matter concentrations for
Hartford, Conn., for the summer of 1980 53
vn
-------
4.6 Scatter diagram of the observed vs. simulated
fine particulate matter concentrations for the
summer of 1980 57
4.7 Scatter diagram of the observed vs. simulated
coarse particulate matter concentrations for the
summer of 1980 57
4.8 Standardized residuals ((0-P)/0) of the fine
particulate matter concentrations for the
summer of 1980 59
4.9 Standardized residuals ((0-P)/0) of the coarse
particulate matter concentrations for the
summer of 1980 59
vm
-------
TABLES
Table Title Page
2.1 Typical, seasonal wet deposition rates for S02,
S04~, fine and coarse particulate matter
for a constant precipitation rate (5mm/h) 21
4.1 Estimates of omitted open source emissions of TSP
for states located within the RELMAP domain 45
4.2 Inhalable particulate network sites used in the
preliminary model evaluation 49
4.3 Statistical evaluation involving fine
particulate matter concentrations 55
4.4 Statistical evaluation involving coarse
particulate matter concentrations 56
IX
-------
Acronyms
LIST OF ACRONYMS AND SYMBOLS
EMSL
ENAMAP
EURMAP
IP
NAPAP
NAAQS
OAQPS
RELMAP
SCC
SSI
TSP
Environmental Monitoring Systems Laboratory
Eastern North American Model of Air Pollution
European Model of Air Pollution
Inhalable Particulate
National Acid Precipitation Assessment Program
National Ambient Air Quality Standards
Office of Air Quality Planning and Standards
REgional Lagrangian Model of Air Pollution
Source Classification Code
Size Selective Inlet
Total Suspended Particulate
Symbols
a
b
CO
HC1
HF
k
Kd
Kt
L
M
NH
NO
S
SO
so
2=
voc
Yc
z
Z«
empirical factor used to calculate wet deposition
empirical exponent used to calculate wet deposition
carbon monoxide
hydrogen chloride
hydrogen fluoride
von Karman constant (0.4)
dry deposition rates
transformation rates
wet deposition rates
Monin-Obukhov length (cm)
mass
ammonia
nitrogen oxides
lead
rainfall rate used to calculate wet deposition (mm/h)
surface resistance to deposition (1.0 s/cm)
Stability category
Sulfur dioxide
Sulfate
friction velocity (cm/s)
dry deposition velocity (cm/s)
volatile organic carbon
stability factor
height above surface (cm)
surface roughness scaling length (cm)
-------
ACKNOWLEDGMENTS
The author would like to express his appreciation to the
many people who have assisted in the preparation of this report.
Specifically Dale H. Coventry of the Data Management Branch, ASRL
and Terry L. Clark of the Atmospheric Modeling Branch, ASRL, who
helped in the development of this publication and also to Thomas
E. Pierce and C. Bruce Baker, of the Environmental Operations
Branch, ASRL whose comments and recommendations proved
invaluable.
-------
SECTION 1
INTRODUCTION
The primary National Ambient Air Quality Standard (NAAQS)
for particulate matter was established in 1970 with the enactment
of the Clean Air Act. The values of the standard were based upon
state-of-the-art information concerning the health effects of
ambient concentrations of Total Suspended Particulate (TSP)
matter and other environmental factors. In 1977, the Clean Air
Requirement Act called for a reappraisal of this NAAQS. One
reason for this reappraisal was a shift in emphasis from TSP,
which ranged in size from 0.0 to 50.0 urn, to smaller, size
discriminate Inhalable Particulate (IP) matter, which ranged in
size from 0.0 to 15.0 urn. The IP was comprised of fine
particulate matter (FINE-10), which included particles less than
2.5 urn in diameter, and coarse particulate matter (COARSE-15),
which initially included particles greater than or equal to 2.5,
but less than or equal to 15.0 urn.
Emphasis was placed on the smaller size particles for two
reasons. First, additional research into the health effects of
particulate matter revealed that the smaller particles were not
only able to penetrate deeper into the respiratory system, but
their expulsion rate was also lower than the larger particles.
Secondly, naturally occurring dust particles with diameters
-------
greater than 10.0 - 15.0 um often made up a large percentage of
the TSP mass collected by high volume samplers. This large
natural contribution, when combined with anthropogenic sources,
often resulted in TSP concentrations exceeding either the 75.0
ug/m3 annual geometric mean, or the 260.0 ug/m3 daily mean
(Hinton et al., 1984).
In 1981, after reviewing EPA's Clean Air Science Advisory
Committee's recommendation and the concurrent International
Standards Organization Task Group recommendations, the Office of
Air Quality Planning and Standards (OAQPS) decided that the
revised standard for ambient air concentrations of IP should be
based upon a 10 um rather than a 15 um criteria (Hinton et al.,
1984). Therefore, COARSE-15 was replaced by COARSE-10, which
included particles greater than or equal to 2.5 um, but less than
or egual to 10.0 um. The proposed new standard would allow
ambient air concentrations of IP to reach an annual arithmetic
average between 50 and 65 ug/m3 and a daily maximum between
150.0 and 250.0 ug/m3 (Federal Register, 1984).
As a result of the revised NAAQS standards for ambient air
concentrations of primary particulate matter, OAQPS has expressed
the need for size discriminate particulate models in order to
assist in regulatory planning. Shifting the emphasis onto the
smaller particles increases the importance of regional scale
models. Much of the mass of the smaller particles results from
gas to aerosol conversion which is a slow process that occurs
over regional spatial scales as opposed to urban scales.
Therefore, in response to the promulgation of the new size
discriminate federal standards for IP, the REgional Lagrangian
-------
Model of Air Pollution (RELMAP) has been modified to include
simple, linear parameterizations which simulate the chemical and
physical processes of FINE-10 (including the conversion of S02 to
S04=) and COARSE-10. Because the contribution of nitrogen
chemistry in the formation of particulate matter is thought to be
negligible at this time, it is ignored by the model. It is
important to remember that this modified version of RELMAP
represents an initial step in the linear, regional modeling of
particulate matter and therefore must be looked upon as an
interim model.
The origin of RELMAP dates back to the mid-1970's, when SRI
International developed a Lagrangian puff air pollution model
called the European Model of Air Pollution (EURMAP) for the
Federal Environment Office of the Federal Republic of Germany
(Johnson et al., (1978). This original version of the model only
simulated monthly concentrations and wet and dry depositions of
S02 and SO4=, for thirteen countries in central and western
Europe. During the late 1970's, the U.S. EPA sponsored SRI
International to modify EURMAP so that it could be applied to
eastern North America. This modified version of the model, now
called the Eastern North American Model of Air Pollution (ENAMAP)
was also capable of simulating monthly concentrations as well as
wet and dry depositions of S02 and S04= (Bhumralkar et al., 1980;
Johnson, 1983).
During the early and middle 1980's, EPA continued to modify
and improve the model to increase its flexibility and its
scientific credibility. Now, at the request of OAQPS, simple
-------
parameterizations involving fine and coarse particulate matter
have been incorporated into the model.
This report examines the incorporation of these new
parameterizations, first through an abbreviated discussion of
their theoretical aspects as presented in Chapter 2. For a more
in-depth discussion of these parameterizations, the reader is
referred to the RELMAP User's Guide (Eder et al., 1986). In
order to determine the sensitivity of the model to these
parameterizations, a sensitivity analysis of RELMAP has been
provided in Chapter 3. Chapter 4 provides a preliminary model
performance evaluation to help assess the model's abilities and
to determine how accurately the parameters simulate the actual
physical and chemical processes of the atmosphere. Conclusions
and recommendations for future work are provided in Chapter 5.
-------
SECTION 2
MODEL BACKGROUND
RELMAP is a mass-conserving, regional-scale Lagrangian model
that performs simulations over 1° by 1° grid cells covering the
eastern two-thirds of the United States and southeastern Canada
as depicted in Figure 2.1. The north-south and east-west
boundaries of the model's domain extend from 25° N to 55° N
latitude and from 60° W to 105° W longitude. Discrete puffs of
S02, S04=, fine and coarse particulate matter are released at
twelve hour intervals from each of the 1350 grid cells that
contain sources. As illustrated in Figure 2.2, the puffs are
then subjected to linear chemical transformation and wet and dry
deposition processes as they are transported across the model's
domain. The suspended mass and deposition for each puff is
apportioned into the appropriate grid cell based upon the
percentage of puff over that grid cell. The rate of change in
the pollutant mass resulting from the transformation and wet and
dry deposition process is directly proportional to the total mass
and is defined through the following linear equations:
S02: dM-L = -Mj^Kt + Kdl + K^); (2.1)
dt
-------
Figure 2.1 RELMAP's 45° x 30° longitude-latitude domain.
-------
»10*
*
'EMISS
EVER
t
CONCE
APPO
_ONTH
/
/[
ION "PI
Y 12 HO
HISSIQ
NTRAT
RTIONE
EBASI
PUFFS AOVECT WITH THE OBSERVED W
FIELD AT EACH 3 HOUR TIME STEP
»
JFFS"F
URSFF
YGRID
ION.WI
DINTC
SOFPE
I
,
ELEAS1
OMEAt
CELlJ
TAND
THE A
RCENT*
t-3
0
:H
ORYD
>PROPR
IGEOF
\*
POSITI
IATE R
PUFFO
t-6
i
}NAMC
ECEPTO
VERE;
V
\.
NO
SOj-SO/ TRANSFORMATION RATE
DEPENDENT UPON SOLAR INSOLATION
t>9
II
UNTSi
RCELL
tCHCEl
\
&RF-
s
L
OIF
SL t"12 8
s,
HO
_J
V(
(
N
FUSIOA
E LINE
RIZONT
.MIXIN
RTICA
MIXING
URING
OMIXIA
ASSUN
»fi INT
AL. UN
BINTH
..UPTC
HEICh
THE D
C ATd
ED TO
HE
FORM
E
THE
T
»Y.
IGHT
Figure 2.2 Depiction of RELMAP parameterizations.
S04~:
dt
Fine Particulate: dM3 = -M3(Kd3 +
dt
Coarse Particulate: dM,,
dt
= -M4(Kd4
K
d2
(2.3)
(2.4)
where M^ is the mass of the respective pollutants (expressed in
ktons), t is the time (h) , and Kt is the transformation rate of
is the dry deposition rate and KW^ is the wet
SO2 into SO4
-------
deposition rate, of the respective pollutants. The 3/2 factor
used in Equation 2.2 represents the ratio of molecular weights
between SO^" and SC^.
Transport and Diffusion
Dispersion generated by small-scale turbulence is not nearly
as significant as long term transport and deposition processes
for regional-scale models such as RELMAP. Because of this, the
model simulates both horizontal and vertical diffusion through
simple parameterizations. RELMAP divides the atmospheric
boundary layer into three layers as seen in Figure 2.3. The
first layer is between the surface and 200 m, while the second
layer is between 200 and 700 m. The depth of the third layer is
variable, depending upon the seasonal-mean maximum mixing
height, and is assumed to be 1150 m during the winter, 1300 m
during the spring and fall, and 1450 m during the summer (Endlich
et al., 1983).
During the unstable regimes of midday periods, pollutants
from both area and point sources become well mixed up to the
mixing height long before they are transported a distance equal
to the spatial resolution of the grid. Therefore, it is assumed
that instantaneous and complete mixing occurs within the three
layers of the model during the unstable daylight hours. However,
after sunset, when mixing is prohibited by stable conditions,
point and area source emissions are confined to the separate
-------
1450
1300
1150
E
i?oo
200
100
-SUMMER-
-SPRING/AUTUMN
--WINTER-
LAYERS
cc
<
- POINT SOURCES
AREA SOURCES-
Figure 2.3 Three layer vertical profile with nighttime
allocation of emissions.
layers into which they are emitted. As again illustrated in
Figure 2.3, all area source emissions remain in Layer 1, within
200 m of the surface, while emissions from point sources are
allocated into Layer 2, accounting for typical plume rise, which
averages several hundred meters (Briggs, 1984).
RELMAP assumes that horizontal diffusion of the puffs occurs
9
-------
at a constant rate so that the size of the puff increases at a
rate of 339 km2/h, and that the distribution of the mass of
pollutant in the puff remains homogeneous at all times. The puff
expansion rate is based upon work by Pack et al. (1978), who
performed calculations on long range trajectories. Each of the
puffs is transported using vertically weighted and horizontally
and temporally interpolated wind fields until the puff is either
transported off the grid, or the amount of mass in the puff falls
below a predetermined minimum value. The puff remains an
indivisible entity. Vertical shear of the puff is not directly
considered as the mass of pollutant in each of the three layers
is transported in the same direction and at the same speed. The
transport velocity of the puffs is determined by integrating the
mass-weighted u and v components of the three layers, which are
derived from the wind velocities from the grid cell containing
the centroid of the puff. Surface winds are used in the lowest
layer, 850 mb winds are used in the third layer, and a weighted
average of (0.2) surface and (0.8) 850 mb wind velocities are
used for the second layer.
Transformation
RELMAP treats fine and coarse particulate matter as
independent non-evolving pollutants, that is physical and/or
chemical transformation of fine particulate matter to coarse
particulate matter are considered negligible. This premise is
10
-------
supported by particle size distributions obtained from monitoring
data (Suggs et al., 1981). As seen in Figure 2.4 the size
distribution of particulate matter generally indicates a bimodal
distribution with peaks in the fine and coarse particle size
ranges and a deep oscillating gap between 1 and 5 um.
-: 1.0
H
CO
5 0.8
Q
3 0.6
ffl c
O 30.4
CC >
a_ ^
Si.
-------
homogeneous component simulates transformation that occurs under
dry conditions.
Seasonal variation in the transformation rates were examined
by Meagher et al., (1983) who determined that average morning
rates ranged from a low of 0.15%/h during the winter to a high
of 1.30%/h during the summer. In a similar study, Altshuller
(1979) concluded that noontime winter transformation rates
averaged about five times less than noontime summer
transformation rates.
Diurnal fluctuations were found to be much larger during the
summer months than during the winter months. Husar et al.,
(1978) found summer diurnal transformation rates that ranged from
a minimum of 0.5%/h during the night to a maximum of between 2.0
to 8.0%/h at solar noon. Conversely, Meagher and Olszyna (1985)
could only detect slight diurnal variation in the transformation
rates during the winter months.
More recent field studies have indicated that in-cloud
processes are also very important in the transformation of
pollutants. The rate of transformation can be increased by an
order of magnitude in saturated conditions, depending upon the
cloud height, precipitation efficiency and mass of S02 in the
mixed layer (Isaac et al., 1983). Based upon a theoretical
algorithm developed by Scott (1982), the magnitude of this
heterogeneous component was set to 7.0%/h during the winter,
11.0%/h during the spring and fall, and 15.0%/h during the
summer.
Figures 2.5 and 2.6 illustrate the relationship between the
composite transformation rate (heterogeneous and homogeneous
12
-------
I
o
<
1'
u.
v>
1 -
: i r i i i i i i i r
SOLAR NOON
JULY
APRIL
OCTOBER
JANUARY
Figure 2.5.
20 25 30 35 40 45 50 55 60
NORTH LATITUDE, degrees
Latitudinal variation in the composite transformation rate
of S02 to S04=.
A ^Tii i i i i i i i r ^ r ,
40 DEGREES NORTH LATITUDE
JULY
O
Ik
t/i
JANUARY
Otr- :
0000 0400 0800 1200 1600 2000 2400
TIME(LST).hr
Figure 2.6. Diurnal variation in the composite transformation rate of
S02 to S04=.
13
-------
components) and the time of day, season and the latitude as
calculated in RELMAP.
From Figure 2.5 it is evident that at solar noon, the
composite transformation rate incorporated into the model is
highest during July (approximately 4.0%/h at 25° N and 3.0%/h at
55° N) and lowest during January (1.2% at 25° N and 0.6%/h at
55°). Figure 2.6 illustrates that the diurnal variation
exhibited by the composite transformation rate at 40° N is also
greater during July (0.9%/h at midnight LST and 3.4%/h at solar
noon) than during January (0.7%/h and 0.8%/h).
Dry Deposition
Dry deposition of SO2, S04=, fine and coarse particulate
matter is a highly variable, complex process that is
parameterized in RELMAP as a function of land use, season, and
stability index. Twelve land use categories, categorized by
surface characteristics and vegetation type (Sheih etal.,
1979), were gridded to RELMAP's 1° by 1° domain. Figure 2.7
illustrates the grid of homogeneous land use types and provides a
Table listing their corresponding surface roughness scale lengths
-
Dry deposition velocities (v^), which represent the downward
surface flux divided by the local concentration, were calculated
as a function of land use type, stability class and season for
S02, S04=, and fine and coarse particulate matter. The stability
classes used to determine the dry deposition velocities are the
14
-------
LAND USE TYPES USED FOR DRY DEPOSITIONS
55555
55555
55555
55555
55555
55555
55555
55555
555
5555
5 5 5 5
5555
55555
55555
5555
5555
55555
2|5 512120.212121212
5 5f 312121212
2 nj 212 2ff2T2i
2 2 2p21212[..22121212
2 2^212121212121212121212
2 2 512121212121212121212
IL212121212I1212121212
5D.21212121212121212121212121
12121212 L212121212a21212121
2 L212121212L212121212L212121212
201212121212 L2121212121212121212
1121212121212121212121212121212
444
22422
22422
22492
44194
2 2 2
2 2 2
2222*
24422
244
212121212120.212121212121212121
2121212 L212121212JL212121212J121212121
21212121212121212123.212121212121212121
2121212121212121212JL212121212J1212121212
21212121212121212121212121212J1212121212
20.212121212 L212121212J12121212123.212121212
2ict21212121212121212120.212121212a212121212
4t212121212!l212121212!l212121212|l212121212
12121212121121212121211212121212
1212121212h.2121212121212121212
212121212R2
212121212
212121212
2121212]
212L212121212
212L212121212
-------
six Pasquill-Gifford categories: (A) very unstable, (B)
moderately unstable, (C) slightly unstable, (D) near neutral, (E)
moderately stable, and (F) very stable (Gifford, 1976). The dry
deposition velocities, measured in centimeters/second, are used
in the model to determine dry deposition rates.
Determination of the dry deposition velocities for S02, S04=
and fine particulate matter were based upon the work of Sheih et
al., (1979) and are discussed below. Coarse particulate matter
dry deposition velocities, which are based upon the work of
Sehmel (1980) and therefore parameterized somewhat differently,
are also presented below.
The algorithm developed by Sheih et al., (1979) was
modified to calculate dry deposition velocities. The
parameterization used in RELMAP for the deposition of S02, S04=,
and fine particulate matter is given by the following:
Vd = ku* (In (z/zo) + ku*rp - Y^'1 (2.5)
where k is the von Karman constant (0.4), u* is the friction
velocity (cm/s), ZQ is the surface roughness scale length (cm)
derived from the twelve land use categories, rp is the surface
resistance to particle deposition ( 1.0 s/cm) and Yc is a
stability factor. Details on the formulation of Equation 2.5 can
be found in Sheih et al. (1979).
More recent studies (Wesely and Shannon, 1984), which are
based upon micrometeorological field experiments, have determined
that dry deposition of S04= calculated by Equation 2.5 was too
16
-------
high by a factor of two. To alleviate this overestimation, the
dry deposition velocities for S04= and fine particulate matter as
calculated from Equation 2.5 were reduced by half. Typical dry
deposition velocities resulting from the calculation range
between 0.05 and 1.15 cm/s for S02, and between 0.05 and 0.50
cm/s for S04~ and fine particulate matter, depending upon the
season, the stability and the land use category.
When considering diurnal variations, use of the equations
derived above is not always recommended. In order to compensate
for the high nocturnal atmospheric resistance, when plant
absorption is minimal, the model assumes that dry deposition
velocities are reduced to 0.07 cm/s for S02, S04= and fine
particulate matter, as recommended by Sheih et al. (1979).
Dry deposition velocities of coarse particulate matter are
parameterized through a very similar approach in order to
maintain consistency within the structure of the model. Using
the same land use categories as described earlier, the model
incorporates the work of Sehmel (1980) who presented plots of dry
deposition velocities of particulate matter as a function u*, zo,
particle density, and diameter. The following equation was used
to determine values of u*, which is a function of stability, wind
speed and ZQ:
u* = ku(ln z/zo) - Yjjj)'1 (2.6)
The stability function, Ym, was determined by using the
appropriate relationships between the Monin-Obukhov length (L),
surface wind speed (u) and stability class, as suggested by
17
-------
Sheih et al. (1979). Determination of u* allows the selection of
an appropriate Sehmel diagram, from which the dry deposition
velocity can be obtained for a given ZQ. Based upon the work of
Mamane and Noll (1985), who analyzed rural particulate matter
characteristics, a constant particle density of 4.0 g/cm was
used in the equation. Unfortunately, Sehmel's study was limited
to surface roughnesses less than 10 cm, while most of the land
use categories used in the model had surface roughnesses greater
than 10 cm, therefore it was often necessary to extrapolate the
appropriate dry deposition velocity.
Unlike S02, S04=, and fine particulate matter, the dry
deposition velocities of coarse particulate matter are much less
dependent upon the time of day and the season, therefore, diurnal
and seasonal variations are considered by the model to be
negligible.
Wet Deposition
The complex process of wet deposition of S02, SO4=, and fine
and coarse particulate matter is thought to be a function of
cloud chemistry, cloud type, pollutant concentration and
precipitation type and rate. RELMAP, however, parameterizes this
process quite simply, treating it only as a function of
precipitation rate and cloud type. The wet deposition rates are
based upon the work of Scott (1978), who presented graphs of
washout ratios between SO4= concentration in precipitation and
S04= concentrations in the air. These ratios are solely
18
-------
dependent upon the precipitation rate and the cloud type, where
the three cloud types considered are Bergeron or cold-type
clouds, maritime or warm-type clouds, and convective-type clouds.
The model assumes that all winter precipitation results from the
Bergeron process, that spring and summer precipitation result
from the convective-type clouds, and that autumn precipitation is
confined to warm-type clouds. The algorithm derived from Scott's
work, which was expanded by SRI (1982) to include S02, is
presented below:
Wet Deposition Rate = a R ; (2.7)
where a and b are seasonal empirical constants derived from the
inherent relationship between the washout ratio and the
precipitation rate R. Because so little is known about the wet
removal processes of nonsulfate aerosols from the atmosphere, the
model currently assumes identical deposition rates for S04=, and
fine and coarse particulate matter. This simplistic approach to
the wet removal processes of nonsulfate particulate matter will
be replaced in the future with more sophisticated
parameterizations as the physics of this process become better
understood.
As a participant in the International Sulfur Deposition
Model Evaluation (ISDME), RELMAP was found to significantly
overpredict SO4= wet deposition amounts during the convective
seasons of spring and summer using the algorithm discussed above
(Clark et al., 1987). Predictions of SO4= wet deposition during
the non-convective months of winter and autumn were, however,
19
-------
more in line with the observed. Further analysis has shown that,
because wet deposition is such an efficient sink of particulate
matter, the length of the simulation period is very critical.
This is especially true during the convective months when
precipitation rates can be very high for short time intervals.
Therefore, in an effort to better simulate the convective type
precipitation event, the time step used to calculate wet
deposition amounts during the spring and summer were reduced from
the nominal 3 h to 1.5 h. Increasing the temporal resolution,
which has produced more favorable results, decreases the amount
of wet deposition occurring for a given amount of precipitation.
Presented below in Table 2.1 are typical wet deposition
rates calculated for a given precipitation amount (5mm/h) and for
each season using the 1.5 h time interval for spring and summer
and the 3 h time interval for winter and autumn.
20
-------
Table 2.1 Typical, Seasonal Wet Deposition Rates for SO2, S04=
Fine and Coarse Particulate Matter for a Constant
Precipitation Rate of 5mm/h.
Pollutant
Season Empirical Constant
a b
Wet Deposition Rate
(% / Time step*)
so2
S04=,Fine
& Coarse PM
Winter
Spring
Summer
Autumn
Winter
Spring
Summer
Autumn
0.009
0.036
0.140
0.036
0.021
0.091
0.390
0.091
0.70
0.53
0.12
0.53
0.70
0.27
0.06
0.27
8.10
17.72
26.36
23.26
18.20
24.31
58.96
36.51
Time Step = 1.5 h during spring and summer, 3.0 h during autumn
and winter
21
-------
SECTION
SENSITIVITY ANALYSIS
The simplified parameterizations, which were recently
incorporated into the model, are designed to simulate the complex
meteorological and chemical process involving fine and coarse
particulate matter. Because of their simplicity, they may be
upgraded or even replaced in the future with more sophisticated
parameterizations as further research is undertaken. As an
initial step in this possible refinement, RELMAP was subjected to
a "local" sensitivity analysis. In this analysis, variations
found in the model's output (concentrations of fine and coarse
particulate matter) due to changes in the model's
parameterizations are examined, while all the other parameters
are held fixed.
The analysis, which employed actual meteorological and
emissions data for July 1980, was performed using the currently
accepted values for all of the input parameters. The
meteorological data were obtained from the National Climatic Data
Center located in Asheville, North Carolina, and included 12
hourly surface and 850 mb wind data and hourly precipitation
data. Because Version 5.0 of NAPAP's 1980 Task Group B emission
inventory was not available at the time of this analysis, the
emissions data were obtained from the Version 4.0 inventory and
22
-------
from Canada's Environmental Protection Service emissions
inventory used in Phase III of the U.S./Canadian Memorandum of
Intent on Transboundary Air Pollution (U.S./Canadian Memorandum
of Intent, 1982).
The parameterizations examined in this sensitivity analysis
included: the transformation rate of S02 into S04~, the wet and
dry deposition rates of S02, fine (including S04=) and coarse
particulate matter. S02 parameterizations are included in this
analysis because it is a precursor to SO4~ and therefore to fine
particulate matter. With each simulation, the values of the
respective parameterizations were allowed to vary +/~ 50% around
their currently accepted or nominal values. A single value of
50% was selected for two reasons. First, choosing a single value
would maintain consistency between and allow intercomparisons of
each of the sensitivity tests. Secondly, the value of 50% was
found to best represent the approximate lower and upper limits of
the realistic changes found in all of the parameters. As a
result, one would expect that the subsequent changes found in the
simulations of fine and coarse particulate matter concentrations
would correctly represent the range in which the actual
concentrations would vary, given that the exact physics and
chemistry had been incorporated into the model.
Results of the sensitivity analysis were recorded along a
specific transect that stretched across the model's domain from
Alabama to Quebec as seen in Figure 3.1. The fifteen grid cells
that comprise the transect were chosen because they provide a
good representation of the actual range of concentration values
found for both fine and coarse particulate matter in North
23
-------
LONG
Figure 3.1 RELMAP domain with sensitivity analysis transect
-------
America. Results from the seven tests are presented graphically
in Figures 3.2 through 3.8. Figures 3.2 and 3.3 illustrate the
sensitivity of coarse particulate matter concentration to changes
in the wet and dry deposition rates of coarse particulate matter,
respectively. Figures 3.4 through 3.8 depict the sensitivity of
fine particulate matter concentrations to changes in the
transformation rate, and the wet and dry deposition rates of
S02, and fine particulate matter. Although the model treats S04=
and fine particulate matter as mutually exclusive pollutants, the
two are combined as one in this graphical analysis and simply
referred to as fine particulate matter, unless otherwise noted.
Each analysis consists of two graphs. The first graph
depicts a transect of the concentration field illustrating the
absolute changes that occur when a parameter is allowed to vary
by +/- 50% around its nominal or base case value. The second
graph of each analysis illustrates the relative, with respect to
the base case, changes that occur along the transect. The
abscissa for each of the plots represent the fifteen grid cells
that form the transect from A to A1. It should be noted, that the
scale of the ordinate, which represents either the actual
concentration (expressed in ug/m ), or the relative concentration
(percent of the base case) can vary significantly from plot to
plot, depending upon the specific parameter and the model's
sensitivity to that parameter.
In each graph, the asterisk represents the concentration of
fine and coarse particulate matter that results when the
parameter being tested is reduced to 50% of its nominal value.
The diamond represents the base case, where the parameter is left
25
-------
at its nominal value, and the circle represents 150% of the
parameter's nominal value. Caution should be exercised when
examining the relative graphs at grid cells fourteen and fifteen.
Concentrations at these two grids cells, which are over Ontario
and Quebec, are so small that even minute changes in the
magnitude of the concentrations result in exaggerated relative
differences with respect to the base case.
Coarse Particulate Matter Concentrations
Examination of Figures 3.2-3.3 provides insight into the
sensitivity of coarse particulate matter concentrations to
changes in the wet and dry deposition of coarse particulate
matter. First, one should note the location of two maxima that
appear in base case concentration field (depicted by the diamond
transect) of the absolute graphs. The first is located in grid
cell 3 over northern Georgia and has a concentration of nearly
1.4 ug/m3. The second maximum,which is located in grid cell 9
over western Pennsylvania, is the largest and has a concentration
of 1.9 ug/m3. A sharp gradient in the concentration field occurs
after this maxima as values fall off quickly to less than 1.0
ug/m3 as the transect enters Canada. Secondly, examination of
the figures also reveals that increasing either the wet or dry
deposition of the coarse particulate matter results, as expected,
in a decrease in the concentration, and that this decrease is
more pronounced in the case of wet deposition. Likewise,
26
-------
to
-J
8
WET DEPOSITION OF COARSE PARTICULATE MATTER
190
175
160
89101112131415
1 2
(a)
10
>OOOOOOOOOOOOO<]
2 3 45 6 7 8 0 101112131415
A A"
(b)
Figure 3.2
Absolute (a) and relative (b) sensitivity of coarse particulate matter
concentration to changes in the v/et deposition rate of coarse particulate
matter. (Asterisk - 50% normal wet deposition, Diamond - 100% normal wet
deposition, Circle - 150% normal wet deposition.)
-------
DRY DEPOSITION OF COARSE PARTICIPATE MATTER
to
00
1 23456789101112131415
m
u.1001
97
91
88
85
82
>00000000000
94(^0-0-*
A A'
(a)
i i i i i i r i r i i i i i
1 23456789101112131415
A AB
(b)
Figure 3.3 Absolute (a) and relative (b) sensitivity of coarse particulate matter
concentration to changes in the dry deposition rate of coarse particulate
matter.
-------
decreasing either the wet or dry deposition of coarse
particulate matter results in increased concentrations.
It is interesting to note that the changes in the magnitudes
of the concentration patterns proved to be non-linear. That is,
the changes in the model output are not directly proportional to
changes in the input, and in most cases are far less than 50%.
This non-linearity can in part, be attributed to the ability of
the parameters to compensate for a given increase or decrease in
a specific parameter. This compensation will to some degree
reduce the response of the model to the forced variation.
This non-linearity is also evident in that the difference
between the base case and the low deposition rates (both wet and
dry) is considerably larger than that between the base case and
the high deposition rates. This is due to the different
"efficiencies" exhibited by each of the parameterizations. Each
parameter has a maximum rate, which once exceeded, will produce
no further changes in the model's simulations. This threshold
value is more easily attained for the more efficient of the
parameterizations, which include all of the wet deposition rates,
the dry deposition of coarse particulate matter, and to a lesser
degree the transformation rate of S02 into SO4=. Therefore,
increasing these efficient parameters by 50% will often result in
this threshold rate being exceeded, thereby limiting the impact
on the concentration.
This phenomenon is well illustrated in the graphs showing
the relative changes in the concentration field. Examination of
the wet deposition graph Fig 3.2.b, shows that for a 50% decrease
in the wet deposition, the concentration increases an average of
29
-------
30 to 50%, but that for a 50% increase in the wet deposition, the
concentration only decreases an average of 15 to 25%. Similar
trends are evident, but to a lesser degree, with the dry
deposition graph as seen in Fig 3.3.b. For a 50% decrease in dry
deposition, the concentration increases an average of 5 to 10%,
but for a 50% increase in dry deposition, the concentration only
decreases between 3 and 6%.
Another interesting feature of the graphs, which is evident
throughout all of the analysis, is that the basic spatial pattern
of the concentration appears to remain the same. That is, the
location of the relative maxima and minima remain the same and do
not shift up or down the transect.
Fine Particulate Matter Concentrations
Examination of Figures 3.4-3.8, which depict the sensitivity
of fine particulate matter concentration to changes in the wet
and dry deposition of S02 and fine particulate matter, as well as
to changes in the transformation rate of SC>2 into S04 = , reveals
many of the same characteristics as noted with the coarse
particulate matter. First of all, two local maxima are again
evident in the concentration field. The first is located in the
second grid cell which falls over northwestern Alabama and has a
value of 3.0 ug/m3. The second and largest maxima, which has a
value of 4.3 ug/m3, is located in grid cell number nine, which is
over western Pennsylvania. As was seen in the concentration
field of the coarse particulate matter, the concentration of fine
30
-------
U)
WET DEPOSITION OF S02
10 H 12 13 14 15
0 0 0 O O
Q Q e-e
T
1 2 3
56
0
7 8
A^^^^ «
n
(b)
Figure 3.4 Absolute (a) and relative (b) sensitivity of fine particulate matter
concentration to changes in the wet deposition rate of SO0.
I 1 1 1 1
10 11 12 13 14 15
-------
WET DEPOSITION OF FINE PARTICIPATE MATTER
N)
1 23456789101112131415
ooooooooooooo<
1 23456789101112131415
A A'
(a)
A A'
(b)
Figure 3.5 Absolute (a) and relative (b) sensitivity of fine particulate matter
concentration to changes in the wet deposition rate of fine particulate
matter.
-------
U)
DRY DEPOSITION OF S02
56789101112131415
23456789101112131415
123
A A'
(a)
A A'
(b)
Figure 3.6 Absolute (a) and relative (b) sensitivity of fine particulate matter
concentration to changes in the dry deposition rate of SO*
-------
DRY DEPOSITION OF FINE PARTICULATE MATTER
6
U)
1 23456789101112131415
$
<102
CD
u.1001
t 98
8 96(
l$94
92
90
88
>00000000000
A A'
(a)
I I I I I I I I I I I
1 23456789101112
A A'
(b)
13 14 15
Figure 3.7 Absolute (a) and relative (b) sensitivity of fine particulate matter
concentration to changes in the dry deposition rate of fine particulate
matter.
-------
TRANSFORMATION RATE OF S02 INTO S04
U)
>0000000000000<
456789101112131415
1 23456780101112131415
A A*
(a)
A A"
(b)
Figure 3.8 Absolute (a) and relative (b) sensitivity of fine particulate matter
concentration to changes in the transformation rate of SO- into SO.=,
-------
particulate matter falls off rapidly as the transect enters
southeastern Canada.
The sensitivity of fine particulate matter concentration to
wet deposition of S02 (which is a precursor to S04=) and to fine
particulate matter each exhibit non-linear behavior. The
influence of S02 wet deposition, however, proves to be minimal as
seen in Figures 3.4 a and b. Examination of the relative graph
shows, that with the exception of the last two grid cells, at
most a 3% change in the concentration field occurs given a 50%
change in the SO2 wet deposition field.
As expected, the wet deposition of fine particulate matter
had a much larger impact upon the concentration field as seen in
Figures 3.5 a and b. For a given 50% increase in the wet
deposition of fine particulate matter, the concentration
decreased an average of 15 to 30%, whereas a 50% decrease in the
wet deposition resulted in a 30 to 50% increase in the
concentration.
Examination of the dry deposition graphs for both S02 and
fine particulate matter, as seen in Figures 3.6 - 3.7, reveals
that the sensitivity of the fine particulate concentration to
these less "efficient" parameterizations proved to be linear.
That is, the difference found between the base case and the low
dry deposition rates is equivalent to the difference found
between the base case and the high deposition rates. This
linearity is evidenced through the "mirror image" effect seen in
the relative graphs about the base case line. The influence of
SO2 dry deposition on the concentration of fine particulate
36
-------
matter proved to be all but non-existent as seen in Figure 3.6.
A +/- 50% change in the S02 dry deposition resulted in at most a
+/- 1% change in the concentration field. The impact of fine
particulate dry deposition on the fine concentration field,
though small, is more noticeable as seen in Figure 3.7.
Inducing a +/-50% change in the dry deposition of fine
particulate matter resulted in a 3 to 6% change in the
concentration field.
And finally, as seen in Figures 3.8, the sensitivity of fine
particulate matter concentration to changes in the transformation
rate is both non-linear and rather significant. A 50% increase
in the transformation rate increases the concentration by an
average of 5 to 10%, while a 50% reduction in the transformation
rate results in a 6 to 12% decrease in the concentration.
The results of separately introducing +/~ 50% changes into
three major parameterizations (wet deposition, dry deposition,
and transformation rate) involving the simulation of fine and
coarse particulate matter concentrations has been examined.
Although the results are preliminary, several important
conclusions can be drawn:
Simulated concentrations of fine and coarse particulate
matter are by far most sensitive to variations in the wet
deposition rates of fine and coarse particulate matter,
respectively. (For a given +/- 50% change in the wet
deposition, a +/- 15 to 45% change occurs in the
concentration). However, concentrations of fine
particulate matter proved to be quite insensitive to wet
deposition of SG>2 (+/- 1 to 4%).
37
-------
Concentrations of coarse particulate matter are somewhat
sensitive to dry deposition of coarse particulate
matter (+/- 5 to 10%). Concentrations of fine particulate
matter are, however, less sensitive to dry deposition of
fine particulate matter (+/- 2 to 6%), and are in fact
highly insensitive to dry deposition of S02 (+/- 1%).
Concentrations of fine particulate matter proved to be
somewhat sensitive to the transformation rate of the
precursor S02 into S04~ (+/~ 5 to 1°%)
38
-------
SECTION
PRELIMINARY MODEL PERFORMANCE EVALUATION
In order to perform an adequate model performance
evaluation, three major components are necessary. First, a
complete and detailed meteorological input data set that
accurately simulates the atmospheric process that are pertinent
to the model simulations is necessary. Second, a comprehensive
emissions input data set which emulates both the anthropogenic as
well as the natural emissions found in the model's domain is
needed. The third, and perhaps most important component, is a
complete evaluation data set that can be used to validate each of
the output parameters simulated by the model over compatible
spatial and temporal scales. Unfortunately, for reasons that
will be discussed later in this section, only the input
meteorological data can be deemed adequate at this time.
Inadequacies inherent to both the emissions input data set and
the model evaluation data set limit the scope of this evaluation,
therefore it must be considered preliminary at this time.
RELMAP was run for the three month period of July, August
and September, 1980 in order to simulate a summer season using
meteorological data obtained from the National Climatic Data
Center located in Asheville, N. C. Included in the meteorological
data are gridded 12-hourly surface and 850 mb wind data and
39
-------
hourly precipitation data. Gridded input emissions data were
obtained from the National Acid Precipitation Assessment Program
(NAPAP) Version 5.0 Emission Inventory. Simulated ambient air
concentrations of fine and coarse particulate matter were then
compared on a monthly and seasonal basis with monitoring data
obtained from the Inhalable Particulate Network (IPN) data set.
This section provides a brief overview of both the NAPAP Version
5.0 Emissions Inventory and the IPN data set and discusses the
many inadequacies encountered when trying to incorporate them in
this model evaluation.
Deficiencies in the NAPAP Version 5.0 Emissions Inventory
Version 5.0 of the 1980 NAPAP Emissions Inventory was
selected for use in this evaluation because it represents by far
the most comprehensive and highest quality emissions data set
available. The Task Group on Emissions and Controls of the
Interagency Task Force on Acid Rain was responsible for
developing the inventory in order to support the modeling needs
of NAPAP.
The emissions inventory contains point source emissions data
for over 14,000 plants comprised of 52,000 source classification
codes (SCC). Area source emissions are reported for 88 emission
categories for over 3,000 counties in the contiguous U. S., and
for 157 emission categories for the 10 Canadian provinces (Wagner
et al., 1986.) In addition to the S02, S04= and fine and coarse
particulate emissions of interest to this evaluation, data are
40
-------
also available for NOX, Pb, CO, HC1, HF, NH3, VOC, and total
hydrocarbons.
Unfortunately for this evaluation, the primary reason for
developing the 1980 NAPAP emissions inventory was to provide an
emissions data base for acid deposition research and modeling,
not regional particulate modeling. Because of this, less
emphasis was placed on the TSP inventory, resulting in numerous
deficiencies in both the fine and coarse particulate emissions.
The total annual emissions of TSP for the entire NAPAP grid area
was estimated to be 74,192 ktons. Of this total, 42,617 ktons or
57.4% was emitted from U. S. sources, and 31,574 ktons or 42.6%
was emitted from Canadian sources. Characteristics of the TSP
emissions found within the boundary of the RELMAP grid and
therefore used in this model evaluation are displayed in Figures
4.1 and 4.2.
TOTAL MASS - 46£60 KTONS
AREA SOURCES
42010
90.22*
POINT SOURCES
4550
9.77«
Figure 4.1 Area and point source emissions of TSP emitted within
the RELMAP domain.
41
-------
These figures illustrate the fractionalization of total TSP,
as well as fine and coarse particulate matter. As seen in
Figure 4.1, 46,560 ktons or 62.8% of the total NAPAP TSP
inventory is emitted within the model's domain. Of this total,
90.23% can be attributed to area sources, and 9.77% can be
attributed point sources. Figures 4.2 a and 4.2 b break these
percentages down even further. Of the 4,550 ktons of TSP
attributed to point sources, 14.20% are emitted as particles with
diameters larger than 10 urn, 7.56% are emitted as fine particles,
7.29% are emitted as coarse particles, and 70.95% cannot be
fractionalized. This last percentage illustrates one of the two
major deficiencies of the NAPAP TSP inventory. A large percentage
of the many point source categories designated by NAPAP do not
have particle size distributions. Because of this, more than 3
million tons of the TSP emitted from point sources can not be
fractionalized, or broken down into the respective size
categories.
Examination of the area source fractionalization reveals
that of the 42,010 ktons of TSP attributed to area sources,
28.71% are emitted as coarse particles, 27.88% are emitted as
fine particles, 42.56% are emitted as particles with diameters
larger than 10 um, and 0.85% cannot be fractionalized. In a
situation similar to that seen with the point sources,
fractionalization was only possible with 64 of the area source
categories designated by NAPAP. Over 356 ktons of TSP from the
area sources were omitted because 24 of the U. S. and 93 of the
Canadian categories could not be fractionalized. When combined
with the point source emissions, a total of 3,584 ktons or 7.70%
42
-------
PONT SOURCE FRACTONALJZAT10N
TOTAL MASS - 4£50 KTONS
NO FRACTIONS
3228
70.951
PH > 10 UB
646
U.20i
(a.)
AREA SOURCE FRACTIONALIZATTON
TOTAL MASS - 42010 KTONS
NO FRACTIONS
356
o.a;.
(b.)
Figure 4.2
Point source (a.) and area source (b.)
fractionalization of TSP emitted within the RELMAP
domain.
43
-------
of the TSP emissions found in the NAPAP inventory cannot be
fractionalized. OAQPS is currently updating their SCC inventory,
but until this is completed, these non-fractionalized emissions
cannot be used as input into the model, which will have
detrimental effects on the model's performance.
Another, even more serious deficiency found with the NAPAP
inventory is the omission of many of the "open" sources of TSP.
Open source emissions, which are defined as sources of air
pollution too great in extent to be controlled by enclosure, are
extremely difficult to estimate. Open sources of TSP include
paved and unpaved roads, agricultural tilling, wind erosion,
construction activity, forest fires (wild and prescribed) and
mining operations. Unfortunately, only roads (paved and unpaved)
and forest fires are included in the inventory. The omission of
agricultural tilling, wind erosion, construction and mining
sources from the inventory reflected the different methodologies
employed by Canada and the U. S. Originally, the Canadian
inventory of TSP (which included 157 area source emissions)
included all of the open sources omitted above; however, because
there were no counterparts in the U. S. for agricultural tilling,
wind erosion, construction and mining, these major sources were
dropped from the inventory altogether in an effort to be
consistent.
As an illustration of the magnitude of this problem,
estimates of the amount of these omitted open sources for states
within the RELMAP domain were calculated and are presented in
Table 4.1. Unfortunately, these totals, which were derived from
Evans and Cooper (1980), are based upon 1976 data and exclude
44
-------
Canadian provinces. The final, annual estimate of 255,646 ktons,
which again excludes the Canadian provinces within the model's
domain, is more than five times larger than the total annual TSP
emissions accounted for in the NAPAP inventory!
Table 4.1 Estimates of Omitted Open Source Emissions of TSP for
States Within the RELMAP Domain.
SOURCE TSP EMISSIONS
(KTons)
Agricultural Tilling 31,446
Wind Erosion 206,776
Construction 15,137
Mining 2,287
TOTAL ' 255,646
It should be noted, however, that open sources tend to emit
larger particles than most anthropogenic sources, and that open
sources tend to be located in remote areas, far removed from
population centers and TSP monitoring sites. Because of this,
Evans and Cooper estimated that a ton of open source emissions
has between 1/101-" and 1/40^" the impact as does one ton of
anthropogenic sources at a TSP monitoring site. Even if this
reduction is applied, between 6,400 and 25,560 ktons, of open
sources that are emitted within the states are being omitted from
the NAPAP inventory. Comparison with the NAPAP inventory
estimates of TSP for the U.S. alone, supports the claim that open
45
-------
sources contributions are equivalent under the most conservative
of estimates, to the anthropogenic sources, yet most are not
included in the inventory.
A third deficiency, which is as detrimental to the
evaluation as the second, is the wide discrepancy observed
between the estimates of open source emissions due to unpaved
roads. Over 70% of the total TSP emissions in the NAPAP
inventory is attributed to unpaved roads. Unfortunately, the
total estimated by NAPAP for this category is much lower than
other independent estimates. For instance, Evans and Cooper
estimate that in the U.S. alone, over 170,000 ktons of TSP
emissions are emitted from unpaved roads, which is almost an
order of magnitude higher than the NAPAP total.
Because of the number and seriousness of these deficiencies,
any model performance evaluation using the NAPAP inventory as a
source ofTSP emissions must be considered preliminary at best.
Until emissions of TSP are given the same consideration as those
of S02, SO4= and other detrimental pollutants, modeling of fine
and coarse particulate matter will continue to lag behind the
other modeling efforts being undertaken today.
Deficiencies In The Inhalable Particulate Network Data Set
The Inhalable Particulate Network (IPN) was developed and
implemented by the Environmental Monitoring Systems Laboratory
(EMSL) in conjunction with the Office of Air Quality Planning and
Standards (OAQPS). The IPN was designed to collect size-specific
46
-------
particulate data in anticipation of the 1977 Clean Air
Requirement Act, which called for a reappraisal of the National
Ambient Air Quality Standard for particulate matter. One reason
for this reappraisal was a shift in emphasis from Total Suspended
Particulate matter (TSP), which ranged in size from 0.0 to 50.0
urn, to the smaller Inhalable Particulate (IP), which ranges in
size from 0.0 to 15.0 urn.
The IPN became operational during April 1979, when 57 sites
located throughout the United States went on-line using hi-vol,
dichotomous and size selective inlet samplers to collect data on
TSP, FINE-10 (0-2.Sum), and COARSE-15 (2.5-15um) particulate
matter. The network eventually grew to 157 sites, when in 1981,
EPA's OAQPS recommended that the revised primary standard for
ambient air concentration should be based on a 10 urn criteria
rather than the 15 urn. The 15 urn limit, which had been suggested
by Miller et al. (1979), had been the subject of debate since the
network's conception. EPA's decision, which was based upon the
recommendations of the Clean Air Science Advisory Committee and
the International Standards Organization Task Group, initiated
efforts toward the deployment of 10 urn size specific samplers at
39 sites in the IPN during 1982. This effort, however, was too
late to provide the necessary data for this model evaluation.
Unfortunately, of the 157 IPN sites that were in operation
at one time or another, only 14 were spatially and temporally
compatible with this evaluation. Table 4.2 provides a list of
these sites along with their respective IPN site numbers,
locations and land use categories. Figure 4.3 illustrates the
location of these sites.
47
-------
DALLAS
STON (2)
TOTAL SITES -
Figure 4.3 Inhalable Particulate Network sites used in the
preliminary model evaluation.
48
-------
Table 4.2 Inhalable Particulate Network sites used in the
preliminary model evaluation.
NAME
Huffman
Mtn. Brook
Hartford
Dover
Boston (Fire Hq)
Boston (S Cen)
Minneapolis (HS)
Minneapolis (Nic)
St. Louis
Kansas City
Buffalo
RTF
Philadelphia
Dallas
SITE #
010570001A07
012540001A07
070420003A07
080020001A07
220240012A07
220240013A07
242260049A07
242260051A07
260030001A07
262380002A07
330660003A07
341160101A07
397140024A07
451310050A07
STATE
AL
AL
CT
DL
MA
MA
MN
MN
MO
MO
NY
NC
PA
TX
LOCATION
Suburban
Suburban
Center City
Center City
Suburban
Center City
Center City
Center City
Suburban
Center City
Center City
Rural
Suburban
Center City
LAND-USE
Residential
Residential
Commercial
Commercial
Commercial
Commercial
Residential
Commercial
Commercial
Commercial
Residential
Commercial
Commercial
Commercial
49
-------
A total of 41 sites were located outside the model's domain,
and 62 sites did not come 'online1 until after the evaluation
period. Of the 54 remaining sites, 33 had insufficient data
(i.e. less than 10 observations during the three month evaluation
period), and 7 were located in areas that were classified as
industrial. Unfortunately, of the 14 sites that were spatially
and temporally compatible, 6 were co-located sites, (i. e. they
were located within the same city and grid cell) which further
reduced the spatial representiveness of the evaluation data set.
These deficiencies, as well as others, are elaborated upon
below.
In order to adequately evaluate a regional scale model such
as RELMAP, which has a 1° by 1° grid cell resolution, one would
ideally have a monitoring network made up of remote locations
that have the same spatial and temporal resolution as the model.
Unfortunately, the IPN was designed primarily to characterize
urban scale concentrations of suspended particulate matter, since
the attainment of air quality standards is evaluated over this
scale (Watson et al.,1981). Because of this, an overwhelming
majority of the IPN sites are classified as either center city or
suburban, where the dominant land use is described as either
industrial, commercial or residential. In fact, of the 157 sites
that make up the network, only 5 are classified as remote, and 9
classified as rural. Of these 14 sites sites available for the
evaluation, only one, the Research Triangle Park (RTP), NC is
classified as rural.
In a study performed in the Detroit area, Wolff et al.,
(1984) concluded that regionally emitted emissions generally
50
-------
dominate the ambient concentration of fine particulate matter
over local emissions. But, they also concluded that ambient
concentrations of coarse particulate matter were dominated by
local sources at all of the sites. With this in mind, a further
criteria was established for sites used in this evaluation in
that they must not be industrial in nature, which further reduced
the number of available evaluation sites.
With few exceptions, the hi-vol, dichotomous and SSI
samplers used in the IPN were only activated once every six days,
at which time 24-hour average ambient air concentrations were
recorded from midnight to midnight (LST). These sixth-day
observations resulted in a dearth of data, which in turn made the
model evaluation very difficult and preliminary at best. The
maximum number of 24 hour observations available for the three
month evaluation period was 16, with 6 of these observations
occurring in July and September, and 5 in August.
This limited number of observations was further depleted
when the amount of "down time" for each site was considered. Of
the 14 stations used in the evaluation, only two, Philadelphia
and Mtn. Brook, Alabama, had the full allotment of 16
observations. The 12 remaining sites averaged between 12 and 14
with a minimum of 10. Using such a temporally inconsistent data
set makes the observations very susceptible to extremes caused by
local sources. The tremendous variability exhibited by the
observed data, whether real or artificial, cannot be modeled by a
regional-scale, long term (monthly) model such as RELMAP. This
incompatibility is best illustrated by Figures 4.4 and 4.5,
which presents the observed and simulated fine and coarse
51
-------
particulate matter concentrations for Hartford, Connecticut
during .the three month evaluation period. On the abscissa, one
finds the date, which ranges from July 1, to October 1, 1980,
while the ambient air concentrations are on the ordinate. At
best, the observed data is inconsistent, with four of the 16
observations missing. Another unfortunate characteristic of the
observed data is its tremendous variance. Fine particulate
concentrations range from 6 to 45 ug/m , while the coarse
concentrations range from 4 to 17 ug/m3. Such variance, which
may be an indication of local sources, is impossible to simulate
by the model.
OBSERVED AND SIMULATED F1NE-PM CONCENTRATIONS
FOR HARTFORD. CONN. DURING THE SUMMER OF 1980
45
40
ff
I"
5»
§25
520
o
§15
u
2
UJ
Z -
\
01JUL80
01AUG80 01SEP80
DATE
010CTBO
4» OBSERVATION H MONTHLY SIMULATION
Figure 4.4 Temporal depiction of the observed and simulated fine
particulate matter concentrations for Hartford, Conn.,
for the summer of 1980.
52
-------
OBSERVED AND SIMULATED COARSE-PI! CONCENTRATIONS
FOR HARTFORD. CONN. DURING THE SUMMER OF 1980
20
^s
E
>«'
3
N^
O
^10
UJ
o
z
8
2 s-
bl
M
$
A
\ /\ >
t \ i \ /
7 \ , / \ /
* / / \ /
^ * \ /
Vx
01JUL80 01AU680 01SEP80 010CTBO
DATE
«« OBSERVATION H MONTHLY SIMULATION
Figure 4.5 Temporal depiction of the observed and simulated
coarse particulate matter concentrations for
Hartford, Conn., for the summer of 1980.
As discussed in the previous section, when the emphasis was
shifted from COARSE-15 to COARSE-10 in 1982, dichotomous samplers
were incorporated into the IPN which measured COARSE-10.
Unfortunately, this transition occurred too late to provide
direct measurement of COARSE-10 needed for this evaluation.
However, several methods have recently been developed that have
allowed full utilization of the COARSE-15 data as a substitute
for the COARSE-10 data. Rodes et al. (1984) examined the
relationships between PM-15 (particles less than or equal to 15
um) and PM-10 (particles less than or equal to 10 urn) data at
eight cities located throughout the United States, and found that
53
-------
PM-10 and PM-15 concentrations exhibited a very strong linear
relationship, making it possible to predict one from the other.
Correlations between the measurements at the eight sites, which
included industrial, rural as well as suburban locations, ranged
between 0.93 and 0.98. The ratio of PM-10/PM-15 was also
consistent, ranging from 0.75 to 0.96, and averaging 0.85. It
should be noted that the lowest value (0.75) was recorded at the
only western site (Phoenix), and that if this outlier is removed
from the data set, an average ratio of 0.87 would result.
In a similar study, Pedco Environmental, Inc. et al. (1984),
examined, tested and evaluated 13 different methods in which PM-
10 and PM-15 could be estimated from PM-15 and TSP, respectively.
Among their final recommendations for estimating PM-10 from PM-15
data was to use a PM-10/PM-15 ratio of 0.88 in the eastern states
and 0.77 in the western states. This method, which produced the
smallest standard errors of any tested by Pedco Environmental,
Inc. et al., (1984), and which was in very good agreement with
Rodes et al.,(1984), was selected for use in this evaluation.
Subsequently, all of the PM-15 data were converted into PM-10
data using the 0.88 ratio. The Coarse-10 fractions were then
determined by subtracting the Fine-10 (which is the same as the
Fine-15) fraction from the PM-10 total.
Model Evaluation for the Summer of 1980
RELMAP was run on a monthly basis for July, August and
September, 1980 in order to produce monthly and seasonal
54
-------
simulations of concentrations and wet and dry depositions of fine
and coarse particulate matter. The monthly and seasonal
simulated values of fine and coarse concentrations (expressed in
ug/m3) were then compared to the 14 compatible sites from the
IPN. The number of stations used each month in the evaluation
varied depending upon the number of which met the minimum
observation requirement of 3 observations/month, or 10
observation/summer season. July and September had 13 stations
fulfill this requirement, while August had 14. Tables listing
the fine and coarse particulate matter concentrations data used
in the evaluation are provided in Appendices A and B,
respectively.
The mean, standard deviation, and minimum and maximum for
each of the three months and the season are presented in Tables
4.3 and 4.4 for fine and coarse particulate matter
concentrations, respectively.
Table 4.3 Statistical Evaluation Involving Fine Particulate
Matter Concentrations.
Month Mean Std. Dev.
Obs. Sim. Obs. Sim.
Minimum
Obs. Sim.
Maximum
Obs. Sim.
July 25.34 5.92 10.16 2.84
Aug. 24.63 5.89 10.65 2.95
Sept. 18.80 9.71 5.77 4.28
11.80 1.52
10.57 1.55
10.58 2.36
45.10 10.95
53.33 10.90
34.02 16.26
Summer 22.71 7.20
6.74 3.32
12.65 1.80
39.86 12.48
Three month mean, weighted by the total number of observations,
55
-------
Table 4.4 Statistical Evaluation Involving Coarse Particulate
Matter Concentrations.
Month Mean
Obs. Sim.
July 15.58 2.46
Aug. 14.54 2.06
Sept. 12.96 3.07
Summer 14.34 2.56
Std. Dev. Minimum Maximum
Obs. Sim. Obs. Sim. Obs. Sim.
8.06 1.66 2.17 1.09 29.15 6.84
8.17 1.17 4.67 1.04 33.43 5.46
5.96 1.57 1.88 1.68 26.46 7.55
6.73 1.43 2.77 1.34 25.74 6.61
Examination of the tables reveals that in all cases, the
model significantly underpredicted the fine and coarse
concentrations. Scatter diagrams, which depict the correlation,
or dependency of the simulated value (ordinate) upon the observed
(abscissa) are presented in Figures 4.6 and 4.7. These too
illustrate that the model simulations were significantly lower
than the observed values. A line of best fit has been included
as a reference. The correlation between the simulated and
observed values of fine particulate matter was 0.533, indicating
that 28.4% of the variance experienced by the observed values
could be accounted for by the simulated values. Likewise, the
correlation between the observed and simulated coarse
56
-------
40
I25
hi
§«i
O
55
01
OBS MEAN = 22.71
SIM MEAN = 7.20
CORR = 0.53
0 5 10 15 20 25 30 35 40
OBSERVED CONCENTRATION (ug/m3)
Figure 4.6 Scatter diagram of the observed vs. simulated fine
particulate matter concentrations for the summer of
1980.
30
f
*S
i20
1151
hi
O
§101
3 »1
W
OBS MEAN = 14.34
SIM MEAN = 2.56
CORR = 0.32
5 10 15 20
OBSERVED CONCENTRATION (ug/m3)
25
30
Figure 4.7 Scatter diagram of the observed vs. simulated coarse
particulate matter concentrations for the summer of
1980.
57
-------
concentrations was 0.322, indicating that 10.4% of the observed
variance could be explained by the simulation.
The standard residuals ((observed-predictedj/observed) for
each of the individual sites for the entire summer are depicted
in Figures 4.8-4.9. These figures indicate that the model is
consistent in its underprediction across the entire evaluation
network. Standardized residuals range between 0.42 and 0.89 for
the fine concentrations and between 0.48 and 0.93 for the coarse
concentrations. This significant underprediction exhibited by
the model is not surprising given the nature of the discrepancies
discussed at the beginning of the section. All of the
discrepancies inherent with the NAPAP emissions inventory would
lend themselves to underpredictions by the model. Nearly 8% of
the total TSP inventory was omitted because size fractions were
not available. And even more significant is the exclusion of
large emissions from open sources.
Several of the deficiencies inherent to the IPN would
likewise result in the model underpredicting the concentrations.
Designed primarily to characterize urban scale concentrations,
the IPN had an overwhelming majority of its sites located within
cities. In fact, of the 14 sites selected for this evaluation,
13 were designated as either center-city or suburban. Although
regional influences dominate over local sources for fine
particles, coarse particles would be adversely affected by such
an arrangement. It is worth noting that the only remote site
available for the study, RTP, actually showed fairly good
agreement between the observed coarse concentrations (2.66 ug/m3)
and that simulated by the model (1.34 ug/m3).
58
-------
Figure 4.8 Standardized residuals ((0-P)/0) of the fine
particulate matter concentrations for the summer of
of 1980.
0.85
0.84
Figure 4.9
Standardized residuals ((0-P)/0) of the coarse
particulate matter concentrations for the summer
of 1980.
59
-------
SECTION 5
CONCLUSIONS AND RECOMMENDATIONS
In response to the promulgation of the new, smaller, size
discriminate National Ambient Air Quality Standards for IP,
RELMAP has been modified to now include simple, linear
parameterizations simulating the chemical and physical processes
of fine and coarse particulate matter. Emphasis was placed upon
the smaller particles for several reasons; first, the smaller
sized particles were found to have a more adverse effect on
health, and secondly, because the larger size particles had a
large contribution from natural sources, attainment of federal
standards was becoming more and more difficult.
Shifting the emphasis to the smaller particles enhances the
utility of regional scale, Lagrangian models such as RELMAP.
In this model, discrete puffs of S02/ S04=, fine and coarse
particulate matter are subjected to linear transformation and wet
and dry deposition processes as they are transported across the
model's domain. RELMAP treats fine and coarse particulate matter
as non-evolving pollutants and assumes physical or chemical
transformation between the two to be negligible. RELMAP does
however, consider the transformation of SO2 into S04~, which it
treats as a function of solar insolation and moisture content.
Dry deposition of S02/ S04=, and fine and coarse particulate
60
-------
matter is treated as a function of land use, season, and
stability. Wet deposition is treated by the model as a function
of cloud type, pollutant concentration and precipitation rate.
Because these recently modified parameterizations are only
accurate to a limited degree, they may be upgraded or even
replaced in the future with more sophisticated parameterizations
as further research is conducted. As an initial step in this
possible refinement of RELMAP, the model was subjected to a
sensitivity analysis. In this analysis, which employed actual
emissions and meteorological data for July, 1980, variations
found in the simulated concentrations of fine and coarse
particulate matter, due to arbitrary +/- 50% variations from
nominal values of the transformation rate and wet and dry
deposition rates were examined.
Results of the analysis were recorded along a transect
consisting of 15 grid cells which stretched across the model's
domain. Each analysis consisted of two graphs, illustrating the
absolute as well as the relative changes, with respect to a base
case simulation. Simulated concentrations of fine and coarse
particulate matter were found to be by far most sensitive to
changes in the wet deposition rates of fine and coarse
particulate matter, respectively. However, concentrations of
fine particulate matter were quite insensitive to changes in the
wet deposition rate of S02. Concentrations of coarse particulate
matter were somewhat sensitive to dry deposition rates .of coarse
particles; however, fine particulate matter concentrations were
less sensitive to dry deposition of fine particles and highly
insensitive to dry deposition of S02- And finally, fine
61
-------
particulate matter concentrations proved to be somewhat
insensitive to the transformation rate of S02 into S04=.
Future research should concentrate upon refining the
parameterizations involving the wet deposition of both fine and
coarse particulate matter. Not only has wet deposition proven to
be the most influential parameterization employed by the model,
it is also currently the least understood. Although the model
proved to be somewhat less sensitive to the other
parameterizations, future research should also address these
areas as well, so that they too will parameterize the essential
physical and chemical processes occurring in the atmosphere
accurately.
In order to determine just how accurately these new
parameterizations actually simulate the physical and chemical
processes of the atmosphere, RELMAP was subjected to a model
performance evaluation. The model was run for the summer of
1980, using actual meteorological data and emissions data from
the NAPAP Version 5.0 emissions inventory. Simulations of
ambient air concentrations of fine and coarse particulate matter
were then compared to data from the IPN. Unfortunately,
inadequacies inherent to both the emissions and validation data
sets limited the scope of this evaluation.
As an illustration of these inadequacies, the NAPAP emissions
inventory was designed primarily to support acid deposition
modeling, not regional particulate modeling. Because of this,
many deficiencies were found with the inventory, including the
following: (1) most open source emissions were omitted from the
62
-------
inventory (which by some estimates exceed 50,000 ktons of TSP),
(2) the estimates of contributions from paved and non-paved
roads, which account for 70% of the total inventory, are much
lower in the NAPAP inventory than other independent estimates,
(3) a total of 8% of the NAPAP inventory cannot be
fractionalized, because particle size distributions are not
available for many source classification codes.
The only way to alleviate these deficiencies is to reduce
the tremendous amount of uncertainties in the estimates of the
open source emissions. Such a solution may be forthcoming as the
NAPAP Task Group II is scheduled to release, in the fall of 1987,
a revised emissions inventory for open source emissions of TSP.
Should this revised inventory include such major open sources of
TSP as wind erosion, agricultural tilling, construction and
mining operations, and should the new estimates include emissions
from paved and non-paved roads which concur with other
independent estimates, RELMAP'S accuracy and therefore its
credibility as a regional-scale particulate model will improve.
A second major deficiency that proved to be detrimental to
the model performance evaluation is the incompatibility of the
IPN data. Of the 157 IPN sites that were operational at one time
or another, only 14 were spatially and temporally compatible with
the requirements of this evaluation. The IPN, like the NAPAP
Version 5.0 emissions inventory, was not designed for regional
scale particulate modeling. Rather the IPN was designed primarily
to characterize the urban-scale concentrations of TSP, therefore,
an overwhelming majority (144 of 157) of the sites were
classified as either center city, or suburban.
63-
-------
Another deficiency inherent to the IPN is the fact that
observations were only recorded once every six days, resulting in
a dearth of data. Since regional-scale pollution episodes have
an average temporal span of several days, 24 h air concentrations
sampled every sixth day are not be sufficient to capture the true
variability of the ambient air concentration data. Therefore,
when combined with the predominantly urban locations of the
sites, the discontinuous sampling records of the IPN render the
data inadequate for regional scale particulate modeling.
At the present time, there are no plans to implement a
network that would fulfill the specific needs of regional scale
particulate modeling. However, in the near future, a network
proposed by NAPAP to assist in the evaluation of acid deposition
models will begin monitoring pollutants on a regional scale at
between 30 and 50 sites located in the eastern United States. If
funded as proposed, continuous 12 h samples of fine particulate
matter concentrations will be obtained, beginning in 1988, along
with wet and dry deposition chemistry data. Although fine
particulate matter is to be sampled, the analysis will
concentrate on sulfate concentrations only. Therefore as
currently proposed, the network fails to address the needs of
regional scale particulate modeling.
Since appropriate data bases to evaluate regional scale
particulate models do not exist, nor are any proposed, and
because the cost of initiating and operating a network are
prohibitive, Clark (1986) has recommended that the
operational/analysis protocol of the proposed NAPAP network be
64
-------
expanded to obtain an appropriate data base for evaluating
regional scale particulate models. Because of its spatial and
temporal distribution, the NAPAP network would provide an
excellent data base. By supplementing the proposed network with
fine and coarse particulate matter monitoring equipment, an
appropriate data base can be generated for particulate modeling
for a fraction of the cost needed to initiate and operate a new
network.
Since the inadequacies discussed above have greatly limited
the scope of this model evaluation, it must be considered
preliminary at this time. Results of the performance evaluation
indicate that RELMAP significantly underpredicted the average
ambient concentrations of both fine and coarse particulate matter
for the three month period. The observed and simulated fine
particulate concentrations were 22.71 and 7.20 ug/m3,
respectively, while the observed and simulated coarse particulate
concentrations were 14.34 and 2.56 ug/m3, respectively. The
correlation between the observed and simulated fine
concentrations was 0.53, indicating that 28.4% of the variance
was explained by the model. The correlation between the coarse
simulated and observed concentrations was 0.32, indicating that
10.4% of the variance was explained.
Considering the nature of the deficiencies discussed above,
such an underprediction by the model, though disappointing, is
not surprising. Each of the deficiencies inherent to the NAPAP
inventory and several inherent to the IPN data would indeed lend
themselves to an underprediction by the model.
In order for RELMAP to become a credible regional
65
-------
particulate model which can be used as a tool in assessing the
effects of various emission control scenarios, it is critical
that: (1) a revised TSP emissions inventory become available
which more accurately emulates both the natural and anthropogenic
emissions, and (2) adequate, regionally-representative, and
continuous measurements of ambient air concentrations of both
fine and coarse particulate matter be obtained.
66
-------
REFERENCES
Altshuller, A. P., 1979. Model predictions of the rates of
homogeneous oxidation of sulfur dioxide to sulfate in the
troposphere. Atmos. Environ. 13: 1653-1661.
Bhumralkar, C. M., R. L. Mancuso, D. E. Wolff, R. H. Thuillier,
K. D. Nitz, and W. B. Johnson, 1980. Adaptation and Application
of a Long-Term Air Pollution Model ENAMAP-1 to Eastern North
America. Final Report, EPA-600/4-80-039, U. S. Environmental
Protection Agency, Research Triangle Park, N. C., 100 pp.
Briggs, G. A., 1984. Plume rise and buoyancy effects. Chapter 8
in Atmospheric Science and Power Production. D. Randerson, Ed.,
U. S. Dept. of Energy DOE/TIC-27601, 347 pp.
Clark, T. L., 1986. Measurements Required for Episodic and Long-
Term Regional Particulate Matter Model Evaluations. Position
Paper, Environmental Protection Agency, Research Triangle Park,
N. C., 6 pp.
Clark, T.L., R. L. Dennis, E. C. Voldner, M. P. Olson, S.
Seilkop, M. Alvo, 1987. The International Sulfur Deposition Model
Evaluation (ISDME). Presented at the 80th Annual Meeting of APCA,
New York, N. Y., June 2.1-26.
Eder, B. K., D. H. Coventry, T. L. Clark, and C. E. Bellinger,
1986. RELMAP: A REgional Lagrangian Model of Air Pollution
User's Guide. EPA-600/8-86/013, U. S. Environmental Protection
Agency, Research Triangle Park,. N. C., 146 pp.
Endlich, R. M., K. C. Nitz, R. Brodzinksy, and C. M. Bhumralkar,
1983. The ENAMAP-2 Air Pollution Model for Long Range Transport
of Sulfur and Nitrogen Compounds. Final Report, EPA-600/7-83-
059, U. S. Environmnetal Protection Agency, Research Triangle
Park, N. C., 217 pp.
Evans, J. S., and D. W. Cooper, 1980. An inventory of
particulate emissions from open sources. J. Air Poll. Con.
Assoc. 30: 1298-1303.
Federal Register, 1984. Proposed Revisions to National Ambient
Air Quality Standards for Particulate Matter to Control Particles
10 Micrometers or Less. Federal Register. 49 FR 10408, March,
20, 1984.
Gifford, F. A., Jr., 1976. Turbulent diffusion typing schemes:
A Review. Nucl. Safety. 17: 68-86.
67
-------
Hinton, D. 0., J. M. Sune, J. C. Suggs, and W. F. Barnard, 1984.
Inhalable Particulate Network Report: Data Listing (Mass
Concentration Only) - Volume II. April 1979 - December 1982.
Final Report, EPA-600/4-84-088b, U. S. Environmental Protection
Agency, Research Triangle Park, N. C., 457 pp.
Husar, R. B., D. E. Patterson, J. D. Husar, N. V. Gillani, and W.
E. Wilson, 1978. Sulfur budget of a power plant plume. Atmos.
Environ. 12: 549-568.
Issac, G. A., P. I. Joe, and P. W. Summers, 1983. The vertical
transport and redistribution of pollutants by clouds.
Transactions of the APCA Specialty Conference on the Meteorology
of Acid Deposition. Hartford-, Conn., October 16-19, 496-512.
Johnson W. B., 1983. Interregional exchanges of air pollution:
Model types and applications. J. Air Poll. Con. Assoc. 33:
563-574.
/
Johnson, W. B., D. E. Wolff, and R. L. Mancuso, 1978. Long-term
regional patterns and transfrentier exchanges of airborne sulfur
pollution in Europe. Atmos. Environ. 12: 511-527.
Mamane, Y., and K. E. Noll, 1985. Characterization of large
particles at a rural site in the eastern United States: Mass
distribution and individual particle analysis. Atmos. Environ.
19:611-622.
Meagher, J. F., E. M. Bailey, and M. Lucia, 1983. The seasonal
variation of the atmospheric S02 to S04= conversion rate. J.
Geophvs. Res. . 88: 1525-1527.
Meagher, J. F. , and K. J. Olszyna, 1985. Methods For Simulating
Gas Phase SO2 Oxidation in Atmospheric Models. Final Report,
EPA-600/3-85/012, U. S. Environmental Protection Agency, Research
Triangle Park, N. C., 76 pp.
Miller, F. J., D. E. Gardner, J. A. Graham, R. E. Lee, W. E.
Wilson, and J. D. Bachman, 1979. Size considerations for
establishing a standard for inhalable particulates. J. Air Poll.
Con Assoc. 30: 1320-1322.
Pack, D. H., G. J. Ferber, J. L. Heffter, K. Telegadas, J. K.
Angel1, W. H. Hoecker, and L. Machta, 1978. Meteorology of long
range transport. Atmos. Environ. 12: 425-444.
Pedco Environmental, Inc., 1984 Estimating PM10 and FP
Background Concentrations From TSP And Other Measurements. Final
Report, EPA-450/4-84-021, U. S. Environmental Protection Agency,
Research Triangle Park, N. C., 75 pp.
Rodes, C. E., and E. G. Evans, 1985. Preliminary assessment of
10 urn particulate sampling at eight locations in the United
States. Atmos. Environ. 19: 293-303.
68
-------
Scott, B. C., 1978. Parameterizations of sulfate removal by
precipitation. J. Appl. Meteor. 17: 1375-1389
Scott, B. C^, 1982. Predictions of in-cloud conversion rates of
SO? to S04~ based upon a simple chemical and kinematic storm
model. Atmos. Environ. 13: 1361-1368.
Sehroel, G. A., 1980. Particle and gas dry deposition: A review.
Atmos. Environ. 14: 983-1011.
Sheih, C. M., M. L. Wesely, and B. B. Hicks, 1979. Estimated dry
deposition velocities of sulfur over the eastern United States
and surrounding regions. Atmos. Environ. 13: 1361-1368.
Suggs J. C.,C. E. Rodes, E. G. Evans, and R. E. Baumgardner,
1981. Inhalable Particulate Network Annual Report: Operation and
Data Summary (Mass Concentration Only). April, 1979 - June,
1980. EPA-600/4-81-037, Environmnetal Protection Agency,
Research Triangle Park, N. C., 238 pp.
U. S./Canadian Memorandum of Intent, 1982. Emissions, Costs and
Engineering Assessment Work Group 3B, Final Report.
Wagner, J. K., R. A. Walters, L. J. Maiocco, and D. R. Neal,
1986. Development of the 1980 NAPAP Emissions Inventory. Volume
I Draft: Final Report. Environmental Protection Agency Contract
Number: 68-02-3997, Work Assignment 9.
Watson, J. G., J. C. Chow, and J. J. Shah, 1981. Analysis of
Inhalable and Fine Particulate Matter Measurements. EPA-450/4-
81-035, U. S. Environmental Protection Agency, Research Triangle
Park, N. C., 334 pp.
Wesely, M. L., and J. D. Shannon, 1984. Improved estimates of
sulfate dry deposition in eastern North America. Environ. Proa.
Vol. 3, No. 2, 77-81.
Wolff, G. T., P. E. Korsog, D. .P. Stroup, M. S. Ruthkosky, and M.
L. Morrissey, 1985. The influence of local and regional sources
on the concentration of inhalable particulate matter in
southeastern Michigan. Atmos. Environ. 19: 305-313.
69
-------
APPENDIX A
Table A.I Fine Particulate Matter Concentrations (ug/m3) for the
Month of July.
STATION
HUFFMAN
MTN. BROOK
HARTFORD
02
41.89
36.05
34.97
08
41.83
33.60
17.48
JULY
14 20
16.40
37.34 10.59
44.30
26
13.63
31.62
AVERAGE
33.37
26.24
32.09
DOVER
BOSTON (Fire St)
BOSTON (S Cen)
MINNEAPOLIS (HS) 9.20
MINNEAPOLIS (N) 9.76
ST LOUIS 35.54
KANSAS CITY 14.04
BUFFALO 53.84
RTF 11.44
PHILADELPHIA 40.31
DALLAS 15.38
14.17 50.95 29.23
17.42 13.25 48.63 29.46
19.70 18.87 8.79 4.74
19.70 22.67 . 4.61
18.69 48.44 26.58 38.58
10.83 7.98 . 17.46
30.67 30.83 67.89 42.29
26.75 20.74 11.61 20.98
19.12 18.39 51.64 11.25
9.14 13.49 .22.39 16.29
31.45
27.19
11.80
14.19
33.57
12.58
45.10
18.30
28.14
15.34
70
-------
APPENDIX A (Cont.)
Table A.2 Fine Particulate Matter Concentrations (ug/m3) for the
Month of August.
AUGUST
STATION 01
HUFFMAN 51.46
MTN. BROOK 34.65
HARTFORD
DOVER 26.88
BOSTON (Fire St) 54.28
BOSTON (S Cent) 48.38
MINNEAPOLIS (HS) 9.81
MINNEAPOLIS (Nic)
ST LOUIS 18.40
KANSAS CITY
BUFFALO 70.25
RTP 31.24
PHILADELPHIA 60.79
DALLAS
07
35
29
15
21
20
9
16
58
18
32
15
.74
.31
.81
.04
.39
.55
.17
.11
.77
.13
.19
13
.
20.
19.
37.
30.
21.
7.
22.
10.
21.
15.
12.
85
93
32
26
62
74
65
61
00
50
48
19
30.
20.
27.
16.
14.
11.
7.
12.
14.
*
15.
14.
25
85
07
95
36
33
59
56
13
68
59
02
45.
31.
21.
40.
21.
20.
27.
39.
19.
28.
26.
33.
01
98
19
05
66
82
30
30
68
23
34
52
3
16
9
28
9
19
7
8
18
10
56
18
7
1
.03
.76
.40
.17
.56
.91
.19
.61
.72
.74
.39
.13
AVERAGE
35.
24.
25.
21.
29.
24.
10.
15.
21.
13.
53.
23.
28.
16.
82
44
43
12
92
73
57
87
64
67
33
67
12
47
71
-------
APPENDIX A (Cont.)
Table A.3 Fine Particulate Matter Concentrations (ug/m3) for the
Month of September.
SEPTEMBER
» * ^ ^ OB ^ «i ^ _ ^ ^ W ^ «~ ^ ^
12 18 24
STATION
06
30
AVERAGE
HUFFMAN 29.61
MTN. BROOK 22.71 29.37
HARTFORD 12.68 20.61
DOVER 33.90 27.08
BOSTON (Fire St) . 18.60
BOSTON (S Cen) 10.00 10.60
MINNEAPOLIS (HS) 11.49 23.30
MINNEAPOLIS (N) 11.82 29.17
ST LOUIS 23.59
KANSAS CITY 14.82
BUFFALO 25.75 18.65
RTP 17.11 26.70
PHILADELPHIA 32.36 27.41
DALLAS 11.04 27.96
14.08 23.66 6.61
12.04 22.68 5.74
7.20 6.71 14.35
8.35 15.99 11.56
7.56 11.82 18.48
7.73 5.37 19.20
19.89
7.50 8.74 31.56
56.04 31.21
13.88 15.57 45.35
16.79
13.58 9.40 18.98
22.31 12.59 8.94
18.49
18.51
12.31
19.38
14.12
10.58
18.23
17.76
34.02
23.84
20.20
20.35
16.57
72
-------
APPENDIX A (Cont.)
Table A.4 Monthly and Seasonal Observed (Obs) and Simulated (Sim)
Fine Particulate Matter Concentrations (ug/m3) for the
IPN Sites.
JULY
STATION
HUFFMAN
MT. BROOK
HARTFORD
DOVER
BOS (Fire S)
BOS (S Cen)
MINN (HS)
MINN (Nic)
ST LOUIS
KANSAS CITY
BUFFALO
RTP
PHILADELPHIA
DALLAS
OBS
33.37
26.24
32.09
31.45
27.19
11.80
14.19
33.57
12.58
45.10
18.30
28.14
15.34
SIM
5.98
5.98
10.95
5.70
5.70
1.52
1.52
6.22
4.07
9.04
4.81
10.15
5.37
AUGUST
OBS
35.82
24.44
25.43
21.12
29.92
24.73
10.57
15.87
21.64
13.67
53.33
23.67
28.12
16.47
SIM
5.86
5.86
10.35
9.21
5.46
5.46
1.55
1.55
5.55
3.22
8.60
4.89
10.90
3.93
SEPTEMBER
OBS
18
18
12
19
14
10
18
17
34
23
20
20
16
.49
.51
.31
.38
.12
.58
.23
.76
.02
.84
.20
.35
."57
SIM
10.
10.
16.
14.
11.
11.
2.
2.
7.
12.
9.
13.
5.
09
09
26
46
57
57
36
36
25
20
25
03
74
SUMMER
OBS
29.
23.
22.
20.
25.
20.
12.
16.
27.
19.
39.
20.
25.
16.
43
15
18
33
41
37
65
10
06
34
86
28
70
13
SIM
7. 28
7.28
12.48
11.79
7.53
7.53
1.80
1.80
5.89
4.82
9.92
6.28
11.34
5.01
73
-------
APPENDIX B
Table B.I Coarse Particulate Matter Concentrations (ug/ro3) for the
Month of July.
JULY
STATION
HUFFMAN
MTN. BROOK
HARTFORD
DOVER
BOSTON (Fire S)
BOSTON (S Cen)
MINNEAPOLIS (HS)
MINNEAPOLIS (N)
ST LOUIS
KANSAS CITY
BUFFALO
RTP
PHILADELPHIA
DALLAS
02
17
10
10
15
17
36
36
30
1
20
15
.05
.46
.52
.06
.54
.68
.09
.22
.00
.15
.86
08
41
6
14
11
13
19
19
30
9
3
14
12
.94
.26
.12
.35
.17
.06
.63
.14
.53
.15
.55
.57
14
11
11
10
24
27
26
26
19
1
13
20
.22
.13
.54
.13
.63
.71
.10
.66
.52
.71
.17
20
28
3
15
6
6
7
18
8
3
10
28
.48
.96
.50
.69
.24
.80
.50
.72
.09
.07
.09
26
3.
8.
12.
9.
6.
12.
21.
19.
10.
0.
9.
14.
34
67
11
66
26
00
76
25
73
91
79
59
AVERAGE
29.
7.
12.
*
9.
9.
13.
19.
24.
27.
15.
1.
13.
18.
15
05
20
98
44
28
06
66
90
77
93
65
26
74
-------
APPENDIX B (Cont.)
Table B.2 Coarse Particulate Matter Concentrations (ug/m3) for the
Month of August.
AUGUST
STATION
HUFFMAN
MTN. BROOK
HARTFORD
DOVER
BOSTON (Fire S)
BOSTON (S Cent)
MINNEAPOLIS (HS)
MINNEAPOLIS (Nic)
ST LOUIS
KANSAS CITY
BUFFALO
RTP
PHILADELPHIA
DALLAS
01
13.71
3.91
16.29
12.59
8.90
24.12
*
19.46
10.59
4.80
8.89
*
07
19.
11.
13.
17.
14.
14.
18.
11.
6.
16.
14.
11
23
43
75
82
20
24
24
33
07
18
13
8
7
10
13
12
5
24
26
2
6
20
.
.56
.23
.82
.22
.34
.86
.41
.97
.87
.81
.48
19
22
7
9
5
7
6
10
24
21
12
23
.83
.88
.43
.28
.94
.29
.60
.32
.15
.17
.22
10
4
10
13
11
13
65
27
22
11
12
18
25
.41
.94
.82
.32
.27
.70
.63
.90
.30
.34
.34
.00
9
2
7
7
16
6
10
38
26
4
9
12
31
.62
.66
.40
.80
.57
*
.35
.34
.31
.50
.34
.88
.68
AVERAGE
15
6
8
10
13
10
12
33
24
25
9
4
11
17
.14
.53
.02
.74
.57
.73
.47
.43
.91
.25
.38
.67
.02
.71
75
-------
APPENDIX B (Cont.)
Table B.3 Coarse Particulate Matter Concentrations (ug/m3) for the
Month of September.
STATION
HUFFMAN
MT. BROOK
HARTFORD
DOVER
BOSTON (Fire S)
BOSTON (S Cent)
MINNEAPOLIS (HS)
MINNEAPOLIS (N)
ST LOUIS
KANSAS CITY
BUFFALO
RTP
PHILADELPHIA
DALLAS
12
4
11
9
13
14
22
38
20
7
0
7
11
06
.23
.79
.10
.59
.83
.73
.61
.31
.79
.38
.82
.89
.59
5
16
19
10
11
9
24
25
12
2
20
14
S
12
.
.03
.82
.39
.92
.61
.79
.33
.91
.39
.89
.75
.14
EPTEMBE
18
14.23
4.13
4.81
16.43
8.17
8.28
18.23
11.72
1.94
4.70
26.45
R
20
5
9
'17
10
9
26
12
20
10
18
24
.98
.08
.93
.69
.87
.40
.46
.04
.88
.85
.64
4
2
14
14
16
16
14
40
21
11
19
6
30
.76
.60
.94
.23
.25
.95
.27
.67
.92
.26
*
.73
.97
AVE
13
4
11
15
11
12
12
26
18
12
1
12
15
RAGE
.05
.33
.52
.46
.55
.02
.93
.46
.25
.73
.88
.78
.59
76
-------
APPENDIX B (Cont.)
Table B.4 Monthly and Seasonal Observed (Obs) and Simulated (Sim)
Coarse Particulate Matter Concentrations (ug/itr) for the
IPN sites.
JULY
STATION
HUFFMAN
MTN. BROOK
HARTFORD
DOVER
BOS (Fire S)
BOS (S Cen)
MINN (HS)
MINN (N)
ST LOUIS
KANSAS CITY
BUFFALO
RTF
PHILADELPHIA
DALLAS
OBS
29.15
7.05
12.20
9.98
9.44
13.28
19.06
24.66
27.90
15.77
1.93
13.65
18.26
SIM
1.54
1.54
1.64
2.14
1.26
1.26
1.51
1.51
4.06
3.72
6.84
1.09
2.62
3.36
AUGUST
OBS
15
6
8
10
13
10
12
33
24
25
9
4
11
17
.14
.53
.02
.74
.57
.73
.47
.43
.91
.25
.38
.67
.02
.71
SIM
1.04
1.04
1.71
2.20
1.41
1.41
1.43
1.43
2.89
2.86
5.46
1.13
2.44
2.34
SEPTEMBER
OBS
13
4
11
15
11
12
12
26
18
12
1
12
15
.05
.33
.52
.46
.55
.02
.93
.46
.25
.72
.88
.78
.59
SIM
1.68
1.68
2.91
3.03
2.65
2.65
2.34
2.34
4.81
4.60
7.55
1.81
3.39
3.30
SUMMER
OBS
15.
6.
10.
12.
12.
10.
12.
25.
24.
24.
12.
2.
12.
17.
42
00
87
89
12
82
86
74
80
21
70
66
39
19
SIM
1.42
1.42
2.08
2.45
1.76
1.76
1.75
1.75
3.91
3.72
6.61
1.34
2.81
3.00
77
------- |