Application of artificial neural networks in the field of geohydrology

Groundwater has been identified as a viable alternative for future freshwater production in South Africa. The management thereof is steadily gaining more recognition from governmental institutions. A significant obstacle in the development of this resource is the conceptual understanding of surface water and groundwater interaction. The availability of reliable data for rainfall, flow volumes in rivers and water levels in boreholes have prompted an investigation into patching incomplete data sets. This study also focused on predicting the influence of rainfall and flow volumes in a river on the surrounding groundwater levels. Neural networks have been used to investigate both data patching and forward prediction of water levels in selected data sets.