Résumé
In the Sudanian region of West Africa, the sub-humid climate is strongly influenced by seasonal and interannual rainfall variability associated with the occurrence of the West African monsoon seasonal cycle. While rivers flow intermittently with a dry season that lasts at least six months and annual river floods of highly variable intensity, several studies along the soil-plant-river-atmosphere continuum in this region have highlighted the important role of vegetation water uptake in shaping regional water pathways. Plants may control recharge through root water uptake in the vadose zone, but also by tapping groundwater year-round, therefore impacting groundwater-fed river regimes as well. Understanding the complex hydrological mechanisms of such systems and their response to climatic and environmental drivers remains a significant knowledge gap. We address this scientific question in the upper Ouémé basin, covering 10,000 km2 in northern Benin. The area is monitored by the AMMACATCH critical zone observatory (www.amma-catch.org), which maintains long-term hydroclimatic data records at several locations, including precipitation, streamflow, soil moisture, piezometric levels and leaf area index. Additionally, isotopic monitoring of precipitation and river water has been conducted since 2018. At this scale where hillslope dynamics combine with larger-scale organization of water pathways, we deployed a processbased, spatially-distributed model (here with a 1-km2 resolution) over the 2018-2021 period. In this modeling approach, the local energy balance directly depends on vegetation foliar dynamics (forced by leaf area index) and shapes water partitioning in the critical zone, while vertical and lateral water routing also tracks stable isotopic signatures (2H, 18O). By constraining the model with different combination of the aforementioned datasets, we applied an automatic calibration procedure with multi-objective cost functions to identify bestperforming model configurations. This approach seeks to identify the various trade-off between process- and scale-specific information content used in these “calibration scenarios”,and specifically the added value of isotopic information to better capture the origin and pathways of water simulated in plant, stream, and groundwater. First results show encouraging performances regarding channel dynamics, for both discharge and isotopic signature. Further calibration is ongoing to obtain a better representation of groundwater depth and relevant output fluxes in this region such as evapotranspiration.