AAU’s approach for assimilating current and future satellite gravity data into high resolution global hydrological models

Terrestrial Water Storage (TWS) data from satellite gravity missions, such as GRACE and GRACE-FO, have been used in past to enhance the representation of water storage in large-scale hydrological models through advanced sequential Data Assimilation (DA). However, these past experiments often relied on monthly gravity solutions, which lack the temporal resolution to capture rapid hydrological processes like floods. Additionally, the significant correlated errors in GRACE(-FO) solutions necessitate low-pass filtering, which can dampen signals and cause spatial leakage. Future gravity missions, including ESA’s NGGM (a single low-low satellite-to-satellite gravity mission on an inclined orbit) and MAGIC (a double low-low satellite-to-satellite gravity mission involving NASA/DLR’s GRACE-C and ESA’s NGGM), aim to provide gravity solutions with 5-day and monthly temporal resolutions and shorter latency. These missions are expected to offer improved spatial resolution and reduced uncertainty compared to GRACE-type missions.
In this study, we introduce the open-access, fine-scale, Python-based Global Land Data Assimilation system (PyGLDA) developed by the Geodesy Group at Aalborg University (AAU) for integrating global satellite gravity data into hydrological models. PyGLDA will be utilized to assess the added value of the NGGM and MAGIC missions in hydrological DA applications, particularly for representing water storage at sub-weekly, sub-monthly, seasonal, and multi-year time scales. This evaluation is crucial for understanding the role of GRACE-C, NGGM, and MAGIC products in future operational hydrological and early warning prediction systems. Our experiments will employ the daily 0.1° resolution W3RA water balance model as basis, driven by ERA5 climate inputs. Observations will include 5-day and monthly GRACE-C-like, NGGM-like, and MAGIC-like TWS simulations of the Technical University of Munich (TUM). The full error covariance matrix of the observed TWS, derived from closed-loop simulations of the ESA SING project, will be considered during the DA experiments, accounting for both instrumental and dealiasing errors. Validations will be conducted against synthetic hydrological water storage used to generate the hydrological truth.