Insights from the DANSk-LSM Project to Advance GRACE(-FO) and Multi-Sensor Data Assimilation for Hydrological Monitoring and Early Warning Systems

Abstract

The DANSk-LSM (Developing Efficient Multi-Sensor Data Assimilation Frameworks for Integrating Earth Observation Satellite Data into Land Surface Models) project, funded by Independent Research Fund Denmark (DFF), developed and demonstrated physically consistent and computationally efficient data assimilation frameworks for hydrological modelling. Between 2022 and 2026, the project integrated multi-source Earth observation data into land surface models to improve monitoring and forecasting of hydrological extremes.A major outcome was the assimilation of GRACE/GRACE-FO terrestrial water storage (TWS) observations together with satellite-derived surface soil moisture and recently the addition of surface water dynamics is being explored. Novel data assimilation settings were developed and evaluated across diverse river basins (e.g., Brahmaputra, Danube and Mississippi), improving the representation of water storage variations, fluxes, and drought and flood evolution. Scientific and implementation challenges were also identified.DANSk-LSM delivered open-access tools, including PyGLDA software, and high-resolution signle sensor and multi-sensor data assimilation tools. The project also fostered collaboration across geodesy, remote sensing, hydrology, and data science communities, laying the groundwork for next-generation Digital Twin Earth hydrology systems and future satellite gravity missions. This poster summarizes the main scientific advances and lessons learned, highlighting the role of gravity-derived water storage observations in (operational) hydrological monitoring and disaster preparedness.Keywords: GRACE and GRACE-FO, terrestrial water storage, multi-sensor data assimilation, land surface models, hydrological extremes, drought monitoring, flood forecasting, Earth Observation, Digital Twin Earth.