|
| 基于机器学习的长江流域磷酸盐源解析及驱动因子识别 |
| Machine Learning Based Phosphate Source Analysis and Driving Factor Identification in the Yangtze River Basin |
| |
| DOI： |
| 中文关键词: 溶解性磷酸盐 空间分布特征 机器学习 源解析 驱动因素 长江流域 |
| 英文关键词: Soluble reactive phosphorus Spatial distribution characteristics Machine learning Source apportionment Driving factors Yangtze River basin |
| 基金项目:国家自然科学基金资助项目（42072201）；安徽省高校协同创新基金资助项目（GXXT-2021-017） |
|
| 摘要点击次数: 982 |
| 全文下载次数: 171 |
| 中文摘要: |
| 采集长江流域内不同季节的84个水样并测定其理化指标，运用自组织映射神经网络和随机森林算法进行溶解性磷酸盐（SRP）的来源解析与驱动因素分析。结果表明，长江流域SRP测定值为0011 mg/L～0771 mg/L，空间差异显著，且中下游地区SRP值较高；上游SRP来源主要受人为源和自然源共同作用影响，自然土地利用类型的贡献率最高，达到55%；中游受到生活污水、农业等人为源影响较大，自然土地利用类型贡献率为45%，人工土地利用贡献率为23%；下游主要受到交通、工矿业和水产养殖等人为源的影响，大气污染贡献率高达56%。 |
| 英文摘要: |
| 84 water samples were collected across different seasons in the Yangtze River basin， and their physicochemical indicators were determined. Self organizing map neural network and random forest algorithm were applied to analyze the sources of soluble reactive phosphorus（SRP） in water and its driving factors. The results showed that the measurements of SRP in the Yangtze River basin ranged from 0011 mg/L to 0771 mg/L， with significant spatial differences. The measurements of SRP were high in the middle and lower reaches. SRP in the upstream was mainly influenced by both anthropogenic and natural sources， and natural land use types contributed the most， reaching 55%. In the midstream section， it was greatly affected by anthropogenic sources such as domestic sewage and agriculture. Natural land use types contributed 45%， while anthropogenic land use accounted for 23%. Anthropogenic sources such as transportation， industry and mining， and aquaculture mainly affected downstream areas， contributing as much as 56% to air pollution. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
