Using Integrated Multivariate Statistical Approaches to Assess the Hydrochemistry of Surface Water Quality, Lake Taihu Basin, China

Date of Award


Degree Type


Degree Name

Master of Science (MS)


Earth Sciences


Donald I. Siegel


Geographical Information System, Hierarchical Cluster Analysis, Hydrochemistry, Principal Component Analysis, Surface water quality, Water-Rock Interaction

Subject Categories

Earth Sciences | Geology | Hydrology


The demand for freshwater needed for increasing crop production, population, and industrialization occurs almost everywhere in China and these conflicting needs have led to widespread water contamination. Because of historical heavy nutrient loading, from all these sources, Lake Taihu (eastern China) notably suffers periodic hyper-eutrophication and drinking water deterioration, which has led to shortages of freshwater for the City of Wuxi and other nearby cities. Contamination does not occur in isolation of the broader geochemical and hydraulic processes which lead to the major dissolved loads of solutes to waters. In this study, I investigated the broad hydrochemical setting of Lake Taihu and its basin, and then assessed how different dominant land use patterns influence the variability of surface water chemistry in the lake and its watershed.

I synoptically collected 26 water samples in 2010 on the lake and 88 water samples on north-western sub-watersheds, and analyzed them for field parameters (e.g. pH, SC, DO, TDS, etc.), nutrients, and major and minor solutes which are useful to fingerprint solute sources and geochemical reactions controlling them.

Graphical methods such as bivariate plots and Piper Diagrams show the waters broadly change throughout the basin from calcium-magnesium-bicarbonate hydrochemical facies type water to mixed sodium-sulfate-chloride type waters. No halite sources occur in the basin so the addition of sodium, chloride and potentially sulfate in the major solute mix logically should be related to land use and potential contamination from it.

Principle component analysis (PCA) of stream and lake water chemical compositions produced three principal components that explained 71% loading of the cumulative variance in the water quality. These three principal components reflect three major types water chemistry related to land use patterns. Agriculture land use is associated with greater concentrations of nutrients; urban areas are associated with high concentration of sodium, chloride, sulfate, fluoride and potassium, and western low hills and northern study areas largely show a calcium-magnesium-bicarbonate water type. Hierarchical cluster analysis produced results similar to that of the PCA analysis, clustering into agriculture, urban area, rural area and forestry areas, lake water and water at tributaries mouths. Broadly speaking, future remediation to reduce nutrient loadings to the lake or industrial contamination could now be focused on specific land use practices, readily identifiable using GIS.

This document is currently not available here.