Salt intrusion is a major, growing threat to the availability of freshwater in coastal areas. Reliable short-term forecasting of salt intrusion is essential to protect the freshwater supplies for human use, agriculture and industry. Currently, real-time one-dimensional forecasting models are not able to optimally model the flow division and salt exchange at tidal channel junctions, due to the complex dynamics that are difficult to capture one-dimensionally. We aim to improve the understanding of dominant mechanisms that govern salt exchange at tidal channel junctions and to derive new parametrisations of the three-dimensional dynamics, to be used in existing one-dimensional models to improve salt intrusion forecasting at the channel network scale.