v>EPA
                  United States
                  Environmental Protection
                  Agency
                               Atmospheric Sciences
                               Research Laboratory
                               Research Triangle Park NC 27711
                  Research and Development
                               EPA/600/S3-85/042 July 1985
Project  Summary
Evaluation  of  Chemical  Reac-
tion  Mechanisms for  Photo-
chemical  Smog:  Part  III. Sensi
tivity  of EKMA to Chemical
Mechanism  and  Input
Parameters
                  Toby B. Shafer and John H. Seinfeld
                   Six chemical reaction mechanisms
                  for photochemical smog described in
                  Parts I and II are used to study the effect
                  of input parameters on volatile organic
                  compound (VOC) control requirements
                  needed to meet the National Ambient
                  Air Quality Standard for ozone. The pa-
                  rameters studied are initial VOC com-
                  position,  dilution  rate,  post-8-A.M.
                  emissions, base-case (present-day) 03
                  levels, entrainment from aloft of VOC
                  and  ozone, initial MONO and initial
                  VOC/NOX ratio. The  Empirical Kinetic
                  Modeling Approach (EKMA) was used
                  to generate ozone isopleths for each
                  chemical mechanism. The VOC control
                  needed to reduce the maximum ozone
                  concentration from, some  present-day
                  value to 0.12 ppm, assuming no IMOX
                  control and a specified initial VOC/NOX
                  ratio, was  calculated  using the  six
                  chemical reaction mechanisms. The ini-
                  tial VOC/NOX ratio is found to have the
                  largest effect of all the parameters stud-
                  ied  on VOC control requirements.
                  Choice of chemical mechanism, ozone
                  entrainment from aloft and the compo-
                  sition of the initial VOC mixture also
                  have a large effect on predicted control
                  requirements. To reduce the degree of
                  uncertainty in control predictions using
                  EKMA it is necessary to establish as ac-
                  curately as possible the composition of
                  urban air in early morning. Also,  be-
                  cause of the substantial effect  the
                  choice of chemical mechanism has on
                               the  predicted control  requirements
                               using EKMA, it is important that future
                               work continue to be directed toward
                               evaluating candidate chemical mecha-
                               nisms with respect to their ability to
                               simulate atmospheric smog chemistry.
                                 This Project Summary was devel-
                               oped by EPA's Atmospheric Sciences
                               Research 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
                                 Currently there are several chemical
                               reaction mechanisms available for use
                               in predicting hydrocarbon and NOX re-
                               ductions needed to meet the National
                               Ambient Air Quality Standard (NAAQS)
                               for ozone for a particular region. Be-
                               cause of the numerous organic species
                               in the atmosphere, it is computationally
                               infeasible to use a chemical reaction
                               mechanism that contains every reaction
                               of every species in carrying out a con-
                               trol calculation. Each chemical mecha-
                               nism reduces the number of organic
                               species in a somewhat different way.
                               This reduction, or "lumping," would not
                               be a concern if each of the  chemical
                               mechanisms gave similar predictions
                               under typical atmospheric conditions.
                               As discussed in Part II, however, several
                               investigators have shown that different

-------
     Dchemical  smog  reaction  mecha-
"nisHs predict  significantly different
 control requirements  under the same
 conditions. In addition, these investiga-
 tors have found that the sensitivity of
 the mechanism's predictions to
 changes  in input parameters varies
 among mechanisms. Because of the im-
 portance  of making intelligent hydro-
 carbon control decisions, it is necessary
 to understand in as fundamental way as
 possible  how the choice  of chemical
 mechanism, as well as the choice of in-
 put parameters, affects control predic-
 tions. The object of this work, therefore,
 was to present a comprehensive,  sys-
 tematic study of the sensitivity of  pre-
 dicted hydrocarbon  control require-
 ments for photochemical smog to the
 choice of chemical mechanism  and
 choice of all input parameters in the
 control calculation.
   To study the sensitivity  of hydrocar-
 bon control predictions to choice of
 chemical  mechanism and input param-
 eters, we used  the U.S. Environmental
 Protection Agency's Empirical Kinetic
 Modeling Approach (EKMA) in which
 ozone isopleths are generated as a func-
 tion of initial volatile organic compound
 (VOC) and NOX concentrations using a
 chemical  reaction mechanism. In the
 EKMA methodology,  a chemical reac-
 tion mechanism  for photochemical
 smog is used to predict the maximum
 ozone concentration  achieved over a
 fixed time or irradiation, say 10 hours,
 as a function of initial precursor concen-
 trations (VOC and NOX). The maximum
 ozone concentrations  corresponding to
 each set  of initial  concentrations are
 then represented as  isopleths, from
 which control requirements are ob-
 tained graphically by assuming that the
 reductions in  initial  concentrations
 needed to lower the  maximum ozone
 concentration from one isopleth  (the
 base-case value) to another (usually the
 NAAQSof 0.12 ppmlare linearly related
 to the emissions of each precursor. The
 chemical  reaction mechanism is inte-
 grated  from  early morning, e.g. 0800
 hours, to calculate the evolution of
 species concentrations in a hypothetical
 air parcel that starts from center city and
 is  advected by  the wind. The air parcel
 may increase or decrease in size due to
 changes  in the mixing height and ac-
 quire  additional pollutants from both
 fresh  source  emissions and entrain-
 ment of aged pollutants from aloft. The
 simulations on which  the  isopleths are
 based can be performed at two different
 levels of  detail: Level  II, which requires
                                  2
a specific meteorological description of
the air parcel trajectory and emissions
along the trajectory, and Level III, which
makes standard assumptions about the
meteorology and which uses a region-
wide average for emission rates.  It has
been shown that the sensitivity of ozone
predictions to choice of chemical mech-
anism does not change  appreciably
when one imbeds the mechanisms in
models that  are more sophisticated
than the simple Lagrangian box model
on which EKMA is based. Therefore, we
expect that the results of the current
study, which we obtained  by use of
EKMA, are a valid indication of the ef-
fects of chemical mechanism and input
parameters on predicted organic con-
trol requirements regardless of the level
of sophistication of the air quality model
in which the mechanism is imbedded.

Results
  We have examined the sensitivity of
predicted VOC control  requirements to
the choice of chemical mechanism and
to the variation of parameters associ-
ated with the chemical and meteorolog-
ical nature of the  atmosphere. Each
chemical reaction  mechanism was
found to respond to a variation of the
parameters  in a somewhat different
way; in many cases the relative sensitiv-
ity of the mechanism can be explained
from an analysis of the intrinsic chemi-
cal features of the mechanism. The sen-
sitivity calculations can be summarized
concisely as follows:

Initial VOC Composition
  The choice of chemical mechanism
has a large effect on VOC control predic-
tions and on the sensitivity of the  pre-
dictions to a change in VOC composi-
tion.  The ERT mechanism is  most
sensitive to  a change  in composition;
the CBM least sensitive.

Aldehyde Content
  Predicted VOC control is quite  sensi-
tive to the aldehyde content of the initial
VOC mixture. Generally, the ERT mech-
anism exhibits the greatest sensitivity
to changes in aldehyde content.

 VOC/NOX Ratio
  The VOC/NOX ratio was generally the
parameter having the greatest influence
on VOC control predictions. Generally,
the ERT mechanism showed the  great-
est sensitivity  to  a  change  in the
VOC/NOX ratio, and the CBM exhibited
the least sensitivity. In addition, the sen-
sitivity of predictions to a change in the
VOC composition  is greatest at low
VOC/NOX ratios.

Base Case Ozone
  Effects of changes in VOC composi-
tion are more pronounced the lower the
base case ozone. The DEM mechanism
is most sensitive to the choice of base
case ozone.

Dilution
  Differences among the predictions of
the mechanisms are somewhat smaller
at higher dilution rates than they are at
low dilution rates.

Emissions
  When emissions  are assumed to  be
present, a given percent VOC reduction
represents a larger reduction in total
VOC in the atmosphere than in the ab-
sence of emissions.

VOC Entrainment from Aloft
  Predicted VOC control requirements
change  by between 4 and 21% when
VOC entrainment from  aloft is included,
regardless of the reactivity of the mix-
ture  assuming  a  base case ozone con-
centration of 0.24 ppm, and between 6
and 40% when a base  case ozone con-
centration of 0.18 is assumed.

Ozone Entrainment from Aloft
  All mechanisms exhibit a strong sen-
sitivity to ozone entrainment from aloft.
Ozone aloft is therefore a key variable in
control studies. The sensitivity of VOC
control to the  VOC/NOX ratio  is influ-
enced more by the entrainment of VOC
than by the entrainment of ozone.

Initial MONO
  Predictions obtained  with the ERT
mechanism are most sensitive to initial
MONO; predictions from the PW mecha-
nism are least sensitive.

Summary  and Recommenda-
tions
  The results obtained in the series of
reports, of which this is the third, paint
a consistent  picture of the behavior of
chemical reaction mechanisms for pho-
tochemical smog.  Several key issues
emerged that are suggestive of future
work:
(1) In much  of the existing  laboratory
   smog chamber data base, chamber
   wall effects are important enough to
   make accurate discrimination
   among  rival chemical  reaction
   mechanisms difficult. Continuing
   extensive analysis of data from out-1
   door chamber  facilities is recom- "

-------
   mended as these systems seem to
   be characterized by smaller wall ef-
   fects than indoor facilities.
(2) We have found that predicted VOC
   control requirements are highly
   sensitive to the assumed VOC/NOX
   ratio and  the VOC composition of
   the initial mixture and emissions.
   Therefore, it is important that these
   variables be as accurately character-
   ized as possible for areas for which
   control strategies are to be de-
   signed.
(3) Because predicted VOC control can
   be  quite  sensitive to entrainment
   from  aloft, the organic and espe-
   cially the ozone concentration of en-
   trained air aloft must be better char-
   acterized  in  areas where control
   strategies are to be designed.
(4) Continued effort should be placed
   on  the evaluation of the perfor-
   mance of  chemical reaction mecha-
   nisms in simulating both laboratory
   and outdoor smog chamber experi-
   ments. These experiments should
   include a  wide range  of conditions
   in order to gain a better understand-
   ing of the effects of input parame-
   ters (for example, initial VOC/NOX,
   emissions and initial VOC composi-
   tion) on control predictions.
(5) When determining control strate-
   gies,  regulatory  agencies should
   use several  chemical  mechanisms
   because at this time it  is impossible
   to determine which chemical mech-
   anism best simulates atmospheric
   chemistry.
(6) It is most  important to know the in-
   put parameters as accurately as
   possible since the sensitivity of one
   input parameter may depend on the
   values assumed for other parame-
   ters.
T.  B. Shafer and J. H.  Seinfeld are with California Institute of Technology,
  Pasadena. CA91125.
Marcia C. Dodge is the EPA Project Officer (see below).
The complete report, entitled "Evaluation of Chemical Reaction Mechanisms for
  Photochemical Smog: Part III. Sensitivity of EKMA to Chemical Mechanism and
  Input Parameters," (Order No. PB 85-210 888; Cost: $11.50, 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:
       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/S3-85/042
      0000329   PS

      U  S  ENVIR  PR£I££TION
                                                                                    &U.S. GOVERNMENT PRINTING OFFICE.-1985—559-016/2711


                                                                                    It  IS I I  II    It   II I

-------