A statistical model for estimating the radiocesium transfer factor from soil to brown rice using the soil exchangeable potassium content

Publication date: Available online 23 August 2018

Source: Journal of Environmental Radioactivity

Author(s): Kohji Yamamura, Shigeto Fujimura, Takeshi Ota, Tetsuya Ishikawa, Takashi Saito, Yoshimitsu Arai, Takuro Shinano

Abstract

To reduce radiocesium uptake by rice, large amounts of potassium fertilizer have been applied to paddy fields contaminated by radiocesium released from the Fukushima Dai-ichi Nuclear Power Plant owned by the Tokyo Electric Power Company. The Fukushima Prefectural Government recommended maintenance of the soil exchangeable K content up to 200 mg K kg⁻¹ before conventional fertilization in rice production. We constructed an equation to predict the transfer factor from soil to brown rice using the soil exchangeable K content. This equation was then used to calculate the appropriate soil exchangeable K content, which ensures a low risk (5%) of brown rice exceeding the standard limit (100 Bq kg⁻¹ for grains) established in 2012. The equations were constructed using field data obtained by an investigation (measurement of the ¹³⁷Cs concentrations in soil and brown rice and measurement of the soil exchangeable K content at harvest) that was performed from 2012 to 2015 in 321 paddy fields distributed all over Fukushima Prefecture. We found that the stochastic fluctuation of the transfer factor approximately follows a lognormal distribution under the given environmental conditions. Four factors are considered in predicting the logarithmic quantity of the transfer factor: (1) the linear influence of the logarithm of the exchangeable K content in soil, (2) the non-linear influence of the logarithm of the exchangeable K content in soil, (3) three districts in the Fukushima Prefecture and (4) the year. The linear model of the logarithm of exchangeable K content in soil was adopted by the R_D criterion, which indicates the absolute goodness of models for prediction. The predictive ability of the model increased by 29% after including the logarithm of the exchangeable K content in soil (factor 1), while the predictive ability further increased by 10% after including spatial and temporal information (factors 3 and 4). The validity of the exchangeable K content recommended by the Fukushima Prefectural Government was re-examined using the proposed equations, which indicated that the conventional recommendation was appropriate at 2012 but is not fully appropriate under the current situation in which the radiocesium concentration in soil has decreased.

from Radiology via xlomafota13 on Inoreader https://ift.tt/2P3Bdcu
via IFTTT