Find a copy online
Links to this item
Find a copy in the library
Finding libraries that hold this item...
Details
Genre/Form: | Thèses et écrits académiques |
---|---|
Material Type: | Document, Thesis/dissertation, Internet resource |
Document Type: | Internet Resource, Computer File |
All Authors / Contributors: |
Laura Rouhier; Pierre Ribstein; Federico Garavaglia; Agnès Ducharne; Isabelle Braud; Roger Moussa; András Bárdossy; Isabella Zin; Matthieu Le Lay; Sorbonne université (Paris / 2018-....).; École doctorale Géosciences, ressources naturelles et environnement (Paris).; Milieux environnementaux, transferts et interactions dans les hydrosystèmes et les sols (Paris). |
OCLC Number: | 1140399058 |
Notes: | Titre provenant de l'écran-titre. |
Description: | 1 online resource |
Responsibility: | Laura Rouhier ; sous la direction de Pierre Ribstein et de Federico Garavaglia. |
Abstract:
To provide reliable simulations, hydrological models usually require the calibration of their parameters over streamflow data. However, the latter are limited and most of the catchments remained ungauged. Consequently, alternative methods termed 'regionalization' are needed to estimate model parameters. The thesis proposes to combine the three classical methods in order to regionalize the parameters of a distributed model over two large French catchments: the Loire catchment at Gien and the Durance catchment at Cadarache. On the basis of the three regionalization methods, the degree of spatialization is adapted to the different model parameters according to their characteristics and their hydrological role. In fine, the proposed multi-method and multi-pattern approach (i) significantly reduces the number of degrees of freedom, (ii) improves the representation of the catchment physical variability, and (iii) significantly improves the performance of the simulations. In the ungauged context, the parameter spatialization allows an improvement of about 10%, and in particular, the multi-method and multi-pattern povides an improvement of about 7% compared to a single regionalization method. Despite these improvements, the impact of the climatic input spatialization remains 6 times greater than th parameter spatialization.
Reviews

