On resorting to a meteorological post-processor to improve ensemble hydrological forecasts - INRAE - Institut national de recherche pour l’agriculture, l’alimentation et l’environnement
Communication Dans Un Congrès Année : 2019

On resorting to a meteorological post-processor to improve ensemble hydrological forecasts

Résumé

This presentation discusses the usefulness and effectiveness of meteorological post-processing in hydrological ensemble prediction systems (H-EPSs). Much work has been devoted to develop meteorological post-processors over the years to improve local biases and reliability flaws in weather ensemble forecasts before using them as input to hydrological forecasting models. At the same time, improvements have been made to the physics and the spatiotemporal resolution of weather prediction systems, reducing, at the source, the same biases that might impact the quality of hydrological forecasts. For hydrologists, it is thus relevant to know if they can directly use the outputs of weather forecasting systems or if they have to post-process them before their use in the hydrological model. In order to answer this question, it is first necessary to better understand the role played by each of the H-EPS components on the final quality of hydrological forecasts, including the meteorological post-processing component. Based on the Censored shifted gamma distribution (CSGD) technique, a state-of-the-art weather forecast post-processor, applied to the 50-member ECMWF weather forecasts, from 2008 to 2018 and over typical watersheds in Québec, Canada, we evaluated the gains in forecast quality that a meteorological post-processor brings to hydrological forecasts. The post-processed and the raw ensemble weather forecasts are used in a multi-model H-EPS (the HOOPLA system) that allows to account separately for the other main sources of uncertainty in hydrological ensemble forecasting. Through the use of EnKF data assimilation and of 20 lumped hydrological models, the influence of the CSGD on the quality of the streamflow forecasts is assessed. Our results contribute to relevant literature on the topic and bring additional insight into the role of meteorological post-processors as a compulsory component in the H-EPS forecasting chain.

Domaines

Hydrologie
Fichier non déposé

Dates et versions

hal-04573601 , version 1 (13-05-2024)

Identifiants

  • HAL Id : hal-04573601 , version 1

Citer

François Anctil, Emixi Valdez, Maria-Helena Ramos. On resorting to a meteorological post-processor to improve ensemble hydrological forecasts. AGU Fall Meeting 2019, Dec 2019, San Francisco, United States. ⟨hal-04573601⟩
6 Consultations
0 Téléchargements

Partager

More