Coming from the Swedish Meteorological and Hydrological Institute (SMHI), Ilias Pechlivanidis explains the elements of the Umeälven river case study and how it is improving forecasting and predictions in Swedish river systems.

‘By having added value in forecasts, we expect to assess the added economical value. That goes back to the objective of the stakeholders, which is to improve their energy production.’


Hydropower in Northern Sweden requires large reservoirs to store most snow’s melted water for next winters production. Short-term weather and runoff forecasts are used to plan the daily production and reservoir management, while seasonal forecasts are used weekly to estimate the remaining runoff volume until the end of summer and keep the long-term production plans updated.

It is important to have the reservoirs filled at the end of Summer, yet, as much water as possible should be used for power production during Spring. Unnecessary release of water that cannot be used for production is recognised as ‘spill’ and is considered as a loss of potential production. Spill may happen when the remaining volumes of Spring flood were underestimated and reservoirs filled up too early or when there was unexpected heavy rainfall.

The case study will investigate to what extent improved hydro-meteorological forecasts from the IMPREX climate services could have reduced the loss of water during of a number of past spill events. The value of the improved forecasts will be estimated by evaluating the gain in potential production versus the historical electricity market. The analysis will be made in collaboration with Sweden’s largest hydropower producer, Vattenfall AB and with Vattenregleringsföretagen AB. The relative importance of the initial snow volume in the forecast models versus the uncertainty in the seasonal forecasts will also be evaluated, to estimate the potential value of the expensive snow surveys conducted by the hydropower companies.

Read more


  • Estimate of the economic value of hydro-meteorological forecasts for hydropower production in general.
  • Evaluation of improved forecast skill in IMPREX climate services compared to forecasts and models used today.
  • Evaluation of methods to estimate the potential to improve seasonal runoff forecasts by snow data assimilation with regard to hydro-meteorological regime and meteorological forecast uncertainty for the particular site.