Meteorological Factor Analysis of NOAA's National Air Quality Prediction Biases

Meteorological conditions are critical in determining both the chemical reactions that transform air pollutants as well as the physical processes that affect their mixing and transport. Uncertainty of meteorological inputs is one of the most important factors causing numerical air quality model prediction biases. The NOAA National Air Quality Forecasting Capability (NAQFC) provides numerical forecast guidance of surface ozone (O₃) and particulate matter with diameters less than 2.5 micrometers (PM_2.5) for the U.S. The NAQFC consists of the U.S. EPA Community Multiscale Air Quality (CMAQ) model and the NOAA North American Model (NAM) Non-hydrostatic Multiscale Model with Arakawa B grid-staggering (NMMB). While the NAQFC shows substantial under-predictions of PM_2.5 in summer and over-prediction in winter, O₃ predictions are persistently higher than observations in summer. Moreover, over-predictions of O₃ are prominent at nighttime and over the Great Lakes and coastal regions. However, meteorological factors causing the NAQFC prediction biases aren’t well understood and quantified. In this study, we present a detailed analysis on the impacts of meteorological factors and related physical parameterization schemes on prediction biases at diurnal and seasonal scales. Forecast biases and related meteorological conditions for different regions are discussed. The NAM performance on predictions of the planetary boundary layer height, temperature, winds, and other important meteorological conditions are verified with observations from different national monitoring networks. We are particularly interested in how the NAM model simulates boundary layer structure including the boundary layer height and eddy diffusivity at night time, transition period, and over lakes and coastal regions and how they are related to the prediction biases. Finally, a bias correction approach, the Kalman Filter to this time-series of ANalog forecasts (KFAN) is applied to the NAQFC for improving PM_2.5 predictions.