Identifying key factors influencing model performance on ground-level ozone over urban areas in Japan through model inter-comparisons

Ozone is one of remaining air quality issues in Japan. Effective strategies are required to reduce ozone at the ground level. Regional chemical transport models are useful in investigating relationships between precursor emissions and ambient concentrations of secondary pollutants including ozone. Model performance on ozone concentrations over urban areas in Japan were thoroughly evaluated in model inter-comparisons conducted in Japan's study for reference air quality modeling (J-STREAM). Specifically, 33 models with different configurations including the Community Multiscale Air Quality Modeling System (CMAQ), the Comprehensive Air Quality Model with Extensions (CAMx), and the Weather Research and Forecasting (WRF) model coupled with Chemistry (WRF-Chem) participated. They realized inter-comparisons with an unprecedented number of different model configurations. All the participating models overestimated ozone concentrations by 22 +/- 4.6 ppb over urban areas in Japan during the summer. Spatial and temporal variations in model performance suggest the influence of background ozone concentrations. Based on differences in the participating model configurations, halogen chemistry and deposition, dry deposition velocity, precursor emissions in other countries, and vertical transport were identified as the key factors influencing simulated background ozone concentrations. Specific sensitivity analyses were conducted to evaluate the effects of the key influencing factors. Halogen chemistry and deposition implemented in recent versions of CMAQ caused more than 10 ppb reduction of simulated ozone over the ocean surrounding Japan, while the original dry deposition schemes used in CAMx and WRF-Chem, without any effect of halogen, can also cause larger dry deposition. Horizontal and downward transport of ozone kept in a residual layer over the continent can spread the effect of precursor emissions in other countries to downwind regions, including Japan. Differences in vertical transport can alter the spatial extent of their effects. It is essential to improve the influence of the aforementioned key factors to realize better model performance on ozone concentrations over urban areas, not only in Japan, but all over the world.