HR: 1340h
AN: A23C-1460 [Abstracts]
TI: Accuracy of the Operational NOAA-EPA National Air Quality Forecasting System during Summer 2005 and 2006 in Philadelphia
AU: * Huff, A K
EM: huffa@battelle.org
AF: Battelle Memorial Institute, 2101 Wilson Blvd
Suite 800, Arlington, VA 22201, United States
AU: Ryan, W F
EM: wfr1@psu.edu
AF: Pennsylvania State University, Department of Meteorology, University Park, PA 16802,
United States
AU: Bahrmann, C P
EM: cbahrmann@psu.edu
AF: Pennsylvania State University, Department of Meteorology, University Park, PA 16802,
United States
AB:
The NOAA-EPA National Air Quality Forecasting System (NAQFS) is a numerical ozone forecast model that
consists of a coupled version of NOAA's North American Mesoscale (NAM)-12 model and
EPA's Community Multiscale Air Quality model (CMAQ). NAQFS runs twice daily at 0600 UTC and
1200 UTC and predicts hourly ozone abundances as mixing ratios in units of parts per billion (ppb). Model output
is typically processed to provide 1-hour and 8-hour average ozone forecasts that correspond to the national
ambient air quality standards (NAAQS) for ozone. Now in its third season as an operational model, NAQFS is
designed to assist air quality meteorologists by providing accurate and dependable forecast guidance. Before air
quality meteorologists are willing to rely upon a new tool like NAQFS to assist them in preparing forecasts for the
public, however, the model must be evaluated in the typical operational setting for air quality forecasting - the
metropolitan scale. To address this need, results will be presented from a pilot statistical study of Summer 2005
and 2006 NAQFS performance in the Philadelphia forecast area. The accuracy, bias, and skill of the model has
been determined by comparing 8-hour average ozone forecasts from the 1200 UTC run of NAQFS to
corresponding observed ozone values. Overall, the model shows the best skill in suburban areas, where ozone
levels tend to peak across the metropolitan region. NAQFS was not reliable in 2005 during Code Orange and
Red events, when ozone levels were greater than 84 ppb. In 2006, NAQFS accuracy increased during these
cases, which was most likely a result of the NAM's changeover from the Eta model to the Weather
and Forecasting (WRF) model. The most probable sources of error that impacted NAQFS performance in 2005
and 2006 will be discussed, including shortcomings in emissions databases and poorly characterized
atmospheric chemistry. Preliminary results from the 2007 summer ozone season will also be provided.
DE: 0345 Pollution: urban and regional (0305, 0478, 4251)
DE: 0368 Troposphere: constituent transport and chemistry
SC: Atmospheric Sciences [A]
MN: 2007 Fall Meeting