Advancing Large-Scale Hydrological Modeling for Climate Studies through Multi-Source Assimilation of ESA CCI Products

Large-scale hydrological models simulating river dynamics for climate studies face limitations from structural, forcing, and process uncertainties. Data assimilation, by integrating observations, mitigates these limitations. In this context, satellite remote sensing (radar altimetry, multispectral sensors) offers long-term spatially distributed hydrological data, compensating for declining in situ monitoring networks.This study builds on the ESA Climate Change Initiative (CCI) River Discharge precursor project, which develops long-term global satellite-derived discharge and water surface elevation (WSE) datasets for climate applications. A first phase of the project (during year 2024) evaluated the assimilation of either CCI discharge or WSE products into regional to global hydrological models (CTRIP and previously calibrated version of MGB). This evaluation highlighted the added value of discharge products in terms of information content, temporal sampling, and uncertainty characteristics (Sadki et al., 2024, HESS Discuss., https://doi.org/10.5194/hess-2024-328). Building on these results, the present work advances multi-source data assimilation strategies using improved and newly developed CCI products.Ensemble Kalman filter experiments, conducted with CTRIP-HyDAS and MGB-HYFAA assimilation systems over the Niger and Congo basins, assess the contribution of increased spatial and temporal sampling and the joint assimilation of discharge and WSE observations. Early results highlight the key role of increased temporal density in correcting model biases and internal variability, while revealing the complementary effects of combining spatially dense WSE observations with hydrologically consistent discharge information. Overall, this work provides new insights into robust multi-observation data assimilation strategies for large-scale hydrological modeling in a climate studies context.