Méthode SCHADEX de prédétermination des crues extrêmes : Méthodologie, applications, études de sensibilité

Abstract : Since 2007, EDF (Électricité de France) design floods of dam spillways are computed using a probabilistic method named SCHADEX (Climatic-hydrological simulation of extreme foods). This method aims at estimating extreme flood quantiles by the combination of a weather patterns based rainfall probabilistic model and a conceptual rainfall-runoff model. The aim of this PhD thesis is to “validate” SCHADEX method by assessing its potential and its limits with a sensitivity analysis of its hypothesis and with a comparison to the widest possible range of data (various regions and climates, different watershed sizes). In the first part of this thesis we describe the stochastic generation process of rainfall events. We introduce a rainfall probabilistic model based on weather pattern classification, called MEWP (Multi-Exponential Weather Pattern) distribution. To validate the MEWP model, we compare it to the standard probabilistic models of extreme values theory. Basing on a wide rainfall dataset (478 raingauges located in France, Switzerland and Spain) and on new specific criteria, we evaluate the suitability of MEWP model in terms of reliability and robustness. In the second part of this work, the hydrological simulation process of volumes, called semi-continuous simulation process, is introduced. We highlight the efficiency and the originality of this process link to its ability to mix various hydrological data for extreme flood estimation while keeping parsimonious extrapolation hypothesis. Basing on a dataset of 32 watersheds, we realize a sensitivity analysis of this process to (i) simulation parameters (i.e. number of simulation, etc.), (ii) to rainfall probabilistic model and (iii) to rainfall-runoff model. This study allows us to fix some simulation parameters and especially to prioritize the simulation steps and options by their impact on final results. The third part is about the passage from extreme volumes quantiles to extreme peak flows quantiles that is done by a peak flows ratio. This ratio is computed on a hydrograms sample that is extract from flood time series by a specific peak over threshold sub-sampling developed in this thesis. Globally, we reformulate, justify and verify the basic assumptions of the method, with a special focus on rainfall risk, hydrological risk and semi-continuous simulation process. In order to make extreme flood estimations more reliable and robust same improvements and simplifications of the method are proposed. To widen the available period of the weather pattern classification from 1953-2005 to 1850-2003, we computed another weather pattern classification based on shade geopotential information. To take into account precipitation-temperature dependency, the simulation process is improved conditioning the generation of rainfall events by weather pattern sub-sampling. Certainly this thesis does not put the end to SCHADEX method developments but ground it on sound and documented basis. Such future investigations about variability of shape flood hydrograms, comparison of hydrological model in extrapolation domain, extreme flood risk estimation on ungauged watershed and in non-stationary context, are planned.
Document type :
Theses
Complete list of metadatas

https://tel.archives-ouvertes.fr/tel-00579507
Contributor : Abes Star <>
Submitted on : Thursday, March 24, 2011 - 10:04:20 AM
Last modification on : Monday, September 2, 2019 - 2:20:10 PM
Long-term archiving on : Saturday, June 25, 2011 - 2:34:13 AM

File

22131_GARAVAGLIA_2011_archivag...
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-00579507, version 1

Collections

Citation

Federico Garavaglia. Méthode SCHADEX de prédétermination des crues extrêmes : Méthodologie, applications, études de sensibilité. Sciences de la Terre. Université de Grenoble, 2011. Français. ⟨NNT : 2011GRENU012⟩. ⟨tel-00579507⟩

Share

Metrics

Record views

3727

Files downloads

4357