LAUSR

Abstract The Green’s function approach, which is an approach to optimize model parameters, was iteratively applied to a large number of pairs of observed and calculated values of surface air concentrations (677 pairs) and ambient dose rates (3279 pairs) to estimate the time variation in 137Cs release rate to the atmosphere from the Tokyo Electric Power Company's Fukushima Daiichi Nuclear Power Station, which was heavily damaged on March 11, 2011. Regarding surface air concentrations, observations suitable for the optimization were selected not to affect the derived release rates due to defects in the modeling framework rather than the release rates. It was found that the release rates of 97 out of a total of 497 segments with a width of 1 hour in the target period were effectively optimized and that the modeled levels using the optimized release rates generally agreed well with the observed ones. The derived time variation was reasonable and consistent with that of preexisting works which were derived using another approach and broadly employed to assess the consequences of the Fukushima accident.