5EPA
United States
Environmental Protection
Agency
Environmental Sciences Research
Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/PS4-80-041 Sept. 1980
Project Summary
Livermore Regional Air Quality
(LIRAQ) Model (Transfer to
EPA)
William H. Duewer, John J. Walton, Keith E. Grant, and Hoyt Walker
Methods were developed to convert
data from the Regional Air Pollution/
Regional Air Monitoring Study
(RAPS/RAMS) into a format
compatible with the Livermore
Regional Air Quality (LIRAQ) models.
Changes made in the LIRAQ models
include (1) improved map plotting
routines, (2) updated chemical
mechanism, and (3) a partial SOz
oxidation mechanism. In the final
report which is available from NTIS
(see last page of this summary),
several appendices serve as a user's
guide to the LIRAQ models and the
associated data conversion routines.
Introduction
The Livermore Regional Air Quality
(LIRAQ) models were developed when,
through the Research Applied to
National Needs (RANN) program, the
National Science Foundation supported
a 21/2-year interagency effort by the
University of California Lawrence Liver-
more Laboratory (LLL), the Bay Area Air
Pollution Control District (BAAPCD),
and the NASA Ames Research Center
(ARC) to develop and evaluate the
LIRAQ-1 and LIRAQ-2 models. These
codes are deterministic two-
dimensional Eulerian grid air-quality
models that compute spatial and
temporal distributions of significant air
pollutants from specified emissions,
meteorological conditions, topography,
and initial and boundary concentra-
tions. LIRAQ-1 deals with chemical
species that are not strongly affected by
chemical reactions and uses a flux-
corrected transport scheme. LIRAQ-2
treats chemically reactive species, but
uses simple upstream differencing to
treat the transport.
In 1976, LLL entered into an inter-
agency agreement with the U.S.
Environmental Protection Agency to
modify and extend the LIRAQ models to
include sulfur species (SO: and
sulf ates) and to prepare data processors
for the. necessary input files to be
developed from EPA "Regional Air
Monitoring Study (RAMS) data tapes so
that the LIRAQ models could participate
in the Regional Air Pollution Study
(RAPS). The improved codes are
designated LIRAQ-1S and LIRAQ-2S.
The present EPA-supported inter-
agency effort consists of a limited
program intended to facilitate the
examination of LIRAQ performance as
part of RAPS. This effort included only
minimal model development; the bulk of
the effort was spent to develop data pre-
processors and operating procedures
for use by remote users of LIRAQ on the
Lawrence Berkeley Laboratory (LBL)
computer system.
A temporary emissions inventory was
provided by the EPA late in 1978. As a
result, less effortthan was intended has
been devoted to model verification and
the analysis of poorly known but
important model inputs such as speci-
fication of upper and upwind boundary
conditions, and an estimate of the
-------
reliability of the procedures used in
estimating mixing depths. As yet, no
final emissions inventory has been
established and no model tuning has
been attempted. A limited set of sensi-
tivity studies have been carried out, but
are not described in this document since
no funding exists to analyze the results.
Conclusions
Under an interagency agreement
with the EPA, LLL has modified the
[ IRAQ models for participation in the
Regional Air Pollution Study (RAPS).
The modification to the [ IRAQ models
have included:
1 Modifications to the chemical
mechanism in LIRAO-2S, so that
the model now includes a repre-
sentation of the homogenous gas
phase photochemistry for
conversion of 5O and SO 4 and
an update of the reaction rates
and mechanisms to reflect the
data available in March 1976.
2. Several input checks (in both
models) to ensure that all inpUt
files (source, meteorology, and
initial condition) refer to the same
day.
3. Changes in the topographic input
designed to improve the versatil-
ity of river and coastline-plotting
routines.
In addition, [ LL has:
1. Developed processors that
generate LIRAG input files from
RAPS data tapes.
2. converted MASCON and VERF
processors to standard FORTRAN
and exported them to Lawrence
Berkeley Laboratory )LBL)
3, Provided documentation.
Because LLL has not yet been
provided with an emission inventory
considered reliable by the EPA, little can
be said about how reliable or suitable
our estimates are for initial conditions,
boundary conditions, or mixing depths
Sensitivity experiments for St Louis
concentrations indicate that model
predictions for 03 are sensitive to the
upper boundary condition. There are
almost no data in the standard RAPS
data base that can be used to guide the
choice of upper boundary conditions,
although pertinent data were taken in
some of the special studies sponsored
by RAPS. Unless data from the special
studies are used, pollutant concentra-
2
tions aloft will provide a set of poorly
known parameters capable of strongly
influencing model predictions. The
treatment of such factors will te a
central issue in questions of verification
and in future applications of complex
models sUch as control-strategy evalua-
tion At present, there is no adequate
basis for judging LIRAQ performance.
Of the changes made in [ IRAQ. for
RAPS, the update of the chemical
mechanism and changes in the treat
ment of the steady-state species for
LIRAO-2S, the consistency checks
added to the input, and the transfer of
VERF and MASCON codes to LBL, all are
likely to be immediately used by oilier
[ IRAQ users. The addition of SO2”S04
chemistry seems less likely to be usef uI.
Recommended future changes in
LIRAQ include a further updating of
model chemistry, removal of SO 2 , SO 4 ,
and HNO 3 from LIRAQ-2S, and the
addition of peroxyacylnitrates, a simpli-
fication of LIRAO-2S numerics, and
possibly an expansion of the LIRAQ-2S
grid. These changes should result in
either a material decrease in running
time or an increase in spatial resolution.
The [ IRAQ model makestheassuriip
tion that the dispersion, transport, and
transformation of pollutants can be
represented by treating only a single
atmospheric layer extending from the
surface to the normal temperature
inversion This assumption, coupled
with the assumed behavior of pollutant
concentration in the vertical, will be
most valid when:
1 Pollutants are emitted primarily at
the surface or below the inversion
(the single-layer model does not
keep track of pollutants emitted
above the inversion).
2. Vertical dispersion is limited by a
strong and persistent inversion.
3. Vertical mixing in the layer is
rapid compared with horizontal
transport across a grid cell.
4. Pollutant concentrations in the
layer result primarhy from
emissions within the layer rather
than prescribed fluxes through
the inversion.
Application of the [ IRAQ model to St.
Louis, rather than the typical west-
coast-maritime, inland-valley regime
for which it was designed, strains the
condition of the single-layer
assumption. Many emission sources,
particularly for SO 2 , are elevated point
sources that persist in the inversion and
then may be mixed to the surface asthe
inversion rises, a process [ IRAQ does
not treat. The atmospheric structure in
St. Louis often involves multiple, weak
or nonexiSting inversions creating
layers that are alternately coupled and
isolated during a diurnal cycle, a
process crudely treated in [ IRAQ.
Vertical mixing can be rapid during
daylight hours, but is often weak, thus
the height of mixing (or single-layer
depth) is not always well-defined The
elevated emissions and the multiple
stable layers coupled with the diurnal
cycle mean that pollutant concentra-
tions in the single layer may often be
strongly influenced by pollutants not
emitted in that layer. Taken together,
the characteristics of [ IRAQ and St.
Louis indicate that expectations of good
verification should not be high and
verification and analysis should be
carefully pursued.
The preliminary results of [ IRAQ
simulations of 03 in St. Louis involve
substantial underprediction. This could
reflect basic model error, improper
boundary/initial conditions, emissions
inventory errors or excessive mixing of
NO from elevated sources down to the
surface.
-------
William H. Due wer, John J. Walton, Keith E. Grant, and Hoyt Walker are wit/ i
Lawrence Liver nore Laboratory, University of California, Livermore, CA
94550
J. H. Shre If/er is the EPA Project Officer (see below)
The complete report, entitled “Livermore Reg,’ona/Air Quality (LIRAQ) Model
(Transfer to EPA),” (Order No. 80221591; Cost: $14.00, subject to change)
wi/I be available from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-557 4650
The EPA Project Officer can be contacted at:
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Re.cearch Triangle Park, NC 27711
3
------- |