Hydrometeorological analysis of floods in the watershed of Mono in West Africa with a rainfall-runoff conceptual model

The recurrence of extreme anomalies of precipitations involving of the floods or the droughts is a normal component of the natural variability of the climate. The harmful effects of the floods these last have strong incidences socio-economic and ecological and involved human losses of life and property damages. The vulnerability with these natural risks is high in West Africa and basin of Mono in particular, where the populations tend to occupy the most exposed zones more and more. The objective is to model the distribution of the risings entering to the dam Nangbeto thanks to hydrological model GR4J with the step of daily time to assess the risks of flood in the low valley of the river.

The data used cover the period 1961 to 2010. They is the daily outputs (m3.s-1), the evapotranspiration (mm/jour, calculated according to method FAO) and average precipitations on the basin (mm/day). The various methods of space interpolation of the rains were tested (linear, cubic, Thiessen, simple average). Model GR4J was useful with modeling of the risings upstream and the stopping of 1988 to 2010 to appreciate the effects of the dam.

The results show a marked variation in rainfall by increasing annual maximum daily followed by an increase in ETP. The rates recorded in output Nangbéto dam are equivalent to those recorded input, especially during floods, showing the weak regulatory role of the dam and the very limited capacity of the reservoir. The annual maximum incoming flow at the dam can be modeled by an extreme value
distribution (GEV ) Fréchet-type, while those measured at the outlet of the dam after a GEV law Weibull (bounded). The dominant runoff generation mechanism in the basin is the gradual rising waters associated with increased saturation of the soil, rather than isolated episodes of rainfall. The analysis of flood flows between 1988 and 2010 showed no increase in flood magnitude during this period. Finally the effectiveness of the model to reproduce GR4J flood flows has been tested showing that the observed floods are simulated during calibration. However, validation tests over several periods with different climatic characteristics show a degradation of performance criteria and the model as one moves away from the calibration period, showing the instability of the model parameters in the time. Under these conditions, the use of such a model to assess future risks future flooding is risky and it is necessary to consider alternative approaches.

The autor

Enseignant-chercheur en Géographie physique appliquée à l’Université de Parakou (Bénin), titulaire d’une thèse de doctorat unique de l’Université de Bourgogne en 2010 axée sur « Variabilité pluviométrique et dynamique hydro-sédimentaire dans l’hydrosystème Mono-Couffo (Bénin/Togo) ». Membre actif et chercheur au Laboratoire Pierre Pagney : Climat, Eau, Ecosystèmes et Développement (LACEEDE), Université d’Abomey-Calavi. Chargé des cours en : ‘’hydrologie’’, ‘’statistique’’ et ‘’modèles de recherche en sciences environnementales’’. Chargé également de cours en : ‘’Hydroclimatologie’’, ‘’Hydrosphère, géosphère et biosphère’’, ‘’Risques et catastrophes naturels et humains’’, ‘’Impacts des changements climatiques sur la dynamique des eaux de surface’’, ‘’Hydrosystème et bassin versant’’ et ‘’Gestion Intégrée des Ressources en Eau’’. En outre, membre du comité de rédaction de la revue Climat et Développement de l’Université d’Abomey-Calavi.

The text

Ce texte a été rédigé dans le cadre d’une bourse Fernand Braudel IFER, du 30 avril au 30 septembre 2012 puis du 15 février au 15 juin 2013, au sein du laboratoire HydroSciences Montpellier (HSM), Université Montpellier2.