Evaluation of the AIRPACT2 air quality forecast system for the Pacific Northwest

Abdullah Al Mahmud

AIRPACT2 is a numerical photochemical air quality modeling system that generates daily forecasts of hourly gas-phase mixing ratios of ozone (O3) and related species, including air toxic compounds, and particulate concentrations for the Pacific Northwest. The performance of the AIRPACT2 system has been evaluated in this work. The current AIRPACT2 system, which employs the second-generation CALGRID photochemical model, is undergoing conversion to utilize the third-generation CMAQ photochemical model. In anticipation of this change, an intercomparison between the CALGRID and the CMAQ model performances has also been completed. For both models, the forecast performance was assessed against observations from September and October 2003 using statistical measures including mean bias (MB), fractional bias (FB), mean error (ME), and fractional error (FE). The system predicted O3 quite accurately during an O3 episode that occurred during early September; but over-estimated O3 during non-episode periods later in the month. Although the CMAQ model produced performance results similar to the CALGRID model for the episode, CMAQ also predicted O3 quite well for non-episodes and for urban sites as compared to CALGRID. The current AIRPACT2 system yielded the following monthly statistics: MB of 14 ppb and FB of 40% for urban sites, and MB of 8 ppb and FB of 22% for semi-urban/rural sites, respectively. Using CMAQ in AIRPACT2 improved these statistics: MB of 0 ppb and FB of -4% for urban sites, and MB of 1 ppb and FB of 6% for semi-urban/rural sites. Both CALGRID and CMAQ captured the temporal patterns of the air toxic compounds benzene, acetaldehyde, formaldehyde, perchloroethylene and 1,3-butadiene correctly; however, both models overestimated ambient levels, except for perchloroethylene. The degree of over-estimation was generally within the bounds of predicted levels within a 3 x 3 grid-cell matrix around the monitoring site. CMAQ predicted mixing ratios of formaldehyde and acetaldehyde more accurately than did CALGRID. For fine particulate matter (PM2.5), both CALGRID and CMAQ captured the PM2.5 concentrations during the morning rush hour reasonably well, but overall under-estimated 24-hr average concentrations. As CALGRID only treats primary emissions of PM2.5, the current system under-estimated PM2.5 at all sites.

Evaluation of the AIRPACT2 air quality forecast system for the Pacific Northwest

Files and links (1)

Abstract

Metrics

Details