NAU Wheat Cd Prediction Model

Wheat Cd Prediction Model by Machine Learning

Model Notes 1. The model dataset consisted of 1339 pairs of soil-wheat samples collected from Jiyuan City, Xinyi City, and Changshu City, China.

2. The model development was based on the extreme gradient boosting machine (XGBoost) algorithm. The dataset was divided into a ratio of 8:2, and then performed 5-fold cross-validation.

3. The model parameters included soil pH (soil-water ratio 1:2.5), soil total Cd, total Mn, and total Zn concentrations (measured by ICP-MS).

4. The model applicable soil pH and total Cd concentration ranges were 4.0 - 9.0 and 0.01 - 2.20 mg/kg, respectively.

5. The model derived the soil Cd thresholds for wheat safe production as follows：

① pH ≤ 5.5：Cd ≤ 0.30 mg/kg

  

② 5.5＜pH ≤ 6.5：Cd ≤ 0.34 mg/kg

  

③ 6.5＜pH ≤ 7.5：Cd ≤ 0.47 mg/kg

  

④ pH＞7.5：Cd ≤ 0.55 mg/kg



How to Cite If you use this model，please cite this website ( http://wheat.cdpredict.cn ) and the following papers：

  • Lü, Q.X., Tang, Z.X., Tang, Z., Dong, G., Xu, Z.R., Zhao, F.J., Wang, P., Interpretable machine learning models to predict cadmium in wheat for safe production and soil management. Fundamental Research, 2025, https://doi.org/10.1016/j.fmre.2025.05.001. [Link]



    • Support Funds 1. National Natural Science Fund for Distinguished Young Scholars, Soil Heavy Metal Pollution and Control, 2024-2028.

    2. Central Universities Business Fees, Soil and Human Health, 2022-2023.

    3. Special Project of Central Fiscal Funds, Jiangsu Province Construction of Safe Utilization of Contaminated Cultivated Land Promotion Area (Changshu), 2021-2023.