The Fukushima Daiichi Nuclear Power Plant accident also contaminates lakes in Japan. Especially in closed lakes, there is a problem of prolonged low-level $$^{137}$$
137
Cs contamination because the activity concentration of $$^{137}$$
137
Cs declines sharply immediately after the accident, but then begins to decrease slowly. In this paper, we derived a long-term prediction formula based on the fractional diffusion model (FDM) for the temporal variation in $$^{137}$$
137
Cs activity concentrations of the water in Lake Onuma on Mt. Akagi, one of the closed lakes, and of pond smelt (Hypomesus nipponensis), a typical fish species inhabiting in the lake. The formula reproduced well the measured $$^{137}$$
137
Cs activity concentration of the lake water and pond smelt for 5.4 years after the accident. Next, we performed long-term prediction for 10,000 days using this formula and compared it with the prediction results of the two-component decay function model (TDM), which is the most common model. The results suggest that the FDM prediction will lead to a longer period of contamination with low-level $$^{137}$$
137
Cs than the TDM prediction.