United States
 Environmental Protection
 Agency
Atmospheric Sciences Research
Laboratory
Research Triangle Park NC 27711
 Research and Development
EPA/600/S8-85/017 Nov. 1985
 Project Summary
Green  River Air Quality  Model
Development:  MELSAR—A
Mesoscale Air  Quality Model for
Complex Terrain
K. J. Allwine and C. D. Whiteman
  Pacific Northwest Laboratory  has
developed an air quality model for ap-
plication in valleys as part of the U.S.
Environmental Protection Agency (EPA)
Green River Ambient Model Assess-
ment program.  The purpose of  the
program is to provide air quality assess-
ment tools applicable to the Green River
Oil Shale Formation region of western
Colorado, eastern Utah, and southern
Wyoming. This region has the potential
for large-scale growth because of its
vast energy resources,  especially oil
shale.
  MELSAR is a  mesoscale  air quality
model for predicting the concentrations
of air pollutants released from multiple
sources. The model is a Lagrangian puff
model for application in complex ter-
rain,  principally  at long source-to-re-
ceptor  transport distances (tens to
hundreds of kilometers) and for short
pollutant averaging times (1 to  24 h).
Terrain influences are treated explicitly
by using a three-dimensional mass-con-
sistent flow model with an account for
the influences of terrain roughness on
diffusion, terrain-following plume tra-
jectories with optional corrections for
the plume ascending hills,  and a pa-
rameterized  treatment of pollutant
sources located  in valleys. The  model
handles releases from both point and
area sources and makes the conserv-
ative assumption that the  pollutants
released are inert and nondepositing.
MELSAR was developed for application
in a region encompassing western Col-
orado, eastern Utah, and southern Wy-
oming. However, the computer code
itself is not site specific and could
conceivably be applied anywhere pro-
vided if the proper inputs were devel-
oped and the parameterizations within
MELSAR are applicable to the region to
be modeled.
  This Project Summary was developed
by EPA's Atmospheric Sciences Re-
search Laboratory, Research Triangle
Park, NC, to announce key findings of
the research project that is fully docu-
mented in two separate volumes (see
Project Report ordering information at
back).

Introduction
  An air quality model was developed to
predict short averaging  time (<24 h)
pollutant concentrations  resulting from
the mesoscale transport of pollutants
released from multiple sources in com-
plex terrain. The  Mesoscale  Location
Specific Air Resources (MELSAR) model
was developed for application  in the
complex terrain of the Green River Oil
Shale formation region of Colorado, Utah,
and Wyoming. Its  purpose is to aid in
assessing the air pollution potential from
development of the extensive oil shale
reserves  and other industrial,  mining,
manufacturing, or power production de-
velopment within this region. The project
report consists of two volumes: volume 1
contains the computer model overview,
technical description, and user's guide,
and volume 2 contains the appendices,
which include listings of the FORTRAN
code.

-------
  MELSAR is a Lagrangian puff model for
application  at  long  source-to-receptor
distances (tens to hundreds of kilometers)
and for  short concentration  averaging
times (1  to  24 h). These distances  and
averaging times are important in esti-
mating the  pollutant concentrations at
the Prevention of Significant  Deteriora-
tion Class I areas in the region.
  MELSAR was designed for application
in a specific 500- by 450-km complex
terrain region  of the Green  River Oil
Shale Formation. It  can compute con-
centrations for up to 20 sources and  two
pollutants at a time. The influences of the
terrain on pollutant transport and diffu-
sion are treated explicitly in MELSAR.
The transport winds  are computed from
measured upper-air and surface weather
data by using a mass-consistent, three-
dimensional flow model. Steering of the
winds around major terrain features
during stable atmospheric conditions is
accounted for, and the effects of terrain
roughness  on pollutant  diffusion  are
treated.  In addition,  sources  located in
valleys are given special treatment.  Pol-
lutants trapped in valley-drainage flows
during the nighttime are ventilated to the
regional winds during the morning transi-
tion period. The pollutants trapped in the
valley are  treated as a  line  source at
sunrise.

Overview
  MELSAR  is a mesoscale air quality
model for predicting  the concentrations
of pollutants released from up to 20 point
and area  sources. Two pollutants can be
treated during  any run.  MELSAR is a
Lagrangian puff model for application in
complex terrain, principally at long source-
to-receptor transport distances (tens to
hundreds of kilometers) and for  short
averaging times (1 to 24 h). The ground-
level concentrations may be computed on
up to four 25- by 25-receptor grids and at
10  individual receptors. The  pollutants
are assumed to be  nondepositing and
inert. This is a conservative assumption
for computing air concentrations of  pri-
mary pollutants.
  MELSAR is set to operate on a 500- by
450-km  region encompassing western
Colorado, eastern  Utah,  and  southern
Wyoming. This region is referred to as the
Green River Ambient Model Assessment
(GRAMA) region. The coordinate system
for the GRAMA  region is  a rectangular
system, and input data can be specified in
either the GRAMA, Universal Transverse
Mercator (UTM)  or  latitude-longitude
coordinate systems.
  Simulation periods from hours to days
(maximum 30 days) can be accommodat-
ed by MELSAR. In addition, intermediate
meteorological  and terrain output files
can  be saved  and cataloged for  later
applications of  MELSAR; this saves the
cost and time of regenerating these files.
A file tracking and cataloging system is
provided with MELSAR.
  MELSAR utilizes a three-dimensional,
mass-consistent flow model to determine
the time-varying and space-varying winds
over the region. The model is diagnostic
and  uses  upper-air data from up to 10
stations and surface weather data  from
up to 15 stations as input. The flow model
also accounts for flow channeling around
major terrain features during stable at-
mospheric conditions by using the con-
cept of a dividing streamline height. This
height is calculated from available upper-
air data and from analysis of the terrain
data. The  winds are specified on up to
nine levels  above  the terrain surface.
Using the same upper-air and surface
weather data that are used to determine
the  winds,  MELSAR produces  hourly
gridded fields of temperature, pressure,
friction velocity, convective velocity, Monin-
Obukov length, mixing height, and stabil-
ity classification. These gridded quantities
are written to disk files and can be  used
during other applications of MELSAR.
  The pollutant puffs diffuse in a Gaus-
sian  fashion  in which  the vertical dis-
tribution is modified by  multiple reflec-
tions from the ground and  an upper
mixing lid. The horizontal and vertical
diffusion coefficients are  computed as
functions of travel time and the standard
deviations in the horizontal and vertical
winds, trv  and erw, respectively. The fol-
lowing  two options are available  in
MELSAR for  computing av and  CTW: (1)
empirical equations relating to 
-------
  For multiple applications of MELSAR, it
may not be necessary to run TERRAIN,
MET, and POLUT for each application.
The output files for one application may
be used during  other applications. An
approach  for naming and tracking  the
input and output files from MELSAR is
provided with the  model. Tracking  the
files will ensure that documentation is
available for future reference and that all
the input and output files will be available
and retrievable for future MELSAR runs.
  The TERRAIN program produces grid-
ded fields of terrain  statistics from  the
base terrain file, which contains 500 by
450 terrain heights; each terrain height is
the average height of a 1 -km2 area. This
program enables users to compute  the
necessary terrain statistics for any sub-
domain in the GRAMA region that they
want to model.  In addition, TERRAIN
provides the necessary flexibility to inves-
tigate the effects of terrain smoothing on
ground-level concentrations during any
sensitivity testing of MELSAR.
  MET is a very flexible meteorological
data processor. It is designed to receive
up to one month's worth of upper-air and
surface weather input data from up to 10
upper-air and 15 surface stations over
different time periods and sampling inter-
vals. It can handle missing data and was
designed  to handle  any  data that  are
available for a specified period of time.
Observations  from National Weather
Service stations, special studies, or in-
tensive field programs can be used.
  POLUT is a Lagrangian puff model in
which  the pollutant  distribution is  de-
scribed in a Gaussian fashion about the
puff's center of mass. The distribution in
the vertical is modified by the treatment
of multiple reflections from the ground
and  an  upper mixing lid. Ground-level
concentrations for  each modeling time
step  for two pollutants can be computed
for pollutants released from  up to five
sources, either point or area sources. The
area sources are treated as virtual point
sources.  Concentrations are  computed
on up to four 25- by 25-receptor grids and
to 10 individual receptors specified any-
where at ground level in the modeling
domain.
  POLPRC  computes moving-average
concentrations of pollutants at  each
receptor for up to  three  user-specified
averaging times. The averaging times can
range from 1 to24h. Up to 20 sources can
be processed at one time. The output data
are presented as tables of highest and
second highest pollutant concentrations
for the duration of the simulation at each
receptor for each  pollutant for each
averaging time. The highest and second
highest moving-average concentrations
resulting from the sum of all sources and
the contribution of each  source to the
highest and second highest sum are
computed.  Highest and second highest
moving average concentrations can also
be computed for each source individually.
Additional tables  list the time of occur-
rence of the highest and second highest
values.

Data Requirements
  The input data required by MELSAR are
the following:

  1.  gridded terrain data
  2.  upper-air weather data
  3.  surface weather data
  4.  points from which  to  interpolate
     wind data (primarily for boundary
     conditions)
  5.  gridded land-use categories (for
     determining surface roughness)
  6.  source valley characteristics
  7.  pollutant source inventory
  8.  receptor layout
  All  but the  upper-air  and surface
weather data must be  specified  within
the modeling region. Data from weather
stations outside the modeling region can
be used by MELSAR.
  The gridded terrain data for the GRAMA
region are provided with MELSAR and do
not have to be developed by the user. Also
included with MELSAR are four files of
surface and upper-air weather data, one
file of wind interpolation points for runs
using the entire GRAMA region, and one
file of  gridded  land-use categories for
runs using the entire GRAMA region.
  New upper-air  and surface  weather
data files can be developed  if meteor-
ological periods other than the ones
provided are desired for model runs. For
applications in which the modeling do-
main is a subset of the GRAMA region, a
new wind interpolation point file and a
new land-use categories file must be
constructed. For each group of sources to
be modeled, a new source inventory file
and a new source valley characteristics
file must be constructed.  New receptor
layout files must be constructed for each
new receptor configuration.

Conclusions and
Recommendations
  The model shows promise for use as a
planning tool and, eventually, as a reg-
ulatory tool. MELSAR has not undergone
any rigorous operational testing, sensi-
tivity  analyses, or  evaluation studies.
Three major steps are required  before
MELSAR is ready for general use in the
region of western Colorado, eastern Utah,
and southern Wyoming: (1) operational
and sensitivity testing,  (2) testing  for
upgrading, if necessary, of specific algo-
rithms by using the limited tracer data
sets available  for the region, and (3) a
regional-scale  tracer experiment to pro-
vide a data set for  further testing and
validation of MELSAR. MELSAR has a
highly modularized computer code, which
facilitates upgrades and additions to the
code. The addition of chemistry and depo-
sition  parameters to MELSAR would be
straightforward  and would  enable
MELSAR to be  used as a tool for estimat-
ing contributions to acid deposition.

-------
     K. Jerry Allwine and C. David Whiteman are with Battelle, Pacific Northwest
      Laboratory, Rich/and, WA 99532.
     Alan H. Huber is the EPA Project Officer (see below).
     The complete report consists of two volumes, entitled "Green River Air Quality
      Model Development MELSAR—A Mesoscale Air Quality Model for Complex
      Terrain:"
         "Volume 1.  Overview, Technical Description and User's Guide." (Order No.
         PB 85-24721 I/AS; Cost: $16.95)
         "Volume 2. Appendices."(Order No. PB 85-247 229/AS; Cost: $28.95)
     The above reports will be available only from: (costs subject to change)
           National Technical Information Service
           5285 Port Royal Road
           Springfield, VA 22161
            Telephone: 703-487-4650
     The EPA Project Officer can be contacted at:
           Atmospheric Sciences Research Laboratory
           U.S. Environmental Protection Agency
           Research Triangle Park, NC 27711
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300

EPA/600/S8-85/017
          0000329   PS
          U  S  F=NVIR  PROTECTION AGENCY
          REGION  5  LIBRARY
          230  S  OEAR80RN  STREET
          CHICAGO               IL   60604

-------