Groundwater and other hydrological models are usually used only in situations where large amounts of historical and spatially distributed data are available. There are however good arguments for using models (a) to test conceptual models and assess the controls on system behaviour, and (b) to assist in planning data collection. The paper describes the application of a finite difference groundwater model to an area of dune fields and oases in northeast Nigeria. Limited field data provided inputs and the basis for calibration of the model, and model output gave useful insights into the functioning of the hydrological system. Conclusions are drawn on overall system behaviour, the relative merits of different recharge hypotheses (specifically addressing the controversy over whether present groundwater levels are the result of current recharge or fossil gradients), and the requirements for future data collection and monitoring.