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

Accueil du site → Master → Pays Bas → Rainfall estimation by remote sensing for conceptual rainfall - runoff modeling in the upper Blue Nile basin

International Institute for Geo-Information Science and Earth Observation (ITC) 2007

Rainfall estimation by remote sensing for conceptual rainfall - runoff modeling in the upper Blue Nile basin

Ymeti, I. (Irena)

Titre : Rainfall estimation by remote sensing for conceptual rainfall - runoff modeling in the upper Blue Nile basin

Auteur : Ymeti, I. (Irena)

Etablissement de soutenance : University of Twente International Institute for Geo-Information Science and Earth Observation (ITC)

Grade : Master of Science in Geo-Information Science and Earth Observation 2007

Résumé
To be able to deal with water problems (floods and droughts), it is common to make use of hydrological models. The successful application of such models depends not only on the model structure, the different time and space scale associated, but also on the accuracy of the rainfall as the main input. However, in many developing countries such as Ethiopia, and specifically in the upper Blue Nile basin, the rainfall observation network is relatively sparse. Therefore, the objectives of this study are : estimation of rainfall from satellite imagery and assessing the performance of conceptual rainfall-runoff model using ground truth and satellite data as input. The advantage of remotely sensed data over data from the rain gauges is that they provide relatively consistent spatial and temporal coverage of rainfall information. However, estimating rainfall from remote sensing is rather complex. The estimation of rainfall in this study is based on the combination of geostationary MSG (Infrared channel) and orbiting TRMM (Microwave channel) satellite data. To combine these satellite data, a regression function associated with a threshold as an upper cloud temperature limit where rain occurs is determined. In this way, in the upper Blue Nile basin rainfall maps at 15 minute intervals with 3 km pixel size from July to November 2005 are generated. By comparing the results established from satellite with the ground truth data, over and underestimation of rainfall is evident. This work is not taking into account the issues associated with calibrating satellite data, but with the direct use of available rainfall data as input to hydrological models. Secondly, two conceptual models SAC-SMA and HBV-96 are chosen to test their performance by applying two different input data in Gumero subcatchment. Based on ground truth input data, both models are calibrated by adjusting the parameters manually. Considering both quantitatively and qualitatively judgment, the performance of the model was not as satisfactory as expected. It should be realized that in any modeling application regardless of the chosen model, there always may be a difference between observed and simulated discharge data. This could be caused by many aspects that relate to model uncertainties. The satellite data are applied to the models to see how they perform. The model responses to the network and satellite data, although some deviations are observed. These deviations are related to the results established from RS, which have direct impact to the hydrological model. Thus, further study in rainfall estimation from satellite should be carried it out.

Mots clés : model ; MSG ; TRMM ; SAC-SMA ; HBV-96 ; Blue Nile ; Gumero subcatchment.

Version intégrale (ITC)

Page publiée le 29 janvier 2018