Surface water quality assessment by the use of combination of multivariate statistical classification and expert information.

The present study deals with the assessment of surface water quality from an industrial-urban region located in northern Poland near to the city of Gdansk. Concentrations of thirteen chemicals including total polycyclic aromatic hydrocarbons (PAHs), halogenated volatile organic compounds (HVOCs) and major ions in the samples collected at five sampling points during six campaigns were used as variables throughout the study. The originality in the monitoring data treatment and interpretation was the combination of a traditional classification approach (self-organizing maps of Kohonen) with PAH diagnostic ratios expertise to achieve a reliable pollution source identification. Thus, sampling points affected by pollution from traffic (petroleum combustion products), from crude oil processing (petroleum release related compounds), and from phosphogypsum disposal site were properly discriminated. Additionally, it is shown that this original assessment approach can be useful in finding specific pollution source tracers. Copyright 2010 Elsevier Ltd. All rights reserved.