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

Accueil du site → Master → Suisse → Rainfall-runoff modelling of Meso-scale catchments in the Upper Ewaso Ng’iro Basin, Kenya

University of Bern (2003)

Rainfall-runoff modelling of Meso-scale catchments in the Upper Ewaso Ng’iro Basin, Kenya

Notter, Benedikt

Titre : Rainfall-runoff modelling of Meso-scale catchments in the Upper Ewaso Ng’iro Basin, Kenya

Auteur : Notter, Benedikt

Université de soutenance : University of Bern

Grade : Master 2003

Résumé
The Ewaso Ng’iro Basin in Kenya is a classical example of a highland-lowland interaction : Mt. Kenya, which receives high precipitation, contributes a major share to the discharge of the rivers in the basin, but growing demand for irrigation water on its foothills produces water shortage in the lowlands. Water scarcity leads to conflicts that can only be solved by community self-help mechanisms like the recently founded River Water Users Associations, but their work is hindered by a lack of knowledge on the present and future availability of water resources. . This study tries to make a contribution to better water management in the area by : - evaluating a new version of the land-use sensitive NRM Streamflow Model (Thomas 1993, McMillan 2003) in three meso-scale perennial catchments on the Mt. Kenya slopes, as a potential tool for water management. The performance of the NRM3 Streamflow Model is tested in predicting discharges, specifically low flows, under various geographic and data availability conditions ; - creating better insight into the hydrology of the study catchments ; - analysing scenarios of land use and climate change ; the NRM3 Streamflow Model is run with modified GIS (for land use scenarios) and meteorological time-series inputs (for climate change scenarios) according to possible trends. As a prerequisite to hydrological modelling, groundwater influences in the Naro Moru, Burguret and Nanyuki catchments are assessed by profile measurements of discharge, electric conductivity and temperature. The results indicate that the lower forest belt and the footzone are groundwater discharge areas and the savannah zone is an area characterized by transmission losses. Hydrological modelling of the catchments is considered possible without major modifications of the NRM3 Streamflow Model structure. The great insecurities on the anthropogenic influences (namely water abstractions) are a serious drawback to the validity of the results. River water abstractions are assessed based on a campaign during the field work period and data from previous monitoring by NRM3, resulting in a naturalised flow series that model outputs can be compared to. In a second step, the NRM3 Streamflow Model is calibrated and validated in the Naro Moru, Burguret and Nanyuki catchments using the time period 1987 - 1991 for calibration and the period 1992 - 1996 for validation. The experiments show a reasonably good performance on the decadal time-step and an acceptable performance on the daily time-step. The model can be used without extensive calibration. GIS input data produce good results up to a spatial resolution of 500 m. It is crucial, however, to use good quality rainfall data from a well-distributed measuring network, which is only given in the Naro Moru catchment. Improvements on the model could mainly be made by modifying the simulation of groundwater discharge (introduction of a second groundwater store or variable groundwater parameters), by introducing an altitude-dependent method of precipitation interpolation and by making the programme more user-friendly. In the last part of the study, the NRM3 Streamflow Model is run in the four nested Naro Moru subcatchments with modified GIS land use maps representing five scenarios of changed land use, and with meteorological time-series modified according to the climate projections of the IPCC SRES illustrative marker scenarios A2 and B2. It is shown that degradation of the land cover would mainly result in higher peak flows ; low flows would be slightly, if at all, reduced. More dramatic is the output of the climate change scenario A2 : With precipitation drastically reduced in June to August, low flows are significantly reduced, increasing the water scarcity ; on the other hand, in the dry season and rainy season months floods are projected to increase dramatically. Scenario B2 causes similar but less severe impacts.

Sujets : Upper Ewaso Ng’iro Basin, Kenya ; Runoff USE Stormwater ; Stormwater ; Rain ; Models ; Catchments USE Watershed ; Watershed ;

Annonce (NCRR)

Version intégrale

Page publiée le 2 février 2017, mise à jour le 26 novembre 2019