Intégration d'un filtre de Kalman dans le modèle hydrologique HBV pour la prévision des débits

Résumé

Un filtre de Kalman standard a été développé en vue de la mise à jour de certaines variables d'état de la fonction du sous-sol et de la sortie du modèle HBV. La version globale du modèle HBV, fonctionnant au pas de temps horaire, a été calée par essais et erreurs sur les bassins tunisiens des Oueds Barbara et Mellila. Le pas de temps journalier a été considéré pour évaluer la qualité de la reconstitution. Le modèle a été ensuite couplé avec le filtre en reformulant le modèle conceptuel en un système d'équations d'état dynamiques et en implémentant la procédure du filtre de Kalman. L'hydrogramme des débits horaires constitue la mesure introduite pour la correction des états. Malgré un choix suivant la littérature des erreurs sur les mesures et les états du système, ce filtre a permis de réduire nettement l'incertitude sur le débit engendré par le processus pluie—débit et la pluie d'entrée.     

Improving generalization of artificial neural networks in rainfall–runoff modelling / Amélioration de la généralisation de réseaux de neurones artificiels pour la modélisation pluie-débit

Abstract

Artificial neural networks (ANNs) are general-purpose techniques that can be used for nonlinear data-driven rainfall–runoff modelling. The key issue to construct a good model by means of ANNs is to understand their structural features and the problems related to their construction. Indeed, the quantity and quality of data, the type of noise and the mathematical properties of the algorithm for estimating the usual large number of parameters (weights) are crucial for the generalization performances of ANNs. However, it is well known that ANNs may suffer from poor generalization properties due to the high number of parameters and non-Gaussian data noise. Therefore, in the first part of this paper, the features and problems of ANNs are discussed. Eight Avoiding Overfitting Techniques are then presented, considering that these are methods for improving the generalization of ANNs. For this reason, they have been tested on two case studies—rainfall–runoff data from two drainage basins in the south of Italy—in order to gain insight into their properties and to investigate if there is one that absolutely gives the best performance.     

Explication de décrues anormales par modèles analogiques de réservoirs

Résumé

Des modèles analogiques de réservoirs ont été élaborés à partir de l'analyse de débits de décrue de deux sources karstiques. La première source est caractérisée par une linéarité entre le débit et le temps; la seconde l'est par une loi en puissance. Pour les deux sources, il n'y a pas linéarité entre le volume d'eau souterraine stocké et le débit de la source pour un même temps. Les modèles analogiques élaborés sont constitués d'une association verticale en série de réservoirs. Pour la première source, il s'agit de réservoirs à parois verticales uniquement. Pour la seconde, il s'agit de réservoirs à parois verticales et d'un réservoir dont la géométrie des parois suit une loi en puissance à dimension non entière. La plausibilité et la reproductibilité des modèles analogiques proposés ont été validées en fonction des données disponibles. Il a été démontré que le cas de réservoirs linéaires en milieu karstique est un cas très particulier. Il correspond à des réservoirs à géométrie hyperbolique ou à des réservoirs à parois verticales à filtre poreux.

Editeur Z.W. Kundzewicz

Citation Carlier, E., Zouhri, L. et Shahrour, I., 2012. Explication de décrues anormales par modèles analogiques de réservoirs. Hydrological Sciences Journal, 57 (5), 913–927.     

Modélisation pluie–débit journalière par des modèles conceptuels et ȁboîte noireȁ; test d’un modèle neuroflou / Daily rainfall–runoff modelling using conceptual and black box models; testing a neuro-fuzzy model

Abstract

Abstract Four rainfall–runoff models were applied on a daily time step and tested in the Cheffia basin, situated in the northeast of Algeria. The models belong to two categories: conceptual models–the GR3j model and the CREC model with eight parameters, and ?black box? models–the ARMAX model and a neuro-fuzzy model, which combines neural structure and fuzzy logic. The models were compared over two periods, one dry and the other wet. This comparison allowed a better model for the rainfall–runoff process to be proposed, on a daily time step, by combining the conceptual approach with a neuro-fuzzy system.     

Water balance of Lake Victoria: update to 2000 and climate change modelling to 2100 / Bilan hydrologique du Lac Victoria: mise à jour jusqu’en 2000 et modélisation des impacts du changement climatique jusqu’en 2100

Abstract

Abstract An annual water balance model of Lake Victoria is derived for the period 1925–2000. Regression techniques are used to derive annual inputs to the water balance, based on lake rainfall data, measured and derived inflows and estimated evaporation during the historical period. This approach acknowledges that runoff is a nonlinear function of lake rainfall. A longer inflow series is produced here which is representative of the whole inflow to the lake, rather than just from individual tributaries. The results show a good simulation of annual lake levels and outflows and capture the high lake level in 1997–1998. Climate change scenarios, from a recent global climate model experiment, are applied to the lake rainfall inflow series and evaporation data to estimate future water balances of the lake. The scenarios produce a potential fall in lake levels by the 2030s horizon, and a rise by the 2080s horizon. A discussion of the application of climate change data to this complex hydrological system is presented.     

Identification of rainfall–runoff models for operational applications / Identification de modèles pluie–débit pour des applications opérationnelles

Abstract

The problem of selecting appropriate objective functions for the identification of a lumped conceptual rainfall–runoff model is investigated, focusing on the value of the model in an operational setting. A probability-distributed soil moisture model is coupled with a linear parallel routing scheme, and conditioned on rainfall–runoff observations from three catchments in the southeast of England. Using an abstraction control problem, which requires accurate simulation of the intermediate flow range, it is shown that using the traditional RMSE fit criterion, produces operationally sub-optimal predictions. This is true in the identification period, when applied to a testing period, and to proxy catchment data. Using a second case study of the Leaf River in Mississippi (USA), where the focus changes to predicting flood peaks over a specified threshold, also suggests that the relevant flood threshold should govern the objective function choice. It is concluded that, due to limitations in the structure of the employed model, it would be counter-productive to try to achieve a good all-round representation of the rainfall–runoff processes, and that a more empirical approach to identification may be preferred for specific forecasting problems. This leaves us with the question of how far hydrological realism should be sacrificed in favour of purpose-driven objective functions.     

Application du réseau des neurones artificiels à la prévision des débits horaires: Cas du bassin versant de l’Eure,France

RÉSUMÉ

Les modèles pluie–débit sont fortement utilisés dans la gestion des risques hydrologiques et la prévision des crues. Dans cette étude, nous présentons un modèle pluie–débit pour la prévision des débits horaires basé sur la technique des réseaux de neurones artificiels (RNA). Ce modèle a été développé et appliqué sur le bassin versant de l’Eure au Nord-Ouest de la France afin de dépasser les problèmes dus à la non-linéarité de la relation pluie–débit et à l’imprécision des données collectées. La création de ce modèle a nécessité plusieurs étapes pendant lesquelles nous avons pu déterminer les paramètres du modèle permettant la compréhension de la complexité hydrologique et la production des informations nécessaires à la prévision. Elles ont abouti à un modèle de réseau de neurones artificiels capable d’effectuer, en quelques secondes, des prévisions des crues efficaces jusqu’à un horizon de prévision de 48 h. Ces résultats confirment que les modèles RNA peuvent jouer un rôle important dans le domaine de la prévision car capables de modéliser la non-linéarité des relations pluie–débit rencontrées sur certains bassins hydrologiques.

Editor Z.W. Kundzewicz Associate editor E. Gargouri     

Clark's and Espey's unit hydrographs vs the gamma unit hydrograph / Les hydrogrammes unitaires de Clark et de Espey vs l'hydrogramme unitaire de forme loi gamma

Abstract

A two-parameter gamma distribution for synthetic unit hydrographs (SUH) is compared with the Clark's and Espey's SUHs. A critical comparison of Clark's and gamma UHs, in terms of recession characteristics and time–area curve, is presented. It is observed that, in principle, a gamma UH can represent the hydrograph recession better than the Clark's UH does. Selection of a time–area curve is needed for obtaining the Clark's UH. The main problem in developing a SUH using the Clark's method is identified as the non-availability of a parametric form of the time–area curve. The time–area curve as represented in the hydrological model HEC-1, for the use in Clark's method, is found inadequate and unjustified. Gamma UHs obtained without optimization, for several examples, are found consistent with their physical meanings and better than the respective Clark's UH in reproducing runoff obtained with optimization. The parameters of Clark's UH (i.e. time of concentration and recession constant), as optimized through the HEC-1 program, are found inconsistent with their empirical origins and physical meanings; these lose their physical meaning and serve only as fitting parameters. This is due to the inappropriate time–area curve. A gamma UH has also the advantage of having fewer parameters than Clark's UH, which makes it more identifiable while still maintaining a connection with the physics of the problem. Espey's SUH for urban watersheds is transmuted to a gamma distribution using the empirical equations for the peak and time to peak of the UH. A numerical UH for a gauged catchment, generally obtained through linear programming or a least-squares approach, can be easily transmuted to a gamma UH and, hence, can be given a conceptual interpretation. Thus, these can also be used for developing a SUH.     

Modélisation du bilan hydrologique de la partie sud de la Mer d'Aral entre 1993 et 2001 / Hydrological balance modelling of the southern Aral Sea between 1993 and 2001

Abstract

Since the beginning of the 1990s, the Aral Sea is made of two separate entities: the Small Aral in the north and the Great Aral in the south. In this study, the water and salt balance of the Great Aral is analysed using newly available data: on the one hand, altimetric data of the sea level provided by the TOPEX/Poseidon satellite; and on the other, climate data provided by the ECMWF and NOAA within the framework of global climate re-analysis. These data indicate that the Great Aral received a mean groundwater inflow of about 4 km³ year^-1 between 1993 and 2001. Without this contribution, the Aral Sea would disappear even faster than is being observed today. Moreover, the temporal resolution of the data shows a systematic phase difference between the model prediction and satellite measurements. This phase shift is considered to be due to the formation of temporary lakes between the last station measuring the Amu-Darya discharge and its mouth in the Great Aral.     

Impact de la variabilité climatique et du barrage Nangbéto sur l'hydrologie du système Mono-Couffo (Afrique de l'Ouest)

Résumé

En Afrique tropicale, de nombreuses études sur la variabilité climatique montrent qu'une tendance à la sécheresse s'est manifestée à partir de la fin des années 1960. Cette sécheresse a affecté aussi l'hydrosystème Mono-Couffo, en Afrique de l'Ouest, avec une persistance remarquable qui s'est répercutée sur les écoulements. Depuis 1988, avec la mise en eau du barrage de Nangbéto, qui coïncide avec une légère reprise pluviométrique dans les années 1990, on assiste à une modification du régime hydrologique du Mono à Athiémé. Les méthodes statistiques appliquées aux données mensuelles de pluie et débit de la période 1961–2000 ont permis de constater que les déficits pluviométriques des décennies 1970 et 1980 ont engendré des déficits de l'écoulement. Mais, avec la légère reprise pluviométrique (2%) de 1988 à 2000, on assiste à un excédent d'écoulement de 37% sur le Mono à Athiémé à régime « artificiel » depuis 1988; et, par contre, pas d'excédent (0,1%) sur le Couffo à Lanta à régime « naturel ».

Editeur Z.W. Kundzewicz

Citation Amoussou, E., Camberlin, P. et Mahé, G., 2012. Impact de la variabilité climatique et du barrage Nangbéto sur l'hydrologie du système Mono-Couffo (Ouest Afrique). Hydrological Sciences Journal, 57 (4), 805–817.     

Application of linear and nonlinear techniques in river flow forecasting in the Kilombero River basin,Tanzania / Application de techniques linéaires et non-linéaires à la prévision des débits dans le bassin de la Rivière Kilombero,en Tanzanie

Abstract

Modelling of the rainfall–runoff transformation process and routing of river flows in the Kilombero River basin and its five sub-catchments within the Rufiji River basin in Tanzania was undertaken using three system (black-box) models—a simple linear model, a linear perturbation model and a linear varying gain factor model—in their linear transfer function forms. A lumped conceptual model—the soil moisture accounting and routing model—was also applied to the sub-catchments and the basin. The HEC-HMS model, which is a distributed model, was applied only to the entire Kilombero River basin. River discharge, rainfall and potential evaporation data were used as inputs to the appropriate models and it was observed that sometimes the system models performed better than complex hydrological models, especially in large catchments, illustrating the usefulness of using simple black-box models in datascarce situations.     

Exercice de scénarisation hydrologique en Afrique de l’Ouest—Bassin du Bani

Résumé

Les pays sub-sahariens basent principalement leur économie sur l’agriculture pluviale. Les projections démographiques à moyen ou long terme montrent que la pression démographique va s’accroître très fortement. A l’avenir, il faudra soutenir encore plus fortement le développement de stratégies agricoles de ces pays : cela nécessite une meilleure connaissance de leurs ressources en eau futures. Cette connaissance de la ressource en eau dans le futur passe par l’élaboration de scenarios climatiques et hydrologiques qui font intervenir différents acteurs, compétents dans des activités très variées, qu’ils soient scientifiques de divers domaines (climatologues, hydrologues, sociologues, …) mais aussi financiers, politiques, acteurs sociaux et décideurs. Ce processus d’élaboration de scenarios s’accompagne de nombreuses incertitudes, plus ou moins bien maitrisées, avec lesquelles doivent s’accommoder les gestionnaires d’ouvrages et planificateurs d’aménagements régionaux. Illustré par l’exemple du bassin du Bani, affluent du fleuve Niger, ce travail propose une réflexion sur la « scénarisation hydrologique » et la prédétermination des paramètres d’un modèle hydrologique dans un contexte global non stationnaire. Le choix de jeux de paramètres a de nombreux impacts sur les résultats de la scénarisation comme l’estimation de la ressource et la variabilité de cette ressource. Ces deux éléments sont fondamentaux pour qui doit conduire un processus de décision en réponse à une demande de stratégie d’aménagement ou à une demande de stratégie de développement. Nous proposons une méthodologie de détermination des jeux de paramètres se basant sur la construction de plusieurs jeux de paramètres issus de calages glissants, qui permet d’aller dans le sens de ce que de plus en plus de gestionnaires préconisent : ne pas essayer de prévoir le futur mais plutôt des futurs envisageables qui forceront des modèles d’impact, ce qui leur permettra de fournir des éléments de décision afin de s’adapter ou d’adapter des solutions d’aménagement aux conditions de ces futurs.

Editeur Z.W. Kundzewicz; Editeur associé C. Perrin

Citation Paturel, J.-E., 2014. Exercice de scénarisation hydrologique en Afrique de l’Ouest—Bassin du Bani. Hydrological Sciences Journal, 59 (6), 1135–1153. http://dx.doi.org/10.1080/02626667.2013.834340     

Estimation of extreme floods of the River Meuse using a stochastic weather generator and a rainfall–runoff model / Estimation des crues extrêmes de la Meuse à l'aide d'un générateur stochastique de variables météorologiques et d'un modèle pluie–débit

Abstract

A stochastic weather generator has been developed to simulate long daily sequences of areal rainfall and station temperature for the Belgian and French sub-basins of the River Meuse. The weather generator is based on the principle of nearest-neighbour resampling. In this method rainfall and temperature data are sampled simultaneously from multiple historical records with replacement such that the temporal and spatial correlations are well preserved. Particular emphasis is given to the use of a small number of long station records in the resampling algorithm. The distribution of the 10-day winter maxima of basin-average rainfall is quite well reproduced. The generated sequences were used as input for hydrological simulations with the semi-distributed HBV rainfall–runoff model. Though this model is capable of reproducing the flood peaks of December 1993 and January 1995, it tends to underestimate the less extreme daily peak discharges. This underestimation does not show up in the 10-day average discharges. The hydrological simulations with the generated daily rainfall and temperature data reproduce the distribution of the winter maxima of the 10-day average discharges well. Resampling based on long station records leads to lower rainfall and discharge extremes than resampling from the data over a shorter period for which areal rainfall was available.     

Multiple criteria rainfall–runoff model calibration using a parallel genetic algorithm in a cluster of computers / Calage multi-critères en modélisation pluie–débit par un algorithme génétique parallèle mis en œuvre par une grappe d'ordinateurs

Abstract

Genetic algorithms are among of the global optimization schemes that have gained popularity as a means to calibrate rainfall–runoff models. However, a conceptual rainfall–runoff model usually includes 10 or more parameters and these are interdependent, which makes the optimization procedure very time-consuming. This may result in the premature termination of the optimization process which will prejudice the quality of the results. Therefore, the speed of optimization procedure is crucial in order to improve the calibration quality and efficiency. A hybrid method that combines a parallel genetic algorithm with a fuzzy optimal model in a cluster of computers is proposed. The method uses the fuzzy optimal model to evaluate multiple alternatives with multiple criteria where chromosomes are the alternatives, whilst the criteria are flood performance measures. In order to easily distinguish the performance of different alternatives and to address the problem of non-uniqueness of optimum, two fuzzy ratios are defined. The new approach has been tested and compared with results obtained by using a two-stage calibration procedure. The current single procedure produces similar results, but is simpler and automatic. Comparison of results between the serial and parallel genetic algorithms showed that the current methodology can significantly reduce the overall optimization time and simultaneously improve the solution quality.     

ANALYSIS ON MOUTH BAR EVOLUTION OF THE YELLOW RIVER ESTUARY

ＡＮＡＬＹＳＩＳＯＮＭＯＵＴＨＢＡＲＥＶＯＬＵＴＩＯＮＯＦＴＨＥＹＥＬＬＯＷＲＩＶＥＲＥＳＴＵＡＲＹＪＩＺｕｗｅｎ；ＨＵＣｈｕｎｈｏｎｇ（Ｅｎｇｉｎｅｅｒ，ＤｅｐａｒｔｍｅｎｔｏｆＳｅｄｉｍｅｎｔａｔｉｏｎＥｎｇｉｎｅｅｒｉｎｇ，ＣｈｉｎａＩｎｓｔｉ...     

Comparison of grey and phase-space rainfall forecasting models using a fuzzy decision method / Comparaison grâce à une méthode de décision floue des modèles gris et d’espace des phases pour la prévision de pluie

Abstract

Abstract This work applies a fuzzy decision method to compare the performance of the grey model with that of the phase-space model, in forecasting rainfall one to three hours ahead. Four indices and two statistical tests are used to evaluate objectively the performance of the forecasting models. However, a trade-off must be made in choosing a suitable model because various indices may lead to different judgements. Therefore, a fuzzy decision model was applied to solve this problem and to make the optimum decision. The results of fuzzy decision making demonstrate that the grey model outperforms the phase-space model for forecasting one hour ahead, but the phase-space model performs better for forecasting two or three hours ahead.     

Impacts anthropiques sur les termes du bilan hydrologique d’un aquifère alluvial dans le Nordeste semi-aride,Brésil / Human impacts on components of hydrological balance in an alluvial aquifer in the semiarid Northeast,Brazil

Abstract

Abstract The exploitation of an alluvial aquifer (2 × 10⁶ m³) has been developed since 1998 in the Valley of Forquilha (Quixeramobim, State of Ceará). For this purpose, 165 wells were drilled along the 23 km of the valley that supplies 500 families and their farms. Monthly monitoring of piezometric and electrical conductivity (2000–2003) show seasonal variations in the water volume (35%) and in the mean value of the conductivity (800–1200 µS cm^-1). A conceptual hydrogeological model was developed. Analysis of the data and simulations highlight that the recharge of the aquifer is mainly due to infiltration from the river in the rainy season, which is of the order of 1% of the rain over the catchment area (195 km²). The abstractions increase the recharge between 30 and 60%. The model makes it possible to propose scenarios of sustainable exploitation of the water resource in the catchment. For the period between 1970 and 1988, other simulations show that it would have been possible to maintain irrigated cultures on 75 ha for 80% of the time. During the remaining 20%, the water level is too low, and farmers would have to reduce the irrigated area.     

A Modified Neural Network for Improving River Flow Prediction/Un Réseau de Neurones Modifié pour Améliorer la Prévision de L'Écoulement Fluvial

Abstract

Artificial neural network (ANN) models provide huge potential for simulating nonlinear behaviour of hydrological systems. However, the potential of ANN is yet to be fully exploited due to the problems associated with improving the model generalization performance. Generalization refers to the ability of a neural network to correctly process input data that have not been used for calibrating the neural network model. In the hydrological context, better generalization performance implies higher precision of forecasting. The primary objectives of this study are to explore new measures for improving the generalization performance of an ANN-based rainfall–runoff model, and to evaluate the applicability of the new measures. A modified neural network model (entitled goal programming (GP) neural network) for modelling the rainfall–runoff process has been developed, in which three enhancements are made as compared to the widely-used backpropagation (BP) network. The three enhancements are (a) explicit integration of hydrological prior knowledge into the neural network learning; (b) incorporation of a modified training objective function; and (c) reduction of network sensitivity to input errors. Seven watersheds across a range of climatic conditions and watershed areas in China were selected for examining the alternative networks. The results demonstrate that the GP consistently outperformed the BP both in the calibration and verification periods and three proposed measures yielded improvement of performance.     

Incorporating groundwater recharge and discharge functions into an existing monthly rainfall–runoff model/Incorporation de fonctions de recharge et de vidange superficielle de nappes au sein d’un modèle pluie-débit mensuel existant

Abstract

Abstract There is an urgent need for an integrated surface water and groundwater modelling tool that is suitable for southern African conditions and can be applied at various basin scales for broad strategic water resource planning purposes. The paper describes two new components (recharge and groundwater discharge) that have been added to an existing monthly time-step rainfall–runoff model that is widely used in the southern African subcontinent. The new components are relatively simple, consistent with the existing model formulation, but based on accepted groundwater flow principles and well understood groundwater parameters. The application of the revised model on two basins in southern Africa with quite different baseflow characteristics has demonstrated that the new components have a great deal of potential, even if the improvement is only to be able to simulate the groundwater baseflow component of total runoff more explicitly. More comprehensive testing and comparison of the results with existing groundwater and geohydrological data is required, while some extensions to the new components need to be considered to ensure that the model can be considered applicable to a wide range of basin and climate types.     

ANALYSIS ON PECULIARITIES OF "92.8"FLOOD EVENT IN THE LOWER YELLOW RIVER

lCOMPOSITIONOFTHE"92.8"FLOODThreerainstormsoccurredfrom7ththrough13,,,August,1992inShaanxiProvince,diStributingfromnorthtosouthinsequence.Therainfallareacoveredtheregionsofintensivesoilerosion,'wheretheaveragerateoferosionis10,000-15',000ton/kmZ'year.Fig.IshowsisohyetsofrainfallintensityinthecatchmentoftheMiddleYellowRiVerdepictingthedistributionoftherainstormsfrom7thto13,,,August,1992(thehydrologicalBureauYRCC,1992).ThecenterofthefirstrainstormwaslocatedattheYikezhaomengPrefec…