Interannual to Multi-decadal streamflow variability in West and Central Africa: Interactions with catchment properties and large-scale climate variability

International audience Droughts and floods are responsible for ~ 80% of fatalities, and ~70% of economic losses related to natural hazards in Sub-Saharan Africa. Rainfall variability which is driven by interannual to multi-decadal climate fluctuations, here underpins these issues but is also crucial to agriculture and livelihoods. However, due to data scarcity, little is known about the impact of these climatic fluctuations and catchment properties on streamflow variability. Therefore, in this study, using a newly reconstructed streamflow dataset, we aim to extend understanding of hydrological variability through investigation of key large-scale controls which modulate climate-river flows relationships at the subcontinental scale. Modes of variability are extracted using continuous wavelet transform and maximum overlap discrete wavelet transform. We first assess the relative importance of catchment properties in modulating streamflow and modes of variability, before examining teleconnections with climate variables from different datasets (ERSST.v5 and 20CR.v2). The results underline the importance of interannual components of river flow along the Gulf of Guinea and the coastal regions of Central Africa where annual rainfall amounts are higher, and runoff is mostly generated from surface and sub-surface processes. In contrast, river flows along the Sahelian band show stronger multi-decadal components: this partly reflects the prominent role of geological formations in modulating high frequency rainfall signals. These modes of variability (also significantly detected in rainfall) are likely related to anomalies in sea-surface temperature (SST) anomaly patterns and associated atmospheric circulations, which together modulate the West African monsoon. Clarifying the picture of streamflow variability in western and Central Africa thus opens up the prospect for improved future predictions for streamflow and water resources in data scarce environments.