Which hydrological signatures of intermittence of rivers and streams in order to identify causes and ecological impacts ?

We now know that more than 60 % of the global river system is affected by flow intermittency. In a context of global change and increased anthropogenic pressures, one needs to understand better this phenomenon to prevent resources issues and potential impacts on biological communities, in order to preserve the biodiversity of Intermittent Rivers and Ephemeral Streams (IRES). The Research Infrastructure of Critical Zone Observatories are relevant to contribute to the understanding of IRES because :•Distributed observatories are helpful to cover IRES diversity in order to build a generaland unified theory ;•Long-term in-situ and remote observations are precious to describe intermittency, to an-alyze its multiple drivers and to predict its probable dynamics under climate and humanforcings ;•The holistic and interdisciplinary approach they promote is highly relevant for complexsocio-ecosystems such as IRESThe aim of this work is to identify the features of intermittency from a sample of OZCAR-RI observatories that include a wide range of climates, geologies, land uses and hydrological dynamics. To this end, we proceeded in three steps. We first proposed a typology of intermittency based on zero-flow metrics such as frequency, duration and calendar of zero-flow periods. We then analyzed the results combining the metrics with additional catchment characteristics on the climate, geology and land uses to investigate for potential causes of flow cessation. Finally, we compared the typology with currently available ecological data in order to analyze the relevance of these hydrological metrics to evaluate the impacts of stream flow intermittency on biological communities.All metrics were calculated from streamflow records retrieved from gauged stations of OZCAR-RI. We computed simple metrics such as number of annual zero-flow days, the durations of zero-flow periods, or the annual julian date of the first occurrence of zero-flow in hydrological years. We also calculated indices to characterize the hydrological dynamics such as annual Base Flow Index (BFI) to identify the relative contribution of base flow to the total flow, and the recession time to determine whether a link could be suggested between zero-flow and groundwater. Climate data records, such as mean daily precipitations and reference evapotranspiration were also included, as explanatory variables, to be linked with the hydrological processes.The results show two types of intermittency in our set of catchments, which differ mainly by the frequency and duration of zero-flow periods. While some sites present long periods without streamflow, in other catchments a large number of very brief flow interruptions occur early in the hydrological year. The distinction between these two types is correlated with BFI and climatic variables. Nevertheless, our method does not appear relevant to identify other patterns related to catchment characteristics which can influence the type of intermittence.To increase the relevance of intermittency typology regarding the biological communities, some additional and spatialized metrics should be defined to determine the type of hydrological continuity along the stream network. For instance, a zero-flow characterized by dispersed pools will not have the same impact as a zero-flow on the entire section from upstream to downstream.