Investigating the impact on modeled ozone concentrations using meteorological fields from WRF with an updated four–dimensional data assimilation approach

Abstract

The four–dimensional data assimilation (FDDA) technique in the Weather Research and Forecasting (WRF) meteorological model has recently undergone an important update from the original version. Previous evaluation results have demonstrated that the updated FDDA approach in WRF provides more accurate wind fields aloft than the original approach, particularly during the nocturnal period when low level jets are a common feature in the Eastern United States. Due to the importance of WRF/FDDA meteorological fields in retrospective air quality applications, a modeling study with the Community Multiscale Air Quality (CMAQ) model was undertaken to ascertain if the improved wind flow fields translate into better performance for ozone. To undertake this objective, separate CMAQ model simulations were performed with meteorological inputs generated by WRF using the original and the updated FDDA approaches for a three month summer period. The evaluation effort focused on observed and modeled surface ozone from a mid–morning hour (10 local daylight time (LDT)). Comparisons of modeled results against concentrations aloft from an instrumented tall tower and from available morning vertical profile measurements were also examined. Surface concentrations near 10 LDT are desirable for evaluating the transport process since they are often representative of ozone that has been transported aloft overnight and has undergone downward entrainment in response to convective mixing the following morning. Statistical results from surface observed and modeled concentration pairs indicated modeled ozone from the CMAQ simulation using the updated FDDA meteorology displayed smaller biases and lower absolute errors at 88% and 80% of monitoring sites, respectively, in the Eastern United States. The CMAQ results with the updated FDDA generally exhibited smaller biases and lower absolute errors at monitoring sites across the northern states than in the southeastern states. The results provide evidence that the more accurate wind flows generated with the updated WRF/FDDA approach improved CMAQ model performance based on the statistical results from 10 LDT ozone concentrations.