Developing Precipitation within Digital Twin Earth Hydrology – Leveraging the individual strengths of multiple products
Paolo Filippucci, Luca Ciabatta, Luca Brocca, Christian Massari
Several Digital Twins of the Earth are being developed in recent years, driven by the growing
interest in integrating the latest advancements in Earth Observation (EO), modeling, artificial
intelligence, and computational power to make them accessible to the scientific community and
interested parties. Such platforms are highly valuable in supporting sustainability efforts and
combating climate change, enabling the visualization, analysis, and prediction of the natural system
- including human activities and their influence.
The European Space Agency (ESA) also shown interest in this framework by launching the DTE
Hydrology project, which focuses on analyzing the water cycle and its key components using the
latest satellite observations and models. A critical aspect of the project involves the development of
high-resolution (at least 1 km, daily) datasets for essential water cycle variables, aimed at
replicating hydrological behavior and understanding interactions with human systems. Among these
variables, precipitation plays a central role due to its impact on agriculture, economic stability,
water resource planning, and disaster risk reduction. Globally, ground-based observation networks
for precipitation monitoring are declining due to political and economic constraints, forcing many
regions to rely on less accurate precipitation datasets, affected by the decreasing gauge density. In
this context, satellite-derived precipitation estimates have the potential to improve precipitation
estimates by filling both spatial and temporal data gaps. However, numerous precipitation products
have emerged over the years, each with their own strengths and limitations, making it challenging
for users to determine the most suitable product for their study area.
To overcome this issue and capitalize on the individual strengths of each datasets, the DTE-
Hydrology initiative has developed a combined precipitation product that merges multiple sources,
including satellite-based and reanalysis datasets, into a unified, enhanced product. Specifically,
precipitation estimates from IMERG-Late Run, SM2RAIN ASCAT (H SAF), and ERA5 Land are
downscaled at 1 km spatial resolution and subsequently merged using pixel-based weights derived
from the application of the Triple Collocation method. The final merged product was thoroughly
validated and compared against a wide range of datasets—both coarse-resolution sources such as H
SAF, IMERG-LR, ERA5, EOBS, PERSIANN, CHIRP, GSMAP, and fine-resolution datasets like
EMO, SAIH, COMEPHORE, and MCM—demonstrating its high reliability and strong
performance.
