La méthode SHYREG a été développée pour la connaissance régionale des quantiles de débits de crue (débit de pointe et lames d'eau maximales écoulées sur les durées de 1 h à 72 h) pour les périodes de retour de 2 à 100 ans suivant une approche spatialisée. Elle associe un simulateur de pluies horaires et une modélisation simple pluie-débit, mis en oeuvre à une résolution kilométrique. Les quantiles de débits se déduisent directement des distributions de fréquence empiriques des valeurs maximales extraites des très longues chroniques de débit simulées. On obtient alors une base de quantiles de crues que l'on peut agréger à l'échelle de n'importe quel bassin versant, moyennant une règle d'abattement avec la surface. La régionalisation de la méthode a été réalisée sur la France métropolitaine, à l'exclusion de la Corse, en exploitant les données hydrométriques de 1 359 stations de jaugeage et des caractéristiques hydro-climatiques et hydrogéologiques spatialisées permettant de décrire la variabilité du paramètre saisonnier du modèle. Au final, cette régionalisation permet la connaissance des quantiles de débits de crue en tout bassin versant de la France métropolitaine avec une bonne restitution des quantiles de débit de pointe et journalier, pour les périodes de retour comprises entre 2 et 10 ans : un critère de Nash minimum de 80 % est obtenu sur les quantiles de débit de pointe pseudo-spécifique et de débit journalier spécifique des bassins versants non utilisés pour la régionalisation.

Different approaches used in hydrological modelling are compared in terms of the way each one takes the rainfall data into account. We examine the errors associated with accounting for rainfall variability, whether in hydrological modelling (distributed vs lumped models) or in computing catchment rainfall, as well as the impact of each approach on the representativeness of the parameters it uses. The database consists of 1859 rainfall events, distributed on 500 basins, located in the southeast of France with areas ranging from 6.2 to 2851 km2. The study uses as reference the hydrographs computed by a distributed hydrological model from radar rainfall. This allows us to compare and to test the effects of various simplifications to the process when taking rainfall information (complete rain field vs sampled rainfall) and rainfall-runoff modelling (lumped vs distributed) into account. The results appear to show that, in general, the sampling effect can lead to errors in discharge at the outlet that are as great as, or even greater than, those one would get with a fully lumped approach. We found that small catchments are more sensitive to the uncertainties in catchment rainfall input generated by sampling rainfall data as seen through a raingauge network. Conversely, the larger catchments are more sensitive to uncertainties generated when the spatial variability of rainfall events is not taken into account. These uncertainties can be compensated for relatively easily by recalibrating the parameters of the hydrological model, although such recalibrations cause the parameter in question to completely lose physical meaning. Citation Arnaud, P., Lavabre, J., Fouchier, C., Diss, S. Javelle, P. (2011) Sensitivity of hydrological models to uncertainty of rainfall input. Hydrol. Sci. J. 56(3), 397-410.

Suite à l’événement exceptionnel de septembre 2002, le Conseil Général du Gard, propriétaire de 5 barrages écrêteurs de crue, a décidé de procéder à la révision des études hydrologiques de ces ouvrages. Les méthodes SHYPRE et SHYREG ont été mises en œuvre par HYDRIS Hydrologie, parallèlement à une approche classique utilisée par BRLi. Sur l’exemple de la retenue de Sénéchas, la communication s’attache à montrer l’impact important de la forme des hydrogrammes de crue sur les cotes atteintes dans la retenue. Afin de lever l’ambiguïté de détermination de la cote des Plus Hautes Eaux de période de retour 5 000 ans, les auteurs introduisent la notion de cotes de projet, notion complémentaire à celle de crues de projet. Les cotes de projet sont déterminées statistiquement par construction de la distribution de fréquence des cotes atteintes dans la retenue. Ces cotes sont obtenues par simulation du fonctionnement hydraulique du barrage,sur la base des scénarios de crue proposés par la méthode SHYPRE. Outre le fait de s’affranchir des hypothèses de construction d’un hydrogramme de projet, l’étude directe des cotes de projet présente différents avantages, notamment la détermination d’une distribution de fréquence des cotes qui peut être comparée avec les cotes observées, l’attribution d’une fréquence d’apparition à une cote donnée, la possibilité de tester différentes hypothèses de dimensionnement des ouvrages évacuateurs de crue...

This paper presents a synthesis of different level and flood references usually used for assessing the safety of hydraulic structures, including dams for flood control or dikes for flood protection. Discussions in the FrCOLD working group on justification for the stability of dams and research conducted by Cemagref in the field of risk analysis applied to hydraulic works show the limits of these definitions and open new perspectives. Proposals are made to clarify terminology. This leads to emphasize the determination of levels reached in the reservoir for different scenarios of flooding. Then it is possible to associate probabilities of occurrence to the maximum water level and to the level of danger of rupture, and parallel to propose minimum targets for these probabilities, depending on the types of hydraulic works and their class.

The hourly rainfall stochastic model SHYPRE generates long hourly rainfall series and enables the estimation of distribution quantiles. Two different uncertainty analyses are proposed, based on frequential and Bayesian methods, to quantify the effect of sampling distribution and parameter uncertainties on the quantile estimations. The results are compared with those of a regional generalized Pareto distribution (GPD) based on extreme value analysis, with a regionally fixed value of the shape parameter. The GPD and SHYPRE are shown to have similar uncertainties. The application of regional approaches is shown to reduce sampling sensitivity in estimations, especially when few data are available. The study is based on a 122-year daily rainfall series in Marseille, France.

The SHYPRE method is a flood risk estimation approach applicable on any point of a territory. It is based on the coupling of an hourly rainfall stochastic generator with a rainfall-runoff model. As a first step, carried out over the whole of France, regionalization of the hourly rainfall model parameters (obtained from 217 reference recording raingauges) was done on the basis of the available daily rainfall information. The information from 2812 raingauge stations was used to create maps of the rainfall model parameters at a scale of a one-square-kilometre pixel; data from 612 additional raingauges were used to validate this regionalization. Starting from these regionalized parameters, the rainfall model was used to obtain a geographic information system providing, at the square kilometre scale for the whole of France, the 1- to 72-hour maximum rainfall quantiles for 2- to 100-year return periods.

This study compares different methods used to estimate extreme rainfall quantiles in an area with a Mediterranean climate. The first method (M1)is basedon a stochastic model of hourly rainfall, which provides hourly rainfall series. The cumulative distribution functions (cdf) of extreme rainfall over various durations are deduced from the simulations. The second method (M2) is a regional approach based on the construction of a regional cdf of maximum annual daily rainfall. In the third method (M3), an exponential cdf is fitted to observed rainfall at a given gauging station. The three methods are compared with their parameters calibrated locally using the information observed at a given gauge, and using regionalised parameters. Comparison of M1 and M2 showed good agreement, with only relatively small differences between the rarest quantiles, less than 20% for the 1000-year return period. These differences can be explained by the regionalisation of the parameters. With both methods, a heavy tail of the daily maxima annual rainfall cdf deviates from the exponential cdf. Moreover, the regionalisation of M I and M2 is more robust than with M3, which is in much more influenced by sampling uncertainties.

The aim of this study was to provide a means of determining reference low flows (the mean annual discharge and the 5-year return period mean monthly flow) at any point in the hydrographic network in the southern half of France, using a continuous monthly rainfall-runoff modelling approach. The model adopted here has two parameters. To take into snow or ice influences in mountainous catchments, a monthly snowmelt modelling was added. The accuracy of the calibrated model was found to be excellent for assessing the mean annual discharge, and limited errors were observed in predicting low flows. A regionalisation procedure was performed for the two parameters. This methodology and the data base were built-in in a convivial software. The user has a general tool which synthesizes the whole of the hydrological information of the network of measurement and it is possible to very quickly treat a great number of hydrological problems related to the water resource, with its seasonal and interannual variability. From the simulated flows time series, it is possible to approach the dimensioning of dams, and the statistical treatments associated propose to deduce the hydrological variables of reference.

La majorite des outils mis en oeuvre pour la prevision des crues utilise en temps reel une information sur les pluies au sol et sur les debits. En absence d'information specifique sur les cours d'eau, les methodes classiques sont inoperantes. C'est pour combler ce manque que le Cemagref et Meteo-France ont associe leurs competences, afin de mettre au point un outil permettant de connaitre la pluviometrie et les debits des cours d'eau en temps reel sur l'ensemble du reseau hydrographique francais. Ainsi a ete developpee la methode AIGA qui produit toutes les heures une carte du risque pluvial a l'echelle du km 2 et du risque hydrologique, et ceci, meme sur les zones geographiques et les cours d'eau sans aucune instrumentation specifique.

Currently, new flood forecasting schemes are being developed. On the one hand, there are watercourses for which real-time information on ground-level rainfall and flow rates are available, and on the other hand, there are the watercourses for which real-time information on the flow rates is either not useable or unavailable. Cemagref and Meteo-France have combined their skills to fill this gap and develop a tool to determine the amount of rain and the streamflow rate in real time on the entire French hydrographical network. Thus AIGA has been developed, which produces a map of rain-related risks on a grid of 1 km(2) and of the hydrological risk, every hour. This is produced even over geographical areas and watercourses with no specific instrumentation.

Click to increase image sizeClick to decrease image sizeCemagref have developed the SHYPRE approach and its regionalized version SHYREG to assess the hydrological risk, anywhere on the hydrographical network.On the base of this method, a stochastic model of hourly rainfall was developed to reproduce the temporal characteristics of the observed rainfall, and to generate long rainfall time series, statistically equivalent to the observed chronicles.The modelling of the temporal characteristics of the rainfall leads to a hyper-exponential asymptote of the frequency distributions of the rains of various durations.Model regionalisation is based on daily rainfall characteristics. In metropolitan France, the regionalization leans on observations of practically 3000 precipitation gages.It leads in the fine spatial knowledge (1 km2 pixel) of the characteristics of the hourly rainfall model. So the model can be used and leads to a Geographical Information System of rainfall risk, which supplies an estimation of rains on the whole range of the frequencies, for 1 to 72-hours durations.Beyond the obvious operational interest, the analysis of the results obtained shows the pertinence of this approach and its robustness.Valeurs rares et extrêmes de précipitations et de débits

Cemagref have developed the SHYPRE approach and its regionalized version SHYPEG to assess the hydrological risk, anywhere on the hydrographical network. On the base of this method, a stochastic model of hourly rainfall was developed to reproduce the temporal characteristics of the observed rainfall, and to generate long rainfall time series, statistically equivalent to the observed chronicles. The, modelling of the temporal characteristics of the rainfall leads to a hyper-exponential asymptote of the frequency distributions of the rains of various durations. Model regionalisation is based on daily rainfall characteristics. In metropolitan France, the regionalization leans on observations of practically 3000 precipitation gages. It leads in the fine spatial knowledge (I km(2) pixel) of the characteristics of the hourly rainfall model, So the model can be used and leads to a Geographical Information System of rainfall risk, which supplies an estimation of rains on the whole range of the frequencies, for I to 72-hours durations. Beyond the obvious operational interest, the analysis of the results obtained shows the pertinence of this approach and its robustness.