La Houille Blanche
Number 1, Février 2008
|Page(s)||109 - 117|
|Section||Prix Henri Milon 2007 - Extrêmes hydrologiques|
|Published online||20 March 2008|
Détection et prise en compte d’éventuels impacts du changement climatique sur les extrêmes hydrologiques en France
Detecting and taking into account possible impacts of climate change on hydrological extremes
University of Newcastle - School of engineering Callaghan 2308, Australia
Auteur de correspondance : firstname.lastname@example.org
Climate change is widely considered as a reality by scientists. Nevertheless, impacts on hydrological extremes are more difficult to observe and to forecast. The aim of this thesis is to answer the following questions: How to detect changes in hydro-climatic series? What are the observed changes for extreme discharges in France? How to take into account possible changes in frequency analysis? These objectives refer to both local and regional scales. This paper describes the developments related to the third question. In a first step, the concept of return period is revisited in a non-stationary context. Frequency analysis methods are then updated in order to account for evolutions in time. This is achieved by modelling trends affecting the distribution parameters. Parameter estimation uses the Bayesian formalism, which is a convenient tool for quantifying the uncertainty related to the stationarity hypothesis. This approach can be generalized at the regional scale, by means of non-stationary regional models. Such models are more general than the model underlying the index flood method. However, results of such a regional analysis are affected by the spatial dependence existing between studied sites. Impacts of this dependence on quantile estimates are highlighted, and a first approach is proposed in order to explicitly model spatial dependence.
Mots clés : Extrêmes hydrologiques / stationnarité / analyse fréquentielle / analyse Bayesienne / dépendance spatiale.
© Société Hydrotechnique de France, 2008
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.