Display Settings:

Format

Send to:

Choose Destination
See comment in PubMed Commons below
Water Res. 2001 Aug;35(12):2881-94.

Pattern recognition techniques for the evaluation of spatial and temporal variations in water quality. A case study: Suquía River Basin (Cordoba-Argentina).

Author information

  • 1Universidad Nacional de Córdoba, Facultad de Ciencias Químicas, Dto. Bioquímica. Cdad. Universitaria, Argentina. dwunder@bioclin.fcq.unc.edu.ar

Abstract

We report a comparative study using three different chemometric techniques to evaluate both spatial and temporal changes in Suquía River water quality, with a special emphasis on the improvement obtained using discriminant analysis for such evaluation. We have monitored 22 parameters at different stations from the upper, middle, and beginning of the lower river basin during at least two years including 232 different samples. We obtained a complex data matrix, which was treated using the pattern recognition techniques of cluster analysis (CA), factor analysis/principal components (FA/PCA). and discriminant analysis (DA). CA renders good results as a first exploratory method to evaluate both spatial and temporal differences, however it fails to show details of these differences. FA/PCA needs 13 parameters to point out 71% of both temporal and spatial changes, consequently data reduction from FA/PCA in this case is not as considerable as expected. However, FA/PCA allows to group the selected parameters according to common features as well as to evaluate the incidence of each group on the overall change in water quality, specially during the analysis of temporal changes. DA technique shows the best results for data reduction and pattern recognition during both temporal and spatial analysis. DA renders an important data reduction using 6 parameters to afford 87% right assignations during temporal analysis. Besides, it uses only 5 parameters to yield 75% right assignations during the spatial analysis of four different basin areas. DA allowed us to greatly reduce the dimensionality of the starting data matrix, pointing out to a few parameters that indicate the biggest changes in water quality as well as variation patterns associated with seasonal variations, urban run-off, and pollution sources, presenting a novel approach for water quality assessments.

PMID:
11471688
[PubMed - indexed for MEDLINE]
PubMed Commons home

PubMed Commons

0 comments
How to join PubMed Commons

    Supplemental Content

    Icon for Elsevier Science
    Loading ...
    Write to the Help Desk