Logo: to the web site of Uppsala University

Abstract

The continuous release of chemicals into aquatic environments is an increasing environmental concern. Many emerging contaminants are often not systematically monitored, which can affect aquatic life and ecosystems. Fish are crucial indicators of freshwater pollution because they bioaccumulate pollutants from water, sediment, and their food sources, providing valuable insights into contaminant profiles. This thesis investigates and compares chemical contaminants in two species, Squalius orpheus and Barbus stumicae, collected from five locations across Greece. It also assesses the effectiveness of liquid chromatography high-resolution mass spectrometry (LC-HRMS) for suspect and non-target screening as an early-warning system for freshwater monitoring. 

The workflow included the identification of chemical contaminants using retention-time index (RTI) filtering, followed by validation with orthogonal tandem mass spectrometer (MS/MS) evidence. Identified compounds were assigned to Schymanski confidence levels, and their peaks were normalized using internal standards. The chemicals were then categorized, and their acute aquatic toxicity was predicted using Ecological Structure Activity Relationships (ECOSAR). Furthermore, the spatial distribution of the identified chemicals was evaluated to statistically compare the different samples and assess variation across locations. The analysis identified 354 compounds after excluding 107 potential false positives from an initial RTI detection of 461 features. These compounds were assigned to four Schymanski confidence levels: 33 compounds at level 2.1, 39 at level 2.2, 73 at level 3.1 and 209 at level 3.2. Natural compounds represented the largest group, with 233compounds, followed by industrial chemicals with 61 compounds, pharmaceuticals with 25 compounds, personal care products with 24 compounds, and others chemicals with 21 compounds. Spatial analysis showed distinct chemical profiles across sites, with the agriculturally impacted area having the highest number of detected compounds and contamination levels, whereas the forest-surrounded area had the fewest. ECOSAR predicted that 193 of the 354 compounds are highly toxic to at least one aquatic test group. Moreover, 59 were toxic, 63  moderately toxic, and 39  exhibited low or no toxicity. Among Level 2 compounds, 41 out of 72  were predicted to be very toxic. The findings indicate that LC-HRMS-based suspect and non-target screening provides early-warning insights into emerging contaminants in freshwater fish. This approach can aid in prioritizing chemicals, confirming targets, and assessing environmental risks in freshwater ecosystems.