Modeling runoff with incomplete data: a comparison of hydrological, deep learning, and hybrid approaches

Key Points

• Runoff modeling accuracy varies significantly between hydrological methods and deep learning approaches.
• Key performance metrics show that hybrid models can outperform conventional methods by 30% in specific contexts.
• Comparative analysis of three distinct approaches highlights advantages of integrating data techniques in runoff prediction.
• Findings emphasize the need for improved data handling methods to enhance operational predictions in water management.