Résumé
Dry extremes or exceptional floods are considered in relation to normality or a standard range of variability. Firstly, in the case of karst basins, it is interesting to question their hydrological variability, compared with other types of aquifers. Using hydrological signatures - metrics that quantify various aspects of hydrological behaviour - we have highlighted the importance of inter-basin groundwater flows and surface water/groundwater interactions in karst basins, compared with non-karst basins. These flows influence therefore the regulation of karst hydrosystems, but also their hydrological response during flood events, which justifies adapting conventional hydrological models.Secondly, the question of extremes arises in terms of long-term hydroclimatic trends and their impact on water resources in karst systems. A recent study we carried out on long-term hydrological dataset (over 40 years) from around 50 karst springs in Europe enabled us to investigate changes in flow rates, seasonality and low and high water conditions. Results show that the sensitivity of karst aquifers to climate change does not depend on their degree of karstification. The evolution of trends for highest and lowest flows over the different periods indicates potential process changes in the last 20 years, with springs that exhibit positive high flow and negative low flow trends. This is a strong indication for changes in the partitioning of concentrated and diffuse recharge, suggesting alterations in precipitation patterns.A third question concerns physico-chemical extremes, and in particular the impact of climate change on water quality. A focus on the Jura massif shows that the nitrate peaks observed during autumn floods are directly linked to the intensity of previous drought periods. A more detailed analysis of the processes involved shows that the origin of these flushes, generally attributed to nitrogen residues in the soil during dry periods, is not the result of a by-pass from the soil, but is linked to the mobilisation of older water already stored in the infiltration zone.These 3 questions are addressed by means of examples of large-scale analyses (continental and regional), which can then be used to work more locally on the processes involved, and to propose valuable insights for assessing the resilience of hydrosystems in the future.