State Environmental Protection Key Laboratory of Soil Environmental Management and Pollution Control, Nanjing Institute of Environmental Sciences, Ministry of Ecology and Environment, Nanjing 210042, China.
State Environmental Protection Key Laboratory of Soil Environmental Management and Pollution Control, Nanjing Institute of Environmental Sciences, Ministry of Ecology and Environment, Nanjing 210042, China.
Sci Total Environ. 2024 Nov 15;951:175787. doi: 10.1016/j.scitotenv.2024.175787. Epub 2024 Aug 24.
The traditional prediction of the Cd content in grains (Cd) of crops primarily relies on the multiple linear regression models based on soil Cd content (Cd) and pH, neglecting inter-factorial interactions and nonlinear causal links between external environmental factors and Cd. In this study, a comprehensive index system of multi-type environmental factors including soil properties, geology, climate, and anthropogenic activity was constructed. The machine learning models of the tree-based ensemble, support vector regression, artificial neural network for predicting Cd of rice and wheat based on the environmental factor indexes significantly improved the accuracy than the traditional models of linear regression based on soil properties. Among them, the tree-based ensemble models of XGboost and random forest exhibited highest accuracies for predicting Cd of rice and wheat, with R in the test dataset of 0.349 and 0.546, respectively. This study found that soil properties, including Cd, pH, and clay, have greater impacts on Cd of rice and wheat, with combined contribution rates accounting for 65.2 % and 29.7 % respectively. Since wheat sampling areas are located in central and northern China, they are more constrained by precipitation and temperature than rice sampling areas in the south. Geologic and climate factors have a greater impact on Cd of wheat, with a combined contribution rate of 49.9 %, which is higher than the corresponding rate of 20.9 % in rice. Furthermore, the Cd of rice and wheat did not exhibit an absolute linear relationship with Cd, and excessively high Cd can reduce the bioconcentration factor of Cd accumulation in crops. Meanwhile, other environmental factors such as temperature, precipitation, elevation have marginal effects on the increase of Cd of crops. This study provides a novel framework to optimize traditional soil plant transfer models, as well as offer a step towards realizing high precision prediction of Cd content in crops.
传统的作物籽粒镉（Cd）含量预测主要依赖于基于土壤 Cd 含量（Cd）和 pH 值的多元线性回归模型，忽略了外部环境因素之间的相互作用以及与 Cd 之间的非线性因果关系。本研究构建了一个包含土壤特性、地质、气候和人为活动等多类型环境因素的综合指标体系。基于环境因素指标的树基集成、支持向量回归和人工神经网络等机器学习模型用于预测水稻和小麦的 Cd，其精度明显高于基于土壤特性的传统线性回归模型。其中，基于 XGBoost 和随机森林的树基集成模型对水稻和小麦 Cd 的预测精度最高，在测试数据集的 R 分别为 0.349 和 0.546。本研究发现，土壤特性（包括 Cd、pH 和粘土）对水稻和小麦 Cd 的影响较大，综合贡献率分别为 65.2%和 29.7%。由于小麦采样区位于中国中部和北部，它们比南部的水稻采样区更受降水和温度的限制。地质和气候因素对小麦 Cd 的影响较大，综合贡献率为 49.9%，高于水稻的相应比例 20.9%。此外，水稻和小麦的 Cd 含量与 Cd 含量之间没有绝对的线性关系，过高的 Cd 会降低作物中 Cd 积累的生物浓缩系数。同时，温度、降水、海拔等其他环境因素对作物 Cd 含量的增加只有边际效应。本研究为优化传统的土壤-植物转移模型提供了新的框架，并为实现作物 Cd 含量的高精度预测迈出了一步。
