Several overarching modeling issues were identified during the simulation of 8-hour ozone concentrations in support of 8-hour ozone designations of the State of Arizona. The simulations were done at the request of Arizona Department of Environmental Quality (ADEQ), which was responsible for recommending science-based ozone designations to be used by USEPA is classifying 8-hour ozone non-attainment areas. Standard community models, CMAQ and MM5 were used, respectively, for air quality and meteorological simulations. The emission inventory for CMAQ was prepared using SMOKE (Sparse Matrix Operator Kernel Emissions) based on Western Region Air Partnership (WRAP) inventory and an extensive GIS database. In general, CMAQ-simulated ozone concentrations showed a good agreement with the observations for the designated episodes for which the simulations were performed.

However, our MM-5 predictions have clearly indicated some deficiencies with regard to nocturnal and transition periods. New sub-grid parameterizations based on the measurements taken in Salt Lake City during the Vertical Transport and Mixing Experiment by Monti et al. (2002, J. Atmos. Sci., , 59, 2513-2534) were attempted with encouraging success for the nocturnal period.

Certain problems associated with transition periods were well evident from the CMAQ simulations, which indicate the formation of an ozone “blob” in the northwestern part of the domain. These elevated levels are absent in the measurements, however, and scrutiny of MM5 and CMAQ simulations clearly indicated that the delayed transition in the simulations has led to high ozone in the northwest. The ozone formed within in the urban core was supposed to be advected to the east by the up-slope flow that initiates following morning transition, but the delayed transition caused the ozone movement to be biased toward northwest. Sound physics and dynamics based parameterizations for morning (and evening) transition ought to be a priority in future model refinements.