Geographically weighted regression Kriging (GWRK) model and regression Kriging (RK) model were employed. On the basis of soil selenium content, the model to predict the soil cadmium content was constructed, which was explored the soil cadmium content prediction method under the condition of only having soil selenium content data. And cadmium pollution prevention and control of risk assessment and selenium-rich agricultural planning were explored. The results indicated that ①Chaihu is qualified for establishing selenium industry as the area of selenium-rich cultivated land reached 5316.57 hm2.②The cadmium background value in soil fits the class Ⅱ standard of Soil Environmental Quality Standard of China,which is safe for plants and the environment. ③There exists significant positive correlation between selenium content and cadmium con-tent in soil. ④GWRK model is qualified for predicting cadmium content in soil based on the data of selenium content in soil, and which is much better than RK does. ⑤Both RK and GWRK model have a similar issue of value range shrinkage, which leads the decline in simulation accuracy, and the accuracy of prediction is significantly higher in the vicinity of the median. Overall, GWRK model could be used for predicting cadmium content based on selenium content, and different plat-ing plans should be carried out for cultivated land according to the differentiation of selenium content and cadmium content in soil.%构建回归克里格模型(RK)和地理加权回归克里格(GWRK)模型,以土壤硒含量为基础,构建对土壤镉含量进行预测的模型,探索在只具备土壤硒含量数据的条件下进行土壤镉含量预测的方法,以及镉污染风险评估防治和富硒农业规划.结果表明,①柴湖镇富硒耕地面积达到5316.57 hm2,具备发展富硒农业的条件;②耕地土壤镉背景值处于《土壤环境质量标准》Ⅱ类中含量偏低的水平,基本不会对植物和环境造成危害和污染;③土壤硒含量与镉含量之间存在显著的正相关关系;④利用GWRK预测的土壤镉含量与检测值较为接近,精度明显高于RK的预测结果,可作为区域土壤镉含量评估的依据;⑤RK和GWRK模型的预测结果都存在值域向中值收缩的现象,是造成模拟精度下降的主要原因,而在中值附近预测的精度明显更高.GWRK模型可以用于在已知土壤硒含量的情况下镉含量的预测,根据不同区域耕地土壤硒含量和镉含量的差异,进行差别化的农业规划.
展开▼
