摘要
Abstract
To comprehensively assess the water quality conditions of Min River basin and recognize tem-poral and spatial variation features of water quality, the SOM( Self-Organizing Mapping) clustering algo-rithm was used to do spatio-temporal cluster analysis eight water quality indexes for nineteen monitoring sections, to recognize the principal pollutants in spatiotemporal dimension by the PCA( Principal Compo-nent Analysis) clustering algorithm, and to analyze water quality variation features of Min River under the spatio-temporal background. The results showed that, SOM network divided Min River's water quality periods into two time stages, i. e. , from April to November and from December to next March, water quality of the former time stage was better than the latter one. The river basin 's nineteen sections were clustered into three kinds, i. e. , S1, S2 and S3. Of which, S1 represented upper reach of Min River's three tributaries, where the water quality was fine; as primary agriculture and forestry production base, S2 represented middle and lower reaches where water quality was affected principally by non -point source pollution; S3 represented downstream reach, where water quality was main affected by sanitary sewage and industrial wastewater. The analysis of PCA indicated that water qualities of middle and lower reach of Min River was mainly controlled by nitrogen and phosphors.关键词
自组织特征映射神经网络/聚类分析/时空分析/水质变化/闽江Key words
self-organizing mapping/cluster analysis/spatio-temporal analysis/water quality change/Min River分类
资源环境
