Hydrological modelling of ungauged catchments : a case study of the lower Kariba catchment

Abstract

In catchment hydrology, it is practically impossible to measure everything we would like to know about the hydrological system, mainly due to high catchment heterogeneity, the limitations of measurement techniques and the cost involved in collecting and processing the hydrological data, hence many catchments around the world remain poorly gauged. The Lake Kariba lower catchment is such a catchment that is not fully gauged. There is however a prospect of a less costly way of estimating lower catchment inflows using hydrological modelling techniques coupled with the use of free satellite data to simulate flows in the Lake Kariba Lower Catchment. RS MINERVE software was used to simulate surface run-off for poorly gauged Lower Kariba Catchment. The integrated rainfall-runoff model HBV was used for calibration and validation, with the model performing satisfactory at monthly time step, with the average runoff of the lower Kariba catchments being approximately 12% of the total yearly runoff recorded at Victoria Falls gauging station. Open source Geographic Information System (GIS) Software QGIS was used to derive drainage networks from DEM for the sub basins, the outflow point of the catchments. The model’s results showed good correlation with observed data giving a Nash Sutcliffe coefficient of 0.82, Pearson correlation of 0.89 and Bias score of 0.94 for the Sanyati flows while Gwayi flows gave 0.64, 0.87 and 0.98 respectively for the same indicators after calibration. The simulation results obtained from the model can be used in many water resources management activities. Keywords: DEM, HBV, Hydrological Model, QGIS, RS Minerve, Satellite Data, Ungauged Catchments.