Large-scale spatio-temporal variability of the Congo Basin surface hydrologic components from space

Despite being the second-largest watershed and tropical forest worldwide, with significant impacts on the global water cycle and in regulating Earth’s climate, the Congo River Basin’s (CRB) hydroclimatology remains among the least studied worldwide due to the lack of situ observations. To better characterize CRB surface hydrology and the variability of its different components at large scale, we jointly used a trove of large records of in situ and satellite-derived observations, specifically, Surface Water Level (SWL) from radar altimetry (a total of ~2,300 virtual stations) and Surface Water Extent (SWE) from the Global Inundation Extent from Multi-Satellite (GIEMS) dataset. A good performance is found between SWL from multi- satellite missions and in situ water height of historical and contemporary observations at different locations. The root mean square error varies from 10 cm for Sentinel-3A to 75 cm for European Remote Sensing-2.  SWL annual amplitude exhibits large spatial variability across the basin, with Northern sub-basins varying more than 5 m while the central and the southern sub-basins vary in smaller proportions (1.5 to 4.5 m). The assessment of SWE also agreed relatively well over a ~25-year period with in situ discharge from sub-basin to basin scale. At the basin scale, SWE shows that cuvette centrale is flooded at its maximum in October/November. The northern part of the basin reaches its maximum in September/October, and the southern eastern one in January/February. Furthermore, SWL and SWE help capture the water travel time across the basin that varies from 0 to 3 months and the regional relative contribution to the flow at Brazzaville station characterized by a bimodal hydrological regime. Northern sub-basins and the cuvette centrale contribute much to the large peak in December-January while the southern sub-basins contribute to both peaks. We further combine these two datasets to estimate the variability of Surface Water Storage (SWS) in rivers, lakes, floodplains, and wetlands across the entire basin over the period 1992–2015. The CRB SWS shows an annual amplitude varying between ~74 km3 and ~112 km3. Moreover, the combination of SWS and the annual variations of GRACE/GRACE-FO-derived terrestrial water storage permits us to estimate the long-term variation of sub-surface water storage. The use of these new long-term satellite-derived observations are an invaluable source of information for hydrological modeling and will allow to properly characterize and reproduce the hydro-climate variability of the CRB, and a better representation of local and regional hydrological processes. These results ensure therefore an improved monitoring of CRB hydrological variables from space, and open new perspectives towards a better evaluation of the impact of climate variability on water availability in the region.