Deep Learning, Epistemic Uncertainties and the Inexact Sciences

Key Points

• This commentary examines how deep learning and machine learning can address epistemic uncertainties in hydrology.
• Analysis of recent literature on deep learning in hydrology
• Discussion on model applications
• Evaluation of uncertainties in existing hydrological models
• Identified potential benefits of deep learning in understanding hydrological processes
• Highlighted challenges due to epistemic uncertainties
• Proposed directions for future research in hydrological modeling