Closing the water balance using multi-source data fusion


Overall goal of the project is to provide improved estimates of water balance components of (sub)watersheds, including precipitation, evaporation, river discharge, and changes in water storage. This provides crucial information about where the water is going, and is important for managing available water resources, especially in water stressed regions.
The proposed approach brings together multi-source, multi-sensor data on all water balance components. This includes data from ground-based sensors (e.g. rain gauges, streamflow discharge gauges) and spaceborne sensors (e.g. satellite-based data on land surface evaporation, changes in water storage, etc).
To deal with data errors, the project develops new methodology based on probabilistic data fusion. The main idea is that, since data contain errors, the data values are not fixed but can be adjusted within a range that depends on magnitude of the data errors (data with larger errors can be adjusted more). Using statistical inference algorithms, these adjustments can be done automatically and simultaneously for all datasets and water balance components. This results in (i) improved estimates of all water balance components that together yield a closed water balance, and (ii) new insights into systematic and random errors of all datasets.
The visitor's travel grant will enable application of the methodology to data-poor and water-stressed regions, specifically hydrological basins in Iran.





Dr. ir. G.H.W. Schoups

Verbonden aan

Technische Universiteit Delft, Faculteit Civiele Techniek en Geowetenschappen, Afdeling Watermanagement


01/12/2019 tot 31/03/2020