Simulating the impacts of land-use change on streamflow time series for southern African rivers

Two rainfall-runoff models (one daily and one monthly time-step) have been tested using catchments where the land-use changes (afforestation and clear felling) in the recent past have been well documented. In general terms, the models performed satisfactorily and the changes in parameter values required to simulate the changes were reasonably consistent across the range of catchments considered. While all the catchments used were small headwater areas, it is suggested that the same scheme of parameter value change can be applied at the larger scale. It is not really possible to determine whether the models are sensitive to differences caused by variations in planting or management practices, largely because these can be masked by errors in, or lack of representativeness of, the input rainfall and evapotranspiration data.