United States Environmental Protection Agency	Office of Research and Development

National Exposure Research Laboratory
FY02 Research Abstract

Government Performance Results Act (GPRA) Goal 1.1.8
APM14

Significant Research Findings:

Update a Faster, More Detailed Approach (Morphecule
Mechanism) for Modeling Atmospheric Chemistry and Work with
EPA Staff To Implement the Mechanism in the Agency's Principle
	Air Quality Model	

Air quality models that realistically describe the photochemical formation of
ozone, particulate matter (PM), and other pollutants, such as air toxics and
pesticides, are needed by the U.S. Environmental Protection Agency (EPA)
and state agencies to develop risk management strategies to decrease their
concentrations to safe levels. The atmospheric chemistry of ozone involves
the complex interactions of thousands of different chemical reactions, while
that of secondary organic particulate matter involves significantly more. For
emerging pollutants that EPA will need to address now and in the future, such
as secondary toxic compounds or organic aerosols, it will be necessary to
account for even more reactions.

Scientific
Problem and
Policy Issues

Including this complex chemistry in a current urban-to-regional air quality
model would swamp even the most powerful computers available; therefore,
the chemistry in current models is always extremely simplified. Air quality
models must represent the thousands of reactions that actually occur in the
atmosphere by only about 80-200 model reactions, and sometimes less.

While these simplifications allow the model to run efficiently, they can result
in inaccuracies in some of the model predictions. To address this problem,
we have developed a way to perform these calculations more efficiently, but
without losing accuracy in the predictions. This method, the Morphecules
approach, can efficiently and realistically describe the complex atmospheric
chemistry of ozone and other photochemical pollutants in regulatory and
research air quality models. This will allow us to develop pollution control
strategies in a more holistic manner, and integrate control strategies for all
important species simultaneously.

Research	The Morphecules approach assigns individual chemical species into a small

Approach	number of groups in the air quality model based on their structure and

reaction characteristics. The computer intensive parts of the air quality
model are performed on these groups, which allows the model to run faster

National Exposure Research Laboratory — November 2002


-------
because the computations are performed on a smaller number of components.
The model remembers the individual chemical species within each group,
their initial concentrations before being added to the group, and their reaction
rates, so that they can be recreated from the group and their concentrations
updated at the end of the time step. This allows the model to retain detailed
information about the atmospheric chemistry of ozone and other
photochemical reactants, yet still efficiently use computer resources.

Under this project, the software to implement the Morphecules approach was
delivered to run on a PC and was ported to a UNIX machine. Substantial
modifications were made to the code to enable it to be executed under a
UNIX-based operating system. The software was linked to a box model
version of the CMAQ/Models3 air quality model. The first step was to
develop, debug, and implement libraries and driver routines to allow the
code to be run in a purely explicit mode, where all species were introduced
to the chemical solver without any grouping. Following that, additional work
was done to utilize the full capacity of the code, using the grouping and
ungrouping routines to speed up the computational solution. A demonstration
test was performed on a detailed chemical mechanism for a family of
aldehyde compounds.

Results and	This work resulted in a highly modified set of compiler, solver, and library

Implications	driver routines that are demonstrated to run within a 1-dimensional box

model version of CMAQ. All code modifications have been documented
extensively within the code, as well as summarized in a final report. The
code and report of modifications were delivered on May 24, 2002.

The software modifications and additional interface codes developed in this
project will allow more complex representations of atmospheric chemistry to
be used in the CMAQ code being developed at EPA for regulatory and
research purposes. By implementing this complex chemistry using Morpho-
type solvers, we will be able to run highly detailed descriptions of the
atmospheric chemistry of ozone, secondary particulate matter, and secondary
air toxics in full, 3-D simulations.

Initial execution of the Morphecules code has shown that it is possible to
produce and utilize a highly efficient, accurate, and user-friendly method for
incorporating complex atmospheric chemistry to represent the chemistry that
occurs in the atmosphere and to incorporate it into air quality models.

National Exposure Research Laboratory — November 2002


-------
Research	The code was developed under collaboration with the University of North

Collaboration Carolina, Chapel Hill funded by contract 68D50129, and implementation of
and Publications ^ C0(je jnt0 CMAQ was performed under contract #68-W7-0055 with
Lockheed Martin Services, Inc.

Future Research The next step of the implementation will be extension of this work from the
1-D version of CMAQ to the fully 3-D CMAQ. Additional tests will be
performed to verify that the computational calculations are accurate and
efficient.

For additional information contact:

Deborah Luecken

U.S. EPA, Office of Research and Development
National Exposure Research Laboratory, Maildrop D205-03
Research Triangle Park, NC 27711

Phone: 919/541-0244
Email: luecken.deborah@epa.gov

Contacts for

Additional

Information

National Exposure Research Laboratory — November 2002


-------