ABSTRACT
Understanding surface-water and groundwater interactions plays a major role in the successful implementation of integrated water-resource management. Challenges have arisen due to the lack of observed data, differences in surface-water and groundwater catchment boundaries and because different methods have traditionally been used for assessing surface water and ground-water. The result is a great deal of uncertainty when modelling interactions. One approach to integration includes the addition of more explicit groundwater algorithms into the commonly used Pitman hydrological model. This study uses the model to explore the main sources of uncertainty identified in the headwaters of the Breede River located in the Western Cape Province, South Africa. The results are presented within an uncertainty framework facilitating a range of possible process descriptions, rather than one hydrological time series. Despite attempting to more clearly identify the dominant sources of uncertainty in the catchment, the range of uncertainty did not vary appreciably. The complexity of the processes in the catchment seems to have resulted in an unresolved amount of uncertainty from a number of sources. While some uncertainties can be reduced (notably water use and recharge), this can only be achieved through substantial investments of human and financial resources. The results suggest large uncertainty in natural and impacted low flows, which has serious implications for water management related to environmental flow legislation.
