5.4 Water column data from two scuba-obtained glass jar samples 



5.4.1 Results of Analysis : 



Two samples obtained in glass jars by Mats Notini during the first 

 few days of the Tsesis incident were analyzed as part of the above 

 mentioned analytical work performed by ERCO. These were samples A-l (1 

 meter depth) and A-2 (0.5 meter depth) taken on 11/1/77 at position D, 

 in the enclosed bay at Lindholmen (see Fig. 7.1.1). Analysis (Boehm and 

 Fiest, 1977) by UV-f luorescence indicated peak concentrations (at 310 

 nm) of 50.9 (Jg/ 1 and 58.2 (Jg/ 1 respectively. Both samples exhibited 

 spectra, classified as type "B", similar to whole Tsesis oil. These two 

 samples were not included in the first set that was selected for GC/MS 

 analysis to identify compositional elements, and as this omission could 

 not be corrected later, it was necessary to rely on the fluorescence 

 spectra to infer that the contaminant was indeed Tsesis oil. 



Comparing these spectra statistically (above 308 run), the corre- 

 lation coefficient between the water sample spectra and the whole oil 

 spectra is about 0.97. The greatest differences, although slight, 

 appear as lower concentrations of the lighter fractions. 



5.4.2 Discussion : 



The two glass jar samples provided the only valid water column 

 hydrocarbon concentration data. This sampling was done underwater, thus 

 avoiding some of the problems of through-the-slick contamination. There 

 is little question that direct sampling underwater incurs the least 

 chance for contamination if careful procedures are followed. Although 

 statistically insufficient, the two samples seem reliable in themselves, 

 and can be useful as indications if assumptions can be made about varia- 

 bility. Analytical variability (including the extraction precedures) 

 can be determined from analysis of replicate extractions of the same 

 sample. Differences between replicates of 21 samples ranged from to 4 

 (Jg/1, implying good analytical and procedural control. 



90 



