关键词: 聚类分析, 判别分析, 水质, 时空变化, 堵河

Abstract:

Multivariate statistical techniques including cluster analysis and discriminant analysis were performed on a water quality data set collected in 9 sites of the Du River in Hubei Province during 2006~2007,aiming at identifying spatialtemporal characteristics of water quality,assisting watershed management and water resource protection and utilization.The results indicate that water quality had significant seasonal variations except Cl-,TDS and Turbidity.Temperature,pH,NO-3,TP is significantly higher in flood seasons than in dry seasons,while SO2-4,HCO-3,NH+4,DO show the opposite trend,that is,in flood season significantly lower levels are observed than in dry season.Grouping the sampling sites based on cluster analysis and recognizing the discriminant parameters based on discriminant analysis which can account for most of the spatial variation in water quality.The sampling sites are classified into 4 classes.Only four discriminant parameters (NO-3,TDS,SO2-4,HCO-3) are identified in discriminant analysis for 4 clusters,which can reflect the overall water quality of the spatial variation.

Key words: cluster analysis, discriminant analysis, water quality, spatialtemporal patterns, Du River, Hubei Province