STORMWATER SAMPLING OF OIL PRODUCTS USING SEMIPERMEABLE MEMBRANE DEVICES (SPMDs) Per-Anders Bergqvist, *Lina Ulčinienė, *Viktoras Račys and *Audronė Žaliauskienė Environmental Chemistry, Umeå University, Sweden and *Kaunas University of Technology, Kaunas, Lithuania Introduction Non-point source (NPS) water pollution is a growing threat to the environment (rivers, lakes and ground water) and public health. It is the accumulation of sediment, chemicals, toxics, nutrients, debris and pathogens that are carried by rain water and snow melt into the nearby water body. Polycyclic aromatic hydrocarbons (PAHs) are among most frequently detected and most important contaminants of urban runoff. Important contribution to PAH contamination is via oil and combustion products in stormwater. Attention must be paid even to low concentration pollutants because of their potential high toxicity. Commonly the PAH pollution of pyrogenic and petrogenic origin is distinguished by additional analysis of methylated PAH (Ngabe et. al. 2000, Science of Total Env.). In this study PAH contamination of Tvärån stream flowing through industrial and urban areas is investigated using semipermeable membrane devices (SPMDs). Materials and methods The quality of Tvärån stream was evaluated using SPMDs. These are diffusive samplers that are used for measuring the truly dissolved concentration of pollutants in water and air. A standard SPMD consists of low-density polyethylene membrane contained 1 ml (0,915 g) of pure triolein. They mimic the uptake of bioavailable organic compounds by organisms. Storm water is one example of sampling condition where flow and concentrations are fluctuating. The SPMD technology performs continuous sampling during an extended period of time (several weeks) and makes it possible to calculate a TWA (time weighted average) concentration of dissolved (bioavailable) compounds. In this study SPMDs were deployed at three sites in the Tvärån stream (Umeå, Sweden) upstream (site 1), after the industrial area (site 2) and after urban area (site 3) during 23 days. Schematic map of sampling places is below. The samples were treated according to scheme presented below. Full scan (FS) and selective ion monitoring (SIM) mass spectrometry was used for the determination of compounds of petrogenic origin and SIM was also used for determining pollutants of pyrogenic origin. Results and discussion The determined concentrations of pyrogenic (19 compounds) and petrogenic (5 methylated compounds) PAHs at three sampling sites are shown in figure (columns in red indicates the methylated PAHs). Overall trend is that concentrations of PAHs in the storm water fed stream are increasing in the industrial and urban area. Heavy PAH compounds are commonly not dissolved in water and thus not sampled in the SPMD. The composition of incoming pollutants is thus different between the two areas. Most of the lighter methylated compounds are added into the stream at site 2 whereas heavier methylated compounds displays an additional increase in the urban area. Figure shows the relationship between the concentration of sumPAHs and the ratio of 1-methyl- PHE and PHE (PHE is phenanthrene). Pyrogenic PAHs increase as the stream is passing industrial and also urban areas. However, after sampling site 2 the concnetration of the petrogenic indicator (1-methyl-PHE/PHE) are increasing more than the concentration of pyrogenic PAHs. This indicates continuos source of petrogenic PAHs in the urban area but the source is different compared to the industrial area. Figure shows the comparison of sum of 18 PAHs and sum of 5 methylated PAHs. At sampling site 1 the petrogenic influence is really low which is in agreement with the forest and agricultural environment upstream site 1. The total amounts of PAHs and methylated PAHs do not significantly differ between site 2 and 3, but there is a change of composition as previuosly described. Figure compares GC-MS chromatograms at the three sampling sites. The chromatograms (m/z 133) shows the influence of a group of unknown oil related compounds (possibly trimethylbenzenes). The amount of these compounds are increasing in the stormwater during industrial area. Another significant increase (2 times) could be seen even at site 3 (after urban area). The sources of these compounds are still to be identified. Also other compounds in the stormwater increased at site 2 and 3. Acknowledgements We thank Umeå university, EU Socrates-Erasmus program and Kaunas University of Technology for financial support. Conclusions All three sampling sites show different pollution profiles related to the probable industrial and urban activities in the drainage regions. When the clean water after site 1 entered the site 2 notable increase of petrogenic and pyrogenic pollution were seen. There were no significant change in the pollutant levels when reaching site 3, however change in composition. Non-destructive sampling with SPMD provides a fingerprinting tool for identifying point-sources. SPMDs is very usefull for testing bioaccumulative pollutants and potentially toxic pollutants independent on flow variations and ocasional factors. The described methodology is suitable for stormwater investigations. SPMDs method was succsesfully used to show the impact of industrial and urban areas on the stream water quality but still it needs futher investigations for general application and specific evaluation. Concentrations (ng/L) of petrogenic and pyrogenic pollutants Comparison of petrogenic and pyrogenic pollutants Influence of unknown oil related compounds Relationship between the concentration of sumPAHs and the ratio of 1-methyl-PHE and PHE