[H-TT18] 環境リモートセンシング

[HTT18-P06] Data-assimilation-based Estimation of Infectivity Parameter of COVID-19 Epidemic Model and its Relation to NO2 satellite observation.

*樺山 修¹、岡崎 淳史²、小槻 峻司¹ (1.千葉大学　環境リモートセンシング研究センター、2.ペンシルベニア州立大学)

キーワード：新型コロナウイルス感染症、二酸化窒素衛星観測、感染数理モデル、アンサンブルデータ同化、データ同化

With COVID-19 spreading around the world, it is necessary to understand the relationship between its infectivity and political decisions. This study uses data assimilation techniques and satellite observations to clarify the relevance. First, modified epidemiological model based on a commonly used compartment model is developed. The model assigns population into five compartments, Susceptible, Exposed, Infectious, Recovered, and Death (here after SEIRD model). In contrast to influenza epidemics, daily reports of deaths are issued by governments in COVID-19 outbreak. Therefore, Death compartment is explicitly separated to utilize the observation information. Ensemble data assimilation of SEIRD model revealed that the time series of infectivity of COVID-19 corresponded with political decisions such as the lockdowns, economic activity resume in the United States and economic restraints in Japan. In addition, satellite observations of NO2 also imply the economic restraint in Japan. We are now investigating the relationship between infectivity and political decisions based on estimated infectivity parameter of SEIRD and NO2 satellite observations. This presentation will include the most recent progress up to the time of the conference.