The EO datasets used in the SEED-FD project to strengthen global flood and drought forecasting are now available for wider use. Covering precipitation, river discharge and soil moisture, these resources can support researchers, forecasting experts and decision-makers working to improve hydrological modelling, early warning and water-related risk management. The datasets were generated for SEED-FD by scientists at CNR-IRPI.
Why EO datasets matter for better forecasts
Flood and drought forecasts depend on reliable observations. In many regions, however, ground-based measurements are limited, especially in the Global South. This can make it harder to anticipate extreme events and prepare for their impacts.
The SEED-FD Horizon Europe project addresses this challenge by using Earth observation data from satellites, which provide repeated measurements across large areas and over long time periods, including in remote or hard-to-access regions with limited monitoring infrastructure. Within the project, these EO datasets complement local observations and help improve flood and drought forecasting by supporting more realistic initial conditions, real-time forecast adjustments and improved error correction.
Precipitation EO dataset: merged rainfall information
Precipitation is a key driver of floods and droughts, and hydrological models use rainfall information to estimate how much water enters a basin and how the hydrological system may respond.
The SEED-FD precipitation EO dataset combines different sources of rainfall information, as individual precipitation products can perform differently depending on the region, terrain, data availability and weather situation.
Key dataset characteristics:
- Coverage in SEED-FD: Paraná, Bhima, Danube, Juba and Niger river basins
- Spatial resolution: 5 km (compatible with the CEMS-GloFAS grid)
- Temporal resolution: daily
- Temporal coverage: 2007–2024
- Public access: Zenodo
The dataset integrates IMERG Late Run, SM2RAIN-ASCAT and ERA5-Land. IMERG Late Run provides satellite-based rainfall estimates from microwave and infrared observations. SM2RAIN-ASCAT adds a complementary perspective by deriving rainfall from changes in satellite-observed soil moisture. ERA5-Land contributes consistent reanalysis data, helping to strengthen the product where satellite-based estimates may be less accurate, for example in mountainous areas. All three of these input datasets are available worldwide and with short latency.
Two product versions are available. The merged IMERG-SM2A product is based solely on satellite estimates, with ERA5-Land used as a reference to guide the merging process. The merged ERA5-IMERG-SM2A product includes ERA5-Land precipitation directly.
By combining complementary data streams, the precipitation EO dataset provides a more robust product that is better suited for hydrological applications than any of the individual datasets alone. While the current dataset covers SEED-FD’s five study basins, the approach is potentially extendable to wider regions worldwide. Beyond SEED-FD, the dataset can support model calibration, validation, evaluation and benchmarking, particularly in regions where local precipitation measurements are limited or inconsistent.
River discharge EO dataset: observing river flow variations
River discharge describes the amount of water flowing through a river. It is an essential variable for forecasting floods and hydrological droughts, as well as for related early warning systems. However, continuous discharge observations usually rely on river gauging stations, which are not available in all regions.
SEED-FD developed a satellite-derived dataset of near-infrared reflectance indices that act as proxies for river discharge variations. In simple terms, the method analyses satellite signals from river areas and periodically flooded zones. Changes in these signals can indicate changes in river flow.
Key dataset characteristics:
- Coverage: selected locations in the Danube, Paraná, Niger and Juba river basins
- Number of locations: 34 virtual stations
- Temporal coverage: 2000–2026
- Public access: Zenodo
The dataset provides river discharge proxy time series derived from MODIS Terra, MODIS Aqua and Sentinel-2 satellite data. While Sentinel-2 offers higher spatial resolution, MODIS provides a longer historical record and more frequent observations.
By integrating complementary satellite data sources, the river discharge EO dataset provides valuable information for hydrological applications. While the current dataset covers selected locations in SEED-FD study basins, the approach is potentially applicable worldwide.
For users beyond SEED-FD, the dataset can support hydrological modelling and forecasting by complementing local measurements, providing additional information on river flow variations and enabling the testing of near real-time update approaches.
Soil moisture EO dataset: a long-term view of drought conditions
Soil moisture is a key indicator for drought development. It shows how much water is available in the upper soil layer and can provide early information, for example on agricultural drought conditions.
For SEED-FD, an ASCAT-derived soil moisture product from the Copernicus Land Monitoring Service was processed for use in project workflows. ASCAT is a satellite sensor onboard MetOp satellites. The product represents surface-near soil moisture as a percentage of saturation.
Key dataset characteristics:
- Coverage: global land areas
- Spatial resolution: 12.5 km in the original Copernicus product; regridded to 0.05° / 5 km for SEED-FD
- Temporal resolution: about one passage per day at mid-latitudes in the original Copernicus product; interpolated to 12-hourly time steps for SEED-FD
- Temporal coverage: 2007–present in the original Copernicus product; 2007–2024 in the SEED-FD processed version
- Public Copernicus product access: Copernicus Land Monitoring Service
For use in SEED-FD, the Copernicus soil moisture product was regridded and temporally interpolated to support GloFAS-compatible hydrological modelling and forecasting workflows.
The wider value of this soil moisture EO dataset lies in its long, consistent and global data record. Based on a uniform satellite sensor record from 2007 to the present, it is particularly useful for identifying soil moisture anomalies that can support drought assessment and forecasting applications.
EO dataset benefits for hydrological applications
Together, SEED-FD’s Earth observation datasets can support large-scale hydrological applications, including flood and drought forecasting, early warning and water resource management. By providing additional satellite-derived information, they can help users develop, test and refine modelling and forecasting workflows.
Potential users include civil protection agencies, water authorities, meteorological offices, agricultural organisations, humanitarian actors, research institutions and consultants working in natural hazard risk management. For these users, the datasets offer practical advantages such as wide geographical coverage, open access and the opportunity to explore how EO data streams can be integrated into modelling and forecasting workflows.
Explore SEED-FD’s EO datasets and share your feedback
Making SEED-FD’s Earth observation datasets publicly accessible is part of the project’s wider effort to strengthen global flood and drought forecasting.
Interested experts are invited to explore the datasets for their own hydrological research, forecasting or operational work. SEED-FD welcomes feedback on where the datasets have proven useful and which applications they have supported. Feedback can be sent to the project coordinator at vp@seed-fd.eu.
To stay informed about the project’s developments, follow SEED-FD on LinkedIn, X and Facebook, or subscribe to the project newsletter.