Spatial variability of radioactive cesium fallout on grasslands estimated in various scales.

摘要

Large releases of radionuclides from the Fukushima Daiichi Nuclear Power Plant have resulted in serious radioactive pollution over a wide region over Eastern Japan and the livestock industry has suffered serious damage as a result. In this study, the spatial variability of fallout radioactive cesium (Cs) on two grasslands (a pasture and a meadow) was estimated in various scales and the number of samples needed to determine the mean deposition density of radioactive Cs was calculated. In both grasslands, variances between sampling points were quite large compared with the repeated measurements. Because deposited radioactive Cs levels were higher in the pasture than in the meadow, the mean square of inter-sampling points and repeated measurements were larger in the pasture than in the meadow. Mean contamination levels of radioactive Cs in the pasture ranged from 143.0 (site A_L) to 172.8 kBq m-2 (paddock Z). The coefficients of variation ranged from 4.8 (site D) to 38.7% (whole pasture). Mean contamination levels of radioactive Cs in the meadow ranged from 59.6 (site Q) to 68.3 kBq m-2 (site O_L). The coefficients of variation ranged from 4.9 (site R) to 10.2% (whole meadow). Sampling numbers that were necessary to get the mean deposition density of radioactive Cs in the range +or-10% with a 95% confidence interval were 57.5 for the pasture and 4.0 for the meadow. The significant relationship between the investigated area and the minimum sampling numbers with an allowable error of +or-10% with a 95% confidence interval estimated by bootstrapping showed that 29 samples are necessary to estimate the reliable mean of fallout Cs on 1 ha of the pasture. For the meadow, the number of samples for reliable mean estimation was 1-5, being independent of the investigated area.