Skip to main content

Informing hydropower energy generation in the Alps with ML approaches

A hands-on Case Study targeting the Energy module of SnowPower, developed by Amigo srl within the I-NERGY 1st Open Call.


Developed by
Business Category
Earth Observation Energy
Technical Category
Machine learning

Knowledge about total water availability, i.e. both the rain and the water contained in the snow cover (Snow Water Equivalent, SWE) is key to optimally control hydropower plants. Nonetheless, energy generation from hydropower plants (especially in the case of run-of-river plants) depends on the natural fluctuations of water availability. Information about such fluctuation are often available with little to no notice, hindering the implementation of informed operational choices at plant level.


To help hydropower plants to take control over the natural fluctuations of the hydrological cycle, in-situ measurements in combination with satellite and meteorological models could improve the estimation of water availability from its storage in snowpacks to water runoff during the snowmelt season. 

SnowPower by Amigo srl is the first commercial climate service of its kind that combines a machine learning approach to assess snow water equivalent with climate data to estimate and predict the energy generated by hydropower plants up to six months in advance, currently targeting the Alpine area (and in particular including Alpine territories of France, Switzerland, Italy, and Austria). SnowPower is one of the projects of the 1st Open Call of I-NERGY.


In this Case Study (here), an example of the use of such satellite and model-derived data in SnowPower is provided, yielding a prediction of hydropower energy generation in the Italian river watersheds in the Alps.
This hands-on Case Study closely mimicks the Energy module of SnowPower, leveraging runoff data from SnowPower's Runoff module and other selected climate variables to produce predictions of energy generation in target areas.