After the Fukushima Daiichi Nuclear Power Station accident, evacuation orders have remained in effect in “difficult-to-return zones,” prioritizing residents’ safety. To lift these orders, it is essential to predict radiation exposure levels realistically under conditions resembling everyday life. 

Previously, Japan’s national assessment assumed that “a resident stays outdoors for 8 hours and indoors for 16 hours per day,” applying a uniform and conservative calculation based on regional averages. Alternatively, direct measurements using personal dosimeters worn by individuals were also used. 

However, the former method could not accurately reflect actual lifestyles, and the latter was limited because it could not be applied to past or future predictions. 

To address these limitations, a research team led by JAEA developed a model that integrates “residents’ location and time of stay (life patterns)” with existing environmental radiation monitoring data and the indoor dose-reduction effects of different building types. This model enables realistic estimation of exposure doses as if residents were actually living in the area. 

Figure 1. Overview of the new external dose evaluation model 

The model estimates indoor air dose rates from outdoor measurements by accounting for the indoor dose reduction factor (the shielding effect of buildings) and natural background radiation (such as radon and radiation from building materials). 

Indoor dose reduction factors were evaluated for 207 buildings in the affected areas. Results showed that the indoor-to-outdoor dose rate ratio depends on the ambient dose rate: when outdoor dose rates are low, the ratio approaches 1. 

Figure 2. Ratio of indoor to outdoor ambient dose rates (measured and estimated values) 

When outdoor ambient dose rates are relatively high, indoor rates are about 40% of outdoor levels. 

As outdoor dose rates approach zero, the indoor/outdoor ratio approaches 1, indicating that background radiation (from radon and building materials) becomes dominant and the shielding effect of houses diminishes. 

By explicitly accounting for this background radiation, the accuracy of indoor dose estimation improved by 32–45% compared with conventional methods. 

The validity of the model was verified using 106 person-days of individual dose data and life pattern records collected from volunteers working in areas including difficult-to-return zones. 

In environments corresponding to an annual dose of around 20 mSv (approximately 3.8 μSv per hour)—the threshold for lifting evacuation orders—the prediction error was about 13%, confirming the model’s practical accuracy. 

Figure 3. Comparison between measured personal doses and doses estimated by the model 

This model has already been utilized in decision-making for lifting evacuation orders in designated “Special Reconstruction and Revitalization Base Areas.” 

It has also been provided to local decontamination verification committees and is now available as a publicly accessible online tool through municipal websites, helping both to support radiation protection and to foster residents’ sense of reassurance. 

The model is also expected to be applied in future for lifting evacuation orders in “Special Return and Residential Areas.” 

Furthermore, in the future, it may serve as a nuclear emergency preparedness tool to predict exposure during evacuation or sheltering-in-place situations in the unlikely event of another nuclear accident.