Multivariate statistical approaches as applied to environmental physics studies

The present communication deals with the application of the most important environmetric approaches like cluster analysis, principal components analysis and principal components regression (apportioning models) to environmental systems which are of substantial interest for environmental physics — surface waters, aerosols, and coastal sediments. Using various case studies we identify the latent factors responsible for the data set structure and construct models showing the contribution of each identified source (anthropogenic or natural) to the total measure of the pollution. In this way the information obtained by the monitoring data becomes broader and more intelligent, which help in problem solving in environmental physics.