Dambos and discharge in central Zimbabwe

Published conclusions on dambo hydrological process indicate that the impedence of drainage and the seasonally high water tables within dambos lead to either i) lower flood peaks and greater duration and magnitude of dry season flows or ii) more rapid flood response and higher dry season evapotranspiration losses, the latter resulting in reduced dry season flows. The absence of empirical research in Zimbabwe has meant that the perception of dambos behaving in accordance with the first model has remained unchallenged. This thesis examines the extent to which river discharges in central Zimbabwe are modified by dambo density and other catchment characteristics, and provides a manual of techniques for estimating flows at ungauged basins.A data base is assembled for 110 gauged catchments covering over 75,000km3, averaging 20 years of data and having dambo density between 1% and 60% of catchment area. Twelve single number indices of catchment characteristics include dambo density, geomorphic and soil indices and climate/rainfall statistics. Eighteen flow statistics include measures of annual yield and hydrograph separation, and statistics of low and flood flows. Multivariate analysis of the data base indicates insignificant relationships between all flow statistics and dambo density at the regional scale and regime variability is attributable more to average annual rainfall, rainfall variability and regolith profiles. However, the reduction of dry season low flow indices and the separation of yield into a higher storm runoffcomponent are significantly related to dambo density amongst basins where dambos occur in association with deeply weathered regoliths. It is concluded that i) rapid response runoff is the dominant process contributing to river flow, but that dambos are insignificant regionally in affecting routing of this response, ii) baseflow contributions are small (approximately 35%) and vary with the regolith profile of the interfluve soils and iii) dambo density is largely insignificant in modifying basin response, except in association with deeply weathered profiles. Multivariate procedures are developed into a practical manual for the estimation of catchment yields, low flows and floods at ungauged sites in central Zimbabwe. Dambo hydrological process rates and models are not appropriate when extrapolated throughout central Zimbabwe where dambo density is largely insignificant in explaining variability in flow regimes.