Résumés
Résumé
La deuxième partie de cette synthèse bibliographique sur la genèse des débits montre comment les connaissances acquises sur le fonctionnement des petits bassins ruraux (cf. Partie 1) peuvent être utilisées pour les modéliser. Elle présente les différents types de modèles hydrologiques (empiriques globaux de type "boîte noire", conceptuels globaux ou semi-spatialisés, physiques spatialisés, physico-conceptuels semi-spatialisés) disponibles pour générer des chroniques événementielles ou continues, et déduit de l'analyse de leurs avantages et limites respectifs certaines recommandations pour leur choix et leur usage. Elle indique ensuite différents problèmes rencontrés dans toute modélisation, et quelques pistes possibles pour les résoudre: incorporation des flux couplés à l'eau dans les modèles hydrologiques, erreurs liées à la structure du modèle (limites et simplifications théoriques, approximations numériques, discrétisations temporelle et spatiale), problèmes métrologiques et méthodologiques limitant la disponibilité des données, hétérogénéités à toutes les échelles limitant l'adéquation des données pour paramétrer les modèles, calage du modèle limitant son aptitude à simuler des scénarios de changement. Elle souligne la nécessité d'une validation multicritère des modèles et d'une estimation de l'incertitude sur les simulations générée par ces diverses sources d'erreurs, ainsi que le besoin d'une meilleure interaction entre expérimentation de terrain et modélisation.
Mots-clés:
- Types de modèles hydrologiques,
- approximation,
- données,
- hétérogénéité,
- intégration spatiale,
- paramétrisation,
- calage,
- validation,
- incertitude
Abstract
The second part of this review on streamflow generation analyses how the knowledge available from field studies (see Part 1) has been used since the 1960s or could be used to improve catchment modelling. After a presentation of the main model types, the various problems encountered during the modelling process are discussed.
The large variety of hydrologic models available for event or continuous simulation can be reduced to a few main types according to the ways the functional, spatial and temporal aspects of the catchment behaviour are represented. Lumped "blackbox" models are useful for many engineering problems but can not be used in "extrapolation" and give no information on the internal catchment dynamics. Lumped conceptual models, which consider a catchment as a system of interconnected reservoirs and simulate the main global fluxes, use empirical lumped relationships and parameters that often have no great physical meaning and are not measurable. Semi-distributed conceptual models use the same reservoir description, but at the scale of "homogeneous" units derived from a space discretisation, which allows one to take catchment structure explicitly into account. Physically-based distributed models, which use theoretical equations and measurable parameters, provide a dynamic explanation of catchment behaviour but require too much information and are too complex to be easily used at the catchment scale. Physico-conceptual semi-distributed models try to overcome the limits of the previous types, while keeping their advantages, by simplifying the dynamic approach and discretization using new concepts.
Physically-based or conceptual models, which describe or explain the water cycle at the catchment scale, are very useful for research, but their use in practical applications comes up against several problems. It is still difficult to incorporate into catchment models the water-coupled fluxes (energy, sediments, solutes, biomass) because of the poorly-known complexity of their interactions. Even sophisticated models are based on many approximations of the reality: lack of suitable theory for some processes, simplification of the theories available, numerical approximation, space and time discretisation all generate simulation errors related to the chosen model structure. Data availability is limited by measurement problems (differences in measurement scale, lack of appropriate measurement techniques), and methodological problems (sampling and interpolation procedures, ...), even though remote sensing is expected to help solve some of them. Data suitability is limited by space and time heterogeneity at all scales, which reduces the representativity of any measurement and complicates the parameterization and upscaling needed. Model calibration (either manual, automatic, or stochastic), which leads to the numerical equifinality of both model parameterization and structure, limits the validity domain of the model, its transposability to other conditions and catchments, and its ability to simulate change scenarios. The effects of these limitations on model quality could be reduced by using multivariable and multiscale validation procedures and should be quantified using stochastic estimation of the simulation uncertainties associated with model and data uncertainties. In order to further progress in catchment modelling, as needed by a large range of environmental issues, field hydrologists and modelers should reinforce their co-operation, especially through interdisciplinary studies on long-term research catchments and carefully designed field experiments.
Keywords:
- Catchment model types,
- approximations,
- data,
- heterogeneity,
- upscaling,
- parameterization,
- calibration,
- validation,
- uncertainty