A WRF-CMAQ study on spring time vertical ozone structure in Southeast Texas

With objective analysis WRF showed considerable skills in simulating temperature and wind profiles. CMAQ did well in replicating daily regional average ozone, with only minor overprediction for most of the days. According to the surface ozone statistics and the ozone biases along the vertical profiles, the model displayed little sensitivity to the two CMAQ model options: (i) fixed night time PBL height vs. WRF calculated night time PBL heights, and (ii) the combination of the Euler-backward iterative chemistry schemes and Yamartino advection (EBI/YAMO) vs. the Sparse-Matrix Vectorized Gear chemistry and Parabolic Method advection (SMV-GEAR/PPM) schemes. CMAQ shows substantial underprediction of ozone in the upper troposphere, but demonstrated fair skills in reproducing the ozone profile in the lower to middle troposphere, with typically a few ppb negative biases which tend to increase with height. However, the model usually cannot reproduce high variability in the ozone profiles due to the fixed default profiles and lack of data assimilation capability. While the default CMAQ static profile setup for boundary condition (BC) does not properly reflect the actual dynamic atmosphere, it remains a reasonable choice, since using downscaled real-time global chemistry model BCs demands considerable efforts and would still depend on the performance of the global model.