During the summer of 2000, emissions, meteorological, and air quality
data were collected using a network of surface and upper air instruments to study
the atmospheric conditions leading to elevated ozone concentrations in Central
California. Of the three ozone episodes captured during this field study, the
episode that took place between July 29 and August 4 was chosen to study in
detail using the MM5 numerical model. The MM5 was initialized using archived
NCEP analysis data and observational nudging using the data collected during
the CCOS, separately, to create a 3-D structure of the atmosphere in time. Aside
from plotting 2-D horizontal and vertical cross sections of common
meteorological variables, the modeled wind and temperatures were also
compared at each station against corresponding hourly observations to evaluate
the spatial and temporal model performance. Overall analysis procedure has
allowed the identification of various errors and/or trends in observational and
modeled data. Examples of such identifications are a) observation data that are
inconsistent with the characterization of a region, b) previously unknown local-
scale features of the atmosphere within certain parts of the CCOS domain, and
c) various errors/deficiencies associated with several modules of the model.
During this seminar, we will share the lessons learned during the modeling of
July-August episode. We will also discuss the need for additional post-modeling
quality assurance of observational data, and the effects of various initialization
parameters on simulated meteorology for this episode.
*The web site for CCOS is http://www.arb.ca.gov/airways/ccos/ccos.htm
|
