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=1—2                                                         (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
                 X—X
                CUD
               D-a
                 A
                 A
                A

             X—X
               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  Laboratory—RTP, 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)

-------