Application of linear discriminant analysis to the study of dew chemistry on the basis of samples collected in Poland (2004-2005)

Abstract

This paper presents the results of determination of selected characteristics (anions, cations, formaldehyde, hydrogen peroxide, phenols, TC, TIC, TOC and metals) in dew samples collected in six different sites in Poland. The influence of local parameters (e.g. wind speed, humidity) was investigated. Discriminant analysis was applied to the study of several dew samples collected from different sampling sites covering six agglomerations in Poland. Discriminant function analysis was used not only for classifying samples into different groups with a better than chance accuracy, but also for detecting the most important variables that discern between the groups of samples considered. In this way it was possible to identify which ions or other physicochemical features are responsible for the similarities or differences observed between different groups of dew samples. A good agreement with their origin and location was observed. It is interesting to note that the classification of all samples was dominated by pH, wind direction, pressure and temperature with a significant contribution of Na+ and Cl- ions. © Versita Warsaw and Springer-Verlag Berlin Heidelberg 2009.