Deep Learning Model With Periodic Embedding Improves Monthly River Runoff Predictions
A new deep learning framework integrating Fourier periodic embeddings improves monthly runoff prediction across climatically contrasting hydrological stations.
A new deep learning framework integrating Fourier periodic embeddings improves monthly runoff prediction across climatically contrasting hydrological stations.
During validation, a variable coefficient convolution combination method retains mean-based streamflow forecast accuracy while producing narrower predictive intervals in the ...
Incorporating uncertainty during hydrological model calibration significantly improves parameter identifiability across diverse catchment conditions in the United States.
A multi-model hydrological study across Japanese river basins demonstrates that Bayesian parameter estimation provides little operational forecasting benefit over deterministic ...
A comparative study of three deep learning models in an Australian alpine catchment shows that LSTM networks can forecast hourly ...
A new study tests a Transformer-based physics-informed neural network against support vector regression for forecasting hydrological drought in the transboundary ...
A new deep learning framework for river flow forecasting puts flood peaks, uncertainty, and interpretability at the center of model ...
A 25-year review of artificial intelligence applications in streamflow forecasting reveals that feed-forward neural networks dominate the field while long-range ...
© 2026 Scienmag - Science Magazine
© 2026 Scienmag - Science Magazine