United States
Environmental Protection
Agency
Environmental Sciences Research
Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S3-83-111 Dec. 1983
f/EPA Project Summary
Effects of Photochemical
Kinetic Mechanisms on
Oxidant Model Predictions
J. P. Killus and G. Z. Whitten
The comparative effects of kinetic
mechanisms on oxidant model predic-
tions have been tested using the Carbon-
Bond Mechanism II (CBM-II) and the
Demerjian Photochemical Box Model
(DPBM) mechanism, in conjunction
with three air quality models, the
OZIPM/EKMA, the Urban Airshed
Model (UAM), and a trajectory model
with the same inputs and chemistry as
the UAM. Simulations were performed
for the Los Angeles Airshed using a
1974 base case and a 1987 emission
inventory (reflecting controls).
To achieve agreement in absolute
predictions among the three models, it
was necessary to treat two significant
processes in the OZIPM/EKMA simu-
lations: surface deposition and the
photochemical reaction of material aloft
before entrainment by the growing
mixed layer. These two processes are
not normally modeled in OZIPM/EKMA
and could be treated only in a general,
averaged fashion. With this treatment,
however, results for the three models
were within 25 percent overall agree-
ment, and within 10 percent agreement
for the peak value when the same
kinetic mechanism was used.
The two kinetic mechanisms pro-
duced different results. Although the
DPBM mechanism produced results
similar to those for the CBM-II with the
1974 base case for all three models, it
exhibited a greater response to the
1987 control scenario (predicting less
ozone). In simulations employing the
CBM-II, the ozone reductions predicted
using the 1987 emission inventory
were 31, 34, and 33 percent for the
UAM, trajectory model, and OZIPM/
EKMA, respectively. The ozone reduc-
tions predicted by the UAM trajectory
model and OZIPM/ EKMA for the 1987
inventory simulations employing the
DPBM were 41. 51, and 53 percent,
respectively. Also, reduction of peak
ozone anywhere in the UAM simulation
grid was 26 percent for the CBM-ll/Air-
shed simulations and 40 percent for the
DPBM/Airshed simulations.
The greater response to hydrocarbon
control exhibited by the DPBM mecha-
nism compared to that of the CBM-II is
contrary to the findings of Jeffries et al.
(1981), which indicated a greater re-
sponse to control for the CBM-II than
for the DPBM. However. Jeffries et al.
examined uniform control of hydro-
carbons, whereas the 1987 emission
inventory used in this study included
the effects of population growth and
differing controls on various source
categories. These factors resulted in a
shift in hydrocarbon speciation, most
notably a decrease in olefins relative to
other hydrocarbons. The DPBM demon-
strated an overresponse to the olefin
component of the hydrocarbon mix.
Thus, the difference in the models'
response to the control strategy was
traced to a specific feature of the DPBM
mechanism.
This Project Summary was developed
by EPA's Environmental Sciences He-
search Laboratory. Research Triangle
Park, NC, to announce key findings of
the research project that is fully docu-
mented in a separate report of the same
title (see Project Report ordering infor-
mation at back).
-------
Introduction
A long-term goal of EPA's research
program is the development of urban and
regional air quality simulation models
(AQSMs) to be used in planning accurate
and scientifically defensible control strate-
gies. Because the pollutants with the
greatest impacts on human health are
secondary pollutants formed by chemical
reactions occurring in the atmosphere,
an understanding of the chemistry that
produces these pollutants is critical to the
development of AQSMs. In recent years,
significant efforts have been made to
explain the chemical transformations
that occur in photochemical smog sys-
tems and to develop chemical kinetic
mechanisms that can be used in the
AQSMs to explain the formation of ozone
and other secondary pollutants. However,
comparisons among the various chemical
kinetic mechanisms that have been de-
veloped for air quality analysis reveal
significant differences in performance.
Furthermore, different AQSMs frequently
give dissimilar predictions, even when
the same kinetic mechanism is used.
It would be useful to be able to deter-
mine the reasons for discrepancies in
AQSM results, i.e., to have procedures for
ascribing these reasons to specific por-
tions of the AQSM such as the kinetic
mechanism, the model formulation and
assumptions, or the data base inputs. The
intent of the project reported here was to
develop a procedure for analyzing the
comparative effects of chemical kinetic
mechanisms used in various AQSMs.
Procedure
Three models were selected to test the
effects of kinetic mechanism substitution
on control strategy assessment: OZIPM/
EKMA, which accepts the kinetic mecha-
nism as an input; the Urban Airshed
Model (UAM), in which the kinetic mecha-
nism is a module that must be receded
when the substitution is made; and the
Airshed Compatible Trajectory Model
(ACTM), which uses the same kinetic
module as the UAM. The UAM and ACTM
already contain the Carbon-Bond II Kinetic
Mechanism (CBM-II) and an extensive
validation base exists for both these
models. Additionally, the CBM-II has
been independently validated using labo-
ratory smog chamber data sets for both
indoor and outdoor smog chambers.
Included in the CBM-II validation set was
the Bureau of Mines (BOM) data base
used in the calibration and validation of
the Dodge/EKMA mechanism.
The Demerjian Photochemical Box
Model (DPBM) mechanism was chosen
for comparison with CBM-II. The DPBM is
a multispecies lumped molecular mecha-
nism that has also been validated with
the BOM data set. Both its BOM data
validation and its compactness make the
DPBM Mechanism a logical candidate for
comparison with the CBM-II in UAM
simulations.
Intermodel comparisons were planned
along a high ozone trajectory in Los
Angeles for which an extensive UAM
data base and numerous control strategy
scenarios already exist. A good match of
predictions between the UAM and the
ACTM along the trajectory was easily
obtained since emission inputs, vertical
dispersion, and chemical kinetics are
identical for the two models, and since
factors that would produce discrepancies
(such as a high degree of wind sheer)
were absent on this simulation day. To
obtain a match of OZIPM/EMKA with the
UAM and ACTM, however, two phenom-
ena not usually treated in OZIPM/EKMA
had to be considered: (1) Reaction of
precursor material aloft, which must be
considered if the chemical process is not
to be retarded by entrainment of unre-
acted material, was simulated by entrain-
ing the average of material as predicted
by the ACTM; (2) surface deposition in
OZIPM/EKMA was included as a first-
order loss process for Oa and
0.42
Results
With these modifications to OZIPM
/EKMA, good matches were achieved for
both the 1974 base case and 1987
control scenario (Figures 1 and 2) among
all three models for both kinetic mecha-
nisms. Also, UAM predictions using the
DPBM were similar to results obtained
with the CBM-II (Figure 3). However,
predictions obtained for the 1987 control
scenario tended to show a substantially
greater reduction in ozone with the
DPBM mechanism compared to that
predicted with the CBM-II (see Table 1).
Intermodel comparisons using the same
mechanism revealed similar ozone re-
ductions with the CBM-II, whereas the
UAM with the DPBM predicted somewhat
less ozone reduction than the ACTM and
OZIPM/EKMA with either mechanism.
Reduction of peak ozone predicted any-
where in the UAM grid was less than the
trajectory-specific reductions.
The greater response to hydrocarbon
control exhibited by the DPBM mecha-
nism compared to that of the CBM-II is
contrary to the results of Jeffries et al.
(1981), who found that the CBM-II pre-
dicted greater response to control than
the DPBM. However, Jeffries et al. exam-
ined uniform control of hydrocarbons,
whereas the 1987 emissions inventory
used in this study was a realistic control
scenario that included both the effects
of population growth and different de-
0.36
0.30
T?
o
0.24
0.12
0.06
• OZIPM/EKMA
•ACTM
• Airshed
1987 Emissions
800 900 1000 1100 1200 1300 1400 1500 1600 1700
Time
Figure 1. Comparison of model results for the CBM-II simulations.
-------
0.42
0.36
0.30
•7)0.24
O
• 0.18
0.12
0.06
£KMA
Traj
Airshed
1987 Emissions
800
900 1000 1100
1200 1300
Time
1400 1500 1600 1700
Figure 2. Comparison of the model results for the DPBM mechanism simulations.
grees of control on various emission
source categories. These factors resulted
in a shift in hydrocarbon speciation,
including a reduction in the olef in fraction.
The greater reduction of ozone pre-
dicted by the DPBM mechanism is attrib-
uted to the greater sensitivity of this
mechanism to olefinic hydrocarbons. In
tests using the DPBM mechanism to
simulate smog chamber experiments, the
mechanism was shown to overpredict the
reactivity of olefins. Correction of this
error in the DPBM would probably result
in closer agreement between the per-
formance of the two mechanisms.
References
Jeffries, H. E., K. G. Sexton, and C. H.
Salmi (1981), "Effects of Chemistry
and Meteorology on Ozone Control
Calculations Using Simple Trajectory
Models and the EKMA Procedure,"
EPA-450/4-81 -034, U.S. Environmen-
tal Protection Agency, Research Trian-
gle Park, North Carolina.
-------
50
40
12
18
20-
10
Downtown LA
1974 CBM2 —
_ 1974 Demerjian—
Observed •
24 0
-nSO 50
12
18
24
i lBBBBB
40 40
30-^30
20w20
o
JO 10
Pasadena
1974 CBM2
1974 Demerjian
Observed a
tat
i i i i
iiiliittililPBIPB
50
06 12 18 24
Time (Hours)
0 6 12 18 24
6 12 18
Time (Hours)
50
40
30
20
10
24
40
30
I
8
20
10
Azusa
1974 CBM2 —
7374 Demerjian—
Observed a
BBBB
50 50
40 40
30^30
20§20
10 10
12
18
24
Pomona
1974 CBM2
1974 Demerjian
Observed
50
40
6 12 18
Time (Hours)
6 12 18
0
24 0
Ban
50
40
30
20
10
I
5?
30
20
10
Upland
1974 CBM2 —
.1974 Demerjian—g
Observed D
50
40 40
30-^30
20 §20
10 10
6 12 18 24
Time (Hours)
6 12 18 24
Z Fontana *
- 1974 CBM2
- 1974 Demerjian-— •
• Observed n • •
50
40
30
20
10
6 12 18
Time (Hours)
24
6 12 18
Time (Hours)
24
Figure 3. Comparison of Oa results for Los Angeles, June 1974, (Airshed base case) for the
DPBM mechanism and CMB-II.
-------
Table 1. Effect of Controls on ACTM, Airshed, and OZIPM/EKMA Model Predictions for the Fontana Trajectory
fa) Carbon-Bond Mechanism
1974 Ozone
Prediction fppm)
Time ACTM
1500 0.376
1600 0.376
1700 0.36
Peak
trajectory
value 0.376
Peak value
anywhere in
Airshed grid
Airshed
0.36
0.35
0.36
0.36
0.39
OZIPM/
EKMA
0.375
0.367
0.36
0.375
1987 Ozone
Prediction fppm)
ACTM
0.22
0.24
0.25
0.25
Airshed
0.21
0.234
0.25
0.25
0.29
OZIPM
0.22
0.24
0.25
0.25
1987/1 974 Ratio
ACTM
0.59
0.64
0.69
0.66
Airshed
0.58
0.67
0.69
0.69
0.74
OZIPM/
EKMA
0.59
0.65
0.69
0.66
Percent Oa Reduction
ACTM
41
36
31
34
Airshed
42
33
31
31
26
OZIPM/
EKMA
41
35
31
33
(at 1700
near
Fontana)
(at 1700
near
Fontana)
(b) DPBM Mechanism
1974 Ozone 1987 Ozone
Prediction (ppm) Prediction (ppm) 1987/1974 Ratio
Time
OZIPM/
ACTM Airshed EKMA ACTM
OZIPM/
Airshed EKMA ACTM Airshed OZIPM
Percent Oa Reduction
ACTM
OZIPM/
Airshed EKMA
1500 0.37
1600 0.36
1700 0.33
Peak
trajectory
value 0.37
Peak value
anywhere in
Airshed grid
0.37 0.40
0.34 0.40
0.35 0.39
0.37
0.40
(at 1400
near
Fontana)
0.40
0.17
0.18
0.18
0.18
0.20
0.20
0.22
0.22
0.24
(at 1600
near
Fontana)
0.17 0.46
0.18 0.30
0.19 0.54
0.19
0.49
0.54
0.59
0.63
0.39
0.60
0.42 54
0.45 50
0.49 45
0.47
51
46
41
37
41
40
58
55
51
53
J. P. Killus and G. 2. Whitten are with Systems Applications. Inc., San Rafael, CA
94903.
Marc/a C. Dodge is the EPA Project Officer (see below).
The complete report, entitled "Effects of Photochemical Kinetic Mechanisms on
Oxidant Model Predictions, "(Order No. PB 84-113 752; Cost: $JO.OO, subject to
change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Environmental 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
&
AU.S. GOVERNMENT PRINTING OFFICE 1984-759-102/817
------- |