Regional analysis of low flow regimes over three river basins in Zambia

For the purpose of planning, design and management of water resources projects, it is necessary that hydrologic data (information) is available at the point where the project is to be based . However it is not always that hydrological data are avaiaible at the point of interest. Some of the data required for planning and design of these water resources projects include the average daily flow, average flows exceeded by different percentages of average daily discharge. One of the techniques that allows the estimation of hydrological information like low flow statistics at ungauged sites is regionalisation. This is a technique which involves the analysing of hydrological information of gauged catchments of an area (country, river basin etc.) in order to establish the hydrological parameters required. When these parameters have been estimated, they are then related to easily measured catchment characteristics like catchment area, slope, length of the main stream, average annual rainfall etc. of the gauged catchments. These relationships, which usually take the form of regression equations, are then used to estimate the required hydrologic parameters at ungauged sites in the area. In this study river flow data (daily discharges) from nineteen hydrometric gauging stations from three main river basins in Zambia have been analysed in a regional low flow anlysis in order to provide methods of estimating low flows at ungauged sites for engineering applications. Flow duration curves for 1, 10, 30, 60, and 90 day durations were established and pooled (grouped) into homogeneous groups using regionalization techniques. Regional regression models (equations) were developed that could be used to estimate/predict such low flow statistics as average flow exceeded by ninety five percent of average discharge. The predictor variables used in this study were area of catchment, average annual rainfall and base flow index (bfi). The results of this study may be used in estimating low flow parameters at ungauged sites: For the allocation of water to competing users; For siting of treatment plants ; For deciding the quantity and concentration of sewage to be disposed of into receiving streams; etc. It is recommended that data from all gauged catchments in the country should be used in order to develop more reliable and practical regression models and the grouping of flow duration curves into groups of homogeneous catchments should be carried out for the whole country. More catchment characteristics should be related to flow parameters instead of just area and rainfall.