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

Accueil du site → Master → Pays Bas → 2011 → Combining different sources of rainfall data for streamflow prediction of Gumara and Ribb rivers, Ethiopia.

UNESCO-IHE Institute for Water Education, Delft (2011)

Combining different sources of rainfall data for streamflow prediction of Gumara and Ribb rivers, Ethiopia.

Wolde Mariam, G.D.

Titre : Combining different sources of rainfall data for streamflow prediction of Gumara and Ribb rivers, Ethiopia.

Auteur : Wolde Mariam, G.D.

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

Grade : Master of Science (MS) 2011

Résumé
In many developing countries, especially in Africa, the number of rain gauges is small and unevenly distributed. The existing gauge networks are deteriorating and the data obtained from these gauges are mostly missing. This research study deals with the use of freely available Tropical Rainfall Measurement Mission (TRMM-3B42V6) rainfall data with a reasonably high spatial (0.25ox 0.25o) and temporal (3 hourly) resolution and the European Center for Medium-Range Weather Forecasts (ECMWF-erainterim.228) re-analysis data to support the scarce ground station rain gauge data. The three data sources are merged together based on the non parametric kernel smoothing method. The accuracy of the merged field is evaluated using two approaches : cross validation on another catchment and simulation result in HEC-HMS hydrological model. The research is applied to a case study area in the Upper Blue Nile, Lake Tana basin with two rivers, the Gumara and Ribb. Reliable rainfall and stream flow information for these rivers is important for flood management. The existing ground stations in the study area are so sparse that only two gauges are available in Ribb catchment while there are none in Gumara catchment. First, the performance of TRMM rainfall estimate is checked without merging, it is very low when compared with observed ground stations. The streamflow simulation result from TRMM is very low compared with observed flow hydrograph. Second, TRMM and gauge are merged to see the effect of merging in the improvement of areal rainfall. The streamflow simulation result was good. Third, the background field is updated by ECMWF reanalysis data and merged with the gauge. Cross validation is made on wetet Abay catchment which is the upper part of Gilgel Abay catchment. Only TRMM and gauge is merged on this catchment since the ECMWF data requested for the study area does not include this catchment. The cross validation results demonstrate that the merged field can efficiently combine information and takes the advantages from each datasets with a reasonable bias 22% and coefficient of efficiency greater than 72%. Finally, in order to evaluate the performance of stream flow estimate from the merged rainfall fields, a rainfall runoff HEC-HMS model is prepared one for Gumara and one for Ribb. Two rainfall datasets, the rainfall from the existing sparse gauges and the rainfall from the merged field are selected to see the effect in the accuracy of stream flow prediction with and without combining the three different data sources.

Sujets  : rainfall data streamflow forecasting rivers Ethiopia

Présentation

Page publiée le 14 février 2020