Additional surface–ground water interaction routines were recently added to the Pitman monthly rainfall–runoff model, widely used in South Africa for quantifying water resources in ungauged catchments. Some evaluations of the model have demonstrated that it can realistically simulate interactions between surface and ground water at catchment scales of approximately 100 to 5,000 km2. The model allows ground water abstractions to be simulated, but no reported evaluations of this component are available. This study uses the model to estimate sustainable abstraction volumes in a semi-arid catchment and includes an assessment of model parameter uncertainties. In recognition of potential spatial scale issues related to the model structure an alternative model configuration, based on splitting the total catchment into recharge and abstraction sub-catchments, was also tested. While the results appear to be conceptually appropriate, there is insufficient available information to quantitatively confirm the model parameters and results. The same would apply regardless of the type of model being applied in such a data-deficient area. Additional geo-hydrological information is required to resolve the model uncertainties and improve the parameter estimation process. This pilot study has highlighted the type of information required, but further work is needed to identify how best to obtain that information.