Recurrent Neural Networks for Routing River Water in Land Surface Models: A Global Evaluation
Recurrent neural networks (RNNs) can effectively model river routing processes and outperform physics-based models when trained on global datasets of runoff and streamflow.