Frysning av vattenprover – effekt på vattenkemiska tidsserier

Abstract: Freezing is considered an effective preservation method for long time storage of water samples, but it may however affect the chemical composition of the sample. The aim with this study is to evaluate the effects of freezing on a restricted number of chemical variables, which are used for both research purposes as well as environmental monitoring of water quality. The catchment Gusseltjärn, located in central Sweden, had water samples freeze stored for approximately 20 years before analysis. As no non-frozen samples were analysed at time of collection this study utilised the catchments Lill Fämtan and Risfallet, as references. In order to evaluate the effects of freezing a regression model was created for each individual chemical variable of interest, using data from the reference catchments. The model was then used to calculate expected concentration of what the variables should have been without freezing and storage. A t-test was utilised to evaluate whether the differences were of statistical significance or not. The models indicate that the concentration for sulphur should have been 0,30 mg/l lower and that pH should have been 0,21 units higher if the samples would not have been frozen. As Gusseltjärns water samples were analysed by different laboratories over time, with changes of laboratory procedures as well, this study could not distinguish whether the differences shown from the models was caused due to long time frozen storage or change of laboratory and laboratory procedures.