Work

Analysis of the relationship between the quality and the economic value of forecasts : application to the hydropower sector.

Year: 2016
Type: dissertation
Abstract: The knowledge of the economic value of hydrological forecasts can be very helpful for decision-makers to lead to a better management.Several studies have investigated the potential value of hydrological forecasts in flood forecasting management.To our knowledge, research on the link between economic value and quality of hydrological forecasts in a hydroelectric dam management context is rare.This study focuses on this link for French watersheds by using a heuristic dam management model.From this model, we estimate the economic income of diverse management rules that depend on hydrological forecasts of different quality.The research is based on a forecasting management modeling chain, where meteorological forecasts are used as input to the hydrological model of the Electricité de France company (MORDOR), which calculates streamflow forecasts up to 7 days ahead.These forecasts are subsequently used as input to a reservoir management model.The output of this model is a management rule indicating when turbines should be turned on to generate energy at the best prices.We also use the observed streamflow (from the banque HYDRO database) and meteorological data (from the SAFRAN database of Météo-France).The study period is defined between 2005 and 2008 and we analyse the results on eleven French watersheds.The reference operational forecast is obtained from the use of ECMWF EPS medium-range meteorological forecasts (50 members up to 7 days ahead) as input to the MORDOR hydrological model.To bring out the link between economic value and quality of a forecast, we create synthetic 7-days discharge ensemble forecasts of different qualities in terms of reliability, sharpness and accuracy.Each streamflow ensemble forecast is used as input to the management model.Then, each output from the manaement model is linked to an economic income.The results from the synthetic ensembles are compared with the results obtained from using observed streamflows, EPS-MORDOR streamflows and streamflow climatology as forecasts.We observe that the income from the management based on observed flows management is better than the income from EPS-MORDOR forecasts and from climatology.In our case study, the management model proved to be sensitive to the quality of each synthetic ensemble.These experiments allowed us to highlight the strengthens and limitations of the reservoir management model, as well as to propose additional steps for further research to improve the evaluation of hydrological forecasts in the context of their economic value for the hydropower sector.
Source: HAL
Language: English
Open Access status: green
OpenAlex ID: W4300580381
Cites: 18
Cited by:
Related to: 9