Thorium-232 Transfer Dataset – Sokoto Basin

This dataset contains paired soil and crop measurements used to evaluate thorium-232 (²³²Th) soil-to-crop transfer in calcareous semi-arid agricultural systems of the Sokoto Basin, Nigeria. The dataset includes activity concentrations of ²³²Th in soil and edible crop components (beans, maize, and pepper), along with calculated soil-to-plant transfer factors (TF). Additional variables include soil physicochemical properties (pH, CaCO₃ equivalent, organic carbon, clay content), crop type, and geographic location (Sokoto and Kebbi States). The dataset was used to develop and validate machine learning models (Random Forest, XGBoost, Artificial Neural Networks, Support Vector Regression, and ensemble methods) for predicting radionuclide transfer and assessing the influence of soil geochemical conditions on thorium bioavailability.本数据集包含成对的土壤与作物测量数据，用于评估尼日利亚索科托盆地（Sokoto Basin）石灰质半干旱农业系统中钍-232（thorium-232，²³²Th）的土壤-作物转移行为。本数据集涵盖土壤与可食用作物组分（豆类、玉米、辣椒）中²³²Th的活度浓度，以及经计算得到的土壤-植物转移因子（transfer factor，TF）。 额外收录的变量还包括土壤理化性质（pH值、碳酸钙当量、有机碳、黏粒含量）、作物类型以及地理位置信息（索科托州与凯比州）。本数据集被用于开发并验证机器学习模型，包括随机森林（Random Forest）、XGBoost、人工神经网络（Artificial Neural Networks）、支持向量回归以及集成学习方法，以实现放射性核素转移的预测，并评估土壤地球化学条件对钍生物可利用性的影响。