Application of multivariate statistics, trend- and cluster analysis to groundwater quality in the Tejo and Sado aquifer

Abstract:

A methodology based on multivariate statistics, classification and trend analysis techniques is applied to describe and classify the groundwater quality monitoring network for the confined aquifer of the Tejo and Sado system. Hydrochemical data consists of information on Ph, conductivity, Na, K, Ca, Cl, HCO3, Mg and SO4 measured in 27 stations and annually sampled from 1980 to 1994. Principal Component Analysis (a factorial method) is used, to define the intercorrelation structures between the variables in different time horizons, in order to identify the differences and similarities of the most relevant parameters. The objective is to generate synthetical indicators (indexes) for a characterization of the hydrochemical changes. The temporal variation of those indexes is analysed by the nonparametric Mann-Kendall test and the global trend is evaluated by means of tau coefficients and robust slopes. Cluster Analysis is used to establish the groups of stations of similar temporal patterns. -from Authors

Año de publicación:

1995

Keywords:

Fuente:

scopus