Comparison of air quality model vs. real-time ozone observations for the summer of 2022 over the Washington DC metropolitan area

This study aims to evaluate the NOAA’s Air Quality Model (AQM) prediction systems in urban and suburban/rural environments. Both AQM v6 and v7 products will be utilized for this study. AQM v6 is NOAA’s current operational AQ system based on an offline-coupled U.S. Environmental Protection Agency (EPA) Community Multiscale Air Quality Model (CMAQ) model and v7 is being developed as NOAA’s next generation AQM system within the Unified Forecast System (UFS) framework. The v7 integrates an up-to-date version of the Finite-Volume Cubed-Sphere Dynamical Core (FV3) atmospheric model, the Common Community Physics Package (CCPP) as well as the online coupled CMAQ version 5.2.1. The AQM v7 is designed for a more accurate representation of the impact of meteorology on atmospheric chemistry.

The focal point of this study will be on the comparison of surface ozone (O₃) variation over the Washington DC metropolitan area. Specifically, the real-time surface measurements from the Howard University Washington DC campus (urban site) vs. the Howard University Beltsville campus (rural site) will be employed to evaluate both AQM v6 and v7 O₃ outputs.

The preliminary v6 version of the model showed that the amount of surface O₃ was underestimated in the urban area (Washington, DC) while overestimated in the rural area (Beltsville, MD). The over the DC downtown was more pronounced during the nighttime. The daytime prediction appears to align well with observations, especially for O₃ peak values. Version 7 of the model corrected the underestimation for our urban site. For the rural site, the overestimation remains during the day and nighttime. It appears that surface wind speed plays a significant role affecting the prediction. Wind speed in the rural area was much slower than the urban area, especially in the nighttime. The momentum flux measured in the Beltsville site decreased as well. It is suspected that the model may not address the vertical mixing in the atmospheric boundary layer properly. Currently we are quantifying the uncertainty through statistical analysis. More comprehensive results will be carried out.