An improved water table fluctuation technique for estimating recharge is applied to a sustained (10-year) groundwater level monitoring record in Uganda, a rare dataset for tropical Africa, and compared against results from soil moisture balance models (SMBMs) and unsaturated flow models. The results show that recharge is directly proportional to rainfall (long-term average rainfall is around 1200 mm/a), even during times when high soil moisture deficits are anticipated. This indicates that preferential and/or localized flow mechanisms dominate the recharge behaviour. SMBMs and unsaturated flow models, based on uniform flow governed by the Richards equation, are shown to be inappropriate for estimating recharge in this location underlain by lateritic soils. Given the large spatial coverage of lateritic soils both globally and in tropical Africa, and despite the convenience of physically based uniform flow models and, in particular, SMBMs, concern is raised over the use of such models for recharge estimation, and thus for exploring future trends due to climate or land use change, unless backed up by sufficient hydraulic data to enable the recharge processes to be confirmed. More research is needed to assess how widespread preferential flow may be within other major soil groups and climate zones.
Keywords: climate change, groundwater recharge, soil moisture balance model, Uganda, water table fluctuation method