Prediction modeling of geogenic iodine contaminated groundwater throughout China
Journal of Environmental Management(2022)
摘要
Geogenic iodine-contaminated groundwater represents a threat to public health in China. Identifying high-iodine areas is essential to guide the mitigation of this problem. Considering that traditional analytical techniques for iodine testing are generally time-consuming, laborious, and expensive, alternative methods are needed to supplement and enhance existing approaches. Therefore, we developed an artificial neural network (ANN) model and assessed its feasibility in terms of predicting high iodine levels in groundwater in China. A total of 22 indicators (including climate, topography, geology, and soil properties) and 3185 aggregated samples (measured groundwater iodine concentrations) were utilized to develop the ANN model. The results showed that the accuracy and area under the receiver operating characteristic curve of the model on the test dataset are 90.9% and 0.972, respectively, and climate and soil variables are the most effective predictors. Based on the prediction results, a high-resolution (1-km) nationwide prediction map of high-iodine groundwater was produced. The high-risk areas are mainly concentrated in the central provinces of Henan, Shaanxi, and Shanxi, the eastern provinces of Henan, Shandong, and Hebei, and the northeastern provinces of Liaoning, Jilin, and Heilongjiang. The total number of people estimated to potentially be at high-risk areas because they use untreated high-iodine groundwater as drinking water is approximately 30 million. Considering the growing demand for groundwater in China, this work can guide the prioritization of groundwater contamination mitigation efforts based on regional groundwater quality levels to enhance environmental management.
更多查看译文
关键词
Groundwater,Iodine,Prediction modeling,Artificial neural network,Population estimation
AI 理解论文
溯源树
样例
![](https://originalfileserver.aminer.cn/sys/aminer/pubs/mrt_preview.jpeg)
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要