Evaluation of chemical and biological quality of rivers: relevance and validity of a range of in situ sampling methods
Evaluation de la qualité chimique et biologique des cours d'eau : pertinence et validité d'une gamme de techniques d'échantillonnage in situ
Résumé
This work has mainly targeted the study of various sampling strategies to assess water quality of rivers in relation to agricultural diffuse pollution. For that purpose, a panel of methodologies was implemented to provide additional knowledge on the dynamics of pesticide concentrations in several small rivers. One of our objectives was to compare the results obtained from spot sampling, automated integrated weekly sampling and passive sampling to evaluate the exposure of biofilms to various pesticides. These methods have involved the development of innovative, more reliable and less expensive sampling techniques for in situ estimates of the time-weighted average concentrations of pesticides. Several types of passive samplers were tested: Polar Organic Chemical Integrative Samplers (POCIS) for hydrophilic organic pesticides, Stir Bar Sorptive Extraction (SBSE) for hydrophobic organic pesticides and Diffusion Gradient in Thin-Film (DGT) for metals. First, the implementation of POCIS and SBSE required analytical developments and laboratory calibrations. Then, the three types of tools were deployed in 2009 and 2010 on several sampling sites on the Morcille and Ardières rivers (Beaujolais area near Lyon), as well as in the Ruiné creek, a sub-basin of the Charente River. These sites are characterized by different agricultural, hydrological, physicochemical and geological contexts, which allowed to study the performances and limitations of the different sampling techniques under various conditions. Furthermore, analytical methods for measuring pesticides and metals accumulated in the river biofilms were developed. Hence, we were able to evaluate the bioaccumulated contaminants and their likely impacts on the periphyton and to compare results with the exposure estimates derived from the different sampling techniques.