Following the Fukushima Daiichi Nuclear Power Plant accident, much of the ¹³⁷Cs deposited in surrounding areas still remains in forests and other land where decontamination has not been carried out. As a result, ¹³⁷Cs continues to be transported from forested areas to the ocean via rivers, and influenced by this, the ¹³⁷Cs concentration in some freshwater fish inhabiting rivers still exceeds the regulatory limit. A complete lifting of shipping restrictions has therefore not yet been achieved. To realize recovery of the water environment from the nuclear accident, it is important to clarify the current status of ¹³⁷Cs concentrations in river water, and how these concentrations are determined under the influence of various environmental factors.  

The National Institute for Environmental Studies conducted monthly sampling and water-quality measurements from 2013 to 2021 at four rivers around the Fukushima Daiichi Nuclear Power Plant (two rivers draining mainly agricultural land and two rivers whose catchments are mostly forest), as well as at a site downstream from a reservoir. Using these data, the Institute analyzed how environmental conditions determine the concentrations of dissolved and particulate ¹³⁷Cs. The analysis showed that water temperature, the concentration of suspended solids (SS), and large-scale watershed disturbances caused by intense rainfall events all play roles in seasonal and long-term changes in ¹³⁷Cs.  

Fig. 1. River monitoring sites. 

Agricultural catchments: (1) Uda River, (2) Mano River Forested catchments: (3) Ota River, (4) Odaka River

Downstream of reservoir: (5) tributary of Ota River (treated separately from forested rivers because SS here is strongly affected by outflow of suspended material from the reservoir) 

Seasonal variation in dissolved ¹³⁷Cs concentrations: linkage with water temperature 

Analysis of ¹³⁷Cs concentration changes in five rivers showed that dissolved ¹³⁷Cs concentrations decreased year by year at all sites, while at the same time exhibiting seasonal variation: concentrations were higher in summer and lower in winter (Fig. 2). One reason for this seasonal pattern is that the balance of concentrations between dissolved ¹³⁷Cs and particulate ¹³⁷Cs (chemical equilibrium) shifts, as water temperature increases, in the direction of higher dissolved ¹³⁷Cs concentrations.  

Fig. 2. Long-term changes in dissolved ¹³⁷Cs concentrations in river water 

A commonly used indicator of this concentration balance is the distribution coefficient, Kd. The larger the Kd value, the more readily ¹³⁷Cs exists in particulate form rather than in dissolved form. 

Distribution coefficient (Kd) 

The larger the distribution coefficient, 
the more easily radiocesium exists in a form adsorbed onto soil particles. 

$$K_d=\frac{\text{Radiocesium adsorbed on soil particles (particulate form)}\left[\text{Bq/kg (dry)}\right]}{\text{Radiocesium dissolved in water (dissolved form)}\left[{\text{Bq/m}}^3\right]}$$

Relationship between water temperature and the ¹³⁷Cs distribution coefficient 

If we can assume that ¹³⁷Cs is in chemical equilibrium in water, the relationship between the distribution coefficient Kd and temperature is known to follow the van ’t Hoff equation from thermodynamics. The research team at the National Institute for Environmental Studies therefore examined whether the observed data could be reproduced using a van ’t Hoff plot of the reciprocal of water temperature (absolute temperature) versus Kd, and calculated the enthalpy of reaction — that is, the strength of adsorption of ¹³⁷Cs onto particles as temperature changes — from the regression equation obtained.  

The results showed that, at the two agricultural rivers (Uda River and Mano River) and at the reservoir outlet site, the observed data were well reproduced by the theoretical van ’t Hoff equation, and the enthalpy of reaction fell in the range −15.6 to −19.6 kJ/mol, almost the same as values reported in previous studies. 

Fig. 3. Relationship between reciprocal water temperature and the logarithm of the ¹³⁷Cs distribution coefficient (van ’t Hoff plot) 

By contrast, in the forested rivers the observed data were not reproduced well by the theoretical equation (the coefficient of determination R² was much lower than in the agricultural rivers). This suggests that dissolved ¹³⁷Cs concentrations in forested rivers are not determined solely by the ¹³⁷Cs concentration balance (ion exchange) between dissolved and particulate forms, but are also influenced by ¹³⁷Cs contained in leaf litter (forest floor litter) and by potassium ions (K⁺), which behave chemically in a similar way to ¹³⁷Cs in water. 

Variation in particulate ¹³⁷Cs concentrations: effects of storm runoff and decontamination work 

Although particulate ¹³⁷Cs concentrations in river water did not show a periodic variation like that of dissolved ¹³⁷Cs, long-term analysis of the data revealed fluctuations caused by rainfall events and by decontamination work. The most marked tendencies were found in the agricultural rivers: as the concentration of suspended solids (SS) increased, the ¹³⁷Cs concentration in SS decreased. One possible explanation is that stronger rainfall increases the proportion of soil from agricultural land that has relatively low ¹³⁷Cs concentrations (because decontamination has been carried out) in the SS. 

In the Mano River, a pronounced decrease in ¹³⁷Cs concentration in SS was observed from 2015 to 2016, when decontamination was carried out in the catchment. This suggests that removal of ¹³⁷Cs from agricultural land through decontamination contributed to lower ¹³⁷Cs concentrations in river water.