Improvement of atmospheric dispersion simulation using an advanced meteorological data assimilation method to reconstruct the spatiotemporal distribution of radioactive materials released during the Fukushima Daiichi Nuclear Power Station accident

Abstract

To improve the atmospheric dispersion simulations of radioactive materials released due to the Fukushima Daiichi Nuclear Power Station (FNPS1) accident, we adopted four-dimensional variational data assimilation (4D-Var) of the data assimilation system (WRFDA) and confirmed the effectiveness of the existing 4D-Var technique for the reproducibility of dispersion simulation during the FNPS1 accident. The simulation was performed by the community meteorological model (WRF) and our atmospheric dispersion model (GEARN). The accuracy of simulated 137Cs deposition patterns in the area closed to FNPS1 and the Ibaraki, Tochigi, and Fukushima Prefectures was increased due to improvements in wind and rain fields in 4D-Var calculations. The results demonstrated that 4D-Var was effective for improving local- and regional-scale atmospheric dispersion simulations.