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

Accueil du site → Master → Pays Bas → Flood simulation using public domain data for a data scarce transboundary basin : the case of Gash River basin, Horn of Africa

UNESCO-IHE Institute for Water Education, Delft (2014)

Flood simulation using public domain data for a data scarce transboundary basin : the case of Gash River basin, Horn of Africa

Rokaya P.

Titre : Flood simulation using public domain data for a data scarce transboundary basin : the case of Gash River basin, Horn of Africa

Auteur  : Rokaya P.

Etablissement de soutenance : UNESCO-IHE Institute for Water Education, Delft

Grade : Master of Science (MS) 2014

Résumé
This study was undertaken to investigate the possibility of developing a reliable rainfall-runoff model for flood forecasting using public domain datasets for the Gash River in eastern Sudan. The Gash is a transboundary river starts in Ethiopia, Eritrea and ends up in Sudan. Five satellite based rainfall estimates (SBRE) viz. TRMM, ECMWF, ARC-2, RFE-2 and TAMSAT were assessed to evaluate which rainfall product represents better the actual rainfall pattern and intensity of the basin. After evaluation based on performance measures such as the : Root Mean Square Error (RMSE), Mean Average Error (MAE) and Correlation coefficient (R2), TRMM, ARC-2 and ECMWF were selected for rainfall-runoff simulations. Model simulations were carried out using HEC-HMS model. Among six years of available discharge data from 2007 to 2012, period of 2007 to 2010 was used for calibration with 2007 as warming up period whereas data from 2011-2012 were used for validation. In additional to continuous daily rainfall-runoff model development using HEC-HMS, an event based flood model was also developed which was used for flood modelling with 72 hours lead time. TRMM and ARC-2 rainfall data were taken as two cases of rainfall inputs for these simulations, based on their superior performance in the earlier analysis. Also simulations based on daily TRMM versus 3-hourly TRMM were compared to evaluate effect of input time step on the results. The study shows that SBRE products provide important insights of the rainfall pattern in this data scarce region of Gash River Basin, both in terms of spatial and temporal resolutions. However, the results indicate better correspondence between observed rain gauge data and SBRE at longer time step such as monthly, while daily time step shows poor correspondence. However, the availability of only one rain gauge data for the whole Gash basin remains a limiting factor for making general conclusions. The SBRE based rainfall runoff model (HEC-HMS) could not simulate the time series of flows, in particular daily peak discharges, with desired level of accuracy. This may limit potential use of these data for flood forecasting before accuracy can be improved further. Nevertheless, the model could be useful in computation of monthly or longer time scale rainfall-runoff process and estimation of water balance components which could helpful to guide water management practice in the basin. However, the event based HEC-HMS model developed using SBRE data with shorter time steps (3-hourly) showed good capability to simulate daily peak discharges. It is expected that accuracy of the forecast could be improved with further enhancement of SBRE against ground measurements.

Sujets  : flood forecasting ; rainfall-runoff modelling ; hydrological modelling ; river basins ; Sudan

Présentation

Page publiée le 7 janvier 2017, mise à jour le 18 octobre 2018