A five day MM5 simulation of meteorological conditions during an important SCOS'97 ozone episode showed that the model accurately reproduced most of the observed features associated with a inland moving sea breeze front, i.e., its movement speed and inland penetration. Statistical comparisons between near-surface observed and predicted meteorological values (using MAPS) showed reasonable agreement, with differences in temperature, however, largest on the day with the strongest large scale forcing. This effect was not reproduced in the current simulation, as it did not include (model) analysis nudging. Future evaluation plans include: saving hourly model values for MAPS comparisons, scaling predicted values to observational levels (to reduce biases in MAPS analyses), comparison of predicted and observed upper level values, and compilation of MAPS statistics on various sub-domains (e.g., coastal, mountain, and desert areas). Future modeling simulations will also sequentially include (model) analysis nudging and the use of the SCOS'97 observations via FDDA. Analysis nudging will allow for the incorporation of large-scale meteorological trends (like the subsidence warming on 5 August), while the FDDA will allow for the incorporation of local (sub-grid) flow features into MM5. Both of these additions should further reduce the differences between observed and simulated fields. Results from the final simulations will be used by CARB modelers as input for photochemical air quality model simulations.
展开▼
