Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Doctorat → Autriche → Event based rainfall-runoff modelling in semi-arid regions

Universität für Bodenkultur Wien (2010)

Event based rainfall-runoff modelling in semi-arid regions

Tulu, Mesay Daniel

Titre : Event based rainfall-runoff modelling in semi-arid regions

Auteur : Tulu, Mesay Daniel

Université de soutenance : Universität für Bodenkultur Wien

Grade : Doctoral Thesis 2010

Résumé
Serious water shortages and poor water management can cause periodic droughts in arid and semi-arid regions and have appeared to limit the socio-economic development and food security. In these areas, geo-hazards, such as, flash floods and landslides triggered by high rainfall intensity and high surface runoff velocity are also very common. For sustainable development and management of water resources and mitigation of flooding, reliable runoff estimation or modelling technique is crucial. Semi-arid areas are hydrologically unique and model adaptation needs great care ; however, hydrological principles, and estimation techniques developed for humid areas have been directly transferred to these areas without proper amendments. The main intension of this study was to fill this gap by giving some insight about primary hydrological processes and parameters, which need special treatment in hydrologic estimation procedures of semi-arid regions. Since modelling has become a standard task in hydrological assessment practices, this study mainly concentrated on rainfall-runoff modelling. The actual event based rainfall-runoff relation assessment was investigated by recording rainfall and river flow data over a period of two years with high time resolution in a representative river watershed called Ilala located in the northeastern semi-arid region of Ethiopia. The investigation showed that rather channel processes dictate the runoff characteristics than the upland processes. It was also found out that rainfall-runoff relation in the region is highly event dependent. The watershed response was more of a function of the state of the antecedent water in the watershed. Therefore, any rainfall-runoff modelling attempt has to be set for specific condition. These findings were transferred to the US Army Corps of Engineers (US-ACE), Hydrologic Engineering Centre Hydrologic Modelling System (HEC-HMS), in which different models have been compared and evaluated statistically, using sensitivity of the models parameters, and finally validated applying a separate dataset. The dominance of specific hydrologic parameters was evaluated based on the sensitivity of the model output for the variation of the parameters. The watershed response was found highly sensitive to the channel infiltration parameters. Since the channel infiltration is related to geology, soil, and morphology, it is concluded that these factors are more dominant than factors related to runoff generation at the uplands. Therefore, hydrologic estimation procedures used in design of hydraulic structures in the region should be revised to consider the channel processes. The calibrated parameters of the runoff transformation models can be used in similar ungauged watersheds in the region. However, it should be noted that the empirical models imbedded in the HEC-HMS are liable to high uncertainty, since the input parameters assignment is highly depend on the expertise of the user. The sensitivity analysis results of this study can be used as a guide to show the level of uncertainty incurred in the model outputs for any deviation from calibrated parameter values. Most of the rainfall events of the two years record did not produce a runoff that can reach the observation point and had a measurable depth. Even from the rainfall events that produced a well developed runoff hydrograph at the observation point only few produced a rainfall excess higher than 1mm. This indicates hydrologic estimations based on mean seasonal and annual water balance approaches are uncertain and invalid. Event based approach in combination with probabilistic techniques to estimate the number of events per year, which produce a runoff, would be a better choice in water management practices. In this study, some directions are given in a separate case study about event-based estimation of water available for groundwater recharge. Even excluding water stored in artificial reservoirs of surface water, it was revealed that high proportion of water resides in channel reaches of the watershed. This in conjunction with the high geologic disturbances and tectonics in the watershed directs to the anticipation of high groundwater recharge potential. The idea can be more developed if farther investigations are made to assess whether the water left in the channel reaches is percolating to recharge the groundwater or gets lost through evapotranspiration

Mots clés : Niederschlag / Niederschlag-Abfluss / semi-arid / Sickerverluste / hydrologische Modellierung / HEC-HMS / Grundwasser-Anreicherung / rainfall-runoff / semi-arid / transmission losses / hydrologic modelling / HEC-HMS / Ilala River / groundwater recharge

Présentation

Version intégrale

Page publiée le 18 janvier 2015, mise à jour le 13 mars 2019