The DeFROST4Hydropower project has demonstrated that the combination of near-real-time geospatial data fused with AI and long-term hydrological models can produce highly accurate seasonal forecasts.
The project was led by Wegaw and integration partner Hydrique, with long-standing partner Alpiq, a Swiss power producer. Alpiq’s portfolio includes conventional and pumped-storage projects spread across continental Europe, with the highest concentration of its fleet residing in Switzerland, where some 60% of all energy consumed is generated by hydropower facilities.
With the support of the Swiss Federal Office of Energy (SFOE), seasonal forecasts were produced for 12 catchment areas, allowing for a wide range of hydrological regimes to be evaluated. After the calibration process, the data was validated based on Ultracam photogrammetry snow height (HS) data collected during four drone flights by the Institute for Snow and Avalanche Research (SLF).
The geospatial data consisted of snow cover extent (SCE), HS and snow water equivalent (SWE) and was fused with AI and then integrated into rainfall-runoff models. This is proving to be especially effective in accurately estimating water supply in spring and summer, according to a release, and resulting in improved input predictability and tangible ROI for long-term trading (from 60-, 90-, 120-, 150- to 180-day periods).
