One of the scientific challenges in ecotoxicology is to contribute to the assessment of the biological quality of hydro-systems by providing information about the impact of pollutants at relevant biological levels, e.g. the population. Under field conditions, population level cannot be easily used to establish a direct relationship between contamination levels and toxicity due to multi-factor environmental influence. One promising approach is the use and the integration of individual responses (survival, growth and reproduction) within demographic models. However this approach involves that robust and reliable in situ bio-assays are available. In general, the interpretation of sub-individual and individual markers used within field experiments is confronted with the absence of reference values; consequently in situ bioassays are mainly used for punctual comparison between assumed control and impacted sites. The assessment and the knowledge of the spatial and temporal variability of biological markers are essential to propose reference values and thus to increase the relevance of in situ bioassays. The aim of this talk is to present our multi-scale approach in the widespread species Gammarus fossarum and particularly 1 methodological development for the assessment of reproductive and feeding rate impairments in in situ exposed organisms 2 - our demarche to propose reference values for these biological responses 3 - how these biological responses could be interpreted in term of impact on population, with population dynamics models and finally 4 - the relevance of this approach for the water quality bio-monitoring through one case-study.