Multivariate classification and modeling in surface water pollution estimation.

The present study deals with the application of self-organizing maps (SOM) and multiway principal-components analysis to classify, model, and interpret a large monitoring data set for surface water quality. The chemometric methods applied made it possible to reveal specific quality patterns of the chemical and biological parameters used to monitor the water quality (relation between water temperature, turbidity, hardness, colibacteria), seasonal impacts during the long period of observation and the relative independence on the spatial location of the sampling sites (water supply sources for the City of Trieste).

Comparison of self-organizing maps classification approach with cluster and principal components analysis for large environmental data sets.

Three classification techniques (loading and score projections based on principal components analysis (PCA), cluster analysis (CA) and self-organizing maps (SOM)) were applied to a large environmental data set of chemical indicators of river water quality. The study was carried out by using long-term water quality monitoring data. The advantages of SOM algorithm and its classification and visualization ability for large environmental data sets are stressed. The results obtained allowed detecting natural clusters of monitoring locations with similar water quality type and identifying important discriminant variables responsible for the clustering. SOM clustering allows simultaneous observation of both spatial and temporal changes in water quality. The chemometric approach revealed different patterns of monitoring sites conditionally named "tributary", "urban", "rural" or "background". This objective separation could lead to an optimization of river monitoring nets and to a better tracing natural and anthropogenic changes along the river stream.

Chemometrics in monitoring spatial and temporal variations in drinking water quality.

This case study reports multivariate techniques applied for the evaluation of temporal/spatial variations and interpretation of monitoring data obtained by the determination of chloro/bromo disinfection by-products in drinking water at 12 locations in the Gdansk area (Poland), over the period 1993-2000. The complex data matrix (1756 observations) was treated with various multivariate techniques. Cluster analysis (CA) was successful, yielding two different groups of similarity reflecting different types of drinking water supplied (surface and groundwater). The locations supplied in general with groundwater could be further classified into two subgroups, depending on whether the groundwater was mixed with surface water or not. Analysis of variance (ANOVA) was used to classify and thus confirm the groups found by means of cluster analysis and proved the existence of statistically significant differences between the concentration levels of CHCl3, CHBrCl2+C2HCl3, CHBr2Cl, and CH2Cl2 in the samples collected. Of all the variables evaluated, only three were characterized by statistically significant correlations (CHCl3, CHBrCl2+C2HCl3, CHBr2Cl). The analysis of correlation coefficients revealed that chloroform formed as the main chlorinated disinfection by-product and, furthermore, the natural presence of bromide in water (both ground and surface) results in the formation of brominated disinfection by-products (DBPs). Temporal variations of volatile organic chlorinated compounds (VOCls) were also evaluated by multidimensional ANOVA. Observation of temporal changes in the concentration of VOCls at the location supplied with both surface and groundwater reveals a steady improvement in drinking water quality. In general, the study shows the importance of drinking water monitoring in connection with simple but powerful statistical tools to better understand spatial and temporal variations in water quality.