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

Titre : Development of Flood Early Warnig System for Flashy Rivers : Case Study of Gash River, Sudan

Auteur : Elhadi Eisa Adam Abaker

Université de soutenance : University of Khartoum

Grade : Ph. D. in Civil Engineering 2020

Résumé partiel
This research presents a Flood Early Warning System (FEWS) developed for Flashy Rivers ; taking Gash river, Eastern Sudan, as a case study. The FEWS was developed based on hydrologic modeling. The Hydrologic Modeling System (HEC-HMS) software of the United State Army Corps of Engineer - Hydrologic Engineering Center was used for rainfall-runoff simulation. The Hydraulic Engineering Center’s River Analysis System (HEC-RAS) was used for routing the outflow hydrograph in the river reach further downstream. The study used near-real-time Satellite Rainfall Estimate (SRE) data from the source “Tropical Rainfall Measuring Mission (TRMM-3B42RT V7)” to generate point as well as area-average rainfall over the catchment. For the first time, ground rainfall data from seven Eritrean stations were used to validate the TRMM data sets. A multiplicative bias procedure was adopted to validate the TRMM data sets. Since the catchment is relatively complex in term of topography, rainfall estimates were checked for elevation dependency. The algorithm was found to be affected by elevation and the SRE were deteriorating with increase in elevation. New reliable discharge measurements at Gera (outlet) and kilo 1.5 stations for the years 2015 to 2018 were obtained. Though this data set are few for calibration of the model, but for the time being it is the only available reliable data. The years 2015-2016 were used for calibration and 2017-2018 for validation. Two statistical performance measures were used to evaluate the model performance ; NSE and RMSE. The average values for NSE were found to be 0.94 for calibration and 0.90 for validation, while for the RMSE the average values were 105.5 m3 /s for calibration and 82.65 m3 /s for validation. The results of hydrological modeling indicate reasonable correspondence between observed discharge and SRE in terms of peak and time to peak. This promising result reflects the great opportunity for HEC-HMS model to be used as a rainfall-runoff simulation tool and further more for Gash river FEWS. Similarly, the SRE data from the source TRMM-3B42RT V7 mimic reasonably the rainfall amount and their spatial and temporal variations in the study region. TRMM data sets are great asset in catchments with limited data or other poorly- v gauged (transboundary) catchments. HEC-RAS was used to route the outflow hydrograph between Gera and Kassala Bridge, but unfortunately the model failed to predict the hydrograph at downstream stations

Présentation

Version intégrale (3,2 Mb)