Development of a high-spatiotemporal-resolution chemical reanalysis dataset of tropospheric ozone over Japan toward an accurate risk assessment for plants using concentration-based metrics

Tropospheric ozone (O 3 ) pollution in Asia is a concerning issue. O 3 has harmful effects on plants, and risk assessment using concentration-based metrics has been established. Because of the inhomogeneity of ground-based observations in time and space, numerical modeling has been used in the risk assessment for plants. However, modeling also has uncertainties in the variation of concentrations. Here, we focus on Japan, located in the downwind region of Asia. According to evaluations with ground-based observations, the original modeling generally overestimated concentration, leading to further overestimation when calculating the metrics of the risk assessment for plants. Thus, we developed a high-spatiotemporal-resolution (approximately 0.375° and hourly) chemical reanalysis dataset of O 3 over Japan by applying the two-dimensional variational method (2D-Var). Compared with the original modeling results, the chemical reanalysis data achieved better bias and error with 20% and 11% improvement, respectively, by conducting the evaluation using the independent observation network in Japan. Based on this developed chemical reanalysis, we calculated the metrics of the risk assessment for plants, and the risk assessment metrics were higher over broad areas of Japan during spring (particularly May) and over the Kanto region, including Tokyo, during summer. Higher values in spring would be influenced by transboundary pollution, while those in summer would mainly result from local pollution. This finding revealed that springtime O 3 levels over western Japan, where wheat (an O 3 -sensitive crop) is cultivated during this season, could cause crop yield loss. Thus, there is an urgent need for field surveys to investigate this issue.