引用本文:林斯杰,齐永强,杨梦曦,等.基于PCA-SOM的北京市平谷区地下水污染溯源[J].环境科学研究,2020,33(6):1337-1344.
LIN Sijie,QI Yongqiang,YANG Mengxi,et al.Source Analysis of Groundwater Pollution in Pinggu District of Beijing Using PCA-SOM[J].Research of Environmental Sciences,2020,33(6):1337-1344.]
【打印本页】   【HTML】   【下载PDF全文】   查看/发表评论  【EndNote】   【RefMan】   【BibTex】
←前一篇|后一篇→ 过刊浏览    高级检索
本文已被:浏览 160次   下载 203 本文二维码信息
码上扫一扫!
分享到: 微信 更多
基于PCA-SOM的北京市平谷区地下水污染溯源
林斯杰1,2,4, 齐永强1,4, 杨梦曦1, 杨庆3, 杨梦凡1,4, 刘毅2, 胡清1
1. 南方科技大学环境科学与工程学院, 广东 深圳 518055;2. 清华大学环境学院, 北京 100091;3. 北京市地质矿产勘查院, 北京 100195;4. 北京环丁环保大数据研究院, 北京 100083
摘要:
为了解北京市平谷区地下水污染物来源,以平谷区2010—2018年监测数据为基础,使用PCA(主成分分析法)识别了地下水水质指标因子,使用自组织映射识别了污染物的空间分布.结果表明:通过监测指标间的Pearson检验发现,平谷区地下水电导率与ρ(Ca2+)(p=0.936)、总碱度与ρ(HCO32-)(p=0.981)、ρ(Mg2+)与总硬度(p=0.944)指标之间显著相关.地下水化学类型主要以HCO3-Ca型为主,其次为HCO3-Mg型.NH4+、SO42-、Cd、Fe(Ⅱ)、NO2指标空间分布离散性和差异性较大,存在局部富集现象.通过因子分析法筛选出影响平谷区地下水水质的8个公因子,首要影响因子为溶滤-富集作用(贡献率为22.398%),次要影响因子为农业、养殖业和填埋场等人为活动作用(贡献率为16.533%),雨水下渗作用(贡献率为8.035%)、工业源人为活动(贡献率为7.466%)对地下水也有一定影响.通过比较各指标的SOM(Self-Organizing Map,自组织映射)特征图像和监测井映射特征图像,发现NH4+受山前地带林业、种植业和平原地带农业、养殖业的双重影响,Na+、Mn受平原地带人为活动的影响;同时,NH4+、NO3-、NO2三者之间及Fe(Ⅱ)与Fe(Ⅲ)之间来源不同,Cd、Al、氰化物三者具有同一来源.研究显示,PCA-SOM(PCA与SOM相结合)可以对地下水化学组分来源进行定性识别与定量分析.
关键词:  自组织映射(SOM)  地下水污染溯源  主成分分析法(PCA)
DOI:10.13198/j.issn.1001-6929.2020.05.28
分类号:X523
基金项目:国家水体污染控制与治理科技重大专项(No.2018ZX07109-002)
Source Analysis of Groundwater Pollution in Pinggu District of Beijing Using PCA-SOM
LIN Sijie1,2,4, QI Yongqiang1,4, YANG Mengxi1, YANG Qing3, YANG Mengfan1,4, LIU Yi2, HU Qing1
1. School of Environmental Science and Engineering, Southern University of Science and Technology, Shenzhen 518055, China;2. School of Environment, Tsinghua University, Beijing 100091, China;3. Beijing Geological and Mineral Exploration Institute, Beijing 100195, China;4. Beijing Huanding Environmental Big Data Institute, Beijing 100083, China
Abstract:
In order to understand the source of groundwater pollutants in Pinggu District of Beijing, based on the monitoring data of Pinggu District from 2010 to 2018, this study used PCA to identify the groundwater quality index factors, and identified the spatial distribution of pollutants using SOM to map the pollution index and the location of the monitoring wells and compare mapping feature by comparison images. The Pearson test between monitoring indicators showed that there was a significant correlation between electroconductibility and ρ(Ca2+)(p=0.936), total alkalinity and ρ(HCO32-) (p=0.981), ρ(Mg2+) and total hardness (p=0.944) in Pinggu District groundwater, and the chemical type was mainly HCO3-Ca, followed by HCO3-Mg type. The spatial distribution of NH4+, SO42-, Cd, Fe(Ⅱ) and NO2 was very discrete and different, and there was local enrichment. Eight common factors affecting groundwater quality in Pinggu District were screened out by factor analysis. The primary impact factor was leaching-beneficiation (contribution rate 22.398%), and the secondary impact factors were human activities such as agriculture, aquaculture, and landfills (contribution rate of 16.533%), rainwater seepage (contribution rate of 8.035%), and industrial source. Human activities (contribution rate of 7.466%) also had a certain effect on its groundwater. By comparing the SOM mapping characteristic images of various indicators, NH4+ was affected by the dual effects of forestry, plantation, and agricultural production in the plain area. The characteristic images of the Na+, Mn response area and the plain area coincide, showing the influence of human activities. At the same time, NH4+-NO3--NO2, Fe(Ⅱ)-Fe(Ⅲ) were not from the same source, and Cd, Al, and cyanide had the same source. The results show that PCA and SOM can be used to identify and analyze the chemical components of groundwater.
Key words:  Self-Organizing Map (SOM)  source analysis of groundwater pollution  PCA