DEVELOPMENT OF THE MESOPUFF II DISPERSION MODEL
by
Joseph S. Scire
Frederick W. Lurmann
Arthur Bass
SCeven R. Hanna
Environmental Research & Technology, Inc.
Concord, Massachusetts 01742
Contract No. 68-02-3733
Project Officer
James M. Godowitch
Meteorology and Assessment Division
Environmental Sciences Research Laboratory
Research Triangle Park, NC 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC 27711
-------
DEVELOPMENT OF THE MESOPUFF II DISPERSION MODEL
by
Joseph S. Scire
Frederick W. Lurmann
Arthur Bass
Steven R. Hanna
Environmental Research & Technology, Inc.
Concord, Massachusetts 01742
Contract No. 68-02-3733
Project Officer
James M. Godowitch
Meteorology and Assessment Division
Environmental Sciences Research Laboratory
Research Triangle Park, NC 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC 27711
-------
DISCLAIMER
This report has been reviewed by Che Environmental Sciences Research
Laboratory, U.S. Environmental Protection Agency, and approved for
publication. Approval does not signify that the contents necessarily
reflects the views and policies of the U.S. Environmental Protection Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
11
-------
PREFACE
This publication contains a technical description of the MESOPUFF II
model and its processor programs. The preprocessor programs require hourly
meteorological surface, twice-daily upper air, and hourly precipitation
(optional) data in the formats archived by the National Climatic Center in
Asheville, North Carolina. The model utilizes the Gaussian puff superposi-
tion approach to simulate a continuous pollutant plume. The model is capable
of multi-day simulations and has algorithms for plume rise, transport, chem-
ical transformations, dry deposition, and wet removal. Terrain variations
are not accounted for in the model.
The puff superposition approach has not been used extensively in air
quality models for the prediction of pollutant concentrations. MESOPUFF II
is being made available to promote testing and evaluation of the methods and
optional features in the model. MESOPUFF II has no regulatory standing and
its application for regulatory purposes should be considered in light of
EPA's Guideline on Air Quality Models.
The model version (1.0) documented in this publication represents an
attempt to utilize recent scientific Information to realistically account for
the relevant physical processes active on the regional to long-range scales.
Modifications may be made in the future based on results by users and findings
from ongoing research programs.
Although attempts have been made to check the computer program code,
errors may be found occasionally. Adjustments to the code to suit different
computer systems may be required. If there is a need to correct, revise, or
update this model, changes may be obtained as they are Issued by completing
and sending the form on the last page of the user guide.
It is anticipated that MESOPUFF II will be made available in the future
on the User's Network for Applied Modeling of Air Pollution (UNAMAP) system.
A tape of this model or the UNAMAP system may be purchased from NTIS for use
on the user's computer system. For information on UNAMAP contact: Chief,
Environmental Operations Branch, MD-80, U.S. Environmental Protection Agency,
Research Triangle Park, NC 27711.
ill
-------
ABSTRACT
The development of Che MESOPUFF II regional-scale air quality
model is described. MESOPUFF II is a Lagrangian variable-trajectory
puff superposition model suitable for modeling the transport,
diffusion and removal of air pollutants from multiple point and area
sources at transport distances beyond the range of conventional
straight-line Gaussian plume models (i.e., beyond -v 10-50 tan). It
is an extensively modified version of the MESOscale PUFF (MESOPUFF)
model (Benkley and Bass 1979). Major additions and enhancements
include: use of hourly surface meteorological data and twice-daily
rawinsonde data; separate wind fields to represent flow within and
above the boundary layer; parameterization of vertical dispersion in
terms of micrometeorological turbulence variables; parameterization of
SO to S0° and NO to NO. conversion, including the chemical
^ *T X J
equilibrium of the HNO-/NH./NH4N03 system; resistance modeling of dry
deposition! including options for source or surface depletion; time-
and space-varying wet removal; and a computationally efficient puff
sampling function. The scientific and operational bases for these
developments are described. The results of a preliminary evaluation
of several model algorithms during a two-day period of the Tennessee
Plume Study are also presented.
This report was submitted in fulfillment of Contract
No. 68-02-3733 by Environmental Research & Technology, Inc. under
sponsorship of the U.S. Environmental Protection Agency. This report
covers the period from February 11, 1982 to March 15, 1983, and work
was completed as of September, 1983.
IV
-------
CONTENTS
Preface
Abstract iv
Figures v*
Tables vii
Acknowledgements viii
1. Introduction 1
1.1 Background 1
1.2 MESOPUFF II Modeling Package 2
1.3 Major Features of MESOPUFF II 4
1.4 Tennessee Plume Study 8
2. Technical Developments 10
2.1 Wind Field 10
2.2 Micrometeorological Parameters 14
2.3 Dry Deposition - Three Layer Model 21
2.4 Chemical Transformations 33
2.5 Wet Removal 30
2.6 Puff Sampling Function 55
3. Demonstration Model Run 62
References 72
Appendices
A. Reactions and rate constants of the Atkinson
et al. (1982) chemical mechanism 77
-------
FIGURES
Number
1 MESOPUFF II Modeling Package
2 Schematic Representation of Puff Superposition
Approach ..................... 5
3 Particle Deposition to Water Surfaces ....... 24
4 S(>2 Deposition Velocities ............. 25
5 Comparison of Source Depletion and Surface
Depletion Models ................. 31
6 Optional Three Layer System Used in MESOPUFF II. . . 32
7 S02 Oxidation Pathways .......... . ..... 35
8 NOX Oxidation Pathways ............... 36
9 NH^N03 Dissociation Constant Temperature
and Relative Humidity Dependence ......... 39
10 Comparison of S02 Oxidation Rates Predicted by
the ERT and Gillani et al. Equations ....... 47
11 Average Plume Sulfur Conversion Rate as a Function
of Mean Time of Day of Plume Transport ...... 48
12 Washout Ratio as a Function of Precipitation Rate
for Different Storm Types ............ 51
13 802 Washout Ratio as a Function of pH and
Temperature for Equilibrium Scavenging
Conditions .................... 53
14 Location of Cumberland Steam Plant, Surface
Meteorological Stations and Upper Air Rawinsonde
Stations on Meteorological Grid .......... 64
15 Boundary Layer Growth and Plume Fumigation ..... 67
16 Observed and Predicted Mixing Heights in the
Vicinity of the Cumberland Steam Plant ...... 68
VI
-------
TABLES
Number Page
1 Major Features of MESOPUFF II 7
2 Options for Lower and Upper Wind Fields 12
3 Solar Radiation Reduction Factor B 17
4 Factors Influencing Dry Deposition Removal Rates. . 22
5 Summertime SC>2 Canopy Resistances as a
Function of Land Use Type and Stability Class . . 28
6 Parameter Variations in the Photochemical Modeling
Simulations 43
7 Default Values of the Scavenging Coefficient,
X (s"1) 54
6 Conversion of Reported Precipitation Type/
Intensity to Precipitation Codes 56
9 Effect of Sampling Rate, N, on Predicted Near-
Field Concentrations for Two Sampling Algorithms. 60
10 Model Run Parameters Used in Demonstration Run. . . 65
11 History of Puff Released 8/23/78 Hour 1 66
12 Observed and Predicted S02 Conversion Rates .... 70
Vll
-------
ACKNOWLEDGEMENTS
The authors wish Co acknowledge Che contributions made by Drs.
A. Venkatram and R. YamarCino Co Che developmenC of MESOPUFF II. The
assistance and advice of Che EPA project officer, James Godowitch, is
appreciated.
Vlll
-------
SECTION 1
INTRODUCTION
1.1 Background
The regional and long-range transport and transformation of sulfur
oxides and nitrogen oxides emitted from major point sources are of
increasing concern. Motivated by the need for easily-used, cost-efficient
mesoscale air quality models suitable for regulatory applications, the
National Oceanic and Atmospheric Administration (NOAA) sponsored a study by
Environmental Research & Technology, Inc. (ERT) to develop, compare, and
evaluate a set of mesoscale models and related processor programs known as
the MESO-models (Benkley and Bass 1979a, b, c; Morris et al. 1979; Scire et
al. 1979). One of these models, the MESOscale PUFF (MESOPUFF) model appears
to be well suited for regulatory use. For this reason, the Environmental
Protection Agency (EPA) has sponsored a second study by ERT to enhance the
capabilities and flexibility of the MESOPUFF model to meet the current and
future needs of EPA in predicting mesoscale transport of pollutants,
especially secondary aerosols.
This report is the first volume of a two-volume set describing the
results of this effort to extent MESOPUFF1s capabilities. Extensive
modifications have been made to MESOPUFF in order to refine and enhance its
treatment of advection, vertical dispersion, removal and transformation
processes. The new model has been designated MESOPUFF II. The objective of
tnis document is to describe the scientific and operational bases for the
most significant modifications made to MESOPUFF. In addition, this document
provides the results of a demonstration run of the model for a two-day
period during the Tennessee flume Study (TPS). The companion report,
entitled "User's Guide to the MESOPUFF II Model and Related Processor
Programs" provides a summary of the basic model equations and includes a
-------
complete set of user instructions for the MESOPUFF II model and its
processor programs (READ56, MESOPAC II, MESOFILE II). The User's Guide also
contains a description of several model algorithms not presented in this
document that were unchanged or only slightly modified (e.g., the puff
trajectory function, dispersion coefficient calculations and the plume rise
algorithm).
In the next section, the MESOPUFF II modeling package is outlined and
the functions of each program are discussed. Section 1.3 contains a summary
of the major modifications made in MESOPUFF II. The TPS is briefly
described in Section 1.4. The second chapter contains a technical
description of model algorithms. A description of a demonstration run of
the model for two days during the TPS and the results are contained in the
third chapter.
1.2 MESOPUFF II Modeling Package
The MESOPUFF II model is one element of an integrated modeling
package. This modeling package (Figure 1) also contains components for
preprocessing of meteorological data (READ56, MESOPAC II) and postprocessing
of predicted concentration results (MESOFILE II). Each component of the
MESOPUFF II modeling package is briefly described below.
READ56 is a preprocessor program that reads and processes the
twice-daily upper air wind and temperature sounding data available from the
National Climatic Center (NCC) for selected stations. READ56 extracts the
data required by the MESOPAC II program from a standard-formatted NCC tape
(TDF5600). READ56 scans the upper air data for completeness; warning
messages are printed to flag missing or incomplete soundings. A file of
processed sounding data is created in a format convenient for possible
editing by the user and it is subsequently input into the MESOPAC II program.
MESOPAC II is the meteorological processor program that computes the
time and space interpolated fields of meteorological variables (e.g.,
transport winds, mixing height) required by MESOPUFF II to describe
-------
REAOS6 Control
Inputs
(TOFS600
Format!
REAOS6 Upper Air
ProprocaMOf Proo/am
Formatted Twiea
Daily Ravnnaonda
Data Pile*
(MESOPAC II
lontrol Paramatar
Inputs
(TO98S7 Formal*
MESOPAC II Metaorotagica
"am
Hourly
Ozona
Maasuremann
CMESOPUFF II
lontrol Paramaiar
Ingun
MESOPUFF II DISPERSION MODEL
Coneantranon
TaMm
(MESOFILE II
^ntrol Paramatar
Figure 1 MESOPUFF II Modeling Package
-------
mesoscale transport and dispersion processes. MESOPAC II reads the upper
air data files created by RE AD 5 6 and files of standard-formatted NCC hourly
surface meteorological data (CD144) and hourly precipitation data (TD9657).
A single output file containing observed and derived meteorological fields
is produced which serves as an input file to MESOPUFF II.
MESOPUFF II is a Gaussian, variable-trajectory, puff superposition
model designed to account for the spatial and temporal variations in
transport, diffusion, chemical transformation and removal mechanisms
encountered on regional scales. With the puff superposition approach, a
continuous plume is modeled as a series of discrete puffs (Figure 2). Each
puff is transported independently of other puffs. A puff is subject to
growth by diffusion, chemical transformations, wet removal by precipitation,
and dry deposition at the surface. Up to five pollutants may be modeled
s imu11 aneously.
MESOFILE II is a postprocessing program that operates on the
concentration file produced by MESOPUFF II. The postprocessing functions
available with MESOFILE II include flexible time averaging of gridded or
non-gridded (discrete) receptor concentrations, line printer contour plots
of concentration fields, statistical analysis of point-by-point or bulk
differences between concentration fields, and summing and scaling
capabilities.
1.3 Major Features of MESOPUFF II
The original MESOPUFF model is a single-layer, two species puff
superposition model. Its meteorological preprocessor (MESOPAC) creates
gridded fields of wind components, mixing height, and stability class from
twice-daily rawinsonde (upper air) data. Chemical transformation of sulfur
dioxide to sulfate is modeled with a spatially and temporally constant
transformation rate. Dry deposition is modeled with a constant deposition
velocity for each pollutant by the source depletion technique.
-------
Figure 2 Schematic Representation of Puff Superposition Approach
I
-------
Table 1 outlines the most important modifications made in MESOPUFF II
and its processor programs. Each of these changes is discussed in detail in
Chapter 2. MESOPAC II supplements twice-daily rawinsonde data with hourly
surface data to construct wind fields at two levels. The greater temporal
and spatial resolution of the surface data allows improved treatment of
plume transport. Wind fields are determined at two user-selected levels; a
lower level to represent boundary layer flow and an upper level to represent
flow above the boundary layer.
The additional information contained in the surface meteorological
observations allows calculation of important micrometeorological variables
that determine the structure of the boundary layer (i.e., surface friction
velocity, u^, convective velocity scale, w^, Monin-Obukhov length, L, and
boundary layer height, z-). These variables are computed by MESOPAC II
from surface meteorological data and surface characteristics (i.e., land
use, roughness length) provided by the user for each grid point.
MESOPUFF II has been expanded to accommodate up to five pollutants:
sulfur dioxide (SO ), sulfate (SO*), nitrogen oxides (N0x = NO + N02),
nitric acid (HNO-) , and nitrate (N0~) . Chemical transformation rate
expressions have been developed from the results of photochemical model
simulations over a wide range of environmental conditions. The rate
expressions include gas phase NO oxidation, and gas/aqueous phase SO^
oxidation. The HNO-j/NH./NH.NO- chemical equilibrium relationship
has also been incorporated into the model.
The dry deposition of pollutants is treated in MESOPUFF II with a
resistance model. The pollutant flux is proportional to the inverse of a
sum of resistances of pollutant transfer through the atmosphere to the
surface. The resistances depend on the characteristics of the pollutant,
the underlying surface, and atmospheric conditions. MESOPUFF II contains
options for the commonly used source depletion method of pollutant removal
by dry deposition (i.e., pollutant is removed from the entire depth of the
puff) or the more realistic surface depletion treatment (i.e., material is
removed only from the surface layer) with a 3-layer submodule.
-------
TABLE 1. MAJOR FEATURES OF MESOPUFF II
Uses hourly surface meteorological data and upper air
rawinsonde data
Wind fields constructed for two layers (within boundary
layer, above boundary layer)
Boundary layer structure parameterized in terms of
micrometeorological variables u*, w*, z^, L
Up to five species (e.g., S02, 804, NOX, HN03,
NO 3)
Space- and time-varying chemical transformations
Space- and time-varying dry deposition; resistance model;
source or surface depletion
Space and time-varying wet removal
Efficient puff sampling function.
-------
Precipitation scavenging is frequently Che dominant pollutant removal
mecnanism during precipitation periods. MESOPUFF II contains a scavenging
ratio formulation for wet removal. The scavenging ratio depends on both the
type and rate of precipitation, and the characteristics of the pollutant.
Improvements in MESOPUFF II have been made in the method which
evaluates and sums the contributions of individual puffs to the total
concentration. The model uses an integrated form of the puff sampling
function that eliminates the problem of insufficient puff overlap commonly
encountered with puff superposition models. This development allows
continuous plumes to be accurately simulated with fewer puffs, thereby
saving computational time and reducing computer storage requirements.
1.4 Tennessee Plume Study
The TPS was conducted in August 1978 as part of EPA1 s Sulfur Transport
and Transformation in the Environment (STATE) program. The experimental
study was conducted in the vicinity of the TVA Cumberland Steam Plant. The
Cumberland Plant is a base load 2600 MW plant which is located in
nortnwestern Tennessee. Although sampling the Cumberland plume was the main
objective of the experiment, plume measurements from other TVA plants (e.g.,
Johnsonville, Paradise and Gallatin) were made when they were transported
near the Cumberland plume. The trajectory of the Cumberland plume was
determined by tracking tetroons, a manned LAMP balloon, and ground and
airborne sampling of a tracer gas (SFg). Dispersion and chemical
measurements were obtained by aircraft and ground-based mobile vans. Four
specific scenarios were studied:
Vertical mixing during highly convective conditions to downwind
distances of 50 km.
Horizontal plume spread during stable conditions with significant
wind shear to downwind distances of 300-500 km. The initial plume
is emitted into the mixed layer.
Dispersion during stable conditions to distances of 400 km. The
initial plume is emitted into a stable layer.
8
-------
Dispersion and chemical changes over a diurnal cycle, with
fumigation in the morning and layering in the evening.
Although a detailed evaluation of MESOPUFF II with the TPS data base is
beyond the scope of the current study, a demonstration run for a two-day
period during the TPS has been made. The purposes of the demonstration run
were to allow a preliminary assessment of the S02 to SO^ chemical
transformation formulation for one of the scenarios (Scenario 4) and to
qualitatively demonstrate the behavior of several other model algorithms.
-------
SECTION 2
TECHNICAL DEVELOPMENTS
2.1 Wind Fields
A principal concern in long range transport modeling is the spatial and
temporal resolution of the data used to construct the wind field for plume
advection. The spatial resolution of routinely available NWS rawinsonde
data is only marginally adequate for long range transport modeling. The
typical distance between rawinsonde stations is 300-500 kilometers. Another
limitation is the poor temporal resolution of the routinely available
twice-daily sounding data. Important variations in the wind field, mixing
height, and atmospheric stability occur on much smaller space- and
time-scales than those resolvable by the NWS upper-air sounding network.
To increase the spatial and temporal resolution of the meteorological
data used in MESOPUFF II and to obtain a better representation of the
boundary layer flow, the meteorological preprocessor program has been
modified to allow the twice-daily upper air data to be supplemented with
hourly surface data from the much denser network of NWS surface stations.
In addition, wind fields are constructed at two levels: a lower level field
representing boundary layer flow, and an upper level wind field representing
flow above the boundary layer. The lower level winds are used to advect
puffs within the mixed layer and to determine the plume rise of newly
released puffs. The upper level winds are used to advect puffs which are
above the boundary layer. At each time step, the appropriate wind field for
advection of a puff is determined by comparison of the height of the puff
center with the spatially and temporally varying mixing height. If the puff
center is above (below) the mixing height at the closest grid point, the
entire puff is advected with the upper (lower) level wind.
10
-------
Considerable flexibility is allowed in choosing Che most appropriate
level or vertically-averaged layer for each wind field. Table 2 contains
the available options. The default instructions are to use the winds
averaged through the mixed layer for the lower level wind field, and the
wind averaged from the top of the mixed layer through the 700 mb level
(v. 3000 m) for the upper level wind field. However, if desired, the user
may select other levels to determine the wind fields (e.g., surface and
850 mb levels). The model may be made a single wind field model by
specifying the lower and upper wind fields to be the same.
The mixed layer averaged winds are calculated from twice-daily
rawinsonde data from upper air stations and hourly surface data from the
typically much denser network of surface stations. Layer-averaged wind
speed and wind direction computed from the rawinsonde data are used to
adjust the hourly surface winds. The following five step procedure, adapted
from Draxler (1979), is used to determine the mixed layer wind at each point:
(1) A representative rawinsonde sounding (00 or 12 GMT) is selected
based upon the stability class at the nearest surface station to
the grid point and the time of day. Neutral/unstable and stable
conditions are assumed to be represented by the 00 GMT and 12 GMT
sounding, respectively.
(2) Using the sounding selected in Step (1), vertically averaged u
(easterly) and v (northerly) wind components are computed through
the layer from the surface to the grid point mixing height.
(3) The ratio, R, of the layer-averaged wind speed to the surface wind
speed at the rawinsonde station, and the angular difference in
wind direction, A9, between the layer averaged and surface winds
are calculated.
(4) The hourly surface wind data are used to calculate spatially
interpolated surface wind components (u , v ) at each grid
S 9
point. Data from all surface stations within a user-specified
11
-------
TABLE 2. OPTIONS FOR LOWER AND UPPER WIND FIELDS
Option Meteorological Data
Vertically Averaged Winds
Surface to mixing ht ' Surface, Rawinsonde
Mixing ht to 850 mb Rawinsonde
Mixing ht to 700 mb(2) Rawinsonde
Mixing ht to 500 mb Rawinsonde
Single Level Winds
Surface Surface
850 mb Rawinsonde
700 mb Rawinsonde
500 mb Rawinsonde
^Default lower level wind field
^Default upper level wind field
12
-------
'scan-radius1 of a grid point are used to compute (ug, vg)
according to
_
t r 2 ' (V V
(u , v ).. - J5 2 (2-1)
s' s ij «_
1 r 2
k s
where u , v are the easterly and northerly components of the
s s
surface wind at grid point (i, j),
u,, v are the easterly and northerly components of the
surface wind at surface station k,
r is the distance from the surface station to grid point
3
(i, j), and
a is an alignment weighting factor (o = 1-0.5 I sin $\>
S 99
where
-------
Vertically averaged winds from the mixing height to the 850 mb, 700 mb
or 500 mb levels are computed in the following manner. The 00 GMT and 12
GMT winds at each rawinsonde station are first interpolated in time, and
then vertically averaged through the layer from the grid point mixing height
to the selected level (e.g., 700 mb). The winds at grid point (i, j) are
obtained by Equation (2-1), with the summation over rawinsonde stations
instead of surface stations. Only rawinsonde stations within a
'scan-radius* of the grid point are considered. The mixing height must be
lower than the pressure level that defines the top of the layer, otherwise,
an error message is printed and execution of the program is terminated.
If one of the single-level upper air wind fields (e.g., 850 mb, 700 mb,
or 500 mb) is chosen, only the wind data at the selected level is used to
construct the wind field. For example, the 850 mb wind at each grid point
is calculated by interpolating the 850 mb winds at each rawinsonde station
over time, and then applying Equation (2-1) with the summation over the
rawinsonde stations.
2.2 Micrometeorological Parameters
Boundary layer turbulence is generated by convective and mechanical
processes. Convective or buoyancy-induced turbulence is produced by a
positive (upward) heat flux at the ground which is driven by solar heating.
Mechanical mixing originates from shear-induced turbulence which is caused
by frictional interaction of the wind with the earth's surface. The
structure of the boundary layer can be described in terms of a small number
of micrometeorological variables; the surface sensible heat flux, H, the
surface friction velocity, u., and the boundary layer height, z.. Many
** L
studies (e.g., Deardorff and Willis 1975, van Ulden 1978) have shown the
importance of these and related parameters in boundary layer meteorology.
MESOPAC II uses simple empirical relationships to estimate
micrometeorological parameters from routinely available meteorological
measurements.- Vertical dispersion and dry deposition of pollutants are
parameterized in MESOPUFF II in terms of these variables. Horizontal and
14
-------
near-field vertical puff growth continue to use dispersion formulas which
require classification of stability into P6T classes.
The following sections describe the methods used to obtain the
micrometeorological parameters needed by MESOPUFF II from routinely
available meteorological data.
2.2.1 Surface Friction Velocity
The surface friction velocity, u^, can be computed from routinely
available meteorological data if the surface roughness characteristics are
known. First, the sensible heat flux is calculated from an estimate of net
radiation. Then u^ is determined from the wind speed, surface roughness,
z , and heat flux.
The sensible heat flux, H, is estimated during daylight hours by the
following equations (Maul 1980):
H = a R + H (2-2)
o
R = 950 8 sin u (2-3)
H = 2.4 C - 25.5 (2-4)
o
where,
_2
H is the sensible heat flux (Win ) ,
H is the heat flux in the absence of solar incoming radiation
(Wm~2) ,
a is a land use constant, (^ 0.3),
_2
R is the incoming solar radiation (Wm ),
0 is a radiation reduction factor due to the presence of clouds,
u is the solar elevation angle, and
C is the opaque cloud cover (in tenths).
15
-------
Table 3 contains default values for the solar radiation reduction factor
(3) due to the different cloud amounts. The values of 6 are adapted
from those used by Maul (1980).
The sine of the solar elevation angle, sin u, is given by:
sin u = sin $ sin K, + cos 4 cos K. cos H. (2-5)
d d A
HA = (ir/12) (T - E ) - X (2-6)
A m
E = 12. + 0.12357 sin (D) - 0.004289 cos (D) (2-7)
m
+ 0.153809 sin (2D) + 0.060783 cos (2D)
D - (d-1) (360.7365.242)(ir/180) (2-8)
KD = sin"1 (0.39784989 sin (ir oA/i80)) (2-9)
OA = 279.9348 + D(180/w) + 1.914827 sin (D) (2-10)
A
-0.079525 cos (D) + 0.019938 sin (2D) - 0.00162 cos (2D)
where
-------
TABLE 3. SOLAR RADIATION REDUCTION FACTOR B
Cloud Cover (Tenths) JL
0 1.00
1 0.91
2 0.84
3 0.79
4 0.75
5 0.72
6 0.68
7 0.62
8 0.53
9 0.41
10 0.23
17
-------
z = z - 4 z (2-13)
m ms o
H/(p cp) (2-14)
« * 3
x. ° u*
Q =-7 (2-15)
0 k 8 z
0.128 + 0.005 In (z /z ) z /z < 0.01 (2-16)
o m on*
0.107 z /z > 0.01
o m
b = 1.95 + 32.6 (z /z )°'45 (2-17)
o m
where,
k is Che von Karman constant
c is the specific heat of air at constant pressure
(996 m2/(s2 deg)),
u^ is the surface friction velocity (m/s),
u is the wind speed (m/s) measured at height z (m),
m ms
z is the surface roughness (m), and
p is the density of air (kg/m ).
During stable conditions, u# is determined by the following method
(Venkatram 1980):
cO-3-
DN In (zm/zo)
(2-19)
IS
-------
4u2
1 2__ oo (2-20)
CDNum
u 2=12 (2-21)
o k A
where y and A are constants with values of 4.7 and 1100, respectively, and
C is the neutral drag coefficient.
ON
2.2.2 Monin-Obukhov Length
The Monin-Obukhov length, L, is defined as:
where T is the observed air temperature and g in the acceleration due to
gravity. During unstable conditions, L is calculated directly from its
definition using values of u* and Q derived earlier. During stable
conditions, L is given by Venkatram (I980b) as:
L =» 1100 u*. (2-23)
2.2.3 Mixed Layer Height
During daylight hours, solar radiation reaching the ground produces a
positive (upward) flux of sensible heat and the development of a well-mixed
adiabatic layer. If the hourly variation of H is known, the mixed layer
height, z., at time t + I can be estimated from z. at time t in a
stepwise manner (Maul 1980).
19
-------
> (Z) 11/2
2 v 2H(l*E)At ^)t(Vt| +n^±i (2.24)
.EHAt^ 1/2
(2-25)
where
i|>. is Che potential temperature lapse rate in the layer above z^,
At is the time step (3600 s) ,
E is a constant (-0.15), and
A9 is the temperature discontinuity at the top of the mixed layer.
The lapse rate, <|»., is determined through a layer Az meters above the
previous hour's convective mixing height. For daytime hours up to 23 GMT,
the morning (12 GMT) sounding at the nearest rawinsonde station is used to
calculate i|» .. After 23 GMT, the evening (00 GMT) sounding is used. To
avoid computational problems, <|», is not allowed to be less than 0.001
°K/m.
The neutral (shear produced) boundary layer height is given by
Venkatram (1980) as:
B u^
z (2-26)
where £ is the Coriolis parameter,
B is a constant (/T) , and
N_ is the Brunt-Vaisala frequency in the stable layer aloft.
The daytime mixing height is the maximum of the convective and mechanical
values predicted by Equations 2-25 and 2-26.
20
-------
la Che stable boundary layer, mechanical turbulence production
determines the vertical extent of dispersion. Venkatram (1980b) provides
the following empirical relationship to estimate z. during stable
conditions.
z.
i
2400 u.372 (2-27)
2.2.4 Convective Velocity Scale
During convective conditions, turbulence is generated primarily by the
sensible heat flux originating from the ground. The appropriate velocity
scale during these conditions is the convective velocity, w^.
The convective velocity can be calculated directly from its definition,
since Q and z. have been determined from Equations 2-14 and 2-24,
respectively.
2.3 Dry Deposition - Three-Layer Model
The rate at which pollutants are deposited at the surface depends on
many factors: the state of the atmosphere, the characteristics of the
surface, and the properties of the pollutant. For example, the rate of
deposition can sometimes be limited by the rate of pollutant transfer to the
surface by atmospheric diffusion processes. Due to the importance of
vegetation as a sink for atmospheric pollutants, the structure of the canopy
and the physiological state of the vegetation are also important factors.
Tne properties of a pollutant such as its solubility, molecular diffusivity
and for larger particles, the size and shape of the particles are additional
factors that influence the rate of deposition. Table 4 contains a listing
21
-------
TABLE 4. FACTORS INFLUENCING DRY DEPOSITION REMOVAL RATES
MicromeCeorology
variables
Aerodynamic roughness:
-Mass transfer
(a) Particles
(b) Gases
-Heat
-Momentum
Atmospheric stability
Diffusion, effect of:
-Canopy
-Diurnal variation
-Fetch
Flow separation:
-Above canopy
-Below canopy
Friction velocity
Inversion layer
Pollutant concentration
Relative humidity
Seasonal variation
Solar radiation
Surface heating
Temperature
Terrain:
-Uni form
-Nonuniform
Turbulence
Wind velocity
Zero-plane displacements:
-Mass transfer
(a) Particles
(b) Gases
-Heat
-Momentum
Depositing Material
Particles
Agglomera t ion
Diameter
Density
Diffusion:
-Brownian
-Eddy equal to
(a) Particle
(b) Momentum
(c) Heat
-Effect of canopy on
Dif fusiophoresis
Electrostatic effects:
-Attraction
-Repulsion
Gravitational settling
Hydroscopicity
Impaction
Interception
Momentum
Physical properties
Resuspension
Shape
Size
Solubility
Thermophores is
Gases
Chemical activity
Diffusion:
-Brownian
-Eddy
Partial pressure
in equilibrium
with surface
Solubility
Surface Variables
Accommodat ion:
-Exudates
-Trichomes
-Pubescence
-Wax
Biotic surfaces
Canopy growth:
-Dormant
-Expanding
Senescent
Canopy Structure:
-Areal density
-Bark
-Bole
-Leaves
-Porosity
-Reproductive structure
-Soils
-Stem
-Type
Electrostatic properties
Leaf-vegetation:
-Boundary Layer
-Change at high winds
-Flutter
-Stomatal resistance
Hon-biotic surfaces
pH effects on:
-Reaction
-Solubility
Pollutant penetration and
distribution in canopy
Prior deposition loading
Water
From: Sehmel (1980)
22
-------
compiled by Sehmel (1980) of Che variables believed to be most important in
influencing dry deposition rates.
The dry flux of a pollutant can be written (Slinn et al. 1978) as:
F = F + v C (2-29)
d z g
where
F is the total (downward) dry flux,
F is the turbulent and molecular diffusive flux,
Z
v is the average drift velocity due to gravitational
O
settling and phoretic effects, and
C is the pollutant concentration.
For larger particles (diameters >1 urn), gravitational settling and
particle inertia become increasingly important effects. Brownian diffusion
dominates the mass transfer of gases and small particles (diameters
<0.1 um) in the near surface quasi-laminar layer. As shown in Figure 3,
a minimum in deposition velocity is observed from particles in the range
0.1-1.0 um where these mechanism are less effective (Hicks 1982). Most
models of dry deposition use the concept of a deposition velocity
(Chamberlain and Chadwick 1953) to express the total dry flux:
F = v. C (2-30)
d d
where v is the deposition velocity (including both gravitational and
diffusive effects) at a reference height.
Due to the number and variability of the factors influencing deposition
rates, reported deposition velocities exhibit considerable variability. For
example, SO. deposition velocities summarized by Sehmel (1980) range over
two orders of magnitude (Figure 4). Although it is not possible to include
the effects of all the variables listed in Table 4 in determining v , it
is possible to improve upon the assumption commonly used in mesoscale models
23
-------
£
o
o
o
^ 0.1
CO
o
a.
LJ
o
0.01
WATER
0.01
0-1 1 10
PARTICLE DIAMETER (/im)
Figure 3 Particle Deposition to Water Surfaces. Solid Circles are
Due to Moller and Schumann (1979), Open Circles to Sehmel
and Sutter (1973). The Dashed Line at the Right Represents
the Terminal Settling Speed for 1.5 g cm-3 Particles.
Source: Hicks (1982)
24
-------
626-ST. LOUIS-1975
621-ST. LOUIS-1973
584-HEDGE
619-WATER LAPSE ATM.
56C - FejQ, MAX RATE
Me - GRASS; 0 STABILITY
54-ALFALFA
610-CRASS. NEUTRAL ATM.
55a - CEMENT MAX RATE
61a - CRASS. LAPSE ATM.
49-GRASS
61h- WATER, NEUTRAL ATM,
SI - GRASS
»b - CEMENT. MAX RATE
52a - FOREST
52(1-GRASS. MEDIUM
55c - STUCCO; MAX RATE
He- GRASS. 0 STABILITY
554 - CEMENT. MAX RATE
610-SNOW. LAPSE ATM.
59 -GRASS
57-GREAT BRITAIN
529 - SOII, CALCAREOUS
Mb-WATER, B STABILITY
Sfta-SOIl ADOBE CLAY-MAX
55«-STUCCO. MAX RATE
Mb-WATER, B STABILITY
55 e - STUCCO. MAX RATE
60a-WHEAT
58f- GRASS. 0 STABILITY
Wa-CRASS, B STABILITY
551-SOII. ADOBE CLAY-MAX
559-SOI I. SANDY LOAM-MAX
56b - SOIL. SANDY LOAM-MAX
Mb-FOREST. 17m
5«9- WATER, 0 STABILITY
52c - GRASS. SHORT
6K-SNOW. NEUTRAL ATM.
61c-GRASS. STABLE ATM.
52b-WATER FRESH
50-SNOW
5' - ICE
611-SNOW. LAPSE ATM.
611 - SNOW. STABLE ATM.
55 h-ASPHALT. MAX RATE
L
I '
REFERENCE
X
A
O-D
O
0
A
a
X
-a
XX
CUD
D-a
A
A
A
XX
a
v
a
x
A "MAXIMUM" RATES
O GRASS
X WATER
V SNOW
O OTHER
J_
'.-I
ur> i 10
DEPOSITION VELOCITY, em/see
Figure 4 S02 Deposition Velocities
Source: Sehmel (1980)
25
-------
of spatially and temporally constant deposition velocities. In MESOPUFF II,
the deposition velocity is expressed as the inverse of a sum of resistances
to transfer of the pollutant through the atmosphere to the surface.
v, = (r + r + r )~ (2-31)
d a s c
where r is the aerodynamic resistance (s/m),
r is the surface resistance (s/m), and
S
r is the canopy resistance (s/m).
The aerodynamic resistance is the resistance to pollutant transfer
through the atmospheric surface layer. It is a function of wind speed,
atmospheric stability, and surface roughness. Except for very large
particles, the aerodynamic resistance for gases and particles is the same.
The surface resistance represents the resistance to transfer across the
quasi-laminar layer surrounding smooth surfaces. Wind tunnel studies have
shown that the thickness of this layer is typically about SO um (Hicks
1982). However, surface roughness elements can sometimes penetrate this
layer, providing an alternative route for the transfer. Therefore, rfe is
an average value of this resistance. The canopy resistance is the
resistance to transfer within the surface or plant constituting the final
resting place for the pollutant. The canopy resistance depends on the
characteristics of the pollutant (e.g., solubility) as well as the
physiological properties of the vegetation.
as:
The aerodynamic resistance, r , is given by Wesely and Hicks (1977)
(k u^)'1 Un(zs/zQ) -*H1 (2-32)
-5z /L 0 < z /L < 1
8 S (2-33)
exp [0.598 + 0.39 In (-z /L) - -1 < z /L < 0
S S
0.090 Un(-zs/L)}2]
26
-------
where z is Che reference height (10 meters in MESOPUFF II),
z is the surface roughness length (m),
o
u^ is the friction velocity (m/s),
(|)u is a function accounting for stability effects,
n
k is the von Karman constant, and
L is Monin-Obukhov length (m) .
as:
The surface resistance, r , can be expressed (Wesely and Hicks 1977)
(k u)" kB* (2-34)
where B~ is the surface transfer coefficient.
For SO., kB~L =2.6 (Wesely and Hicks 1977). The other gaseous
pollutants in MESOPUFF II (e.g., NO , HNO-) are assumed to have similar
-1 x J
values of kB . For particles, r is a complex function of many
factors. Depending upon the pollutant size distribution, particle inertia
and gravitational settling effects may be important. Given current
uncertainties regarding r for particles, r is simply assumed constant
3 - 8 9
for SO, and NO- with a default value of 10 a/cm. Although Wesely
and Hicks (1977) suggest r may be as low as 1 s/cm, the larger value is
presently used in the model to be consistent with deposition velocities of
tO.1 cm/a found in other studies (e.g., Garland 1978) for sulfate.
Shieh et al. (1979) estimate canopy resistance for SO. as a function
of land use and stability class for summertime conditions. These values,
contained in Table 5, are used as default values in MESOPUFF II. It should
be noted that these values are based only on expected midsummer conditions.
More appropriate values (e.g., for snow covered surfaces) may be entered for
model applications during other seasons.
Based upon its high solubility and reactivity, r for HNO.J is
assumed equal to zero (Hicks 1982). Canopy resistance for N0x are
27
-------
TABLE 5. SUMMERTIME S02 CANOPY RESISTANCES AS A
FUNCTION OF LAND USE TYPE AND STABILITY CLASS
Category Land Use Type
1 cropland and pasture
2 cropland, woodland and grazing
land
3 irrigated crops
4 grazed forest and woodland
5 ungrazed forest and woodland
6 subhumid grassland and semiarid
grazing land
7 open woodland grazed
8 desert shrubland
9 swamp
10 marshland
11 metropolitan city
12 lake or ocean
0.20
Stability Class
A.B.C D E
100. 300. 1000.
0.
0.30
0.05
0.90
1.00
0.10
0.20
0.30
0.20
0.50
i-aj
10 4
100.
100.
100.
100.
100.
100.
200.
50.
75.
1000.
0.
300.
300.
300.
300.
300.
300.
500.
75.
300.
1000.
0.
1000.
1000.
1000.
1000.
1000.
1000.
1000.
100.
1000.
1000.
0.
0.
0.
0.
0.
0.
0.
1000.
0.
0.
0.
0.
From: Shieh, Wesely, and Hicks (1979).
28
-------
1.3 s/cm (A-C stability), 5 a/cm (D stability), and 15 s/cm (E-F
stability). Uptake of the particles S0~ and N03 by plant stomata
is less relevant; therefore, total resistance for SO^ and NO-j is
determined by r and r (i.e., r = 0).
d S C
With knowledge of the concentration and the deposition velocity, the
pollutant flux is determined by Equation 2-30. MESOPUFF II has two options
for treating the removal of pollutant from a puff. The first option is the
commonly used source depletion approximation. This method assumes that
material deposited is removed from the full depth of the puff. The change
in mass is:
Q(t) exp " ' ' - (2"35)
Where Q(t), Q(t+l) is the mass (g) of pollutant in the puff at the
beginning and end of the time step,
s, s + A s is the position of the puff at the beginning and
end of the time step, and
g(s) is the vertical term of Che Gaussian puff equation as given
by Equation 2-59. For a puff uniformly mixed in the vertical,
g(s) - l/zi.
The source depletion model effectively enhances the rate of vertical
diffusion of the pollutant because mass removed at the surface is
immediately replaced with material from above. However, in the atmosphere,
the rate of deposition can be limited (usually only during stable
conditions) by the rate of pollutant mass transfer through the boundary
layer to the surface layer. This overall boundary layer resistance is not
included in the aerodynamic resistance. Horst (1977) suggests that the
source depletion model may introduce a bias in the deposition flux.
Excessively high deposition fluxes and concentrations may be predicted by
the source depletion model in the near-field, and as a result, the
concentrations and deposition fluxes may be underpredicted further
29
-------
downwind. This effect is illustrated in Figure 5 where the source depletion
model results are compared to those of the surface depletion model of Horst
(1977).
To account for the effect of boundary layer mixing, MESOPUFF II has the
option to treat puffs that have become vertically well-mixed with a 3-layer
model (see Figure 6). The surface layer is a shallow layer (10 m) next to
the ground that rapidly adjusts to changes in surface conditions.
Pollutants in the middle layer are uniformly mixed up to the top of the
current boundary layer. The upper layer consists of pollutant material
above the boundary layer dispersed upward during previous turbulent
activity. The pollutant flux into the surface layer is:
Flux = K (C - C )/(«.-«) v. C (2-36)
m s is d s
2
where K is an overall boundary layer eddy diffusivity (m /s),
C is the concentration in the middle layer, and
m
C is the concentration at the top of the surface layer.
During stable conditions, K is given by Brost and Wyngaard (1978) as:
K - k u^ z. (2-37)
and during neutral or unstable conditions K is:
K » Maximum {k, u., z. , k, w. z.} (2-38)
l i i " i
The constants ls~ and k_ have default values of 0.01 and 0.1,
respectively.
The term v. C can be written as v. C , where v^ is an
effective deposition velocity taking into account boundary layer mass
transfer.
30
-------
^v
§
o
8
Surface Depletion Model
z= 1 m
103
Downwind Distance,
Figure 5 Comparison of Source Depletion and Surface Depletion Models.
For vd/u = 10'2, Stable Thermal Stratification (F Stability)
from Horst (1977)
31
-------
«IO»I20
'max
N)
Nonturbulent Atmosphere
Mixed Layer
1
5 Surface Layer
Figure 6 Optional Three Layer System Used in MliSOPUI-F fl
-------
, K V
= S r (2-39)
d < + vd(z.- zs)
In Che 3-layer model, only material in the surface layer is available
for deposition at the surface. The effective deposition velocity, vrf
is used to evaluate the change in pollutant mass in the puff due to dry
deposition. The model predictions are those corresponding to Cg in
Figure 6.
2.4 Chemical Transformations
The accuracy of air quality models for chemically reactive species
depends strongly on the chemical submodel, as well as the transport,
diffusion, and deposition formulations. The fidelity of atmospheric
chemical mechanisms is often limited by the availability of kinetic and
mechanistic data for the species of concern and sometimes the model's
structure. For example, often the atmospheric chemistry of emitted
compounds depends on numerous other compounds formed and destroyed in other
chemical reactions. Not only are the rate constants and products uncertain
for many of these reactions, but also the model's formulation may not allow
for inclusion of intermediate species and/or second-order reactions. The
latter is true for the puff transport/dispersion formulation used in the
MESOPUFF II model. Thus, chemical mechanisms for models such as MESOPUFF II
must be formulated as pseudo-first-order reactions. The accuracy of the
first-order reaction mechanism may be enhanced by parameterization of the
rate constants so as to reflect the characteristics of the higher-order
reaction system.
The chemical process of concern for the MESOPUFF II model are the
conversions of sulfur dioxide (802) to sulfate aerosol (SO.) and
oxides of nitrogen (NO ) to nitrate aerosol (NO.). Although the
X J
atmospheric chemistry of these compounds has been studied for nearly two
decades, substantial uncertainties exist in the current chemical knowledge
of SO and NO reaction pathways and rates under ambient conditions.
33
-------
Laboratory and field studies have shown that chemical transformation rates
for these species can vary several orders of magnitude under different
environmental conditions (Calvert et al. 1978; Wilson 1981; Richards et al.
1981; Newman 1981). It is, therefore, important for the chemical submodel
to incorporate the dependency of transformation rates on environmental
conditions.
A first order reaction mechanism consisting of the following reactions
has been formulated for MESOPUFF II:
S02 * S0° (2-40)
N0x * HN03 (2-41)
NO * RNO- (2-42)
A J
NH.
HN03 * N0~ (2-43)
The rate constants have been parameterized in terms of environmental
conditions such as solar radiation, relative humidity, temperature, and
background ozone concentrations. The parameterizations have been developed
from laboratory data, field data, and analysis of nonlinear chemical
mechanisms for SO and NO oxidation. The following subsections
xx
describe the rationale for and development of the MESOPUFF II chemical
transformation scheme.
2.4.1 Chemical Pathways for Sulfate and Nitrate Aerosol Formation
Research performed during the last twenty years has identified many of
the important pathways for SO. and NO oxidation. Laboratory and field
£ X
studies have shown fine particulate matter to be a major product of 862
oxidation and a minor product of NO oxidation under ambient conditions.
Figures 7 and 8 illustrate the chemical pathways for S02 and NO^
oxidation, and aerosol formation. Oxidation may occur by gas and aqueous
34
-------
ROG,
HO,
Photo-
chemical
eactions
Aqueous
reactions
Aerosol with
Metal Ions
and Carbon
Water Vapor
^
V H2°2
Evaporation
Cloud Water
r
-
r
w
\
Figure 7 S02 Oxidation Pathways
35
-------
ROG
H02, R02, OH,
Figure 8 NOX Oxidation Pathways
36
-------
phase reactions. The gas phase reactions for both SO and NO involve
X «
free radical photochemistry and, thus, are coupled to the oxidation of
reactive organic gases (ROG). The aqueous phase oxidation reactions for
SO and NO are less well understood than the gas phase reactions,
X Jfc
however, photochemical products such as ozone (0^) and hydrogen peroxide
(H-O,) are believed to be the principal oxidants for SOg.
Homogenous gas phase reactions are believed to be the dominant SO^
oxidation pathway in the presence of sunlight and absence of clouds or fog
(Calvert et al. 1978). Three of the most important reactions for S02 are:
S02 + OH + M * HS03 + M (2-44)
S0 + CH0 * S0 + CB0 (2-45)
S03 + CH3CHO (2-46)
In the presence of trace amounts of water vapor, HSO_ and S03 rapidly
form a sulfate aerosol or attach to pre-existing aerosols. Reactions of
HSO- with NO in the presence of 0. may occur but the mechanism remains
uncertain (Calvert et al. 1978). The reaction with the hydroxyl radical
(OH) is believed to be most important. The reactions with the Criegee
biradicals, formed from ozone-alkene reactions, may be important at high
alkene concentrations in urban environments (Atkinson and Lloyd 1980) .
These reactions can produce SO oxidation rates of up to 5% per hour.
SO oxidation may also occur via reactions of the dissolved S(IV)
constituents, primarily bisulfate and sulfite, with dissolved ozone and
hydrogen peroxide (H202) (Maahs 1982; Penkett et al. 1979). The aqueous
phase oxidation may also be catalyzed by Mn , Fe , and/or elemental
carbon (Martin 1982). Recent reviews suggest that oxidation by H^ may
be the dominant process under acidic conditions (Schwartz 1982). Although
there is considerable uncertainty whether H 02 production in the gas or
aqueous phase is sufficient to sustain this reaction, relatively small
amounts of H.O. can produce transformation rates of up to 100% per hour
37
-------
locally in cloud or rain water. Since cloud water is believed to be
recycled (condensed-evaporated) rapidly, the aqueous phase reactions may be
an important pathway for sulfate aerosol formation under cloudy or foggy
conditions (Uegg and Hobbs 1981).
The oxidation of NO is strongly dependent on gas phase
X
R06/NO /O photochemistry and is generally more rapid than SO
oxidation. As shown in Figure 8, NO can be oxidized to nitric acid
X
(HNO3) and organic nitrates (RNOO including oxygenated nitrates such as
peroxyacetylnitrate (PAN). Nitric acid formation occurs primarily by the
reaction of NO- with OH (at a rate ~8 times faster than S02 + OH).
Oxidation of NO to N-Oe (with involvement of 0.) followed by a
heterogeneous reaction with water and reactions of N0« with aromatic
hydrocarbons may also lead to HN03 formation. HNO, is, in turn,
destroyed very slowly by photolysis and reaction with OH.
Nitric acid combines with ammonia gas to form solid or aqueous ammonium
nitrate (NH,NO_). Unlike S0~ formation, the N03 formation process is
reversible. Equilibrium is established between nitric acid, ammonia, and
ammonium nitrate:
NH4N03 (s or aq) * HNO.j(g) + NH3(g) (2-47)
The equilibrium constant
[NH.J [HNO-]
- - 3 .- ^3 (2-48)
is dependent on temperature and relative humidity in a nonlinear manner as
shown in Figure 9 (Stelson and Seinfeld 1982). The equilibrium constant can
vary several orders of magnitude over a typical diurnal cycle. Given fixed
amounts of total nitrate, ammonia, and water vapor, higher NH^NO-j
concentrations are expected at night, due to lower nighttime temperature and
38
-------
0.08
50 60 70 80 90 100
RELATIVE HUMIDITY , %
Figure 9 NlfyNOs Dissociation Constant Temperature and Relative
Humidity Dependence
Source: Stelson and Seinfeld (1982)
39
-------
higher relative humidity. Thus, the nitrate aerosol cannot be considered a
stable product like sulfate. Also, unlike SO^, its ambient
concentrations are limited by the availability of ammonia which is
preferentially scavenged by sulfate (Stelson et al. 1983).
The formation of organic nitrate such as peroxyacetylnitrate (PAN) and
PAN analogs is the second major pathway for NO oxidation. The organic
nitrates are formed primarily by reactions of NO- with RCO., radicals
(such as acetylperoxy). Organic nitrates may also be formed by reactions of
NO with RO radicals and NO with RO radicals. The RCO- radicals are
formed from acetaldehyde and higher aldehydes which are emitted directly by
sources and photoeheroically formed from organic gases. PAN formation rates
are, therefore, strongly dependent on hydrocarbon loadings. The stability
of PAN, and therefore, its net formation rate, strongly depends on
temperature. PAN decomposes into NO. and RCO- at a rate which increases
with temperature. This has lead many scientists to view PAN as more of a
temporary reservoir for NO- than a permanent sink. The results of
multi-day simulations of HC/NO /0_ systems with diurnally varying
temperature and radiation performed at ERT suggest the cumulative formation
of PAN greatly exceeds its decomposition, hence we believe PAN is a major
NO sink. However, in contrast to the HNO pathway, there is no
C J
evidence of nitrate aerosol formation from PAN.
Little is known regarding aqueous phase oxidation of dissolved NOj to
N0~. NO- has low solubility and at this time the only reaction of
importance involves S(IV)aq + N(III)aq (Schwartz and White 1982; Martin et
al. 1981). Its reaction products are unknown. Based on the current
information, the aqueous phase oxidation pathways appears to be far less
important than the gas phase oxidation pathway for NO .
X
2.4.2 Development of a Pseudo-First Order Reaction Mechanism
Rate constant expressions for reactions 1-3 (Equations 2-40 to 2-42)
were developed to represent SO, and NO oxidation under different
£ X
environmental conditions. The gas and aqueous components of the SC
40
-------
oxidation rate were developed separately. Only gas phase oxidation was
considered for NO . The HNO-j/NH-j/NH^NO-j equilibrium relationship
(Equation 2-43) was incorporated directly into the mechanism. Since the gas
phase chemistry for SO. and NO is better understood than the aqueous
* X
phase, greater emphasis was placed on developing the gas component of the
psuedo-first order rate expressions.
Since the oxidation of SO and NO depends strongly on gas phase
^ X
photochemistry, rate expressions were derived from the results of
photochemical model simulations. A variable volume Lagrangian photochemical
box model was exercised over a wide range of environmental conditions. The
model was designed to simulate plume gases dispersing into and reacting with
background air. The Atkinson et al. (1982) photochemical mechanism was
employed for the calculations. It incorporates the important gas phase
reaction pathways for the ROG/NO /SO chemical system shown in
X X
Appendix A.
Five groups of parameters expected to influence photooxidation rates of
plume gases were allowed to vary in the photochemical model runs. These
surrogate parameters included season, background reactivity, dispersion
conditions, time of emissions release, and plume NO loadings. A total of
144 model runs were made representing parameter combinations for 3 different
seasons, 4 different background reactivities, 2 different dispersion
conditions, 2 different times of emissions release, and 3 different plume
NO loadings.
X
Solar radiation and ambient temperature data for the photochemical
model runs varied with season. Diurnally varying clear sky solar radiation
for a latitude of 40° and daily average temperatures of 30, 20 and 10°C were
employed for the summer, fall, and winter seasons, respectively. The
background air concentrations included 11 classes of ROG compounds and
ozone. The background ROG concentrations of 0.05, 0.25, 0.50, 2.0 ppmC were
employed. The composition of ROG was assumed to be 60% reactive alkanes,
10% alkenes, 25% aroma tics, and 5% aldehydes on a carbon basis in all
cases. Background ozone concentrations were varied between 0.02 and
41
-------
0.20 ppm. Plume NO loadings were varied from 0.007 to 1.4 ppm. Relative
humidity of 60% and a constant initial SO. concentration of 2.0 ppm were
employed for all the calculations. All parameter values used for the
simulation are summarized in Table 6.
The runs generated a data base with 1224 hourly conversion rates and
associated environmental conditions. The data base included the conversion
rate of SO, to SO?, NO to all products and NO to nitric acid. The solar
2 4 x x
radiation and concentrations of NO , ROG, and 0, at the midpoint of the
X *
hour were stored along with the time, temperature, stability, release time,
etc. for each hour.
Stepwise linear regression on the logarithms of the output variables
was performed to determine the controlling variables and the best regression
equations. Solar radiation, background ozone concentrations, and
atmospheric stability were found to be important parameters controlling
daylight gas phase SO oxidation rates. Background ozone concentration,
atmospheric stability, and NO concentrations were found to be most highly
X
correlated with the predicted NO oxidation rates. The following hourly
transformation rate expressions were determined :
36 R0'5^0'7^"1-29 (gas component) (2-49)
1G
1206 OZ1'50 S-^W0-33 (2-50)
1261 OZ1*45 S-1'34 NOX'0-12 (2-51)
where k-_ is SO. to SO, transformation rate (% per hour);
k is NO to HNO_ + RNO_ transformation rate (% per hour)*;
k is NO to HNO- (only) transformation rate (% per hour);
*The rate constant for NO * RNO- is k.-k..
42
-------
TABLE 6. PARAMETER VARIATIONS IN THE PHOTOCHEMICAL MODELING SIMULATIONS
Model Input Parameters and Variations
Surrogate
Parameter
Number of
Variations
Season
Background Air
Reactivity
Dispersion
Release Time
Plume NOX Loading
Ambient temperature and solar radiation
varied with season. Ambient temperatures
of 30, 20, and IO°C were employed for the
summer, fall, and winter cases, respec-
tively. Diurnally varying clear sky solar
radiation for 40° latitude in the 3 seasons
were employed.
Background air concentrations of ozone
and BOG were varied together to represent
background air reactivity. Background
ozone concentrations were assumed to be
correlated with season. Ozone concentra-
tions of 0.02, 0.05, 0.08, and 0.20 ppm
were employed for summer. Fall and winter
ozone concentrations equal to 75 and 50% of
the summer values were employed. The
corresponding four ROG concentrations
employed were 0.05, 0.25, 0.50, and 2.0
ppraC. The composition of the ROG was
assumed to be 60% reactive alkanes, 10%
alkenes, 25% aromatics, and 5% aldehydes.
The rate of plume dilution varied with
atmospheric stability and wind speed. A
stable case with 1.5 m/sec wind speed and a
slightly unstable case with 5 m/sec wind
speed were assumed. Dilution rates were
based on the time rate of change of plume
cross-sectional area (oy°z) from an
initial area of 10,000 m.
Sunrise and noon time were used as emission
release times.
Initial (at oyoz = 104m2) plume NOX
concentrations of 0.007, 0.35, and 1.40 ppm
were employed. The NOX was partitioned
as 90% NO and 10% N02 on a volume basis.
43
-------
2
R is total solar radiation (kwatt/m );
OZ is background ozone concentration (ppra);
S is atmospheric stability index ranging from 2 to 6 (e.g., 2 for
PGT stability classes A and B, 4 for class D, and 6 for class F);
-4
NO is ambient NO concentration (ppm), minimum value is 10 ppra.
2
The correlation coefficients (R ) for these regression equations are 0.80,
0.89, and 0.87, respectively, which indicate good correlations.
The dependency of these transformation expressions on environmental
parameters is consistent with physical expectations. The rates are
inversely proportional to the stability index which is consistent with the
expectation that higher background air entrainment rates (i.e., low
stability index) should result in higher conversion rates. They are
proportional to the background ozone concentration. Since ozone can be
thought of as a surrogate for OH concentration, this is consistent with
expectations from the gas phase chemistry. The S02 rate expression is
also dependent on solar radiation. Since the rate of the photolytic
reactions, which generate the free radicals, depend directly on the
radiation intensity, this result is expected. The NOX expression has a
weak inverse dependence on the NO concentration which is consistent with
the expectation that higher NO concentration impedes oxidation rates. A
similar dependence was expected for the S02 oxidation rates. However, the
statistical analysis indicated NO concentration was not nearly as
significant as the three other parameters in the S02 rate expression.
An aqueous phase SO. conversion rate expression was determined
empirically. Since the amount of aqueous phase S(1V) available for
conversion depends on the amount of water present (as well as condensation/
evaporation rates and SO. solubility which in turn depends on pH, etc.),
relative humidity was selected as a commonly available surrogate for liquid
water content. Although conversion may occur rapidly in the aqueous phase,
only a small portion of the SO is in the aqueous phase. For this reason,
a relatively low maximum aqueous oxidation rate (to be applied to SO (g))
was selected: three percent per hour. Since observations of SO-/SO,
44
-------
in plumes suggest overall oxidation rates increase dramatically at high
relative humidity (Gillani et al. 1981) a higher order dependence on
humidity was selected. The aqueous phase component of k, is
k1A = 3x10 RH4 (aqueous component)* (2-52)
where RH is the relative humidity (in percent). A minimum of 0.2% per hour
is used for k... The accuracy of this expression is highly uncertain.
Depending on environmental conditions, such as the availability of hydrogen
peroxide or metal catalysts, the actual conversion rate may be an order of
magnitude higher or lower than indicated by the expression.
Other researchers have formulated SO, oxidation rate expressions
based on observations. Henry and Hidy (1981, 1982) employed principal
component analysis of urban aerometric data to derive expressions for the
homogeneous and heterogeneous components. This analysis showed that a large
portion of the variance in SO. oxidation rate was explained by the
variance in ozone concentration. This photochemical component was generally
much larger than the heterogeneous component. The following gas phase SO^
oxidation rate expressions (in % per hour) represent the average for all the
stations examined:
k = 34 [0 ] for St. Louis (2-53)
k = 85 [0_] for Los Angeles (2-54)
wnere [0.] is the average hourly ozone concentration in ppm. These
expressions generally predict SO. oxidation rates greater than predicted
by ERT expression.
*NOTE:
-------
Gillani et al. (1981) derived a SO. oxidation rate expression which
is applicable when relative humidity is less than 75% based on plume
chemistry studies. This expression is
kL = 0.03 R h [031 (2-55)
where
k is SO. oxidation rate in % per hour,
2
R is total solar radiaton (kw/m ),
h is plume depth (m) (minimum of z. or 3c ), and
[0-j] is background ozone concentration (ppm) .
It is important to note that this expression identifies essentially the same
variables controlling S02 oxidation rate as the ERT expression. Figure 10
shows a comparison between SO. oxidation rates predicted with the Gillani
et al. and ERT expressions for a range of conditions. These conditions
include 8, 10, and 12 A.M. clear sky radiation in summer, fall, and winter;
ozone concentration of 0.02 to 0.12 ppm; and a range of stability/mixing
heights. The results show the two expressions predict comparable (within
+30%) SO oxidation rates. This result is quite significant since each
equation was derived in an entirely different manner: Gillani et al. from
observations, and ERT's from the kinetic model. The good agreement between
the results provides additional confidence in both equations. However, since
the ERT expression was generated from a wider range of conditions than the
Gillani et al. expression, it is the preferred mechanism in MESOPUFF II.
The parameterized rate constant expressions discussed above apply only
to daylight conditions. The gas phase free radical chemistry turns off at
night. The expressions which employ ozone concentration and radiation
levels as surrogate for OH concentration are inappropriate at night (Zak
1981). Nighttime oxidation of S02 and N02 to sulfates and nitrates,
respectively, is believed to be slow due to the absence of OH. Observations
of plume chemistry confirm this expectation. Figure 11 shows observed
hourly conversion rates of SO- to SO^ from eight plume studies as a
46
-------
4.0
3.5
3.0
o
i
£2.0
UJ
x
o
01
0.5
WINTER
FALL
SUMMER
1.0 1.5 2.0 2.5
S02 OXIDATION RATE - ERT
3.0
3.5
Figure 10 Comparison of SC>2 Oxidation Rates O per Hour) Predicted by
the ERT and Gillani et al. Expressions
47
-------
24 68 ID "12 14 16 18 20 22 24
TIME OF DAY
Figure 11 Average Plume Sulfur Conversion Rate as a Function of Mean
Time of Day of Plume Transport. Only Data Corresponding
to Plume Age Greater than 1.5 Hours are Plotted.
Source: Wilson (1981)
48
-------
function of time of day (Wilson 1981). These data show S(>2 oxidation
rates are generally less than 0.5% per hour at night. Observed nighttime
NO oxidation rates are also low. Forrest et al. (1981) found NO to
x x
total inorganic nitrate conversion rates in plumes of 0.1 to 3% per hour at
night and during early morning hours. These low oxidation rates are
presumably the result of heterogeneous reactions. Since these reactions are
not well understood and, in general, are less important than daytime
oxidation rates, constant oxidation rates are used in the model at night.
Based on the results of plumes studies, oxidation rates of 0.2 and 2% per
hour for SO and NO , respectively, were selected for nighttime
* X
conditions in the MESOPUFF II chemical submodel.
2.4.3 Implementation of Chemistry
It is important to design air quality models with flexibilty and
options to accommodate different applications and future improvements in
scientific knowledge. Several options have been incorporated in the
MESOPUFF II chemistry submodel to provide flexibility. First, in addition
to the ERT expressions for SO and NO transformation rates, the
X A
submodel includes the Gillani et al. and Henry and Hidy expressions for
SO, oxidation rates as options. Second, the model includes the option for
external (user) specification of hourly transformation rates of reactions
1-3 (Equations 2-40 to 2-42). The user also has the options to specify
hourly ozone data from a network of stations, a single ozone concentration
for all hours, or use the default value of 80 ppb. Similarly, the user may
use the default NH concentration (10 ppb) and nighttime oxidation rates
(see above) or specify values more appropriate for the application. Thus,
the MESOPUFF II chemical submodel has ample flexibility to accommodate
different applications and even different pollutants.
One of the problems in implementing chemistry in the puff modeling
framework is that the model keeps track of puffs individually, yet
atmospheric chemistry is a function of the concentrations from all puffs at
a given location. This is particularly important for the NO^ chemistry,
since the parameterized oxidation rate depends on N0x concentration and
49
-------
Che NH4N03 concentration depends on the total NHj and nitrate
concentrations. Clearly, in a situation where puffs overlap, it would be
incorrect to calculate the NO oxidation rate solely on the puff NO
x x
concentration and/or to calculate the particulate nitrate assuming all the
ambient NH-j would be available for one puff. Thus, the model has been
designed to employ the local average NO concentration from all puffs in
the NO oxidation rate expression for a single puff and apply the
X
HNO./NH /NH, NO. equilibrium relationship using the sum of total
nitrate concentrations from all puffs and ammonia (total ammonia minus
sulfate) at the location of interest.
2.5 Wet Removal
Numerous studies (e.g., Scott 1978, 1981; Garland 1978) have shown
precipitation scavenging to be an efficient pollutant removal mechanism,
especially for particulate pollutants such as SO^ that serve as cloud
condensation nuclei. Wet removal of soluble and reactive gaseous pollutants
such as SO and HNO- is also very important. Plumes can be nearly
completely washed out by moderate rainfall within a few hours. During
precipitation events, wet removal can easily dominate dry deposition in
pollutant removal. On an annual basis, the average wet removal rates for
SO and UNO, in Eastern North America and Europe are comparable to those
due to dry deposition (Scriven and Fisher 1975; Levine and Schwartz 1982);
for SO, wet removal appear:
(Scriven and Fisher 1975).
for SO, wet removal appears to be the more important removal mechanism
Wet removal includes both in-cloud scavenging (rainout) and below cloud
scavenging (washout). The scavenging process is a complex one involving
many factors. Scott (1978, 1981) has found precipitation scavenging of
sulfate to be a strong function of the mechanism of precipitation formation
and storm type (Figure 12). For example, the ratio of sulfate concentration
in precipitation to that in air (i.e., the washout ratio, W) is 10-50 times
larger for precipitation with growth due primarily to accretion than for
precipitation growth due to vapor deposition. The scavenging efficiency of
gases is a function of pollutant's solubility in water and reactivity.
50
-------
105
vf
O
103
102
I
a 01
0.1 10 10
PRECIPITATION RATE (mm h"1)
100
Figure 12 Washout Ratio as a Function of Precipitation Rate for
Different Storm Types. Curve 1 Represents Predictions for
Intense Convective Storms or from Clouds Whose Tops are
Warmer than 0°C; Curve 2 Represents Predictions for Storms
Where Rain Develops Without the Assistance of an Ice Growth
Stage; Curve 3 is for Storms Where the Ice Growth Process
is Necessary for Initiating Precipitation; Curve 4 is from
Observed 24Na Concentrations in Rainwater at Quiilayute,
Washington on 5, 6 April 1970; and Curve 5 is the Same as
Curve 4 Except Curve for Data from 11 December 1969.
Source: Scott (1978)
51
-------
Barrie (1981) has shown SO. washout ratios to be strongly dependent on the
pH of the rain and temperature (Figure 13).
However, a simple parameterization of wet removal suitable for order of
magnitude wet removal estimates and using only routinely available
meteorological variables is required in MESOPUFF II. A convenient approach
compatible with the puff superposition principle is the scavenging
coefficient formulation:
Q(t + 1) = Q(t) exp [- A At] (2-56)
where Q(t), Q(t + 1) is the mass (g) of pollutant in the puff at the
beginning and end of the time step,
A is the scavenging ratio (s ), and
At is the time step (s).
Haul (1980) expresses A as:
A = x (R/RL) (2-57)
where R is the rainfall rate (mm/hr),
R. is a reference rainfall rate of 1 mm/hr, and
L -1
X is a scavenging coefficient (s ).
The rainfall rate used in Equation 2-57 is that observed at the closest
surface station to the center of the puff. Table 7 contains the default
values of the scavenging coefficient used in MESOPUFF II. Different values
of X are considered for liquid and frozen precipitation. Slinn et al.
(1978) note that snow scavenging of gases is generally negligible. The
scavenging coefficients for SO. and SO, removal by rain is based on
Maul (1980) and Garland (1978). The scavenging coefficient for N0~ is
believed to be roughly comparable to that for S0~. Maul (1980) showed
-5 -1
that values of X * 3-6 x 10 s can be derived for sulfate with
the assumption of full removal from air entrained into the clouds. The high
solubility and reactivity of HNO- suggests that a scavenging coefficient
52
-------
10a -
SO, WASHOU
0
I*
0
K
10
pH
Figure 13 S02 Washout Ratio as a Function pH and Temperature for
Equilibrium Scavenging Conditions
Source: Barrie (1981)
53
-------
TABLE 7. DEFAULT VALUES OF THE SCAVENGING COEFFICIENT, (s-1)
Liquid Frozen
Pollutant Precipitation Precipitation
SO 3 x 10~5 0.0
2
SO? 1 x 10~4 3 x 10~5
4
NO 0.0 0.0
x
HNO 6 x 10~5 0.0
N0~ 1 x 10~4 3 x 10"5
54
-------
similar to SO** is appropriate for HNO_. Levine and Schwartz (1982)
emphasize the sensitivity of HNO. removal to the raindrop size
distribution, especially to the lower radii limit of the distribution
because of the dominant contribution of the smaller drops to the removal
rate. Their recommendations suggest a scavenging ratio of
6.5 x 10~5 s"L for a rainfall rate of 1 mm/hr. Based on the low
solubility of NO , a negligible scavenging coefficient is expected.
x
A precipitation code determined from the surface observations of
precipitation type/intensity is used to determine if the value of X for
liquid or frozen precipitation is most appropriate. Precipitation
observations are converted to precipitation codes as shown in Table 8. The
liquid precipitation values of X are used for precipitation codes 1-18;
the frozen precipitation values are used for codes 19-45.
2.6 Puff Sampling Function
Puff superposition models such as MESOPUFF II represent a continuous
plume with a number of discrete puffs. The concentration at a receptor is
calculated by summing the contributions of each nearby puff, generally
evaluated by taking a "snapshot" of each puff at particular time intervals
(sampling steps) specified as a program input. The concentration at a
receptor due to a horizontally symmetric with a Gaussian distribution is
given by:
C(s) - 8 g(s) exp
2*oy2(s)
I". r2(s) '
L 2o 2(s) .
(2-58)
*2"*) 1
«.2(.) J
55
-------
TABLE 8. CONVERSION OF REPORTED PRECIPITATION TYPE/INTENSITY TO PRECIPITATION CODES
Liquid Precipitation
Frozen Precipitation
Precipitation
Code
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Type
Rain
Rain
Rain
Rain Showers
Rain Showers
Rain Showers
Freezing Rain
Freezing Rain
Freezing Rain
Not Used
Not Used
Not Used
Drizzle
Drizzle
Drizzle
Freezing Drizzle
Freezing Drizzle
Freezing Drizzle
Intensity
Light
Moderate
Heavy
Light
Moderate
Heavy
Light
Moderate
Heavy
-
-
-
Light
Moderate
Heavy
Light
Moderate
Heavy
Precipitation
Code
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
Type
Snow
Snow
Snow
Snow Pellets
Snow Pellets
Snow Pellets
Not Used
Ice Crystals
Not Used
Snow Showers
Snow Showers
Snow Showers
Not Used
Not Used
Not Used
Snow Grains
Snow Grains
Snow Grains
Ice Pellets
Ice Pellets
Ice Pellets
Not Used
Hail
Not Used
Not Used
Small Hail
Not Used
Intensity
Light
Moderate
Heavy
Light
Moderate
Heavy
-
*
-
Light
Moderate
Heavy
-
-
-
Light
Moderate
Heavy
Light
Moderate
Heavy
-
*
-
-
*
-
Intensity not currently reported Cor Lee crystals, hail and small hail.
56
-------
where, C(s) is Che ground-Level concentration,
s is Che distance travelled by Che puff,
Q(s) is Che mass of pollutant in Che puff,
o (s) is Che standard deviation of Che Gaussian distribution
y
in Che horizontal,
o (s) is Che standard deviacion of Che Gaussian distribution
z
in Che vertical,
r(s) is Che radial disCance from che puff center,
z. is Che mixed-layer height, and
H is Che effeccive height of Che puff center.
The vertical term, g(s) reduces Co Che uniformly mixed limit of
for a /z. > 1.6. In general, puffs within Che daytime mixed layer
Z L ^
satisfy Chis cricerion abouC an hour or two after release.
Wich Equation 2-58, an accurate representation of Che continuous plume
depends upon the puff release race and sampling race being sufficient to
ensure that adjacent puffs overlap. Ludwig ec al. (1977) have shown chat if
puff separation distances exceed *2o , inaccurate results may be
obtained. The frequenC sampling and/or puff release necessary Co satisfy
the 2o cricerion in Che near-field of a source (where ic is most
y
restrictive) has a negative impacC on model sCorage and computational
requirements. Ludwig eC al. (1977) recommend uniform space rather than
uniform time release of puffs with a puff merging scheme to reduce the total
number of puffs on Che grid. However, frequenC sampling or puff release is
still necessary for near-field recepCors. An alternate approach suggested
by R. Yamartino (personal communication) and used in MESOPUFF II is to
integrate Equation 2-58 over the distance of puff travel, As, during one
sampling step.
57
-------
s + As _, v ( x r 2 , . -i
q(8) |(8) exp | "r U) ds (2-60)
8 2ir ay (s) L 2oy (s)J
If it is assumed chat Che most significant s dependence during the sampling
seep is in the r(s) and Q(s) terms, this integral can be evaluated.
Assuming the trajectory segment is a straight line and transforming s to a
dimensionless trajectory variable, p, results in:
r2(p) = (Xt - Xr + pAX)2 + (Yt - Yf + pAY)2 (2-61)
where p = 0,1 correspond to the beginning and end points of the trajectory
(Xc, YC) and Ut+1, Yt+1), respectively, (Xr> Yr) are the
receptor coordinates, and AX, AY are the incremental X and Y distances
travelled by the puff during the sampling step.
Equation 2-60 becomes:
C = ^ ^ Q(p) exp | -^-Hp- I *P C2-62)
2na '
y
L
The exponential variation of Q due to removal and chemical transformation
processes is expressed as a linear function over the sampling interval:
Q(p) - Qt + P
-------
P exp
[- *>>
L 2o ^ -I
(2-64)
The integrals in Equation 2-64 can be solved analytically and expressed in
terms of error functions and exponentials.
[ Vi + (V V
(2-65)
I. = - exp [1/2 (b/a - c)]
1 /2a
- erf
l/Ia" J /2a
" « Jl * exp I" 1/2 (b2/a - c)l
exp[~- | b2/al- expf-
(a*2b+b/a)
Ux2 » Ay2) / o 2
- xr)
-xr)
- yr)2l/a 2
(2-66)
(2-67)
(2-68)
(2-69)
(2-70)
The vertical term, g, and a are evaluated at the midpoint of the
trajectory (p=0.5).
Because the integrated contribution of each puff over the sampling step
is computed, this sampling function eliminates the problem of insufficient
puff overlap. Table 9 contains the results of sampling tests performed with
the conventional sampling algorithm (e.g., as in MESOPUFF) and the new
sampling function used in MESOPUFF II. The analytic (straight-line)
Gaussian solution is also shown. As expected, the conventional algorithm
produces inaccurate results when the puff separation exceeds 2o . The
puff separation is u6t, where u is the wind speed (5 m/s), 6t is the
59
-------
TABLE 9. EFFECT OF SAMPLING RATE, N, ON PREDICTED NEAR-FIELD (<50 km) CONCENTRATIONS FOR TWO
SAMPLING ALGORITHMS. PRESENTED ARE VALUES OF C/Q xlO7. N IS IN SAMPLES PER HOUR.
(Wind Speed - 5.0 m/s, PGT Stability D, Mixing Height = 1,000 m, Uniform Vertical Distribution)
Distance °y Straight Line Conventional Sampling Algorithm MESOPUFF II Sampling Algorithm
(km) (m) Gaussian Eqn. N=l N=2 N=4 N=8 N°16 N=l N=2 N=4 N=8 N=16
10 518 1.54 0.00 1.32 0.69 0.57 1.50 1.69 1.09 1.33 1.51 1.55
20 966 0.83 0.54 0.28 0.23 0.82 0.80 0.59 0.72 0.81 0.83 0.83
30 1,392 0.57 0.00 0.11 0.45 0.56 0.56 0.63 0.54 0.59 0.57 0.57
40 1,803 0.44 0.11 0.09 0.45 0.43 0.43 0.39 0.44 0.45 0.44 0.44
50 2,203 0.36 0.25 0.16 0.36 0.35 0.35 0.41 0.37 0.36 0.36 0.36
-------
sampling interval («t=3600s/N), and N is the sampling rate (samples per
hour). The 2a criterion is not satisfied for N=l or 2 even at 50 km,
y
resulting in 'gaps' in the concentration distribution. More accurate
results are obtained with the MESOPUFF II algorithm. Acceptable results
(i.e., within-v 5%) are obtained beyond 20 km with N=2 vs. N=8 for the
conventional algorithm. The major source of error in the MESOPUFF II
algorithm is due to the assumption of constant o during the sampling
interval. The value of o is evaluated at the midpoint of the
trajectory segment. Thus, during the first half of the trajectory o is
somewhat overestimated; during the second half, o is underestimated.
Because the length of each trajectory segment is proportional to the wind
speed, this error may be minimized by increasing the sampling rate at higher
wind speeds. MESOPUFF II offers the option to dynamically determine the
sampling for each puff as follows:
N = 1 + (2-71)
uc
where u is a reference wind speed specified by the user. For example,
c
for a u of 2 m/s, N will be assigned values of 1, 2, or 3 for values of u
of 1.0, 2.0, and 4.0, respectively. The value of N given by Equation 2-71
is then compared to a user-specified minimum sampling rate; if lower, N is
set equal to the minimum rate.
61
-------
SECTION 3
DEMONSTRATION RUN
The MESOPUFF II model has been run for a two-day cest period to
allow a preliminary evaluation of the SO- * SO? transformation
mechanism and to qualitatively demonstrate the behavior of several other
model algorithms. The modeled period was-during the Tennessee Plume Study
(TPS) which was conducted in August 1978 in the vicinity of the Cumberland
Steam Plant in northwestern Tennessee (Schiermeier et al. 1979). The TPS is
part of a larger EPA field program, called the Sulfur Transport and
Transformation in the Environment (STATE) program, which was designed to
examine the effects of SO emissions on regional scale sulfate
concentrat ions.
Detailed measurements from aircraft traversing the Cumberland plume
provided data on chemical processes and dispersion. Plumes from other TVA
plants were also sampled when they were transported near the Cumberland
plume. Plume trajectories were determined with aircraft and ground-level
measurements of an injected tracer gas (SFg), and by tracking tetroons and
a manned LAMP oalloon. Four specific scenarios were studied: (1) vertical
mixing during highly convective conditions to downwind distances of 50 km;
(2) horizontal plume spread during stable conditions with significant wind
shear to downwind distances of 300-500 km; (3) dispersion during stable
conditions to distances of 400 km; and (4) dispersion and chemical changes
over a diurnal cycle, with fumigation in the morning and layering in the
evening. The two-day study period (August 22-23) chosen for the test run
falls under Scenario 4.
The model runs were made with a 24 x 30 grid covering the area
encompassed by latitudes 35°-39° N and longitudes 86°-90° W (approximately
from Memphis, Tennessee in the southwest corner of the grid to southern
62
-------
Indiana in Che northeast part of the grid). A grid spacing of 15 km was
used. Only SO emissions from the Cumberland Steam Plant were modeled.
Meteorological data from nine surface stations and three rawinsonde stations
were processed (Figure 14). Special meteorological observations available
during the TPS which would not be available for an operational application
of the model (e.g., soundings made at 6-hour intervals) were not used.
Mixed layer averaged winds were used to advect puffs within the mixed
layer. Puffs above z. were advected with vertically averaged winds
through the layer from z. to 700 mb. The surface depletion (3-layer) dry
deposition model was used. Table 10 summarizes the model run parameters.
The time history of a particular puff through a diurnal cycle is
presented in Table 11. The puff was released at 0100 CST on August 23 from
the Cumberland stack. The puff quickly rose well above the shallow
nocturnal boundary layer to a height of about 850 m. The rapid growth of
the convective boundary layer and eventual fumigation of the puff is shown
in Figure 15. Following the period of relatively slow puff growth, the puff
quickly becomes uniformly mixed in the vertical after entrainment into the
mixed layer. The puff growth rate while above the mixed layer is given by
the E stability Turner dispersion curves.
Observed and predicted mixing heights in the vicinity of the Cumberland
stack are presented in Figure 16. The model appears to correctly predict
the growth of the morning convective mixed layer, although afternoon mixing
heights are overpredicted. However, general conclusions regarding the
quantitative performance of the mixing height algorithm should not be drawn
from two case studies.
Modeling of dry deposition begins when the puff is fumigated into the
mixed layer for the first time (1000-1100). Dry deposition and chemical
transformation are of about equal importance as depletion mechanisms for
S02 (loss rates of 1.5 - 2.5% per hour during daytime). The significance
of boundary layer mixing as a limit to the deposition of pollutants is seen
in the ratio v1 ,/v,. During the day when buoyancy-induced turbulence
d d
causes vigorous mixing, v'd/vd is nearly unity. However, during stable
63
-------
Little Rock
30
28
26
24
22
20
18
16
14
12
10
8
4
Salem
Evansville
} Cape Girardeau
Paducah
Bowling
Green 9
£ Fort
Cambell
Cumberland 4
Nashville
iJackson
Memphis
Louisville
2 4 6 8 10 12 14 16 18 20 22 24
Figure 14 Location of Cumberland Steam Plant ()» Surface
Meteorological Stations () and Upper Air Rawinsonde
Stations (A) on Meteorological Grid
64
-------
TABLE 10. MODEL RUN PARAMETERS USED IN DEMONSTRATION RUN
Pollutants S02, S04
Grid Size 24 x 30
Grid spacing 15 km
Time step 1 Hour
Sampling Rate Variable - u£ = 2.0 m/s (see Equation 2-71)
minimum rate = 4 (Aug. 22), 2 (Aug. 23)
Puff Release Rate 1 puff/hour
Background Ozone
Concentration 80 ppb
Puff Growth Rate above
Boundary Layer E stability
65
-------
TABLE 11. HISTORY OF PUFF RELEASED 8/23/78 HOUR 1
Hour
(CST)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
?o
Xt
(km)
0
7
17
26
37
47
58
69
79
88
99
117
135
155
180
201
219
235
248
9S9
°y
(m)
1
278
603
915
1226
1543
1862
2164
2444
2712
3577
5527
7326
9127
11002
12801
14527
16327
18127
19777
°z
(m)
1
55
82
102
118
134
147
159
170
179
1703
*
*
*
*
*
*
*
*
*
z.
(m)
11
29
11
U
17
124
298
551
782
955
1120
1311
1417
1485
1510
1519
1497
1389
S4
z . (MAX)
(m)
-
-
-
-
-
-
-
-
-
955
1120
1311
1417
1485
1510
1519
1519
1519
1519
Puff(il
Code
3
3
3
3
3
3
3
3
3
3
I
2
2
2
2
2
2
2
2
6
' K/10
(%/hr)
0.2
0.2
0.2
0.2
0.2
2.0
2.4
1.9
1.4
2.4
2.4
2.2
2.1
2.0
1.8
1.0
0.9
0.6
0.2
Vd
(cm/s)
-
-
-
-
-
-
-
-
-
0.82
0.81
0.83
0.83
0.85
0.75
0.83
0.80
0.28
0.28
Vd
Vd
^
-
-
-
-
-
-
-
-
-
1.00
0.96
0.96
0.96
0.96
0.96
0.95
0.94
0.47
0.2fi
Kd(ii°
(%/hr)
^
-
-
-
-
-
-
-
-
-
2.9
2.4
2.3
2.1
2.0
1.6
1.8
1.8
0.3
0.1
^
Qt
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
1.0
1.0
1.0
1.0
1.0
1.0
1.0
0.99
0.91
0.03
*Puff Uniformly Mixed in Vertical - Q not calculated
Z
( ' \
Puff codes: 1 = puff within mixed layer and Gaussian
2 = puff within mixed layer and uniform
3 = puff above mixed layer and Gaussian
6 = puff currently above (but previously below) mixed layer and uniform
I = S02 -> 804 transformation rate
/ " \
Kd = S02 dry deposition depletion rate
-------
Height
(m)
1600-r
1400-
1200-
1000-
Vertical Puff
Spread
600-
400-
200-
V* <
800-T*^<^
'*
6
8 10 12
Time (Mrs)
14
i
16
I
18
20
Sunrise
Sunset
Figure 15 Boundary Layer Growth and Plume Fumigation
67
-------
August 22. 1978
zi
(m)
2000
1500
1000
500
6 8 10 12 14 16 18
Time (CST)
20
(m)
2000
1500
1000
500
August 23. 1978
6 8 10 12 14 16 18 20
Time (CST)
Figure 16 Observed () and Predicted (-) Mixing Heights in the
Vicinity of the Cumberland Steam Plant
63
-------
nighttime conditions, v1 ,/v, « 1, indicating the importance of the
boundary layer mixing rate as an additional resistance to mass transfer. At
2000, the surface concentration (Cg) is only about 1/4 of the
layer-averaged concentration (C ). This lower near-surface concentration
reduces dry deposition flux, thus increasing the lifetime of SO. in the
atmosphere.
Predicted SO. to SO, conversion rates have been calculated and
compared to observations reported by Gillani et al. (1981). The average
conversion rate between the time of puff release and sampling is given by
Gillani et al. (1981) as:
where
=
k- is the average SO. to SO, conversion rate,
Cf is a correction factor accounting for changes in puff
mass during the period of travel,
Q is the total initial sulfur mass in the puff (S02 + S0~)
weighted as SO,,
Q is the total sulfur mass in the puff at the sampling time,
Q - is the mass of SO. in the puff at the sampling time,
and,
Q . is the mass of SO. in the puff at the sampling time.
The MESOPUFF II conversion rates as well as those predicted by Gillani
et al. (1981) are shown with observed rates in Table 12. Both schemes
predict rates generally within the range of observed transformation rates
during August 22. This day was generally sunny and relatively dry (low
relative humidity). SO oxidation was probably dominated by gas phase
reactions. During August 23, however, a maritime tropical air mass
characterized by high humidity and hazy conditions existed in the region.
-------
TABLE 12. OBSERVED AND PREDICTED S02 CONVERSION RATES
Transport Time
DATE
8/22
8/23
Obs. Plume
(CST)
2:55 -
2:50 -
11:30 -
11:00 -
6:30 -
6:00 -
5:15 -
6:15 -
6:25
6:50
13:30
14:30
10:30
11:25
15:15
16:15
Plume
Age
(Hours)
3.5-4.0
2.0
3.5
4.0
5.4
10.0
10.0
Observed Predicted Rate
Transport Time
Rate (Gillani et al. 1981) Pred. Plume*
(%/hour) (%/hour) (CST)
O.I -
1.2 -
1.5 -
1.7 -
2.1 -
2.7 -
2.4 -
0.45 0.05
1.7 1.4 - 1.7
1.8 1.9 - 2.0
2.8 0.3
2.5 0.4
2.9 0.9 - 1.0
3.3 1.0 - 1.1
3:00 -
11:00 -
11:00 -
6:00 -
6:00 -
5:00 -
6:00 -
6:45
13:00
14:30
10:00
11:30
15:00
16:00
Plume Predicted Rate
Age Eq. 2-49 and
(Hours) Eq. 2-52
3.75
2.0
3.5
4.0
5.5
10.0
10.0
0.59
1.5
1.7
2.0
2.1
1.9
2.0
*Puffs released at hourly intervals
-------
Low clouds increased during Che day and dissipated after sunset.
Significant plume-cloud interactions were reported by Gillani et al. (1981)
during the late'morning and early afternoon hours. The conversion rates
predicted by MESOPUFF II are within the observed range during the morning
transition period. However, the conversion rates averaged through the day
are underpredicted, probably due to enhanced aqueous phase reactions
associated with plume-cloud interactions. The empirical aqueous phase term
of the rate equations is based on surface relative humidity measurements and
is not able to account for these interactions. The Gillani et al. (1981)
relationship is intended for conditions when gas phase reactions dominate
(i.e., relative humidity < 75Z). Therefore, it cannot account for the
high observed rates during August 23 when liquid phase reactions contribute
significantly.
In summary, MESOPUFF II modeling results for a limited two-day period
during the TPS have been presented. The qualitative behavior of several
model algorithms, including plume growth, development of the convective
boundary layer, plume fumigation, and deposition processes have been
presented. In particular, an encouraging preliminary evaluation of the
SO to SO, chemical transformation algorithm has been presented.
These results represent only a brief and limited evaluation. Further
evaluation with TPS data and additional regional-scale monitoring/
experimental measurement studies are recommended.
71
-------
REFERENCES
Atkinson, R. and A.C. Lloyd 1980. "Evaluation of Kinetic and
Mechanistic Data for Photochemical Smog Chamber Modeling", EPA
Contract No. 68-02-3280, ERT Document No. P-A040.
Atkinson, R., A.C. Lloyd, L. Winges 1982. An Updated Chemical
Mechanism for Hydrocarbon/N0x/S0x Photooxidations Suitable
for Inclusion in Atmospheric Simulation Models, Atmos. Environ^,
16., 1341.
Barrie, L.A. 1981. The Prediction of Rain Acidity and 502
Scavenging in Eastern North America. Atmos. Environ., \5_, 31-41.
Benkley, C.W. and A. Bass 1979a. User's Guide to MESOPLUME Mesoscale
Plume Segment Model. EPA 600/7-79-xxx, U.S. Environmental
Protection Agency, Research Triangle Park, NC. 141 pp.
Benkley, C.W. and A. Bass 1979b. User's Guide to MESOPUFF (Mesoscale
Puff Model. EPA 600/7-79-xxx, U.S. Environmental Protection
Agency, Research Triangle Park, NC. 124 pp.
Benkley, C.W. and A. Bass 1979c. User's Guide to MESOPAC Mesoscale
Meteorology Package). EPA 600/7-79-xxx. U.S. Environmental
Protection Agency, Research Triangle Park, NC. 76 pp.
Briggs, G.A. 1975. Plume Rise Predictions. Lectures on Air
Pollution and Environmental Impact Analyses. American
Meteorological Society, Boston, MA, pp 59-111.
Brost, R.A. and J.C. Wyngaard 1978. A Model Study of the Stably
Stratified Planetary Boundary Layer. J. Atmos. Sci. 35.
1427-1400.
Calvert, J.G., F. Su, J.W. Bottenheim and O.P. Strausz 1978. Mechanism
of the Homogeneous Oxidation of Sulfur Dioxide in the Troposphere.
Atmos. Environ., 12, 197.
Chamberlain, A.C. and R.C. Chadwick 1953. Deposition of Airborne
Radio-Iodine Vapor. Nucleonics. 8_, 22-25.
Deardorff, J.W. and Willis, G.E. 1975. A parameterization of
diffusion into the mixed layer. J. Appl. Meteor., 14:1451-1458.
Draxler, R.R. 1977. A Mesoscale Transport and Diffusion Model.
National Oceanic and Atmospheric Administration Tech. Memo.
ERL-ARL-64, Air Resources Laboratories, Silver Springs, MD.
Draxler, R.R. 1979. Modeling the Results of Two Recent Mesoscale
Dispersion Experiments. Atmos. Environ. 13, 1523-1533.
72
-------
Fisher, B.E.A. 1980. Long-range Transport and Deposition of Sulfur
Oxides. CERL internal report, Central Electricity Research
Laboratories, Leatherhead, Surrey, United Kingdom.
Forrest, J., R.W. Garber and L. Newman 1981. Conversion Rates in
Power Plant Plumes Based on Filter Pack Data: The Coal-fired
Cumberland Plume. Atmos. Environ., _lli 2273.
Garland, J.A. 1978. Dry and wet removal of sulfur from the
atmoshpere. Atmoa. Environ.. 12, 349-362.
Gifford, F.A. 1981. Horizontal Diffusion in the Atmosphere: A
Lagrangian-Dynamical Theory. LA-8667-MS, Los Alamos Scient.
Lab., P.O. Box 1663, Los Alamos, NM, 87545, 19 pp.
Gillani, N.V., S. Kohli and W.E. Wilson 1981. Gas-to-Particle
Conversion of Sulfur in Power Plant Plumes: I. Parameterization
of the Gas Phase Conversion Rates for Dry, Moderately Polluted
Ambient Conditions. Atmos. Environ., 1^5. 2293-2313.
Hefter, J.L. 1965. The Variations of Horizontal Diffusion Parameters
with Time for Travel Periods of One Hour or Longer. J. Appl.
Meteor.. 4, 153-156.
Hegg, D.A. and P.V. Hobbs 1981. Cloud Water Chemistry and the
Production of Sulfate in Clouds. Atmos. Environ.. 15, 1597-1604.
Henry, R.C., D.A. Godden, G.M. Hidy, and N.J. Lordi 1980. Simulation
of Sulfur Oxide Behavior in Urban Areas. ERT Document
P-A070-200. Prepared for the American Petroleum Institute.
Henry, R.C. and G.M. Hidy 1981. Discussion of Multivariate Analysis
of Particulate Sulfate and Other Air Quality Variables, Part I.
Annual data from Los Angeles and New York. Atmos. Environ., 15_,
424.
Henry, R.C. and G.M. Hidy 1982. Multivariate Analysis of Particulate
Sulfate and Other Air Quality Variables by Principle Components -
II. Salt Lake City, Utah and St. Louis, Missouri. Attnos^
Environ.. 16. 929-943.
Hicks, B.B. and J.D. Shannon 1979. A method for modeling the
deposition of sulfur by precipitation over regional scales.
J. Appl. Meteor.. 18. 1415-1420.
Hicks, B.B. 1982. Critical Assessment Document on Acid Deposition,
Chapter VII - Dry Deposition. ATDL contribution file no. 81/24.
Horst, T.W. 1977. A Surface Depletion Model for Deposition from a
Gaussian Plume. Atmos. Environ.. 11, 41-46.
73
-------
Levine, S.Z. and S.E. Schwartz 1982. la-Cloud and Below-Cloud
of Scavenging of Nitric Acid Vapor. Atmos. Environ., 1£,
1725-1734.
Ludwig, F.L., L.S. Gasidrek, and R.E. Ruff 1977. Simplification of a
Gaussian Puff Model for Real-Time Minicomputer Use. Atmos.
Environ., 11. 431-436.
Maahs, H.G. 1982. The Importance of Ozone in the Oxidation of Sulfur
Dioxide in Nonurban Tropospheric Clouds. 2nd Symposium on the
Composition of the Nonurban Troposphere, Amer. Meteorological
Society, Williantsburg, VA, May 1982.
Martin, L.R., D.E. Damschen and H.S. Judeikis 1981. The Reactions of
Nitrogen Oxides with S02 in Aqueous Aerosols. Atmos. Environ.
J.5.: 191-195.
Martin, L.R. 1982. Kinetic Studies of Sulfite Oxidation in Aqueous
Solution, to be published in "S02, NO and NO2 Oxidation
Mechanisms: Atmospheric Considerations, ed. J.G. Calvert, Ann
Arbor Science, Woburn, MA.
Maul, P.R. 1980. Atmospheric Transport of Sulfur Compound Pollutants.
Central Electricity Generating Bureau MID/SSD/80/0026/R.
Nottingham, England.
Morris, C.S., C.W. Benkley, and A. Bass 1979. User's Guide to
MESOGRID (Mesoscale Grid) Model. EPA-600/7-79-xxx. U.S.
Environmental Protection Agency, Research Triangle Park, NC.
118 pp.
Newman, L. 1981. Atmospheric Oxidation of Sulfur Dioxide: A Review
as Viewed from Power Plant and Smelter Plume Studies. Atmos.
Environ., 15, 2231.
Penkett,< S.A., B.M.R. Jones, K.A. Brice and A.E.J. Eggleton 1979. The
Importance of Atmospheric Ozone and Hydrogen Peroxide in
Oxidizing Sulfur Dioxide in Cloud and Rainwater. Atmos.
Environ. . U., 123-137.
Richards, L.W. et al., 1981. The Chemistry, Aerosol Physics, and
Optical Properties of a Western Coal-Fired Power Plant Plume.
Atmos. Environ.. 15, 2111.
Schwartz, S.E., 1982. Gas-Aqueous Reactions of Sufur and Nitrogen
Oxides in Liquid-Water Clouds; to be published in S02, NO and
NO2 Oxidation Mechanisms: Atmospheric Conditions, ed. J.G.
Calvert. Ann Arbor Science, Woburn, MA.
Schwartz, S.E. and W.H. White 1982. Kinetics of Reactive Dosclution
of Nitrogen Oxides into Aqueous Solution, in Advan. Environ.
Sci. Techno 1. . Vol. 12, S.E. Schwartz, Ed. (New York: Wiley and
Sons, Inc., 1982).
74
-------
Scire, J.S., J. Beebe, C. Benkley, and A. Bass 1979. User's Guide to
the MESOFILE Postprocessing Package. EPA-600/7-79-xxx, U.S.
Environmental Protection Agency, Research Triangle Park, N.C.
72 pp.
Scire J.S., F. Lurmann, A. Bass, and S. Hanna 1983. User's Guide to
the MESOPUFF II Model and Related Processor Programs. U.S.
Environmental Protection Agency, Research Triangle Park, NC.
Scott, B.C. 1978. Parameterization of sulfate removal by
precipitation. J. Appl. Meteorol., 1^7, 1375-1389.
Scott, B.C. 1981. Sulfate Washout Ratios in Winter Storms.
J. Appl. Meteor., 20_, 619-625.
Scriven, R.A. and B.C.A. Fisher 1975. The long-range transport of
airborne material and its removal by deposition and washout.
Atmos. Environ.. 9_, 49-58.
Sehmel, G.A. 1980. Particle and Gas Dry Deposition - A review.
Atmos. Environ. 14. 983-1011.
Shexh, 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 (10), 1361-1368.
Slinn, W.G., L. Hasse, B. Hicks, A Hogan, D. Lai, P. Liss,
K. Munnich, G. Sehmel, and 0. Vittori 1978. Some Aspects of the
Transfer of Atmospheric Trace Constituents Past the Air-Sea
Interface. Atmos. Environ.. 12; 2055-2087.
Smith, F.B. 1981. The significance of wet and dry synoptic regions on
long-range transport of pollution and its deposition. Atmos.
Environ.. 15, 863-873.
Stelson, A.W. and J.H. Seinfeld 1982. Relative Humidity and
Temperature Dependence of the Ammonium Nitrate Dissociation
Constant. Atmos. Environ., 16, 983-992.
Stelson, A.W., M.E. Bassett and J.H. Seinfeld 1983. Thermodynamic
Equilibrium Properties of Aqueous Solutions of Nitrate, Sulfate
and Ammonium. Acid Precipitation, Chemistry of Particles, Fog
and Rain Volume. J. Teasley ed., Ann Arbor Science, Woburn, MA
(in press).
Turner, D.B. 1964. A Diffusion Model for an Urban Area. J. Applied
Meteorol.. 3, 83-91.
Turner, D.B. 1970. Workbook of Atmospheric Dispersion Estimates.
U.S. Dept. of H.E.W., Public Health Service, Pub. 999-AP-26,
88 pp.
U.S. EPA 1978. Guideline on Air Quality Models, OAQPS Guideline
Series No. 1.2-080, EPA report No. EPA-450/2-78-027 (NTIS No.
PB288783), 84 pp.
75
-------
Van Ulden, A.P. 1978. Simple estimates for vertical diffusion from
sources near the ground. Atmos. Environ.. 12. 2152-2129.
Venkatram, A. 1980a. Estimating the Monin-Obukhov Length in the Stable
Boundary Layer for Dispersion Calculations. Boundary-Layer
Meteorology 19. 481-485.
Venkatram, A. 1980b. Estimation of turbulence velocity scales in the
stable and the unstable boundary layer for dispersion
applications. In: Eleventh NATO-CCSM International Technical
Meeting on Air Pollution Modeling and its Application 54-56.
Venkatram, A., B.E. Ley and S.Y. Wong 1982. A statistical model to
estimate long-term concentrations of pollutants associated with
long-term transport. Atmos. Environ., 16, 249-257.
Wang, I.T. and P.O. Chen 1980. Estimations of Heat and Momentum
Fluxes Near the Ground. Proc. 2nd Joint Conf. on Applications of
Air Poll. Meteorology. New Orleans. LA, March 24-27. pp 764-769.
Wesely, M.L., and B.B. Hicks 1977. Some factors that Affect the
Deposition Rates of Sulfur Dioxide and Similar gases on
Vegetation. J. Air Poll. Control Assoc. 27. pp 1110-1116.
Wilson, W.E. 1981. Sulfate Formation in Point Source Plumes: A
Review of Recent Field Studies. Atmos. Environ.. 15_, 2573.
Zak, B.D. 1981. Lagrangian Measurements of Sulfur Dioxide to Sulfate
Conversion Rates. Atmos. Environ.. 15, 2583.
76
-------
APPENDIX A
REACTIONS AND RATE CONSTANTS OF THE
ATKINSON et ml. (1982) CHEMICAL MECHANISM
Reaction
Rate Constant (ppm min units)
Inorganics
02
1. N02 » hv ** NO
2. NO
* N02 + 02
3. 03 + hv * 2 OH
4. OH + NO 3 HONO
5. OH + N02 HN03
6. HONO + hv * OH + NO
7. H02 + NO + OH + N02
8. H02 + N02 2 H02N02
9. H02N02 ^ H02 + N02
10. H0 + H0 » H0 +
11. H90,
hv * 20H
°
12. OH + CO -> H02
13. N02 + 03 * N03
14. NO + N03 » 2N0
15. N02 + N03 5 N2
16. N0 3 N0 + N
17.
2HN0
18. N03 + hw -» 0.3 NO + 0.7 N02
+ 0.7 00
k- variable
. 1.0 , 106 I'1 e-1450/T
k3 S
x 7'6 x
=8.7 x 108 T"2
= 1.5 x 109 T"2
k6 " P2kl
k? = 3.7 x 106 T"1
kg = 1.5 x 108 T"2
k9 = 7.8 x 1015 e-10420/T
k1Q . 3.4 x 10* T'1 e110°/T
* [H20] 5.8 x lO'5 T-2 e58°°/T
= 1.3 x 105 T"1
= 5.3 x 104 T'1 e-
= 8.4 x 106 T"1
7 -1 -HOO/T
= 3.1 x 10 ' T L e 11UUM
,.18 -12280/T
= 3.5 x 10 e
k1? =
6.7 x 10"4 T"1
!8
S Bk
4l
77
-------
APPENDIX A (Continued)
Reaction
19. OH + 03 -» HO
20. H02 + 03
Formaldehyde
2
OH
21. HCHO + hw +' H02 + H02 + CO
22. HCHO + hv * CO + H,
°2
23. OH + HCHO »* H02 + CO
Acetaldehyde
24. CH3CHO + hv V
25. OH + CH3CHO
+ H0
CO
26. CH3C03 + N02 * PAN
27. PAN * CH3 CO^ + N02
28. CH3C03 + NO -» N02 + CH,
29. CH302 + NO + HCHO + H02
Propane
30. OH + PROPANE * P02
31. P02 + NO * H02 + N02
CH3COCH3
Alkanes
32. OH + ALKANE * A02
Rate Constant (ppm min units)
-940/T
'68°/T
= 7.0 x 105 T'1 e-940/T
. 4.8 x !03 I' e
k21 * P5kl
k22 ' P6kl
k = 4.4 x 106 T"1
k24 =
k25 =
k =
27
3.0 x 106 I'1 e250/T
2.1 x 106 T"1
, in!8 -13543/T
1.2 x 10 e
= 3.1 x 106 T'1
= 3.1 x 106 T"1
k3fl . 6.6 x 106 I'1 e-680/T
k = 3.1 x 106 T"1
= 8.0 x 106 I'1 e-560/T
for explicit n-butane
, 6.6 x 10* T'1 «-4°°/T
for lumped > C, alkane
78
-------
Reaction
APPENDIX A (Continued)
Rate Constant (ppm min units)
33. A02 + NO -> 1.3 N02 - + -0.4 NO + 0.9 H02 + 0.6 CH3CHO + 0.1 RCHO
0.5 MEK
Explicit n-butane
* 1.7 NO, + -0.8 NO + 0.9 H02 + 0.15 HCHO + 0.3 CH3CHO
+ 0.1 RCHO + 0.3 CH3COCH3 + 0.45 MEK
Lumped > C, alkane mechanism
k = 3.1 x 106 T"1 for both
Higher Aldehydes
systems
34. OH + RCHO ** RC03
35. RC03 + N02
36. PPN -» RCO.
PPN
37. RCO, + NO -» C,H,0, + N02
3 0
38.
NO
HO, + NO
39. RCHO + hv -> C2H5°2 * C0
+ HO.,
k34 = 9.2 x 106 T"1
k35 2.1 x 106 T"1
, ,A18 -13543/T
k36 = 1.2 x 10 e
k3? « 3.1 x 106 T"1
k3g = 3.1 x 106 T"1
k39 =
Ketones
40. OH + MEK *2 X02
41. X02 + NO * N02 + CH3CHO
= 4.4 x 106 I'1 e-330/T
= 3.1 x 106 T"1
42. MEK + hw
k42 =
43. CH3COCH3 + hw
CH3°2
k43 = P10kl
79
-------
APPENDIX A (Continued )
Reaction Rate Constant (ppm min. units)
Alkenes
0_ e _i 380/T
44. OH + Ethene -2 2HCHO + H02 k44 = 9.7 x 10 T e
N02 - NO
0, g _! 540/T
45. OH + Propene -»TICHO + CH3CHO k^ = 1.8 x 10 T e
+H02
+ N02 - NO
46. OH + Butene +21.8 CHgCHO +0.9 H02 k46 = 5.0 x 106 l"1 e5 °'
+ 0.9 N02 - NO
o 1 2S60/T
47. 03 + Ethene * HCHO + 0.4 CH262 k4? = 4.2 x 10 T" e
+ 0.4 CO + 0.12 H02
48. 0. + Propene -» 0.5 HCHO + 0.5 CHgCHO + 0.2 CH262 + 0.2 CHgCHOO
+ 0.3 CO + 0.2 H02 + 0.1 OH + 0.2 CH302
3 -1 -
kA8 = 3.1 x 10J T e
49. 03 + Butenes * CH^CHO +0.4 CH3CHOO + 0.3 H02 + 0.2 OH
+ 0.45 CH302 + 0.2 CO
. 3.3 x 10* I'1 e-l050/T
50. CH262 + NO - HCHO + N02 k5(J = 3.1 x 106 l"1
51. CH262 + N02 * HCHO + N02 k51 = 3.1 x 105 l"1
52. CH262 + S02 * HCHO + SO^ k52 = 3.0 x 10 T" T"
53. CH262 + H20 * Product kJ3 = 1.5 T
54. CH3CHOO + NO * CH3CHO + N02 k54 = 3.1 x 106 T*1
55. CH3CHOO + N02 * CH3CHO + N03 k55 = 3.1 x 105 l"1
56. CH3CHOO + S02 -» CH3CHO + 80^ k5fi = 3.0 x 10 T"
57. CH3CHOO + H20 * Product k5? = 1.5 T"
80
-------
Reaction
APPENDIX A (Continued)
Rate Constant (ppm min units)
Aromatics
58. OH + Benzene * 0.25 Cresol
+ 0.25 H02 + 0.75 ADD
59. OH + Toluene + 0.15 AR02
+0.20 Cresol +0.20 HO,
+ 0.65 ADD
60. OH + Xylene + 0.25 Cresol
+ 0.25 H02 + 0.75 ADD
61. ADD + NO » 0.75 N(>2 + 0.75
(CHO)
0.75 DIAL + a
2 CH3COCHO
62. OH + DIAL » El
63. El + N02 » El N02
64. El N02 -» El + N02
65. El + NO * 3 N02 -2 NO + a3(CHO)2
+ cu CO + a-»HO_
02
66. OH + (CHO)2 + H02 + CO
67. (CHO), + hv -» HCHO + CO
°2
68. OH + CH^COCHO »* CH.CO, + CO
3 0 J J
69. CHgCOCHO + hu +* C&^CQ +
H02 + CO
= 5.3 x 105 T"1
ksg = 2.7 x 106 T"1
k6(J = 7.9 x 106 T"1 for
lumped xylene
= 1.05 x 107 T"1 for
explicity m-xylene
k,, = 3.1 x 106 T"1
= 1.3 x 107 T"1
= 2.1 x 106 T"1
= 1.2 x 1018 e-13543/T
CH3C03 + au CH3COCHO
k65 = 3.1 x 106 T"1
k66 = 8.8 x 106 T"1
k67 = Pllkl
k68 = 6.6 x 106 T"1
k69 = P12kl
81
-------
APPENDIX A (Continued)
Reaction
70. OH + Cresol * ADD2
71. ADD2 + NO * 0.75 N02
+ 0.75 H02 + 0.75 DIAL
72. N03 + Cresol -» HN03 + Phenoxy
73. Phenoxy + N02 * Products
(o-, p-nitrophenols)
74. ARO, + NO » 0.75 NO,
75.
76.
77.
78.
79.
80.
§o2
+ 0.75 H02 + 0.75 ARCHO
ARCHO + hv * Products
°
OH + ARCHO
ARC03 + N02
PBZN » ARC0
ARCO, + NO -» Ph02 N02
ARC0
PBZN
PhO.
+ NO > Phenoxy + N02
81. OH + SO,
M
* soT
Rate Constant (ppm mia units)
^70
= 1.9 x 107 T"1
kyi = 3-1 x 106 T"1
k?2 = 6.6 x 106 T"1
= 6.6 x 106 T"1
k?4 = 3.1 x 106 T"1
k75 S P13kl
k?6 = 5.7 x 106 T"1
k?7 2.1 x 106 T"1
k?8 - 1 x 1017 e-13025/T
k?g = 3.1 x 106 T"1
k80
106 T'1
kgl = 1.5 x 10
13 "
NOTES
2) a, =
proportionality of photolytic rate for the ith photolytic
reaction rate to k.. P. are a function of solar zenith angle.
variable stoichiometric coefficients which depend on the
benzene, toluene, and xylene concentrations.
82
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
i REPORT NO.
2.
3. RECIPIENT'S ACCESSION»NO.
I. TITLE AND SUBTITLE
DEVELOPMENT OF THE MESOPUFF II DISPERSION MODEL
5. REPORT DATE
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO
J. S. Scire, F. W. Lurmann, A. Bass, S. R. Hanna
9 PERFORMING ORGANIZATION NAME AND ADDRESS
Environmental Research & Technology, Inc.
696 Virginia Road
Concord, Massachusetts 01742
10. PROGRAM ELEMENT NO.
CDTA1D/02-1607 (FY-84)
11. CONTRACT/GRANT NO.
68-02-3733
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental -Sciences Research LaboratoryRTP, NC
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/600/09
15. SUPPLEMENTARY NOTES
16. ABSTRACT ' '
The development of' the MESOPUFF II regional-scale air quality model is described
MESOPUFF II is a Lagrangian variable-trajectory puff superposition model suitable
for modeling the transport, diffusion and removal of air pollutants from multiple
point and area sources at transport distances beyond the range of conventional
straight-line Gaussian plume models (i.e., beyond ^ 10-50 km). It is an extensively
modified version of the MESOscale PUFF (MESOPUFF) model. Major additions and enhance-
ments include: use of hourly surface meteorological data and twice-daily rawinsonde
data; separate wind fields to represent flow within and above the boundary layer;
parameterization of vertical dispersion in terms of micrometeorological turbulence
variables; parameterization of S0? to SO* and NO to NOZ conversion, including the
chemical equilibrium of the HN03/NH3/NHJ103 system; resistance modeling of dry deposi-
tion, including options for source or surface depletion; time- and space-varying ,
wet removal; and a computationally efficient puff sampling function. The scientific
and operational bases for these developments are described. The results of a
preliminary evaluation of several model algorithms during a two-day period of the
Tennessee Plume Study are also presented.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDEIMTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
3. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report!
UNCLASSIFIED
21. NO. OF PAGES
20. SECURITY CLASS (This page)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
-------
INSTRUCTIONS
1. REPORT NUMBER
Insert the EPA report number as it appears on the cover of the publication.
2. LEAVE BLANK
3. RECIPIENTS ACCESSION NUMBER
Reserved for use by each report recipient.
4. TITLE AND SUBTITLE
Title should indicate clearly and briefly the subject coverage of the report, and be displayed prominently Set subtitle
°** report " -* fc "-S - A2
5. REPORT DATE
Indicate "" basis on which u WM selected "*
6. PERFORMING ORGANIZATION CODE
Leave blank.
7. AUTHOR(S)
Give name(s) in conventional order (John R. Doe. J. Robert Doe. etc.). List author's affiliation if it differs from the performing organi-
8. PERFORMING ORGANIZATION REPORT NUMBER
Insert if performing organization wishes to assign this number.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Give name, street, city, state, and ZIP code. List no more than two levels of an organizational hirearchy.
10. PROGRAM ELEMENT NUMBER
Use the program element number under which the report was prepared. Subordinate numbers may be included in parentheses.
11. CONTRACT/GRANT NUMBER
Insert contract or grant number under which report was prepared.
12. SPONSORING AGENCY NAME AND ADDRESS
Include ZIP code.
13. TYPE OF REPORT AND PERIOD COVERED
Indicate interim final, etc., and if applicable, dates covered.
14. SPONSORING AGENCY CODE
Leave blank.
16. SUPPLEMENTARY NOTES
SSS s * wch M: ***** fa cooperation -* Transiation <* *- - «-« «*
16. ABSTRACT
18. DISTRIBUTION STATEMENT
AeTbE^tn^d^^rpn«.0r/liinitati0n "' KaS°'a °**' ^ SBCUrity for exampte "Retease Unlimited." Cite any availability to
19. & 20. SECURITY CLASSIFICATION
DO NOT submit classified reports to the National Technical Information service.
21. NUMBER OF PAGES
Insert the total number of pages, including this one and unnumbered pages, but exclude distribution list, if any.
22. PRICE
Insert the price set by the National Technical Information Service or the Government Printing Office, if known
EPA Form 2220-1 (S-73)
------- |