Gacu et al. (2025) Application of Artificial Intelligence in Hydrological Modeling for Streamflow Prediction in Ungauged Watersheds: A Review

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Water
• Year: 2025
• Date: 2025-09-14
• Authors: Jerome G. Gacu, Cris Edward F. Monjardin, Ronald Gabriel T. Mangulabnan, Janelli M. Mendez
• DOI: 10.3390/w17182722

Research Groups

This paper is a review and synthesis of existing literature; therefore, specific research groups involved in conducting a primary study are not applicable. The review synthesizes findings from numerous research groups globally.

Short Summary

This review synthesizes recent advancements in artificial intelligence (AI) for streamflow modeling in ungauged watersheds, demonstrating that AI-based models, particularly deep learning architectures, consistently outperform traditional models in capturing nonlinear hydrological responses.

Objective

• To synthesize recent advancements in artificial intelligence (AI) for streamflow modeling in ungauged watersheds, focusing on machine learning, deep learning, and hybrid modeling frameworks.

Study Configuration

• Spatial Scale: Global (synthesizing literature across diverse climatic and physiographic settings).
• Temporal Scale: Recent literature (focusing on contemporary advancements in AI for hydrology).

Methodology and Data

• Models used: The review examines Machine Learning (ML), Deep Learning (DL), and hybrid modeling frameworks, specifically highlighting Long Short-Term Memory (LSTM) networks and attention-based architectures. It also discusses traditional conceptual and physically based models for comparison.
• Data sources: The review discusses the use of proxy variables (e.g., Normalized Difference Vegetation Index (NDVI), soil moisture, digital elevation models) and remote sensing data to generate synthetic data for streamflow modeling, particularly in data-scarce regions.

Main Results

• AI-based models, especially LSTM networks and hybrid attention-based architectures, consistently outperform traditional conceptual and physically based models in capturing nonlinear hydrological responses across diverse climatic and physiographic settings.
• The integration of AI with remote sensing significantly enhances model generalizability, particularly in ungauged and human-impacted basins.
• Key research gaps identified include inconsistencies in model evaluation protocols, limited transferability across heterogeneous regions, a lack of reproducibility and open-source tools, and insufficient integration of physical hydrological knowledge into AI models.

Contributions

• Provides a comprehensive synthesis of recent advancements in artificial intelligence for streamflow modeling in ungauged watersheds.
• Highlights the superior performance of AI-based models, particularly deep learning architectures, over traditional hydrological models.
• Identifies critical research gaps and proposes future research directions, including the development of physics-informed AI frameworks, standardized benchmarking datasets, uncertainty quantification methods, and interpretable modeling tools.

Funding

The provided text does not contain information regarding the funding sources for this research.

Citation

@article{Gacu2025Application,
  author = {Gacu, Jerome G. and Monjardin, Cris Edward F. and Mangulabnan, Ronald Gabriel T. and Mendez, Janelli M.},
  title = {Application of Artificial Intelligence in Hydrological Modeling for Streamflow Prediction in Ungauged Watersheds: A Review},
  journal = {Water},
  year = {2025},
  doi = {10.3390/w17182722},
  url = {https://doi.org/10.3390/w17182722}
}