Discrepancies between MICS-Asia III simulation and observation for surface ozone in the marine atmosphere over the northwestern Pacific Asian Rim region

Abstract

Abstract. In order to identify the causes of overestimate of the surface-level O3 mixing ratio simulated by three regional chemical-transport models, NAQPMS v.3 (abbreviated as NAQM in this paper), CMAQ v.5.0.2, and CMAQ v.4.7.1, compared to the EANET observational data at a marine remote site at Oki in July 2010, analyses of hourly O3 mixing ratios and net ozone production were made in the context of MICS-Asia III. In addition to Oki, model-simulated and observational data for two other EANET marine sites, Hedo and Ogasawara, were also examined. Three factors, i.e., long-range transport from the continent, in situ photochemical formation, and dry deposition of O3 on seawater, have been identified as contributing to the overestimate by these regional models at Oki. The calculated O3 mixing ratios during long-range transport from the continent were much higher for all three models than those of the observation. In situ photochemical formation, demonstrated by a distinct diurnal variation which was not discerned in the observational data, was seen in the simulated data of all three models and ascribed to the virtual transport of NOx from the southern urban areas of the main island of Japan. The overestimate of the O3 mixing ratio in the background oceanic air mass has been discussed referring to dry deposition velocity (Vd) of O3 over oceanic water. Sensitivity analysis of the dry deposition velocity to the concentration of O3 was made for Oki in July. An increase in Vd from 0.0005 to 0.001 cm s−1 used in the standard runs for CMAQ by a factor of 10 decreases the O3 mixing ratio by more than 20 ppbv on an event basis in certain periods of time and by ca. 4.9 ppbv as a monthly mean in July. The dry deposition velocity of O3 in Bohai Bay and the Yellow Sea has been assumed to be comparable to that of the open ocean in all three models, which could have resulted in the overestimate of O3 mixing ratios in this area and also in the long-range transport of O3 from the continent to Oki. A higher value of dry deposition velocity in this marine area is expected considering the higher content of organics in the surface sea layer brought by rivers and atmospheric wet deposition. Empirical measurements of the mixing ratios and dry deposition flux of O3 in this area are highly recommended, since they would affect the simulated mixing ratios in the downwind region in the Pacific Rim region.