Résumé
In West Africa, surface water dynamics of the Senegal river floodplain remain poorly understood by hydrological monitoring and modelling due to data scarcity and the flat, heterogeneous topography. Understanding their variability under anthropic and climate changes are essential to support ecosystem services including flood recession farming, biodiversity and groundwater recharge. Multi-sensor earth observations from Landsat, MODIS and Sentinel-2 are classified and extracted through cloud-computing geoprocessors (Google Earth Engines) to explore surface water dynamics from 1999-2020 over this 2250 km² floodplain, and combined with ground-based observations to derive a statistical model between flooded areas and upstream flow data. The impact of climate change and upstream dam construction on surface water areas are simulated based on a combined GR4J rainfall-runoff and WEAP modelling approach. Results identify large variations in peak flooded areas between 150,000 ha and 450,000 ha (mean 226,800 ha). Strong correlations obtained between monthly maximum flow at Bakel gauging station and annual peak flooded areas (R² = 0.87) allow simulations of future trends. Projections from 7 Cordex RCM models under RCP4.5 and RCP8.5 highlight significant divergences between scenarios, however results concur towards a decline in peak flooded areas (mean 164,000 ha over 2020-2065). WEAP simulations of four planned dams reveal inundated areas may decline under 100,000 ha, but also highlight how dam management including translucent releases would increase flooded areas (>42%) with minimal impact on hydropower production. Outputs can help design adequate river basin development strategies and optimise water allocation between growing, competing demands in the Senegal river basin.