Regional Air Quality Prediction using Global Model as Lateral and Top Boundary Conditions: Results for the ICARTT Period

This study presents results for regional air quality prediction during ICARTT (international Consortium for Atmospheric Research on Transport and Transformation) period using global model's real-time forecast as the lateral and top boundary conditions. In this period, a regional chemical transport model, STEM-2K3, was coupled with a global model, MOZART by utilizing the latter's prediction as lateral and top boundary conditions over North America domain. The STEM model ran in 2 domains: 60km primary domain, and 12km nested domain. In this study, we tested the 60km simulation using MOZART time-varied prediction and time-averaged fixed boundary conditions. The nested 12km domain also imported time-varied and fixed boundary information from the 60km mother domain and time-averaged values. The measured data from NASA DC-8 and NOAA WP-3 aircrafts were used to examine the simulations. Our sensitivity study shows upper-layer O3 is very sensitivity to external inflow and boundary conditions. For most other species, the sensitivity of 12km domain to the lateral boundary is higher than the sensitivity of 60km domain. Elevated layers are more sensitive to the boundary conditions than the surface sites. In most scenarios, pollutant concentrations over surface sites, especially those near urban areas, are not sensitive to the variation of lateral boundary, but to the background concentrations of relatively long-lived species, such as CO and O3. Other than the simulation scale, the influence of boundary conditions is also determined by the relative strength of the inner pollutant emissions compared to external forcing.