NERL Research Abstract

EPA's National Exposure Research Laboratory
GPRA Goal 1 - Clean Air
APM # 501

Significant Research Findings

Operational Evaluation of
Models-3/Community Multiscale Air Quality for Ozone

Purpose	Model evaluation is an important component of the quality assurance of models

developed to support decision making in the regulatory process and for
establishing acceptance of a model by the scientific community. The goal of
evaluation is to determine a model's degree of acceptability and usefulness for a
specified application or task. The central task of air quality model predictions
for ozone is to quantify the relative and absolute effectiveness of controlling
emissions of either or both of ozone's two precursors, nitrogen oxides (NOx)
and volatile organic compounds (VOCs), to reduce ambient levels of ozone.
The purpose of this research was to produce an initial, operational evaluation of
Models-3/Community Multiscale Air Quality (CMAQ) modeling system for
ozone using the 1995 field study databases of the North American Research
Strategy for Tropospheric Ozone (NARSTO) Northeast and Southern Oxidant
Study (SOS) Nashville- Middle Tennessee Field Studies.

Research Model evaluation of a complex, process-based model, such as CMAQ, is itself
Approach complex. The task is to challenge the science in the model, a challenge that

cannot result in black and white answers of model acceptability, however. The
evaluation methodology followed has two major components: operational and
diagnostic testing. Operational performance is determined with direct
comparison of model-predicted pollutant species against ambient observations
of the few species of regulatory interest, generally at the earth's surface. This is
the most common form of model evaluation. Diagnostic testing has the
objective of providing information to better understand operational
performance and to account for model behavior using process-level
explanations. Diagnostic testing often involves combinations of species other
than those of regulatory interest that are expected to provide process-level
insight into the functioning of the model. This study covered only the
operational testing of CMAQ against ambient ozone mixing ratios. In this
operational evaluation of CMAQ, model performance was judged over a
number of attributes of regulatory interest, including: ability to reproduce the 1
hour peak and 3 hour and 8 hour average daily maximum for ozone. To better
challenge the model and test it for intermediate ozone levels, a two-week period

National Exposure Research Laboratory - September 2000

-------
from 1995 was used that included some of the highest and lowest daily
maxima. Statistical and graphical comparisons were used.

Major	These first results indicate that the model is generally functioning in a way that

Findings can accounted for, both where model fits to the observations are good and
where they are not as good. Wide-spread use of the model is encouraged.
Specifically, CMAQ does well on high ozone days, but less-well on low ozone
days and on days with larger mean-to-maximum ranges. This behavior may be
explained by the fact that there is little difference in the urban predictions using
either the CB4 or the RADM2 chemical mechanism. The 36-km grid is likely
too coarse for most city and some regional model applications. The 12-km
resolution is an improvement nearly everywhere nearly every day, but 4-km
resolution is not a small step down from 12-km, and improvement is not
ubiquitous. Supporting evaluation comparisons against formaldehyde time
series from Photochemical Assessment Monitoring Systems sites are consistent
with ozone results, lending support to the overall determination of performance.

This research was conducted by EPA's National Exposure Research Laboratory
Atmospheric Modeling Division at Research Triangle Park, NC. It has been
presented at various scientific meetings and will be published in peer reviewed
scientific literature.

Arnold, J.R., Dennis, R.L., Tonnesen, G.S. Advanced techniques for evaluating Eulerian air
quality models: Background and methodology. Preprints of the 10th Joint Conference
on the Applications of Air Pollution Meteorology with the AWMA; January; Phoenix,
AZ. Boston, MA: American Meteorological Society. 1998.

Arnold, J.R., Dennis, R.L. First results from operational testing of the U.S. EPA Models-

3/Community Multiscale Model for Air Quality (CMAQ). In: Air Pollution Modeling
and Its Application. Volume 24. Gryning, S.E., Schiermeier, F.A., Eds. New York,
NY: Plenum Press. In press.

Future research will focus on extending the CMAQ evaluation for ozone to
diagnostic testing against the SOS 1995 Nashville-Middle Tennessee Field
Study data.

Questions about Models-3/CMAQ for ozone may be directed to:

Robin L. Dennis

U.S. Environmental Protection Agency
National Exposure Research Laboratory (MD-80)

Research Triangle Park, NC 27711
Phone: (919)541-2870
E-mail: dennis.robin@epa.gov

Research

Collaboration

and

Publications

Future
Research

National Exposure Research Laboratory - September 2000

-------