Abstract:Groundwater discharge from the Riverine Plains of the southern Murray-Darling Basin is a major process contributing salt to the Murray River in Australia. In this study, data from an irrigated 60 000 ha catchment in the Riverine Plains were analysed to understand groundwater discharge into deeply incised drains, the process dominating salt mobilization from the catchment. We applied three integrated methodologies: classification and regression trees (CART), conceptual modelling and artificial neural networks (ANNs) to a comprehensive, spatially lumped, monthly data set from July 1975 to December 2004. Using CART analysis, it was shown that rainfall was the most important variable consistently explaining the salt load patterns at the catchment outlet. Using the conceptual model representing spatially lumped groundwater discharge into deeply incised drains, we demonstrated that salt mobilization from the study catchment can be well represented by a rainfall contribution, influenced by the hydraulic head in the deep regional aquifer and potential evapotranspiration. Using ANNs, it was confirmed that rainfall had a much higher impact on salt loads at the catchment outlet than irrigation water use. All these results demonstrate that under conditions similar to those experienced from 1975 to 2004, it is rainfall rather than irrigation water use that governs salt mobilization from the study catchment. Management of salt mobilization from irrigated catchments has traditionally focussed on the improvement of irrigation practices but it could be equally important to further understand the scope for management to control groundwater discharge in these irrigation areas.