Validation of a Lagrangian atmospheric dispersion model against middle-range scale measurements of⁸⁵Kr concentration in Japan

Abstract

The Lagrangian atmospheric dispersion model of the computer-based nuclear emergency response system WSPEEDI-II was validated with the measured 85 Kr concentrations in tens-of-to hundreds-of-km (middle-range) scale area by conducting dispersion simulations using the release rate from the nuclear fuel reprocessing plant in Rokkasho, Japan. The calculated weekly concentrations of 85 Kr in two simulation cases during April and September in 2008 agreed with the measurements within a factor of two at the sampling sites 170 to 2000 km away from the plant. However, the sensitivity analysis of horizontal grid resolution of the meteorological model ranging from 2 to 54 km showed that the calculated results had the dependency on the grid resolution, i.e., the calculated concentrations became low compared with the results with the grid resolution of 54 km as the grid resolution became high. An empirical modification of the horizontal diffusion parameter used for long-range dispersions by WSPEEDI-II was attempted based on the sensitivity analysis to reduce the redundant diffusion effect in dispersion simulations with relatively high grid resolution. The modified horizontal diffusion parameter contributed to reduce the dependency of calculated concentrations on the horizontal grid resolution.

Horizontal turbulent diffusion in a convective mixed layer

“…Such local circulations are not present in our study, since we have assumed statistical homogeneity in the horizontal direction. In fact, Terada, Nagai & Yamazawa (2013) suggests that eddy diffusivity of Gifford (1982) should be reduced if resolution of numerical model is improved to resolve such local circulations.…”

Model Intercomparison of Atmospheric 137Cs From the Fukushima Daiichi Nuclear Power Plant Accident: Simulations Based on Identical Input Data

“…The emission inventory has been estimated by various studies (Katata et al, ; Saunier et al, ; Terada et al, ; Winiarek et al, ; Yumimoto et al, ), but it remains a major cause of the uncertainties in the simulated atmospheric 137 Cs concentrations because of the limited acquisition of direct measurements of radionuclide concentrations during their release. Internal elements refer to the components included in the numerical models, such as the physical processes of wet and dry deposition of 137 Cs (Morino et al, ; Quérel et al, ), clouds and rain (Saito et al, ), grid resolution (Sekiyama et al, ), and horizontal diffusion (Terada et al, ). Overall, the results derived from numerical models have large discrepancies.…”