Skip to Main content Skip to Navigation

Analyse du lien entre la qualité des prévisions hydrologiques et leur valeur économique pour le secteur hydroélectrique

Abstract : The quality of hydrological forecasts is a widely-studied field. It has been shown that probabilistic or ensemble forecasts (i.e. with several scenarios) are often better in terms of quality and decision support than deterministic (single-valued) forecasts. Probabilistic forecasts are therefore increasingly used in operational forecasting. However, the contribution of these forecasts in terms of economic value for the users remains a subject that has so far received little attention. In the literature, the evaluation of a forecast is mainly based on its quality (comparison with an observation). However, a forecast can also be assessed with regard to its contribution (or value) to the decision-making process. In the case of reservoir operation for hydroelectric generation, studies on the value of forecasts are also less frequent, particularly when it comes to the short-term range (forecasts ranging from a few days to a few weeks).The work carried out in this thesis aims to examine the link between the quality and the economic value of a forecast for the hydropower sector. In other words, we want to know in which cases a better forecast impacts the management revenues. We investigate the value of hydrological forecasts at several forecasts ranges: short (up to seven days) and medium (up to seven months). Finally, we investigate whether short-term reservoir management can be improved when coupled with medium-term management. The last research axis of this thesis thus consists in setting up a coupled forecast-reservoir management system, where medium-term management information is used to inform short-term reservoir management.The results of this thesis contribute: (1) to shed light on the links between the quality and the value of hydrometeorological forecasts and (2) to the modelling of hydropower reservoirs for optimal management. The modelling tools established in this study allow complex hydroelectric systems to be represented in a simplified way. With their help, we have highlighted the existence of a link between the quality of hydrological forecasts and their economic value. The economic value of short-term hydrological forecasts depends on their quality: the best forecast system in terms of forecast quality corresponds to the forecast system with the best management revenue and conversely. However, this relationship also depends on how the forecast information is taken into account in the reservoir management model, and on the size of the reservoir with regard to the average inflow volumes. In the medium-term management context, the link between the quality of medium-range forecasts and their economic value may also exist, but it is less obvious. Finally, the revenue obtained from the short-term management of the reservoir can be improved by more than 10 % by taking into account long-term management information.
Complete list of metadatas
Contributor : Abes Star :  Contact
Submitted on : Monday, November 30, 2020 - 12:40:18 PM
Last modification on : Tuesday, December 1, 2020 - 3:45:23 AM


Version validated by the jury (STAR)


  • HAL Id : tel-03031033, version 1


Manon Cassagnole. Analyse du lien entre la qualité des prévisions hydrologiques et leur valeur économique pour le secteur hydroélectrique. Environnement et Société. AgroParisTech, 2020. Français. ⟨NNT : 2020AGPT0001⟩. ⟨tel-03031033⟩



Record views


Files downloads