Regional scale atmospheric dispersion simulation of accidental releases of radionuclides from Fukushima Dai-ichi reactor

Abstract

This paper presents the results of regional scale atmospheric dispersion simulation of accidental emission of radionuclides from the Fukushima Daiichi Reactor, Japan following the Tohoku earthquake and tsunami event on 11 March 2011. The objective was to study the temporal behaviour of plume trajectory, concentration, deposition and radiation dose pattern over an 80 km range around the reactor. The time-varying meteorological parameters during the release period were simulated with a multi-scale nested atmospheric model WRF ARW and the trajectory, plume dispersion were computed with Lagrangian Particle Dispersion models HYSPLIT, FLEXPART using the available information on accidental source term. The simulations indicated that the wind flow over Japan during the release period was driven by the large scale extra-tropical westerly waves and associated low pressure systems. In the lower levels, the flow was influenced by the local topography/sea breeze causing occasional landward wind shift on the east coast of Japan. Simulated airflow trajectories revealed that the plume stayed over the ocean by westerly winds on most days and the radioactivity dispersed over sea surface. Landward trajectories were found on a few days due to southeasterly, easterly and northeasterly flow (15–17, 19–21 March 2011) during which much of the radionuclides deposited over the land region. The hotspot of depositions occurred over east Pacific Ocean near to Japan. Over the land relatively high depositions were simulated in a narrow zone of 20 km width and 80 km length in the northwest sector in agreement with monitor data. Simulations showed wet depositions over the land to be higher than the dry depositions during 12–30 March due to occurrence of rainfall on some days. Comparison of activity deposition and air dose values with available observations confirmed that the plume pattern in a finer length scale around the site could be simulated realistically and agree with the measurements within the limitations of the uncertainty in source term.