Source term estimation of atmospheric discharge during the Fukushima Daiichi Nuclear Power Station accident by Bayesian inversion with multi-scale dispersion simulations

Terada, Hiroaki   ; Nagai, Haruyasu  ; Tsuzuki, Katsunori  ; Furuno, Akiko   ; Kadowaki, Masanao   

We attempted to estimate source terms of radionuclides discharged into the atmosphere during the Fukushima Daiichi Nuclear Power Station accident by Bayesian inference using multi-scale dispersion simulations and various environmental measurements compositely. To make source-receptor matrix for the Bayesian analysis, we preliminary constructed a database of air concentration and surface deposition from dispersion calculations for multiple scales; several hundred km-square area, East Japan area, and the northern hemisphere area. We used various measurements such as dust sampling data, surface deposition map by airborne monitoring, fallout, hourly air concentration derived by analyzing suspended particulate matter collected at air pollution monitoring stations, and air concentrations at the Comprehensive Nuclear-Test-Ban Treaty Organization stations. We conducted a test estimation using these prototype databases and confirmed the validity and effectiveness of the method.