The water available to a hydropower plant results from a complex hydrological balance within the upstream basins. Multiple factors influence the time that water takes to reach the withdrawal point. More specifically, factors such as hydrography, soil characteristics, evapotranspiration, snowmelt, and anthropic structures (e.g., dams, pipelines and intakes), among others, need to be considered.

Hydropower companies seek to ensure that water abstraction is coherent with hydropower production planning. In fact, companies try to do this while making informed decisions on extreme weather events. However, climate change and its effects are becoming increasingly frequent. This results in water discontinuity and extreme weather events. Consequently, convenient predictors, such as the historical mean, are no longer adequate for accurate water availability forecasts. At the end, this affects production optimisation with inefficiencies. Nonetheless, weather forecast applied to hydropower needs is possible, and we are doing it in Di-Hydro.

Weather Forecast supporting Hydropower Optimisation

Seasonal meteorological forecasts are numerical weather predictions that are issued by meteorological institutions (such as ECMWF, DWD, NOAA). These forecasts provide, at monthly intervals, weather projections for the next six months. Nowadays, seasonal forecasts have become increasingly important. The reason is simple: positioned between short-term weather forecasts and climate projections, they offer valuable insights on the months to come. These projections also account for inflow forecasting.

Is it possible to translate seasonal meteorological forecasts into seasonal inflow forecasts? The answer is yes. In fact, inflow forecast is possible by using properly calibrated hydrological models that account for the basin saturation conditions, snowpack status and meteorological forecasts. As a result, this service provides water inflow or hydropower production forecasts for the upcoming six months with a monthly update. More specifically, it quantifies medium-term water availability with a probabilistic approach. By providing forecasts within confidence intervals to account for the uncertainty. This will improve plant management through optimised reservoir storage and early identification of potential critical situations like water scarcity.

Waterjade Forecast Model

“What are the expected inflows in the coming weeks? What is the achievable production in the coming months based on snow and meteorological forecasts?”. Waterjade addresses these questions through an innovative modelling approach: the Digital Twin of the catchment.

In short, this technology combines the strengths of physical modelling and Machine Learning to reconstruct the hydrological cycle at the basin scale. In other words, this Digital Twin, fed with seasonal meteorological forecasts, provides water inflow forecasts for the next six months. The application of this modelling and forecasting approach to the A2A use case is part of Waterjade’s contribution to Di-Hydro.

Stay tuned with Di-Hydro and follow our research through our LinkedIn and X pages. Also, do not miss the Resource page on our website: it feature the project public reports, materials and much more!

Authors: Fabio Pilotti (Hydrologist), Stefano Tasin (Hydrologist), Anna Paola Lonardi (Meteorologist), Francesco Bressi (Hydrologist), Matteo Dall’Amico (Hydrologist) – Waterjade