Multi-Sensor Bayesian Data Assimilation for Large-Scale Drought Monitoring System (MuSe-BDA)

Informazioni relative al progetto

MuSe-BDA

ID dell’accordo di sovvenzione: 101068561

DOI

10.3030/101068561

Progetto chiuso

Data della firma CE 1 Agosto 2022

Data di avvio 8 Settembre 2022

Data di completamento 7 Agosto 2025

Finanziato da

• Marie Skłodowska-Curie Actions (MSCA)

Costo totale

Nessun dato

Contributo UE

€ 254 333,99

Coordinato da

AALBORG UNIVERSITET
Denmark

Periodic Reporting for period 1 - MuSe-BDA (Multi-Sensor Bayesian Data Assimilation for Large-Scale Drought Monitoring System (MuSe-BDA))

Periodo di rendicontazione: 2022-09-08 al 2025-08-07

Sintesi del contesto e degli obiettivi generali del progetto

Reliable assessment of droughts and freshwater availability requires accurate knowledge of how water is stored and exchanged between the land surface, soil, and groundwater. However, current global monitoring systems still face major challenges. Hydrological models simulate these processes continuously but often miss important spatial details and underestimate uncertainty. Satellite missions such as GRACE and GRACE-FO observe large-scale variations in total water storage, yet they cannot separate the contributions from individual components such as soil moisture or groundwater. Remote-sensing products, while increasingly available, differ in coverage, depth sensitivity, and noise characteristics. These inconsistencies make it difficult to generate high-resolution and physically consistent pictures of the water cycle needed for reliable drought assessment.

The MuSe-BDA project (Multi-Sensor Bayesian Data Assimilation for Large-Scale Drought Monitoring Systems) was established to overcome these limitations. It developed an innovative Bayesian Data Assimilation (BDA) framework that merges multi-sensor Earth-observation data with a hydrological model to improve the estimation of terrestrial water storage and its components. The system integrates satellite gravimetry (GRACE/GRACE-FO), soil-moisture observations, and climate-driven model simulations within a probabilistic optimization scheme that quantifies uncertainty and ensures physically realistic and continuous estimates of water-storage changes through time.

Although the approach is globally applicable, validation focused on Europe and the United States, where dense observational networks exist for independent evaluation. These regions were used to calibrate and test the method before extending it globally. MuSe-BDA ultimately produced global drought maps and indices that reveal how drought intensity, extent, and recovery vary across continents. By advancing global hydrological monitoring capabilities, the project contributes to international climate-resilience efforts and supports European leadership in digital Earth-system modeling and sustainable water management.

Lavoro eseguito dall’inizio del progetto fino alla fine del periodo coperto dalla relazione e principali risultati finora ottenuti

The MuSe-BDA project developed a new computational framework to merge multiple Earth-observation datasets with hydrological models for improved monitoring of terrestrial water storage and drought. The work was organized into four main components. First, satellite and climate datasets—including GRACE/GRACE-FO water-storage changes, soil-moisture observations, and ERA5 meteorological inputs—were processed and harmonized to quantify their uncertainties. Second, the W3RA hydrological model was enhanced to better represent water-storage processes and to operate at a finer spatial resolution of about 10 km.

A central achievement of MuSe-BDA was the development of a consistent strategy to bridge the differences in spatial scale and resolution between satellite observations and model simulations. This innovation made it possible to integrate information from both sources more effectively, providing refined water-storage estimates while distinguishing between surface, soil, and groundwater compartments.

The core of the project, the Constrained Bayesian (ConBay) framework, merges these diverse datasets within a single probabilistic system. ConBay manages the uncertainties of each data source and maintains realistic relationships between the different components of the water cycle.

The framework was applied to analyze major droughts in Europe and the United States, such as the 2018–2020 events in Europe and the 2014–2016 and 2021 California droughts. Validation against in-situ observations confirmed that MuSe-BDA improves the accuracy of groundwater and soil-moisture estimates and enhances the detection of drought intensity and recovery. The system was then extended to a global scale, producing drought indicators such as the Standardized Soil-Moisture Index (SSMI) and Standardized Storage Index (SSI), which better capture the timing and severity of droughts worldwide.

Progressi oltre lo stato dell’arte e potenziale impatto previsto (incluso l’impatto socioeconomico e le implicazioni sociali più ampie del progetto fino ad ora)

Prior to MuSe-BDA, many data-assimilation frameworks had been developed to combine satellite and model data for hydrological applications. However, most of these systems were tailored to specific models or sensors and faced limitations in scalability and uncertainty treatment. Two main challenges remained:
(1) the lack of a flexible assimilation framework that can integrate multiple satellite datasets with various hydrological models, and
(2) the absence of computationally efficient methods that balance model complexity with realistic representations of uncertainty in both models and observations.

The MuSe-BDA project addressed these challenges by developing a model-independent Bayesian data-model integration framework capable of merging information from different Earth-observation sources—such as GRACE/GRACE-FO, GNSS, and soil-moisture missions—with diverse hydrological models. The system is globally scalable and can operate at any spatial or temporal resolution, while explicitly accounting for the spectral, spatial, and temporal mismatches between datasets.

This flexible design goes beyond previous assimilation schemes by enabling consistent estimation of terrestrial water-storage components and their uncertainties across regions and scales. MuSe-BDA thus provides a foundation for next-generation drought-monitoring and early-warning systems that can be adapted to local, regional, or global applications.

Future work will focus on transferring the framework into operational drought-warning and water-management platforms, integrating socio-economic indicators, and strengthening collaboration with international initiatives on climate-resilient water resources and digital-Earth infrastructures.