Radiocesium dispersed by the Fukushima Daiichi Nuclear Power Plant accident (primarily cesium-137, with a half-life of 30 years) is an important factor affecting the usability of mushroom bed logs. A guideline value of 50 Bq/kg (dry weight) is used for shipment screening, but the degree of variability within a stand had not been well understood, leaving the appropriate number of samples uncertain. 

Photo 1. Mushroom bed-log forest in Tamura City, Fukushima Prefecture (May 2014) 

A research team from FFPRI analyzed the radiocesium concentrations of 418 current-year shoots collected from mushroom bed-log forests. Using a statistical model, they evaluated the variability among individuals within each survey plot (predictive distribution and the 95% prediction interval). 

The analysis revealed that for every survey plot, the ratio between the upper and lower bounds of the 95% prediction interval was contained within approximately 23-fold, demonstrating that the width of within-plot variability can be treated as constant. 

The prediction intervals are extremely wide when no measurements are available for a plot (up to 386-fold; Fig. 1(a)). However, as actual measurements are added, the prediction interval narrows (Fig. 1(b), (c)). 

Considering these results along with previous studies, the team determined that collecting five individuals per plot is sufficient to estimate the concentration distribution with practical accuracy. 

Figure 1. Examples of radiocesium concentration predictions for current-year shoots 

(a) Predictive distribution (blue line) and 95% prediction interval (blue shaded area) when no measurements are available. The histogram shows the distribution of 418 measured values. 

(b) Predicted distribution assuming five measurements were obtained in a low-concentration case (red points: measured values; red line: predicted distribution; red shading: 95% prediction interval).