Local-Scale Atmospheric Dispersion Modelling of Radionuclides Following the Fukushima Daiichi Nuclear Accident Using SWIFT-RIMPUFF

Shuhan Zhuang,Sheng Fang,Xinwen Dong
DOI: https://doi.org/10.1115/icone29-90748
2022-01-01
Abstract:Abstract Air dispersion modeling is an important tool for emergency response following a nuclear accident, such as the Fukushima accident. Current researches mainly focus on global- and regional-scale modeling with wind field data derived from different methodological models and observations. However, the capability of the local-scale atmospheric dispersion model hasn’t been discussed in detail for the Fukushima accident. In this paper, the local-scale modeling of radionuclides following the Fukushima accident was investigated with the combination of the wind diagnosed model SWIFT and the radionuclide transport model RIMPUFF. The coarse input wind field of SWIFT is prepared with WRF using the European Centre for Medium-Range Weather Forecasts (ECMWF) Meteorological data. The SWIFT-diagnosed wind field was used to drive RIMPUFF for calculating the dispersion of radionuclides and gamma dose rates around the Fukushima Daiichi Nuclear Power Plant (FDNPP) site. The diagnostic wind fields were validated to the on-site meteorological observations, whereas the dispersion and gamma dose rates were validated the onsite gamma dose rate monitoring data and the observations of suspended particulate m? near the FDNPP. The plume patterns were also analyzed to help understand the transport behaviour of the radionuclides. The validation demonstrates that, with the 1km-resolved ECMWF meteorological data, SWIFT fairly reproduces the wind field. The speed is slightly overestimated, with a Normalized Mean Squared Error (NMSE) below 6. The wind direction is well simulated at some specific moments, which is critical for reproducing some peaks of the dose rates. However, RIMPUFF underestimates the dose rates around the FDNPP, partly because of the overestimated wind speed. The concentration simulations better agree with observations in the Naraha station than Futaba station, with the Fractional Bias (FB) under 0.3 and NMSE under 6 at the Naraha station.
What problem does this paper attempt to address?