Radiocaesium soil-to-plant transfer: a meta-analysis of key variables and data gaps on a global scale

Area of Science:

• Environmental Science
• Radiological Science
• Agricultural Science

Background:

• A lack of harmonized, accessible data on radiocaesium transfer from soil to plants hinders model development and risk assessment.
• This gap is particularly critical for regions with nuclear facilities and limited research.

Purpose of the Study:

• To create a harmonized, publicly accessible database of radiocaesium soil-to-plant transfer factors (CR).
• To facilitate the evaluation and establishment of transfer models for radiocaesium.
• To support impact assessments of food chain contamination.

Main Methods:

• Systematic literature screening for radiocaesium CR data.
• Data extraction based on experimental soundness, relevance, and traceability.
• Compilation of a harmonized database including 7,182 CR data points from 139 sources.
• Statistical analyses (univariate, multivariate, mixed-effects models) and machine learning (random forest).

Main Results:

• The database contains 7,182 CR data points, with values ranging from 0.000028 to 380 kg kg⁻¹.
• CRs were highest in tropical soils and lowest in temperate soils, though tropical and arid climate data remain limited.
• Study-specific methods, soil properties, and plant species significantly influenced CRs.
• Semi-mechanistic models showed moderate fit (R²=0.42-0.50), while random forest (R²=0.51) and mixed-effects models (R²=0.58) performed better.

Conclusions:

• The developed harmonized database is a valuable resource for radiocaesium transfer research.
• The findings highlight the influence of study design and environmental factors on transfer rates.
• Further data completion and analysis can enhance machine learning applications for improved contamination assessments.